This is one of my articles where I connect seemingly unrelated topics, such as Carbon, Hydrogen, Trojans And Javalins, perhaps VTOL, UFOs, Crop Circles, Graffiti, MIBs, Mandela, Octrees or the frankly alarming Pressure Cooker, Induction Heater, systemd, PHP and MySQL. In this article I will describe common functionality found in analog game controllers, analog servo control, WS2812 and WS2812x NeoPixel™ LEDs and other circuits. In all cases, a capacitor is filled via a current limiting resistor over a period of time. The time taken to fill the capacitor conveys useful information.
In games like Pong, an analog potentiometer controlled by a player forms the current limiting resistor which controls the rate at which a capacitor is filled. If there are two or more players, circuitry is duplicated for each player. If a potentiometer is turned to one extreme then a corresponding racket within game is at the uppermost position. If the same potentiometer is turned to the opposite extreme, the racket is at the lowermost position. How does this work? The games console is connected to a television. The games console transmits a serialized picture to the television. This is rastered like reading a book, from left to right, from top to bottom. Unlike a book, frames may be sent 25 times per second or much more frequently. Also unlike a book, a frame may be split into two separate fields in which the odd and even lines are sent separately. This provides a good compromise between signal quality and high speed action.
At the end of line's transmission there is a horizontal retrace period. At the end of each field or frame's transmission, there is a vertical retrace period. This is provided for two purposes. Firstly, it provides sufficient time for CRT electron beam deflection hardware to adjust to a new line (or field or frame). It also provides a synchronization pulse in the otherwise continuous signal. Early game consoles extended this technique to incorporate low level changes with each line of the display and high level changes. Low level changes include a decision whether to draw a racket or not, and modification of the playfield. High level changes include scoring during a game and, between games, selection of game parameters. It is functionality between these levels which is of interest.
At the beginning of display, a transistor empties the capacitor. During the display of lines, the capacitor fills until a threshold Voltage level is reached. The values for the resistor, capacitor and threshold are chosen to ensure this happens during display or a little afterwards. Furthermore, a logarithmic potentiometer is often used to compensate for the exponential curve of capacitor charging/discharging. The threshold may be detected by polling or interrupt and segment may be repeated over successive display lines, as determined by the difficulty level of the game, to represent one player's racket. On systems which are particularly starved for resources, such as an Atari 2600, a flag register records intersection between different types of display object, including wall, sprite, ball and missile. 50 or 60 times per second, this register may be queried to determine high level game behavior. In the case of Pong, it allows the velocity of a ball to be reflected before its position is updated.
This crude extension to rastered television created an unbelievable craze in the late 1970s and it continued to build for about seven years. Some of the early systems used dedicated logic and no micro-processor. Others, such as the popular Atari 2600, were supplied with a 6507 (a 6502 in a cheaper 28 pin package) and a generous 128 bytes of RAM. It only occurred to me recently that 128 bytes is insufficient to hold a full bitmap display but exceeds the memory which can be accessed during one scanline. Regardless, this is definitely an environment where it is critical to count clock cycles. A 1.19MHz 650x in a system displaying 262 pseudo-interlace NTSC has approximately 76 memory cycles per scanline. This is sufficient to make trivial Boolean choices and poke registers but little else.
Servos
Open loop analog servos use a similar arrangement of capacitor, resistor and discharge transistor to convey information. In this case, the system typically works faster. A smaller capacitor may be filled 400 times per second or more frequently. A common standard for analog servo signal is 1ms pulse for servo at one extreme of movement and 2ms pulse for servo at the other extreme of movement. This applies to a wide range of servos irrespective of size, power, angular range or linear movement. It was disappointing to discover such detail. My reaction was "That's it? That's your standard?" This is the lowest common denominator of protocol standards. It is easy to explain. It is easy to implement. It is easy to calibrate. It is easy to repair. It is an interoperable standard across a wide range of hardware vendors. More importantly, have you got anything better to compete against this entrenched standard?
Unlike Pong, there are two sources of pulse. One is transmitted from a controller which may be a hand controller or recording. The other is generated locally. The positive (rising) edge of the incoming signal begins the charging cycle of the local capacitor. This is moderated via a local potentiometer which is positioned at the end of a motor and gearbox. From here, one of two cases occurs. Either the capacitor reaches a threshold before the incoming signal pulse ends or the incoming pulse ends first. (In this analog arrangement, exclude the possibility of simultaneous change.) The signal mis-match nudges a electric motor in the middle of an H-bridge. The nudge is proportionate to the duration of the imbalance. Indeed, the motor is nudged with every pulse cycle. However, with a clean signal, the nudge may be sufficiently brief that the motor and/or gears do not overcome friction.
Remote Control Models
Remote control models typically require a minimum of two channels: locomotion and steering. Model helicopters require control for a minimum of three channels: main rotor, trim rotor and steering. Models may have other functional or decorative movement. While it is possible to modulate pulse trains with separate carrier frequencies, it is easier to implement a synchronization pulse and stagger the encodings for, maybe, eight channels. A standard encoder module and decoder module was available with such functionality and allowed encode/decode of up to eight servo pulse trains. I would be highly unsurprised if it was used to control Omnibot's numerous functions. It was definitely used in Teddy Ruxpin, an extremely creepy toy from the 1980s which is like the furry version of a Chucky doll. Teddy Ruxpin had a compact cassette player. Tapes had one channel of read-along story which used the standard modulated audio encoding. The other channel encoded eight channels of servo pulses. This controlled Chucky, erm, Teddy's eyes and suchlike and there was dedicated channel which could wired to the murderer's accomplice, erm, Teddy's little friend. Unfortunately, this channel did not use the standard modulated encoding and thus discouraged unofficial spoof tapes.
NeoPixels™
I've been working with LEDs and micro-controllers. It was inevitable that multiple friends suggested NeoPixels™. This is a brand of AdaFruit which consists of some marketing, some more marketing, marketing in the guise of tutorial videos, a completely borked library, overpriced circuit boards and some otherwise undifferentiated open-loop, serial, daisy-chainable, programmable brightness, WS2811 RGB LED controllers and WS2812 and WS2812x RGBW controllers with integrated LEDs. If you look carefully, unbranded parts have the same component numbers and are 1/3 of the price.
Independently, two of my friends physically connected a string of these NeoPixels™ to an Arduino™ and compiled the standard LGPL NeoPixel™ library with the standard LGPL Arduino™ library before flashing the resulting BLOB via a poorly defined protocol spoken by avrdude. From here, it is easy to tinker with example programs which generate rainbow patterns and similar. But how does it work? My friends didn't know.
I like to investigate more thoroughly. A WS2811 datasheet says that eight bits per channel of brightness data should be sent MSB [Most Significant Bit] first, with one byte for red then green then blue. For reasons which are completely unfathomable to me, a WS2812 re-arranges this to green then red then blue. Also, the protocol lacks any forward compatibility. So, there is no padding byte to maintain compatibility between RGB and RGBW LEDs. The WS2812 datasheet says that a zero bit should be encoded as a 350ns±50ns pulse and that a one bit should be encoded with a minimum pulse of 700ns and no maximum. Data is automatically set live after a minimum of 50000ns inactivity. However, electronics forums widely repeat the factoid that many components have a less tolerant timeout of 6000ns. That's a broad and unhelpful range. It is also a cockeyed encoding. What's wrong with serial UART with a start bit? I tried explaining this encoding to a friend. Before I finished my explanation, my friend said "So, it works like Morse code?" I was surprised. A fixed number of bits are sent with fixed spacing but, yes, it otherwise works like Morse code. In this case, dot and dash represents zero and one. But why choose this encoding? It retains a very useful property which is shared with Morse code. It is a self-clocking signal. The receiver doesn't require an oscillator or UART over-sampling. Transmission speed may differ by a factor of two and it still works. Furthermore, it requires very simple circuitry which is very small.
Rather than working at the scale of 50Hz/60Hz television or the 1ms timescale of servo control, this works at a 350ns timescale. For any given value of resistor, a capacitor can be shrunk and filled proportionately faster. The positive (rising) edge of the signal begins capacitor changing. The negative (falling) edge of the signal determines the sampling time for the capacitor. There is no maximum period for a one (dash) because the capacitor remains full whenever it is sampled. This leads to the opposite question. Why is this type of protocol not used everywhere? Due to signal skew, it is only suitable for short-range communication. Morse style encoding is only suitable for high speed communication or long distance communication. Choose one. In this case, it is high speed communication over a short range and each LED acts as a repeater to the next LED. For this reason, it differs from typical LAN/WAN encodings.
Unfortunately, there are multiple, incompatible variants of WS2811, WS2812, WS2812B and other LEDs. Most variants don't receive red values first, although there are exceptions. Also, there is no placeholder byte and therefore RGB and RGBW LEDs are mutually incompatible - unless they are used exclusively with very smooth monochromatic values or completely random values, which is unlikely. In the case that you receive the wrong variant, be ready to patch code, re-compile code and re-deploy code.
In practice, there are two modes of operation for WS2812 LEDs. If 24 bits of data are sent (or 32 bits for RGBW variants), data is shuffled along by one LED. If this is repeated 20 times per second or suchlike, it creates a soothing trail along an unrestricted length of LEDs. Or, at least, until a chain breaks or becomes intermittent. Given that LEDs are powered in parallel, problems are easier to locate than series Christmas lights. However, it may be difficult to repair if the chain is within a water-proofed strip.
The other mode of operation is to blat serial data at (approximately) 1Mb/s and flush through all of the previous values. This is the most practical method to make a progression of lights run backwards through a chain or otherwise allow arbitrary change per frame. The upper limit for this process is chain length reliability, power considerations and the product of chain length × frame rate - which is adversely affected by the vague timeout parameter. In practice, 512 RGBW LEDs can be updated at 60Hz. To improve signal quality or exceed this limit, it is possible to divide a chain into segments. Each segment may have a separate micro-controller and they may be co-ordinated via uni-directional broadcast messages over RS-232 UART or RS-485 multi-drop serial. The latter was successfully used to implement interactive Christmas lights at King's Cross Station and a one dimensional retro game which is quite addictive.
There are separate problems with Voltage and Ampage. A WS2812 LED requires a 5 Volt supply. Power is automatically pulse limited for 2.7V red, 3.0V green and 3.3V blue channels. However, the blue channel is most susceptible to Voltage drop. This creates an effect where color temperature through the chain may drop from white hot, to yellow hot, to amber, to nothing as it extends from a power supply. In the general case, Voltage drop can be minimized by affixing a power feed in the middle of a chain rather than at one end. This doubles the effective span of power collection/distribution.
The next problem is current. I ordered a huge quantity of RGB LEDs. However, after it arrived, I realized that it required a 5V × 8A current correcting PSU. I then sheepishly ordered a 16 LED ring for development and testing. If you are brave enough, it is possible to connect multiple switched mode PSUs in parallel and hope that the Voltage drop makes the arrangement self-balancing. If this doesn't work, it'll make a very loud bang, destroy equipment and/or catch fire.
Some display signage uses WS2812 protocol. Square and rectangular panels are typically sold with 2^N LEDs with 3mm or 5mm spacing. However, to minimize distance between LEDs in a square grid matrix, they are wired in a serpentine arrangement. Therefore, even rows of LEDs are left to right and this is alternated with odd rows of right to left. Any transfer of an arbitrary bitmap from a host requires alternate lines to be sent backwards. Panels are now available in fine strips which can be fitted to the inside or outside of a gentle and varying curve. However, the joints between rows are most susceptible to damage.
NeoPixels™ Library
I downloaded the current version of the library used by my friends. I unpacked the example programs, the 88KB library and the 2KB addendum which is required to compile on ESP8266 with different compiler flags. My first reaction was "Oh, fuck me. It's amateur hour." 90KB? To blink lights?? That's either very comprehensive documentation, a large number of example programs or someone is doing it wrong. After writing a replacement library, which is 4KB, runs on more micro-controllers and - using the optional bit matrix transpose - drives up to eight chains of LEDs, I conclude that all of the contributors to the NeoPixels™ library are doing it wrong.
In the official NeoPixel™ library, the bulk of the source code is inline assembly; mostly variants of code so that a subset of supported micro-controllers, via a subset of supported pins, may work with a WS2811's legacy half-speed mode. I wondered if it was possible to replace most or all of the platform specific code with something dependent upon the de facto F_CPU compiler parameter. I also wondered if it was possible to have a delay function which works on a larger timescale; ideally beyond one second. My initial strategy was as follows:-
32 bit ARM is notoriously poor with 16 bit integers. It effectively has an 8/32 bit field on some instructions. Much of the remainder is handled with bitmasks. This is relatively tractable because many other instructions have a 2 bit field which allows an 8 bit constant 0xRS to be loaded as 0x000000RS, 0x0000RSRS, 0x00RS00RS or 0xRS00RS00. This allows the bitmask 0xFFFF to be handled relatively concisely. Regardless, the 16 bit loop isn't required on ARM. Actually, it isn't required on any other platform. After a few more tweaks, chose an arrangement which includes:-
This works very well. For long pauses, a C pre-processor is able to calculate the residual number of clock cycles required for the shim. Meanwhile, for short pauses, a C optimizer can omit the 32 bit busy loop. In all cases, the shim is output as a sequence of idle cycles. The official NeoPixel™ code isn't completely terrible. I was able to salvage one line and use it for a space saving optimization. Specifically, for AVR, it is possible to replace a pair of one cycle NOP instructions with RJMP .+0 which takes the same time but uses less ROM. Similar practice covers other architectures. My contribution to portable assembly may extend to a volatile empty string within the 32 bit loop. This appears to work on avr-gcc but I cannot guarantee that it is universal.
The advantage of my technique is a portable, (mostly) clock-cycle accurate busy wait which works over a wide timescale. It should work on a superset of micro-controller architecture which has NOP instruction, lacks instruction cache, executes exclusively from internal storage and is supported by GCC or other C compiler with inline assembly. The disadvantage of this technique is a power hungry busy loop which does not count elapsed time within interrupts. It is possible to save 10mW or more, for a brief period, while waiting for a hardware timer. However, this energy consumption is a minor consideration when I target automotive or agriculture; handling kiloWatts or MegaWatts.
Arduino™ + NeoPixels™ doesn't cover every case either - and this is a horrible example of leaky abstraction. Arduino™ covers a small subset of micro-controllers. NeoPixels™ covers a different subset and is very specific about clock frequency unless you want to modify the inline assembly for each processor architecture. NeoPixels™ provides clock-cycle accurate timing for one specialist case, outside of interrupt, unless you want to modify the inline assembly for each processor architecture.
My code busy waits but it doesn't busy wait until a hardware timer interrupt sets a global flag. Arduino's camel case delay(short) and delayMicroseconds(short) may use AVR's gloriously inconsistent 8 bit TIMER0, 16 bit TIMER1 or 8 bit TIMER2, with no fall-through case, using a worked example from an Atmel, erm, Microchip Corporation datasheet rather than the best accuracy. If you require a re-occurring interrupt, you are hopefully using a different timer.
Remember that long pulses have no upper bound? This is particularly ugly. On some processor architectures, the NeoPixels™ library only works outside of interrupt but to maintain responsiveness, only disables interrupts around the short pulses. Regardless, interrupts should not exceed the official timeout of 50000ns or the unofficial timeout of 6000ns. This possibly requires empirical testing with each batch of LEDs because the alternative is counting clock cycles of disassembled output or setting another hardware timer. Neither is pleasant. Some of the cruft in the NeoPixels™ library is for compatibility with Wiring® which was secretly registered by one of its contributors. This subsequently incurred a fork of the project which was named after a bar which was frequented by some of the contributors. The bar itself was named after Arduin Of Ivrea, who was King Of Italy from 1002 to 1014. The NeoPixels™ library compiles with either branch.
For my suggested implementation strategy:-
For Arduino™ + NeoPixels™:-
This is an unsatisfactory situation. There are many applications where neither implementation is suitable. Arduino™ + NeoPixels™ is most suitable for trinkets, although mass production is inadvisable. My implementation is most suitable for high power control and could be extended to low power control.
Re-implementation, with the described strategy, may be most suitable for a kernel module. This would allow multiple chains of WS2812 LEDs to be driven directly from a credit card computer; typically running Linux or BSD; most typically, a Raspberry Pi running Raspbian. This is not possible from an application due to privilege and timing problems. Furthermore, the NeoPixels™ library is not compatible with Broadcom's ARM I/O implementation. Most micro-controllers and implementations of ARM micro-processors have registers to specify if a pin is digital input or output. This may be subsequently overridden by enabling a specialist unit, such as UART, I2C, PWM or analog. It is quite typical for a pin to have five seperate functions specified across three or more registers. Broadcom takes a refreshingly sane approach and provides 30 bit registers which allow the function of 10 pins to be specified in octal. It could be improved by having 32 bit registers which allow *consistently* numbered functions to be specified in hexadecimal. Regardless, it is unusual and therefore neglected.
Feature Request
I would dearly like a variant of WS2812 LED with a restricted palette suitable for automotive use. Although it is possible to restrict color in software, I would also like a hardware interlock. This would prevent anything outside of typical light cluster colors. This can be easily achieved with a small amount of logic. Imagine a RGB color cube. Black in one corner. In adjacent corners are the primary colors: red, green and blue. Opposite to black is white. In adjacent corners to white are the secondary colors: cyan, magenta and yellow. If R ≥ G ≥ B then only 1/6 of the volume (a skewed tetrahedron) is available. This limits color within the range of black, red, yellow and white. It prevents green, blue, purple or any permutation (due to transient error, corroded wiring or other condition) which may be mistaken for emergency services. It also prevents citation for a vehicle with illegal lights while allowing matrix LEDs with user configurable indicator patterns.
Edit 1: By chance, I encountered the author of the one dimensional game. Said author confirms that the NeoPixel library is junk and recommends FastLED. I skimmed through the code and I confirm that it is far more competent. However, it uses C++ templating for clock cycle delays which is quite hairy.
I wrote a 13000 word epic about vehicle design and I could have written more. In this post, I'm going to try a more graphical approach to explain ideas. In particular, I'm going to explain why I've rolled sheets of paper into tubes and taped them together. Specifically:-
Deprecated Design
I previously considered making a pyramid car because the triangle shapes should be strong. Apparently, this is an old idea. A friend presumably searched for 'pyramid car' and found a 1980 concept car, Citreön Karin, with a glass pyramid roof and three seats. The interior is lovely and very much in the style of the NCC1701-D's bridge. The steering wheel and keypad on the steering wheel is dodgy but it is otherwise a fantastic design. Where can I buy one? I'd definitely drive it.
Proposed Design
Unfortunately, a pyramid vehicle is only safe for one occupant in a center position. I would say I'm working on a more conventional design. However, it is more like BigTrak, shuttlecraft Galileo or Africar.
I'm working on a hilariously small development budget. Regardless, this is more than sufficient to make prototypes of a space-frame design from rolled paper at 1:10 scale and 1:4 scale. Apparently, 1:4 scale prototyping is standard in the car industry. This is understandable because this scale is quite convenient to carry through doorways. However, I'm not sure that it is common practice to use Build-A-Bear plushies as crash test dummies. Regardless, I take safety sufficiently seriously to order safety clothing for bears. This includes crash helmet, goggles and dungarees.
Future Developmemt
I'm currently working on a 1:4 scale metal frame. Unlike paper models, this is suitable for a small combustion engine. (The full size design is parameterized but the default configuration has enough space for a Mercedes AMG engine.)
It is possible to make considerable progress with vehicle development by working entirely on scale model prototypes. Indeed, it is possible to install a credit card computer and miniature displays. From here, it is possible to have a reversing camera, music system or similar features found in modern cars. I won't spend much time on conveniences. I just want to show that it is practical and affordable to develop most features at scale. I'm more concerned about functional matters, such as energy efficiency. I also want to work on the killer feature of the proposed design. Specifically, the dashboard is three columns of 19 inch racking. The allows after-market functionality to be easily added. However, I'm working at 1:4 scale and therefore I'm currently investigating 1:4 scale 19 inch racking. This may be of general interest for projects which don't involve vehicles (or bears).
Given the miniaturization of electronics, miniature racking may be of interest for hi-fi separates, server farms or other applications. For example, it is possible to fit a credit card computer into a 1:4 scale rack. It is also possible to fit a 12 channel, surround sound amplifier into the same space. Indeed, it is possible to fit quite a powerful and flexible entertainment system into the space of maybe half an IKEA pigeonhole shelf. I don't mean half of the shelfing. I mean half of one pigeonhole.
At full scale, I might have a glider ready for London's next Red Bull Soapbox Race. Previous events in London have been on odd years, on the first or second Sunday of July. Assuming the event continues, it is unlikely to be on Sun 4 Jul 2021. It is more likely to be on Sun 11 Jul 2021. Admittedly, if I'm entering a glider into a comedy cart competition, I haven't made sufficient progress.
How You Can Help
The instructions, diagrams and a spreadsheet to calculate sizes are at the end of the previous post. Example pictures of construction may also be useful. Please verify the design or use the techniques to make a competing design. The objective is to make safe, affordable vehicle which is easy to repair (and also cheap to prototype). If you have better a technique or a cheaper, safer design then we will readily adapt to your methods.
Work on electronics or firmware. There is essential and non-essential functionality. The six channel agricultural light (hardware, firmware) is easily adapted into a automotive light cluster controller which can drive 55W headlights. Anyone who has driven a strip of NeoPixels with an Arduino will find the frivolous functionality in a Mini trivial to replicate.
Request parts from us. Collect in London or we'll list them on EBay at cost. For some reason, you may want a 1:10 scale paper model rather than making your own from instructions. To send this within the UK it would cost £5 or less. Larger models and stronger material is likely to cost considerably more. Light clusters (5V, 12V, 20V or 48V) without PSU would be less than £10 per pair. Cost of 1:4 scale racking is currently unknown.
Welding. Technically, I'm trained to arc-weld but I know my limits. All welding tasks paid at commercial rate.
Looking for opportunity to promote this design and I'm not too fussy. If someone has an *actual* production budget for a music video and wants something unique, interesting and visually arresting, well, we've got it. And whoever pays for the hardware keeps the hardware.
If you're in London then come to the weekly geek meetings where we discuss this and other topics.
The weekly geek meeting, which occasionally incorporates Sewing For Physicists, has changed day, changed location and gained additional attendees. We now get 3-7 people who meet at the back of The Foxglove Public House, 209-211 Liverpool Road, Islington, London, N1 1LX, which is approximately 400 metres North of Angel Underground Station. This is not to be confused with The Regent on the North West corner of Richmond Avenue and Liverpool Road.
Unfortunately, we now meeting at an independent pub where drinks are more expensive due to the lack of volume purchasing or volume selling provided by (politically divisive) companies such as Wetherspoons. Restaurant options are slightly more limited. There is a Vietnamese restaurant between The Foxglove and The Regent. Alternatively, The Regent serves two pizzas for £14 on Mondays and Tuesdays. The Regent also has an unusual item on the desert menu which will be sampled at some point. Specifically, Nutella and banana calzone. The Regent also has a comedy open mic night in its (rather small) basement. Previous attendees include Rob Beckett, Sara Pascoe and dozens of other professional comedians. So, if you get bored with our geek conversation, there are other options within the immediate vicinity.
We will possibly discuss:-
Any reference to the UK leaving the European Union, colloquially known as Brexit, will initiate a game of Mornington Crescent or reference to Brawndo. Political comedy is two doors along and in the basement.
(For previous text about vehicles, see Part 1, Part 2 and Goodwill. Also see index of linked ideas.)
For more than 10 years, I've considered a pyramid car design. This impractical design is deprecated. However, since Nov 2018, I have been gathering requirements for a more practical implementation. The result is an angular, squashed hexagon cylinder with a truncated front. This text explains some of the design decisions and explains how you can make a remote control scale model of the design.
Enumeration Of Constraints
There are different arrangements for vehicles:-
Personal transportation. In the most extreme case, this includes the Sinclair C5 (and the planned C10 and C20 which look astoundingly similar to the Smart ForTwo and ForFour). It also includes one seat, three wheel bubble cars with a door at the front and no reverse gear.
Three wheel vehicles with one motor wheel and two turning wheels - or two motor wheels and one turning wheel. Designs with less than four wheels are regarded as notoriously cheap and unreliable.
Sub-compact vehicle. Typically has four wheels, two doors and four seats. Chassis may egg shape; similar to a bubble car. Examples include Fiat 500 and Morris Mini (which shares parts with a Morris Minor and common Dalak).
Sedan/hatchback/station-wagon with engine at front, four wheels, 2-5 doors, 4-7 seats. Often has 60:40 split rear seats which fold down for extra storage.
Variations include Jeep (not derived from Just Enough Essential Parts or similar), SUV [Sport Utility Vehicle] and cross-over.
Mid-engine vehicles. Typically have token storage at front, two seats and engine ahead of rear axle. Provides best balance and handling.
Variations include race cars (Formula 1, IndyCar) and cars inspired by race cars (McLaren F1, Ferrari Enzo). Template is an aerodynamic teardrop cockpit with swept back containing mid-engine. Most extreme example is Koenigsegg Agera RS1, a US$10 million vehicle with wrap-around glass cockpit where windscreen wiper defaults to vertical position to reduce wind resistance. I quite like the aesthetic and vertical integration of Koenigsegg with the exception that the electronics are twee and the exploded view (courtesy of car reviewer, Doug DeMuro) looks like a glass teardrop on a go-kart chassis.
Another example is the Bugatti Veyron which is a teardrop swept mid-engine has difficulty sourcing air for the engine intake and 11 cooling radiators. The lesser known successor, the Bugatti Chiron (see Doug DeMuro's review), is an acquired taste where the interior and exterior has a dividing line like a fin of a sea monster. Other variants, such as the Porsche 918 Sypder, appear to take inspiration from various iterations of the BatMobile. The blandest example of the practice is the 10th anniversary Audi R8 Decennium (see Doug DeMuro's review) which looks like a 1990s Volkswagen Beetle modded to look like a BatMobile. (As Alfred the Butler might say: "Will you be wanting the BatPod, sir? ... The Lamborghini then. Much more subtle.")
Variation has engine behind rear axle; popularized by Ferdinand Porsche for teh real Nazis.
My deprecated square based pyramid design was an attempt to economically compete with sub-compact/sedan/station-wagon designs:-
In the smallest version, I considered using the tip of the pyramid for storage. In the smallest two versions, I considered one door at rear. In the largest two versions, I strongly considered 19 inch racking between the engine and passengers. However, beyond that, I hadn't really considered the practicalities of steering, doors, windows - or even changing a tire.
I abandoned the design after reading about a failed prototype car. The gas turbine car featured in the television series, UFO was a commercial failure because it was introduced around the time of the 1973 oil crisis. Furthermore, one of the actors from UFO, Ed Bishop, who portrayed Colonel Edward Straker, often bumped his head against the sloped glass of the door window. This made me very concerned that flat glass windows were impractical and/or dangerous. I'd like to make a very safe design and too many vehicle manufacturers knowingly compromise safety. However, a "safe" chassis is worthless if occupants are more likely to be injured or killed by glass. I want a safe design which can be manufactured at home; possibly as a single unit for myself. One of the many barriers to vehicle manufacture is automotive glass. It is generally curved and designed for a specific vehicle. Sourcing such glass constrains a design until it is redundant, dangerous or tied to the obsolescence of the intended vehicle. I wrongly assumed that the era flat glass in vehicles was over until I began watching car reviews.
From Nov 2018 to Dec 2019, I watched more than 40 hours of Doug DeMuro's car reviews. These are generally 20-30 minutes and are quite detailed. Each video covers atypical placement of handles and switches - and even strange menu items in car computers. Doug also has 10 category DougScore™ which is summed as a percentile and collated as a spreadsheet. From this, I have iteratively gathered design constraints, sketched an outline design, gathered more constraints and refined the design. I settled on a mid-engine design for very simple reasons. Designs with engines at the front have been explored extensively and there is a glut of examples to purchase. Indeed, I'd be arrogant to compete with these designs. However, mid-engine and rear-engine designs are significantly more sparse and generally more expensive. At the very least, this could improve the margins of a commercial venture. We are also at the twilight of internal combustion engines. With the proliferation of high power semi-conductors, re-chargeable batteries and cameras (often mandated by regulations), the conventional arrangement of engine-block-at-front is open to review. This is especially true if a vehicle doesn't have an ICE [Internal Combustion Engine].
Designing a mid-engine vehicle is relatively unconstrained. Fortunately, the process is greatly simplified by copying previous designs, as is common throughout the industry. The process can be summarized as:-
Toyota's MR2 is an almost a perfect car. It is easy to drive and easy to park. However, the dinky two seat mid-engine design relies upon a Wankel engine. Historically, the extended path of the triangular pistons required frequent engine re-builds. Alloys may have improved but a Wankel engine remains an exotic choice to the extent that Doug DeMuro reviewed 99 cars in 2017 and only one obsolete model had a Wankel engine (a Mazda RX-7, see here). Would prefer to allocate space to more conventional choices. This does not exclude use of a Wankel engine.
McLaren's F1 has a center driving position. Admittedly, this involves swapping the conventional number of rear-view mirrors and sun visors. Surprised this is not more common because the additional seat incurs minimal additional length. Placement of doors suitable for driver and passengers means that passengers are recessed from doors. This allows better crash protection because passengers may be fully shielded from side impacts by fixed beams. Likewise, the driver is always half a seat away from a side impact. Perhaps it is not popular because the official US version involved "deletion of side seats" to obtain safety certification. The US version of the McLaren F1 is therefore single seat, driver only.
General observation is that flat glass was common in the 1950s but also that designs become more rounded over time:-
Even within Lamborghini designs, there is a distinct progression of style:-
Vehicle bodies tend to get more rounded over time. Compare:-
Therefore, iterations of a blocky design become more rounded over time. However, design elements of a Lamborghini Countach are of particular interest. Is it possible to make something like a Countach but more practical, with the footprint of a commodity sedan without specialist car industry CAD software? Oh, definitely yes.
It is possible to make a flat glass, irregular pentagon/hexagon windscreen with a single recessed wiper which sweeps the vast majority of the windscreen in 90° or so. The steering column and wiper motor assembly is also compatible with two columns of 19 inch racking. In addition to Lamborghini's deprecated pentagonal windscreen, it is possible to retain the extremely popular wedge shape of a super-car. Although it is possible to have storage at front - and some cars have the ridiculous constraint that it should be possible to carry one full set of golf clubs - it is possible to abandon this constraint entirely and have a front access hatch which is primarily for rack cable access, fuses and air conditioning.
The vehicle is primarily intended to be four wheel drive diesel electric. Over a sufficient number of wheels and a sufficient number of motor phases, 5kW MOSFET or GaNFET transistors are more than sufficient to make a small vehicle accelerate fast. Regardless, I want redundant engine mounts so that it is possible to fit a Chevrolet Straight-6 engine, other common engine and (if internal combustion engines become illegal) retro-fit as fully electric. It is intended that the chassis can be retained for 30 years while the engine and electronics are re-fitted every 10 years or so.
A hill-billy home-brew SUV F1 MR2 Countach may look ugly. If the external shell of a hypothetical Countach van (without swept back) is stretched to the height, width and (almost) length of a Ford Focus, it'll look midway between a street cleaning machine and a super-car. However, with the exception of seating and public liability insurance, it'll moderately out-perform a Ford Focus on safety, depreciation, fuel efficiency, range, speed, acceleration, ease of repair and entertainment options while steadfastly not looking like a family sedan.
Home-brew construction requires tubular steel. It is possible to make a tube chassis from vanadium steel, aluminum, titanium, magnesium or other material. However, use of stainless steel is assumed for ease of construction and repair. Unfortunately, a bonded chassis has become more popular than a tube chassis due to cost (and repeat sales). This has occurred due to the following progression:-
Compounding this arrangement:-
The way out of this trap is to use a different construction technique: tubular steel. It has the benefits of corrugated steel, is easy to repair and avoids premature corrosion. So, how should the tubes be arranged? The deprecated pyramid design has uninterrupted tubes which run from front to back and from side to side. The remainder is the minimum number of triangles to hold them all apart. This provides integrity for front impact, side impact, rear impact and roll-over. However, placement of the wheels is unspecified. I would like to retain bars or tubes; especially lengthwise. A more moderate design places wheels beyond two parallel tubes. This provides protection for driver in a central position but does not provide protection for passengers. I considered various arrangements including squares, triangles and even a squashed dodecahedron inspired by Richard Buckminster Fuller. Perhaps this is not the best inspiration, given RBF's failed, fatal vehicle venture. Regardless, the windscreen is a relatively weak open pentagon and I hoped to increase its integrity by constraining everything around it. One fruitless avenue was 12 pentagons where 11 of them are each split into three rigid triangles.
The proposed solution is a squashed hexagonal cylinder with one end sliced off to make an angled, regular hexagon for a super-car style wedge front. The regular hexagon at the front is sub-divided with cross tubes into an irregular pentagon windscreen (like the outline of the SuperMan logo), front light clusters and an access hatch for electronics. The wheels may be placed under the corresponding tapering hexagonal body. Between the axles, the chassis may be extended to a rectangular bottom to accommodate two passengers who flank the driver's center position. However, this does not explain how passenger safety is maintained to a standard which exceeds a McLaren F1. A cross section of the hexagon cylinder between the engine compartment and the cabin is a mesh of tubes. This acts like the roll cage in, for example, a Porsche 911 GT3RS 4.0 except that it is integral part of the base design rather than a retro-fitted cage. The cabin's rear wall and floor are both a mesh of triangles. The doors have side impact bars which lock into the side triangles. Passengers are protected on four sides of a hypothetical cube and are sufficiently protected if a vehicle rolls. The driver is a different case and the driver's feet (near the very front of the vehicle and between the front wheels) are particularly vulnerable to front impact and side impact injury. In particular, should avoid type injury incurred by Alex Zanardi. In these cases, the broader wedge with two wheel wells (and reduced speeds) permit a design which buckles downwards and avoids crushing legs between metal.
This is not a full enumeration of mass market safety tests applied to all occupants. It certainly doesn't exceed specification. However, this is sufficient to iterate a design to a scale model prototype. Before making a model, an outline idea of doors and windows is particularly useful.
It is possible to affix doors in a variety of configurations. I prefer doors with vertical hinges and three stop positions. However, it is becoming common for car doors to work like laptop hinges and stop at any position. Also possible to have near vertical hinges to avoid curb scrape. This may be combined with "puddle lights" which shine a logo on the ground and the new trend is to shine a color logo. Possible to have a horizontal hinge and gull wing doors, although this is quite impractical. A feature of the Toyota Sera design (which was copied in the McLaren F1 design) is doors fixed at the side and top of the windscreen to make a more practical version of gull wing doors. Other variations for doors include the highly impractical scissor doors; commonly found on a 1980s Lamborghini. Koenigsegg has a more practical version of scissor doors where the doors push out and spiral upwards. This has the exoticism of scissor doors and the practicality of a sedan. I presume this feature is patented and not available for license. Further options with doors include the downward motorized doors of a BMW Z1 and the pointless over-hang of a T-top Ford GT.
Safety features on doors may include rear reflectors and/or warning lights. Some vehicles also have a cross hand latch which encourages an occupant to avoid door collision with an approaching cyclist. This Dutch reach-around technique (repeatedly featured in London's Evening Standard) emphasizes the importance of industrial design. This practice, enforced by latch design, could save an estimated eight lives per year in the Netherlands - none of which are connected the driver or occupants.
Doors and roof may be partially or fully glass. This looks exotic on a McLaren Senna (see Doug's review) and aids parking. However, it may encourage comparison to a street cleaning vehicle. The roof may be fitted with liquid crystal to obscure sun. (See Doug's fascination with this feature.) Polarizing planes around the liquid crystal may be chosen such that the roof defaults to clear or opaque when electricity is not applied. Like a Lamborghini Countach (see Doug's review), flat angled glass restricts lowering a window fully, if at all. However, like a Lamborghini Espada, it may be possible to open a pane with a lever.
The proposed design is not intended to be made as an open top model. However, the design has space for eight or more steel tubes which may run the full or partial length of the vehicle. Therefore, it is trivial to stiffen the floor of the chassis. Similar opportunity exists to stiffen the windscreen frame. Actually, a sneaky ruse played by multiple super-car manufacturers is as follows:-
Anyhow, we confirm that there are absolutely no plans to make an open top model despite changes to the blue-prints being trivial. Likewise for a gull-wing T-top. It definitely won't happen.
I semi-jest about Audi, BMW, Porsche and Volkswagen designs all being interchangeable (and Porsche and BMW drivers taking 50% more priority because they paid 50% more for their interchangeable sedans). However, there is one marketing gimmick definitely worth copying from Audi R8: each model is only available in one color. I therefore propose the following colors:-
Actually, I might save British Racing Green for a version with the best feature. By using the same materials and components (but in different quantities), it is possible to greatly improve traction by adding extra wheels. The most obvious method is to make a longer vehicle with an extra axle. However, it is also possible to remove passenger seats and add another axle behind the doors. This has several advantages. An all wheel drive vehicle has approximately the same weight but with six motors each powering a wheel. Traction is improved by half. For safety certification and insurance, the engine is unchanged, vehicle length is unchanged, wheelbase is unchanged and side impact handling is effectively unchanged for the driver. It is also very distinctive. Six wheel vehicles are relatively rare. Examples include the six wheel Land Rover design and the Formula 1 Tyrrell P34 which raced in 1976 and used four small front wheels in an attempt to improve aerodynamics.
A single seat design and a lengthened design remain compatible options. Therefore, it is possible to have more than six wheels to further improve power delivery. If the design is particularly fast, may want to stagger wheels like the BloodHound with the intention of reducing shocks. As a marketing gimmick, may want proposed design painted green, with rotary street cleaning brushes, billed as The World's Fastest Street Cleaning Machine.
Requirements
Mandatory requirements include:-
Essential requirements:-
Disability access:-
Desirable requirements:-
Optional safety/anti-theft features:-
Less important options:-
The interior of the proposed design similar to the interior of a Boeing CH-47F Chinook (or Vector W8 "aeromotive" design, see Doug DeMuro's review) with rack mount displays and two columns of switch panels. Difference to Chinook design is additional rack for displays and the two columns of switch panels separated by a middle seat. The panels are intended to be multiples of 4 inch × 2 inch (10cm×5cm) and be of the following types:-
Some features were not relevant to this design but may of particular note for competing designs which incorporate a tailgate. Specifically, ensure that the design has a extensible rotating cage like a Subaru Baja and seats and cup holders like a Pontiac Aztek while conforming to regulations about lights and license plate.
Optional packages include:-
The floor to ceiling cuboid of space behind the driver and between the passengers is a miscellaneous volume which may vary significantly:-
Target Customers
I assumed this would be a niche product (possibly for a market of one) where the requirements are for a safe, configurable vehicle where styling is secondary. Specifically:-
Friends suggest a utility vehicle to extend a person's ability by 10×10×10. Specifically, to allow one person to transport 10 times what they can carry, 10 times further than they can walk at 10 times the speed. This would cover light industrial tradesperson (carpenter, plumber, electrician, cable layer), fast food delivery or mobile beautician. However, traditional vehicles are sufficient for these rôles and low-range, fully electric vehicles already replace them.
After considering the performance version, people considering a premium vehicle may consider a more distinctive vehicle at a reduced price. However, after considering specification in more detail, realize that it is the modern equivalent of a Citroën 2CV. I believe that the use case for this vehicle was to climb a 45° dirt track and allow a farmer to get eggs to market. That explains much about this quirky (but easily maintained) vehicle including the soft suspension. Unfortunately, a 2CV doesn't meet modern safety regulations and has ceased manufacture. A safer design with similar constraints would be ideal in rural France. Indeed, with the social unrest in France, Iraq, Chile and elsewhere triggered by fuel costs (distinct from political unrest in Portland, Germany, Hong Kong and elsewhere), mass appeal may be outside of the relatively calm Anglosphere. The French Government planned to ban diesel vehicles by 2030. This is particularly onerous when the French are heavily taxed and regulated. They can barely afford their current vehicles. The proposed design provides a migration path from the current vehicle to fully electric, at the least cost, over an indeterminate time-scale:-
Many steps of this migration are difficult, time consuming and resource intensive. In particular, it requires expertise in diverse skills including welding, electronics and possibly writing software to handle network protocols. The latter skill is required to migrate a modern engine management computer and maintain vehicle emission standards. This requires work which may be specific to each year of each model of each manufacturer.
For customers who depend upon private transport, there is considerable scope to offer products and services up to and including delivering a roadworthy vehicle to customer specification on 100% credit. More modestly, it is possible to offer blue-prints and software license for US$100, redeemable against products and services. For example, a welded chassis, etched circuit boards, flashed micro-controllers, cut glass, wound motors, batteries, power control boards, ECU bridging and re-conditioned engines. There is also considerable scope to offer upgrades to customers. The most significant upgrade is migration from internal combustion engine to fully electric. At present, mass market adoption of electric vehicles is not possible due to:-
An electric vehicle is not currently economic for the average person, even with 100% interest free credit. However, the *option* to migrate would build a significant number of repeat customers. The popular (and sometimes violent) opinion is that migration by 2030 is optimistic. Therefore, this business is a very long-term proposition over 30 years or more. This requires:-
This is manufacturing business which is:-
Thankfully:-
One of the most trusted competitors is unprofitable and likely to cease trading.
Independent resellers are trusted less than manufacturers. However, we offer a long-term solution and therefore exposure to independent resellers may be restricted to annual safety checks.
Increasingly large vehicles are buoyed by predatory lending practices. Henry Ford distinguished between luxury products and commodity products. If employees cannot afford a product then the product is not a mass market product. From this, Ford instigated purchase plan where it was possible for an employee obtain a car on credit and own it outright after 47 weeks of labor. This was the norm for many decades and vehicle loans became the most profitable part of Ford. Unfortunately, we now have the situation where 8-10 year compound interest loans are offered on vehicles with a median lifespan of 7 years or less. Furthermore, lenders are willing to roll debt into subsequent a vehicle loan, for a fee, because it is preferable to maintain security on the debt, even if it is negative equity.
When fuel is unaffordable (and French protesting in the streets every weekend for a year firmly indicates such a position), it is small modest cars which became international favorites and, in some cases, maintain profitable sales for 30 years or more. Although we are currently in a oil glut which encourages large, heavy vehicles, this only endures while producers accept payment. After the US Dollar was removed from the gold standard in 1971, the subsequent stock market crash led to an oil crisis:-
The oil crisis pushed West European car buyers away from larger, less economical cars. The most notable result of this transition was the rise in popularity of compact hatchbacks. The only notable small hatchbacks built in Western Europe before the oil crisis were the Peugeot 104, Renault 5 and Fiat 127. By the end of the decade, the market had expanded with the introduction of the Ford Fiesta, Opel Kadett (sold as the Vauxhall Astra in Great Britain), Chrysler Sunbeam and Citroën Visa.
Buyers looking for larger cars were increasingly drawn to medium-sized hatchbacks. Virtually unknown in Europe in 1973, by the end of the decade they were gradually replacing sedans as the mainstay of this sector. Between 1973 and 1980, medium-sized hatchbacks were launched across Europe: the Chrysler/Simca Horizon, Fiat Ritmo (Strada in the UK), Ford Escort MK3, Renault 14, Volvo 340/360, Opel Kadett, and Volkswagen Golf.
In the US:-
Before the energy crisis, large, heavy, and powerful cars were popular. By 1971, the standard engine in a Chevrolet Caprice was a 400-cubic inch (6.5 liter) V8. The wheelbase of this car was 121.5 inches (3,090 mm), and Motor Trend's 1972 road test of the similar Chevrolet Impala achieved no more than 15 highway miles per gallon. In the 15 years prior to the 1973 oil crisis, gasoline prices in the U.S. had lagged well behind inflation.
The crisis reduced the demand for large cars. Japanese imports, primarily the Toyota Corona, the Toyota Corolla, the Datsun B210, the Datsun 510, the Honda Civic, the Mitsubishi Galant (a captive import from Chrysler sold as the Dodge Colt), the Subaru DL, and later the Honda Accord all had four cylinder engines that were more fuel efficient than the typical American V8 and six cylinder engines. Japanese imports became mass-market leaders with unibody construction and front-wheel drive, which became de facto standards.
From Europe, the Volkswagen Beetle, the Volkswagen Fastback, the Renault 8, the Renault LeCar, and the Fiat Brava were successful. Detroit responded with the Ford Pinto, the Ford Maverick, the Chevrolet Vega, the Chevrolet Nova, the Plymouth Valiant and the Plymouth Volaré. American Motors sold its homegrown Gremlin, Hornet and Pacer models.
Some buyers lamented the small size of the first Japanese compacts, and both Toyota and Nissan (then known as Datsun) introduced larger cars such as the Toyota Corona Mark II, the Toyota Cressida, the Mazda 616 and Datsun 810, which added passenger space and amenities such as air conditioning, power steering, AM-FM radios, and even power windows and central locking without increasing the price of the vehicle. A decade after the 1973 oil crisis, Honda, Toyota and Nissan, affected by the 1981 voluntary export restraints, opened US assembly plants and established their luxury divisions (Acura, Lexus and Infiniti, respectively) to distinguish themselves from their mass-market brands.
Compact trucks were introduced, such as the Toyota Hilux and the Datsun Truck, followed by the Mazda Truck (sold as the Ford Courier), and the Isuzu-built Chevrolet LUV. Mitsubishi rebranded its Forte as the Dodge D-50 a few years after the oil crisis. Mazda, Mitsubishi and Isuzu had joint partnerships with Ford, Chrysler, and GM, respectively. Later, the American makers introduced their domestic replacements (Ford Ranger, Dodge Dakota and the Chevrolet S10/GMC S-15), ending their captive import policy.
An increase in imported cars into North America forced General Motors, Ford and Chrysler to introduce smaller and fuel-efficient models for domestic sales. The Dodge Omni/Plymouth Horizon from Chrysler, the Ford Fiesta and the Chevrolet Chevette all had four-cylinder engines and room for at least four passengers by the late 1970s. By 1985, the average American vehicle moved 17.4 miles per gallon, compared to 13.5 in 1970. The improvements stayed, even though the price of a barrel of oil remained constant at $12 from 1974 to 1979. Sales of large sedans for most makes (except Chrysler products) recovered within two model years of the 1973 crisis. The Cadillac DeVille and Fleetwood, Buick Electra, Oldsmobile 98, Lincoln Continental, Mercury Marquis, and various other luxury oriented sedans became popular again in the mid-1970s. The only full-size models that did not recover were lower price models such as the Chevrolet Bel Air and Ford Galaxie 500. Slightly smaller models such as the Oldsmobile Cutlass, Chevrolet Monte Carlo, Ford Thunderbird and various others sold well.
Economical imports succeeded alongside heavy, expensive vehicles. In 1976, Toyota sold 346,920 cars (average weight around 2,100 lbs), while Cadillac sold 309,139 cars (average weight around 5,000 lbs).
Federal safety standards, such as NHTSA Federal Motor Vehicle Safety Standard 215 (pertaining to safety bumpers), and compacts like the 1974 Mustang I were a prelude to the DOT "downsize" revision of vehicle categories. By 1977, GM's full-sized cars reflected the crisis. By 1979, virtually all "full-size" American cars had shrunk, featuring smaller engines and smaller outside dimensions. Chrysler ended production of their full-sized luxury sedans at the end of the 1981 model year, moving instead to a full front-wheel drive lineup for 1982 (except for the M-body Dodge Diplomat/Plymouth Gran Fury and Chrysler New Yorker Fifth Avenue sedans).
Comparison To Tesla Motors CyberTruck
I'm not a fan of Elon Musk's ventures:-
Elon Musk is also a magnet for controversy:-
Regarding vehicles, the progression has been:-
Naming products in the style of Henry Ford is sensible. However, naming them SX3Y (or BFR) is puerile.
Regarding the Real Model Y, I have doodled in a similar space and I am definitely unimpressed with Tesla Motors. I stated on Tue 9 Oct 2018 that it was possible to avoid car industry CAD software but the design would look angular; like an F-111 Aardvark or the USS Zumwalt. I have worked further in this area. Despite that, I was not prepared for the style of the Real Model Y. My first impression was "J*sus, F**k! That's uglier than my design!" Please note, that's from someone planning to sell an angular, unpainted car with bullet-proof glass option and - due to small volume production side-stepping crash testing - may have a product ready before the Real Model Y. And you know the Real Model Y is an ugly vehicle when a competitor jokingly compares its *own* design to a street cleaning vehicle and still thinks it has the better design.
The major influences of the Real Model Y appear to be a Lamborghini Countach and a Hummer. More specifically, a Countach Hummer Zumwalt DeLorean iPhone. Perhaps with a hint of Atari ST mouse. (Does anyone remember the Atari ST mouse? Or, more pertinently, does anyone *fondly* remember the Atari ST mouse? I quite liked it but I'm definitely in the minority.)
Admittedly, if I placed more emphasis on passenger seats, cruise speed and smart-phone aesthetics, my design would be much closer to Tesla Motors' most recent effort. However, I specifically rejected the swept back. It increases efficiency at speed but it is also possible to gain significant volume with very little metal. I also rejected the upward sweeping line. It is supposed to make a vehicle look like an angry, aggressive, road hugging wedge. However, it reminds me of:-
A famous mantra is that "form follows function." Unfortunately, the main function of an upward sweeping line is to signal idiocy and, more recently, aggressive idiocy.
Onto more substantive matters, the Real Model Y is only offered with one, two or three electric motors. Each improvement to the specification is accompanied with a price increment of US$10,000 or US$20,000. However, the all-terrain vehicle is not offered with one motor per wheel option. This greatly compromises acceleration and braking.
Elon Musk claims zero market research prior to launching the Real Model Y and this is quite believeable. The failure of bullet-proof glass was the most notable clanger of the Real Model Y's launch. Worryingly, I may have researched bullet-proof glass more thoroughly than Elon Musk and my findings are quite concise. Bullet-proof glass in retail environments is typically recessed from direct sunlight because ultra-violet light may cause the glass to de-laminate and become counter-productive and/or dangerous. The window frame of a moving vehicle is a significantly more demanding environment and Tesla has failed to specify suitable material in the most high profile instance.
Prototyping
People may find fault with my development process but I have made considerable progress making a 1:10 scale model and then progressing to a 1:4 model (which is suitable for three Build-A-Bear occupants). My process is to doodle in a vector drawing package then tape a tube chassis from rolled paper and then repeat the process after obvious weaknesses have been eliminated. The most beneficial part of this process is that crash testing is trivial. Merely hold the paper model at the desired angle and drop it onto floor. It generally makes a noise like a wicker basket being dropped. More importantly, it is a quick method to diagnose faults.
The most objectionable part of my process is that I have been doodling in OpenOffice. People may suggest numerous CAD programs but OpenOffice files are relatively trivial to share. Additionally, it fits into my workflow for business plans and video presentations. Anyhow, onto physical prototyping. I considered using old copies of London's Evening Standard as a substitute for structural steel. Indeed, this is common practice in London where, for example, the partial collapse and subsequent demolition of Ronan Point Tower Block was due to similar substitution. However, it is more generally applicable to use A4 (or US Letter) paper rolled into 30cm (or 11 inch) batons. With moderate practice, it is possible to roll paper into tight tubes with fairly consistent diameter. The actual diameter doesn't matter too much - nor does the consistency. For spans longer than any piece of paper, the obvious technique is to tape two or more pieces of paper together before rolling. However, a superior technique is to roll individual batons then tape them lengthwise before splinting them with a second layer of paper rolled around the first. The wider tube can be made on demand, cut to length without waste and has the appearance of bamboo. It is sufficiently strong despite the inner tube sliding freely within the outer tube. Alternatively, use drinking straws as the construction material. (Or perhaps K'Nex which was inspired by the concatenation of drinking straws by two drunken brothers at a wedding.)
The 1:10 scale prototype was made without wheels. It has a foot-print slightly larger than a pad of paper. It is approximately 32cm×22cm×15cm - and it looks *much* better than expected. It looks like a StarTrek Next Generation run-about shuttle-craft except with symmetrical bulges for side impact safety. The regular hexagon front window at 45° looks particularly good. Actually, a hexagonal white frame could plausibly be a prop from StarWars, Silent Running or Battlestar Galactica. It has a late 1970s futuristic vibe influenced by the mid 1970s Lamborghini Countach design.
Some observations from physical modeling:-
1:4 scale is extremely practical because the result fits through an ordinary doorway.
Trivial to convert scale model into remote control car. Indeed, this reduces cost and would be a particularly good educational gift.
It is easy to complain about the arrangement of a humble sedan but it is quite impressive to fit an engine block, five adults and one large suitcase into 10m3 box (or less), made from a modest amount of metal, self-propelled for several hours at a rate much faster than sprinting speed. It is a small miracle which is deemed unremarkable. Unfortunately, manufacturers treat sales as final - even when there is a duty of care to inform customers about fatal flaws.
It is manageable (but not trivial) to design a vehicle as an irregular polyhedron with 60 vertexes or less.
It is trivial to design wheel well size and position but it is significantly more difficult to design a safe and responsive steering column. Doug DeMuro has repeatedly mentioned a dead zone when driving old cars (and a lack of dead zone in new cars). Even for non-drivers, this may be familiar concept when playing driving games, such as Sega Rally or Scud Race which has a widening dead zone as a penalty for minor collisions. In the real world, it is widely known that steering is not designed with a perfect turning circle and is instead biased with deliberate under-steer/over-steer to compensate for rear/front wheel drive. This may extend to bias in power steering. In particular, CANBus power steering may be non-linear to reduce the apparent slack in steering. This leads to the question about automatic trim of the steering potentiometer. Presumably, dead center is derived from the weighted average of the steering position over a sufficiently large distance. Even for people who enjoy driving in circles for a considerable period, obvious out-liers can be excluded. Regarding the universal joints in the steering column:-
Physical model indicates that front to back balance is excellent.
Building architects are a bunch of weenies. Especially the ones who design unimaginative glass and metal buildings. I've been making white paper models, like architects. Except my design is weight optimized, on wheels and has to survive collision with similar designs.
Numerous problems can be solved by making a longer vehicle. For example, and five or six seat version can be made by duplicating the rear of the cabin. Passengers in the last row would sit between two bulk heads. A more unusual arrangement would be six wheels with four seats in a diamond configuration.
The short six wheel configuration is definitely incompatible with passenger seats. However, the reason is divided into two cases. In a long, flat sport-car, seat would intersect wheel well. In a tall, short configuration, seat would be above wheel well. However, the removal of side impact tubes mean this arrangement works until it doesn't. This is an unacceptable compromise to passenger safety.
Open front, optional glass in lower doors and optional sun roof creates astounding visibility. It is like a motorized green-house. Ignoring windscreen wipers and tachometer, it is possible to have a giant hexagonal window with visibility to the floor. However, probably wouldn't want to drive this with a short skirt and definitely shouldn't drive it while wearing a kilt in the traditional manner! More seriously, front has to be divided but if it is divided as a door, it is possible to provide access for three wheel-chairs. Multiple electric wheel-chairs can be charged while driving.
Physical model has given me a good appreciation of the mass production crash impact tests. They only cover common cases and should definitely be exceeded. Current iteration of model likely to pass full front impact, side impact and rear impact. Partial front impact is an awkward case but many designs which passed full front impact caused catastrophic injury in this less common case. Current design likely to fail roll test. Sufficiently securing the upper windscreen is relatively difficult.
It is relatively easy to design a cabin as a convex hull which resists deformation leading to catastrophic injury. Major problem is that force at one point may deform in a beneficial manner but force of the same magnitude and direction at the next vertex may crush occupants to death. Graze impact is quite insidious. First iteration was very fragile at front to the extent that clipped front corner led to buckled chassis and possible write-off. Obviously, this is now resolved. However, graze impact can occur from anywhere, including clipping an open lorry door at top of chassis. (From personal experience, a Volvo 345 handles this case surprisingly poorly.)
Easy to design vehicle where eight tubes meet at one point. Unfortunately, this is difficult to weld (or repair) with certainty. May require cast iron joints and rivets. If the joints are designed to be symmetrical, some spurs may be used for seat belt fixings. However, it may be preferable to avoid an accumulation of tubes at one point. An early design, from the side view had a radial arrangement of horizontal, vertical and diagonal bars welded to the central cross impact bar. It was possible to change the upward diagonal across the storage compartment to a downward diagonal. This also provides a hatchback option. ("Form follows function.") A further tube can be removed from this knot by hinging doors to open forward. This changes the diagonal sweep of side impact protection so that it does not affix to center. Moderate change is to set constraint that no more than six tubes may meet at a point.
Engine compartment may be 90cm×90cm×50cm or thereabouts. This is sufficient for:-
Diesel electric allows transverse installation of any engine because mechanical coupling is not required. It is also possible to install multiple engines; adjacent or stacked. In the most extreme configuration, may have three or more engines working simultaneously.
Not in original specification but front wheel drive engine may be installed with mechanical coupling in mid-engine vehicle. In this arrangement, it is possible to have only rear wheel steering and stiffened front for additional impact protection.
Possible to have rear view and pillars similar to Toyota MR2 or Jaguar XJS. This option significantly improves access to engine but eliminates trunk.
Air intake for mid-engine design is a traditional problem. Possible to take air from underside of vehicle. However, this creates a virtual cushion of air which reduces handling - and requires energy to suck the air under. Design has space for air intake ahead of rear wheels. Unfortunately, intake will be triangular and may therefore require a pyramid number of small, relatively ineffective fans. Side grilles also have similar appearance to diesel electric train.
It is possible to fit a heavy engine sideways, towards to the cabin. It is also possible to fit a light engine at rear. Both options maintain excellent front to back balance. Indeed, it may be desirable to have tapering mount points to maintain engine balance. Tapering at base of rear compartment is also structurally compatible with a tow fixing.
Possible to make matching trailer (or camper or caravan) which is half height, full height or compliments hatchback. In the latter case, air which sweeps down from a hatchback then sweeps up the matching trailer.
Matching trailer requires two or four matching rear light clusters, rear view color/infra-red camera (possibly a ninth channel), parking distance sensor and optional strip lighting. Maintaining upward and downward compatibility with the dumb seven pin connector is also useful.
Matching trailer may also have brakes. However, if the brakes are fail-safe, it may be difficult to move an unattached trailer.
Light clusters are all within elongated right-angle triangles. Where the rear is a squashed hexagon, it is possible vertically mirror the light clusters for a distinctive look by duplicating the symmetrical parts. Door on front for disabled access may require re-location of the front light clusters. However, these can be affixed over front wheels and are also elongated right-angle triangles, although smaller. Side turn indicators can be mounted inside glass panels such that the remainder of the panel aids parking.
Optional storage area is above engine. Therefore, have problem of thermal separation. This is a matter of conduction, convection and radiation. The shelf is a suspended structure and it is possible to line the underside with a reflector to reduce radiation and insulating foam to reduce convection. This fulfills the use case of "eggs to market" but a 10 hour drive would create hard-boiled eggs. Conduction of heat through the frame is the most awkward part of the problem. Many super-cars have the related problem that wind resistance over the front compartment (hood? trunk? trood? hunk?) is sufficient to cook an egg. Perhaps there is no good solution.
Open tubing may be suitable to prevent heat conduction to storage compartment but this may create a water trap and premature corrosion. Open tube at windscreen would allow air to pass through door roll bar and continue to rear before exit. This may prevent upward conduction of heat from engine to storage compartment but may also incur relatively high thermal stress at welding weak point. Solution may require a full scale ball-and-stick construction but with cast iron joints, rivets and ducting through or around the joint.
Tubing may contain two, four or six pulse jet engines for turbo boost. Aardvark's Doodlebug pulse jet design is suitable with the exception that a heart valve design avoids the requirement for a flared air intake section. This is suitable to periodically accelerate and safely negotiate traffic at speed. However, it is not suitable for sustained use due to the clanking pulse jet noise which would propagate through the chassis. Remember to include pulse jet thrust in brake caliper calculations!
I may have discovered why a center driving position is so rare (in addition to "deletion of side seats"). The more common off-center driving position typically uses two universal joints to a central rack and pinion. For crash protection, the middle rod works as a lever allowing the steering column to collapse into a Z shape. The universal joints fail to rotate in this collapsed arrangement but the effect is to withdraw the steering wheel from the driver prior to imminent deformation of the cabin. The may cause permanent injury to hands but it is preferable to permanent brain injury or death. A center driving position requires an off-center rack and pinion to re-create this safety arrangement. In a front-engine design, this would be placed under the engine. In a mid-engine design, the driver is further forward and there is less space to connect a steering wheel to a pinion. A vertical off-center steering axle (horizontal pinion) is possible but it requires a curved rack which correctly meets the pinion while taking into account the parallelogram which places each front wheel in the correct turning circle. Historically, that was awkward to calculate. However, it should be possible to parameterize such a shape in suitable CAD program. (This is why I doodle and make physical models. Suitable CAD software has yet to be selected.)
Possible to have open chassis with no side panels and no glass. Under UK regulations, a vehicle with three or more wheels does not require crash helmet. (See Ozzy Osbourne's antics for further details.) However, it remains desirable protect eyes with goggles or, more sensibly, a crash helmet and neck brace.
Under UK regulations, may be possible to have open engine compartment if covered with bird net sufficient to prevent gerfingerpoken and prevent struck pedestrians from encountering cam belt. Open rear may be restricted to upper one or three panels.
Hoped to have two columns of full depth 19 inch racking. However, this is not possible for two reasons. Firstly, center mount post and center drive position is a safety hazard. Secondly, full depth racking intersects front wheel wells. Fortunately, maximum width may be 175cm (69 inch) and three columns of 19 inch racking is only 57 inch (145cm) excluding mount posts. Therefore, it is possible to have one column of full depth racking flanked by two full columns of shallow racking suitable for buttons, displays and trivial equipment.
Previously assumed that two columns of 4U rack would be approximate limit. However, racking has to exceed radius of a possibly after-market steering wheel plus thickness of gloved hands plus a comfortable clearance. Therefore, three columns of 6U rack is minimum. 2U or more may be reserved for mandatory gauges but these do not have to be in the central (full depth) column.
Optional glove box(es) may hang from racking and bolt over or under rack panels.
The industrial designers of the cordless kettle suggest looking for design elements which remain unchanged. In addition to front-engine design, rectangular door window with clipped upper corner is also common to the extent that deviation looks aesthetically wrong. It appears to be a compromise between roll safety, visibility and material cost. Even if safety is the primary concern, it may not be beneficial to affix vertical pillar to top of windscreen because it reduces visibility and the longer, weaker pillar provides no sacrificial section of windscreen frame for roll safety.
I wrongly assumed that an open top model mostly required a chassis to be stiffened around the bottom plane. However, it is quite a feat to suspend a large sheet of glass at an angle such that it is unaffected by the pressure of wind at speed. Also, I assumed that, for example, the Bugatti Chiron is not available as an open top model because few people like the wind in their hair at Mach 0.7. However, the constraint appears to be technically difficult.
It is possible to "delete" the center seat and have a somewhat more conventional wedge, mid-engine vehicle with two seats. Optionally, this may be tall and short or low and long. For safety, it requires corresponding changes to the console and steering. Unfortunately, this requires deletion of full depth racking, deletion of half of the accessory panels and deletion of retro-fitted wheel-chair access. Anyone considering this arrangement should first consider purchase of a Toyota MR2.
While many cars exceed length of 4.4m, desirable to keep vehicle under 4m (13 foot) because it is possible to fit three such vehicles into an ISO 6346 40 foot (12m) shipping container. Especially desirable to keep vehicle under 3m (10 foot) because is is possible to fit four such vehicles into a 40 foot (12m) shipping container, two vehicles into a 20 foot (6m) shipping container or one vehicle into a 10 foot (3m) shipping container. Wedge front allows partially assembled vehicles to exceed 10 foot with some penalty. Cannot angle vehicles upwards and total height should be less than 7.4 foot (2.3m). These constraints can be relaxed if transport is restricted to over-size containers only.
For this design, solid tubes occupy approximately 2.6% of Length×Width×Height cuboid and hollow tubes occupy approximately 0.8% of the volume. As explained by a lead researcher of aerospace grade titanium, welds are weak points and therefore an extruded chassis is lighter and/or stronger. In this case, approximately 99% of a solid metal block would be discarded. Alternatively, processes (such as sand casting) would incur lesser waste at the expense of greater error. Extrusion allows more amenable shapes but otherwise provides minimal benefit. Additionally, solid block extrusion is prohibitively expensive and significantly more energy intensive. Subtractive manufacturing remains preferable for engine blocks, super-sonic aircraft and air-lock hatches. However, welded tubes provide the safest, affordable, repairable option for wheeled vehicles.
Further Development
A previous attempt to collaborate on a diesel electric vehicle made minimal progress because it required participants to purchase a similar model of roadworthy vehicle. This is obviously a resource intensive exercise which eliminates the vast majority of potential contributors. The current proposal reduces the initial resources to a pad of paper, some tape and a few hours to craft a scale vehicle chassis. At present, this does not require a common chassis design. Following a common template is encouraged because results are more certain. However, for anyone who wishes to deviate, the recommended steps are:-
Many tasks can run in parallel. For example:-
In particular:-
A commercial product can be sold with only two of these features because customers can upgrade after purchase. In some cases, upgrades may be a free firmware update. The most difficult tasks come last:-
Until this occurs, it is no loss and, indeed, only advantageous for people to prototype the suggested design (or competing designs) until commercial production can be established. On this basis:-
Eventually, I hope that a subset of collaborators (most likely determined by proximity) will formalize a commercial venture which sells everything from a "blue-prints and firmware" package to complete, roadworthy vehicles with long-term support for spares. You can participate and innovate in this process by building a 1:10 scale prototype from 20 sheets of paper and then a 1:4 scale prototype using proportionately more paper. (Approximately 100 sheets in total.) From here, it is possible to make a remote control car with rack servo control steering.
It is fun and educational to make scale models and significantly more fun to make your own remote control model. Initially, the electronics can be purchased. However, it is possible to swap with own parts as experience grows. This includes 3D printing motor frames and winding your own motor coils. Components using micro-controllers may use the same firmware as a real car; differing at most by the choice of numerical constants which account for scale.
This may provide you with nothing more than an affordable and overlooked method to make your own remote control car. However, it may segue into a due diligence exercise and vehicle maintenance training. Even if you intend to purchase an existing vehicle from an established competitor, I hope this helps you consider your purchase in more detail. In particular, I hope you re-consider or defer purchase given that many options are possible but currently unavailable.
Suggested design:-
begin 644 projection0-0-8pub.odg
M4$L#!!0```@(`"Z]GD^-D8T(O!0``,L/`0`+````8V]N=&5N="YX;6SM/=F2
MX\AQ[_X*!"4K+'L(XCY:VSVQDKVAAYZ58F>T4OAE`@V";.R"``6`?>C1[PX_
MZE/\`?X3?X"_P74`8!50``H7FV1C(H9LHK*`K+PJ*RLK\<W'EUT@/'EQXD?A
M[4(6I87@A6ZT]L/M[>)/7[Y;6HN/=__P3;39^*YWLX[<P\X+TZ4;A2GX%D#O
M,+G!K;>+0QS>1$[B)S>AL_.2F]2]B?9>F/>Z(:%OT+/PE21]#;B[(V"R=^J]
MI+R=(2S5UWG@?S(")GNO8^>9MS.$!40ENV\BWLXO2;#<1(#JN[V3^B4L7@(_
M_/EV\9BF^YO5ZOGY67Q6Q2C>KF3;ME>HM4#8+>#VASA`4&MWY04>?%BRDD5Y
ME</NO-3AQ0_"DBB%A]V#%W.3QDF="E?WL9<`$#!<*)A\-R+[4/+UM.66KJ=M
M#9G=1R?FEC,$3(N*NN87%75-]MTYZ6,-?ZW5)]"(/C[='^4JWO$^"\)2I')C
M?\\]3`Q-]H^BJ$`5=L#*CM!5)$E;X=\$]',C^'/LIUY,@+N-X*X3N`7%HQV+
M:`!.7@&(I?<$1;Y0(DB(I*:#LL+-!7"RKKWU7S[=?W8?O9US!/;;@9=^F*1.
M>*1,LO,#;BX`V!JA=4*?6Q0@;$5U8B@,M1375[&WC^*T8-"&?Q(`3U$*&CVF
MNZ#>A,'6''0;K]=,4(".N@+F#!B3Y9/O/?]B0<U.S8)IEP03F?JV+@B(G`L:
M.\C2"L(4Y@2(ZG'BBK?%W+J)#N$:\P$3T'O9>[$/FYP`=;NA[D!R*XAZW#*;
MCXD[4).4[P6YY2J&Q+Q-%"UW"1!DH+#1_H;H3<]Y\>Z%[W902:/UIGS'DL%R
MDT1-6?+PY8<5;%O"*1],:MF3"%='6=SE?@VV8\FJN+`!_LURX[C><NVY07+W
M#9Z?BLL"_@WQOEW\`0SD\^ON(0KDA0`FD!P,J.4KV;H0B+O`&2+Q`.8ORP2W
MKGB?HO`_Q5E'<8CF]MO%#][V$#CQ,"P^ITY</]9CZUA/88VUYBFCC_5?X\A?
M"Y^=,!&^<X+@P7%_9J'S*V<?);]A`.,&"H>]G[I@+G]R8A\;CR'X?1=['GP@
M"ZF\;;JG?^_O'@X)'O%]/5THL*DI0FD!IXJ,B@`EFIQR.QU76"K*QQ4,O=QZ
M(;#(P.>*HYT33H<U0WGJ<>?3M/((DM<D]7;3*R,+;Z8RGAI!BN7::)*1//M)
MTNG9>N]G,RR\\+LH7'MAXJU'P_('*.O"]Y$MW+=A2H".HD1=R'A"W>ZNO>K8
MVHOX-P#'7`59B/&IYU`,*+;4TZ>?_@WAGO>3\^-!R`A0P[0CS#BFMKM$G7XZ
MX)2H">T]P&!5MR#)KCN'-(*K=G>)[E.L5-`GA>]Z+Q</R_#,8I++O;,EAEOM
MNHTK7;>QLW_TW?SRWHEA6!C]6.).,)ZQ=N+U(K]OUF6Y!XLY+TY]+Q$@`N!Y
M<?0S1-!)'A?DI26\`H0!?'KKKXKT]:MB?07$^:K87S'5,[AG?PTC9)(HZ;*[
MHYK<*(CBV\4O)/0ON_W.B7_V8H"M$Z?'SJJNP\XDA!>NV>T;/PB*>]OV9F/;
M60M<;@)B.,O'*/;_%L&U[-()_"U8<OYT2%)_\UH&?(+$<(]@.W^]AI,]@H+\
M76[CZ'GYZ/G;1S#?;YP@`:V;"!!]#>/TRS3:0_QDA!YQ_2%*4Q@$8S0%WB:%
M#8I9;HGQ4_(F*(*$6#3)B#++R/N4D8P?C\`]@]%<@)H7=A$<=1:<>L'9@'^&
M,0L.2W"T67#>I^!TD!%]&ADY'SY?\\S2@<_&M?-YUF?$9_,$-C^,0N_<3/EF
M<UK6=V")=>VJYZ)_[U[U['?@;DFFT2@C9/LL(XQ`D70AQD#6E49&D^WO9AX>
MNB"3WT.<<):<*22G1_0P>OC)<]-G/WV,#BFD]%E[$[E;R<?4,[3L/<)TG3AT
M5DK^[KA[BEC:.:ZKKM4@=^']U<?(WF(!?8Z,OI0@V;P*&\CH4T3)WGJ*[BTD
M,%Y>1,RCO>/Z*:"0J?_C=<H-Z9HOJ>%VD:A+"?+-4G%*J9@H)'@^GL,U[ZX-
M7;,KEQ+KF_W&*;@_4:SO?+@_.Y.8T1,E]IT/HZ\Y#C!8S>?LO-E$M`K)1-'#
M\V'T-?N!71C]+D*%\UQ0P_T>\<-SV`*:]WF.+)PH,GA.<9SK5>`NC.X1L!MK
M0WW$Y(F+U[<>$;(3L.&]F3UUHE#5O(JY(I.ISA&M"V;T4.=6/>\PU^WBVQ@0
MMU$N=$MCR`6S)R4O9+^+,^Y_K&@M8)2#&%&PI;A",@8\&PXV'Y_KP3IC'9Y;
MD9>AS\5]4!L-C@[_/V=:%4;QS@DZX%D)X)T43[*^`6Y:.HGOA$T`L)9AX+WT
M&&LE#C7M6!/_;^"IBK9/%Y7Q/T0!79,&`N=CQUUJ2</N6Y"EKG<!@/IS4^U+
M18/@>!=3">27BN)P/0[36I99M&:($DEMW*F6G*B9'_V*/O6AUDG5XDM%+;I0
M7&$1+Z<LL_%(6:4392O!Q0YH7KP25F)KC8,GGLLJF4BWYB-JA"GP)J'XT:_$
ME;JA7ZD0QAY!&U@QB!(@_S@J89-NXZ"K=+('T0A3C("$JJ"/ZI@'?I*YH]00
M[N4%"1!X3UZ0>:T/AR#P4@$WPNM`11?X)VY"9>5N%__[]_\JADS<A!@XZK/S
MPV7@/("6PI4TL.O6K+'U!1$/"?220YB3A*O<8;<XC0_>HJ3P6I:JU#32#G10
M!M`AV</"3`_>)HJ]G`C71R%U-`K)HG*5%-)&I)!UE1321Z.0(FI722%C-`JI
M5TD?<SSZ7*F=MD:CD':E=MH>D4+7::?A,?Z12*1?EJ$FFC-?>U5;:C-K>(C6
MK\4/&$3UP^W=-RB:"BMKXK@J]LWA;[G8(WHM`LFH.&<6GTU2+T8E.9=Y@=:-
M<PC@&@0!N(<$X+%,'IV]5[T1*M59!'*I8#4,6**6P'F%+^<!7]$AQ5'HC!V&
MJ.IY8#J/XJNB502KX9)>%PVI``(\D4T11;,SZ=IG,D`^N%B0`,D(J^U?Y"-?
M]E=TF[L_(\',[@7!B/LB3GCA(WSIRWJY]:*=E\:OB*CP[26_C0"M)4$2%-F0
MLL^<L:][</_8<U,GW!;;+\6M]NAM09\$V/D^ZT[=1CA^2\*_"]^CX59$*Q?@
M*'C=1B%+T)1A@J;994F315O2FD1-D455R7]35#)471)D6]*S)^\C'U4)ESZ`
MU@^P0=`T3<=_0>`/TC"!;>1M[_LJE.BM2`ZT\T,=P`]-!.3IS@]55!2=R1!(
M[1:&*+I&LN;(D!5[Z&V&3ZL=O]PZ?M+(U1D^2Y0UEC02:/=7ZIT?QU%,;,7F
M<QAJW08';UF0$"FN+%F$*N-?I)*3U]$G21VXP0?N!(@/&A7S^&F;93.#8III
M['/;FNXVIHVQ^@#!YF`L<%"T>4+C,GI9.VYTX6*UN#,V@6\R(<ZZ,XWN<'B#
MMJCHL^Z\%V>P_PS<+FRF:"@,8:,<$##CVE5Q,TXS!V>M>5K,;&,*MAOCLUUN
MY?J$3,_RUMA<)QO?-]O-`?ZV(EJZ65UO*";I<(N&2K'=$!6M=6[Y[*\]X4?`
M\ZNSKD/HK9$^;@V]5464Y>X$_^2D[J,?;H4OL>,'7GQU=)]@5I,;R?Z6YFV>
MU*Z4ZTV>S#RGC>'*R)*(?-%&7T8738-DNRG:EGX.OLQ9F5QK"!N(68V7"X!U
MBF2.Q8;W1$I5$PW*.;=%R[Y*0I[:W54L497T5O?KCX$3SO[N*/XN)\6OWN$=
M0GA5E,TVPE<D7;0YR/Y=#*8T).G"3'%*U.6JE2Y37!7-S@3_71PEB?`94,B/
MPIGD'<UYQ<?@(_H/GA-?ISFW!S@BK%V\,KWAJ4J2W+:HFZ-YUN^9DJHE:L9E
M4K)(PZK/FI"E089`5RMI+-3%E^,2CY+,8EN#)I>NX8]2U@1<,G_`+1\0#/YJ
MRYB8:MAP7X8U;(W*EY!%TZZ,6V6.6[5D-#JE/&Y@,N%(%02!4T?>:M#-O,:W
MC9TPV43Q#@AXE#JI)_S3$CA$FJS;BJ&KNF6;]J\%!!6@5K#NU&$"C:#:H@F3
M27[=72B@**C643;.32A4:P!]LHPC3!^-39],>%2;+3RJ#:4F$Z!S$QYR=[.\
ME\YA)V262,#K6!S>5%_@2J"24E8:<B[_/&-6P<*B=LRX&0\=4^#<.#W(3&2F
MM%$-&LU$-FED4\@;V@BV3`RR$6AMTVPAL.BP+41F0>U<@E1K"'7JLZ\'6-`!
MHI-OGF/ZZ(T6M"HZV?0KC61!AQ&'PYR4EKD]C`F6B-R82",8DRD&K9H#],40
M+;/-Y\@UQFKT.7*-,<]-8RBAJ.3M-HA%I@A,L4"3RSA^Z)`Q\[@2U4SM47R)
M\U-^:L&1"3;7@B/3_\J"@W8FE#?6_TF<B4P?SL*9F,0X#G`FLMC'N7@3DZS'
MR-A$;=2LHY.`F\\Y-C%DRE3`38U<*M3F56AEQB2DPLJIU"X5_1,5.*R**=H5
MQ=%%R6K8HT&;960RV.4$`0,_9!^4K#_`UD##%[E\8D0FJ(-!E'+0#UQ035&V
M2P1<%?A-ABJ-!Q-5>C1GC2KEX.>8ZB?'E$)#+N\D3\#^^J-]@S&E9X#!-.V-
M*8T'$]4SX3X'IMF$\?:H4GB,R/_60^A##F,RDS?**5P`R"Y-4*9&9AZQ]U>_
M//J)D)3WM)MZN(<8UGT-7H5#N":S/9HZK6%E@6B_`QTO:2,77]OR,KG^<!G7
MTJ;B@U0/?AR/>Y\@5W+.?3X18TMQNE'Y>MV<:U@U#3R%)5<63=1*BG4.J[X<
M@VDH>$546@Q)']`R*0NB0:C\"P$/7!,-277E."=>F7**\@?C"VXNLJILJ/@C
M%U#T0[8TU:0D.;N"/BEY1E=@)X90'^].=*=DV@L"?Y_4"/.?2)VA/@&:Q@AF
M:(A(<R2ZJ(2ENNZ,H1-3\H(SAMZ"DH;=:&5FL>PDELVK!"B9EQG&:B5F?49K
M.S'93IML&J37IHB:2M(2'III7W&UU47X[#GI55=%F(0Q6A-C=-&F_)*9,6S&
M#'$8ZQBC$XPI9W2IJB@9YLR8MSZ%6=[<>)^',)MVX.I3S%4N#NB5U215\8_!
M`XW4'8K$L@2&A(KVE5:365Y"MHQ$8'A]J8-%R`=4$K"TL.RA;VISY9Y3KU,Y
MCF26*3OFB<R+$>ZWYD$YL75FPMO'&7,3,\>/Y_CQ7`3L=#QOMZAD\DD=S[-B
M1),H\[GX7<KH)P_*0?PR&:E$A6H*(#.&+Q&E>_DB]U,-&<B-V5Y&FGO(T&^L
M*UI\'#+R+M]LR'Q<MD33&L1E-%SD3)/5FB^+T9G/Q<]GQ%EBM-G8A^Y+G7J?
M?%33.,]JIV3=%*[,[*R,&F#B\%6*0C*7[ZRT45(9\KJ1/J0L:TB_$.J_A5L_
M]$9^>P-H?SHF?%UF:+9?4+#G!%YQ6VKBLC4S>(USBF;QX@`+1P!PQ<7SHNB]
MUK4D.%4N_QX./>G]DH_IF6"(ICR,"=A!%L9D0@N+,%&%,KWU(:]3::`TU^9T
M]11*F=+E%T_4YU,565,E2A>G$;.<JJ'![A%>\4#)]WB6];R*^0\0*Z57,G>7
M/+U*>;7.:7I%1M[QN.L%Y.>5T[GF[+Q)L_,FKOI?K<XX\>N6YJK_[Y+M<W#_
M#"QWN8CP;+LO+[.:RF*M<A1FL8[VIIZS"H%,GL5:(>85YU<K0]YBR7.(PR+.
MOJ/(G#7X?65W?]C#XX1.,'X\";3$$X2IGA\]KP;;V<R.9V:5(:=@6;L'U,E8
MN-331)5^1U*6"CR.82`H_7$#83YN%/!?IRB\CIY#`!5'S[D;%:W]C>_%4`K`
M0E(2=*VX9<6JX/O>H^^/&UG`GU5G"E[%3\??F#GYZ/YZ<*`&XF=@"F_R1\)"
M'X?`N5W\4A86JZ8><J6'U-)#*?=`R"Y;GZ2R^\'!-7?4RAT_;E3AGR'Q5MC%
M;.RM5WO+PK]\W&@M3S58_;B?:K)Z@[$:QZ<"H5@'F0KAOX$A27QX*\2W7^;,
MS!IC,%-YRY?ES@_]W6$'!+BFW7G![:3_3<&\MMSC]7@/K!J%/:BH5,^4CX$[
MC.W)`+7OG.F1\\$=SJW$$55&['NB<'"/W(3)F:"JHFH-8\*Q#!M?6G--4%SA
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MD5K,^_KJ[/>`V5%?W/WE1\.0Y-7O[X5?_?40I;_YO??B@`D/_[CHS+-:T>J5
M\51K7HXO[F)K`=!JT^RK!>.B2EO"O(`_@>N0HH)CXEJQA`Q<&?->]]EZZGBJ
M(HD&M329XZES/+6^]QQ/K<"<)IXZP7*)?F,"<[E$@[SATHY^HP435QJ$%]>I
MTF&IX@]UQ[<J[[^JS5M$I1UJSNEE![@T8-CQ7Q"X+3K4NF':5`*B]WV'18CJ
M\TC;^0'GO<IDV,J//"FGRA!([1:&',]/(M8,3B8=4@"?(T^@^O;"T^6AO>]<
M,[7^G.@HI3E*JW3"5EY>2*;!/JBCOU=I0#U8CO,+7%M:^-A6)=1N<D>-X?FL
M_MU[[7F=O#R<)5HF]2J/TF%P6+=O>,[3Q1<AJ_/+U-[5]C59E%I\R!((]-1,
M46_E1LL.U)^A,#0RS%C<_<_?N8_QF(L[A1O8`K?^[__[C__L<G>CJ9@6A^_<
MGT=T5*["-'88MX`9PB3%DMR=$,5"$(&%4'S:45=7-]2YVIH=EAQFR+#SW,CJ
M2Q=.,.YJ(+S*[FJL:A1V?QN^"H$7;H%BGG+$I20'&>4(VG3,,P^>'VU0!C)D
MO&@WN<=0^\<A2WK*,K?5G8,<9%`8<NH<Y?)+V<JG(<XM2]EY"6J2E/O[`N\J
MFSCS&YTM_#O:;'S7NX&7_'![O/`0K5^)YL@]P-?++%VP@`3?=_\/4$L#!!0`
M``@(`"Z]GD_5-'*CN`D``.,_```*````<W1Y;&5S+GAM;-U;7V_;.!)_OT\A
M:+&'NP=%ENRTL:_.XFX7!0YHNH=V]UX#6J)DMI0HD%0<]TO<XWV_^R3'(?6'
MLB5'3NRDFQ9((LZ0G/G-'VI&TKN?[C/JW&$N",N7;G`Q<1V<1RPF>;IT?__M
MO7?E_G3]IW<L24B$%S&+R@SGTA-R2[%PU.1<+`QQZ98\7S`DB%CD*,-B(:,%
M*W!>3UK8W`N]E1G1BXV=KIGMV1+?R[&3@;<S%ZW&[ZR9[=DQ1YNQDX%786I/
M3]C8R?>">@GS(I852)(=*>XIR;\NW;64Q<+W-YO-Q69ZP7CJ!_/YW-?41N"H
MX2M*3C57'/F88MA,^,%%X->\&99HK'S`:XN4E]D*\]'0((GVK%IP+!2+4A?\
M<MQ"]IR.?]VEH[WK+AV`.5HC/MK/-'/75:;Q>%>9QO;<#,GU@'VO_!M%U#]N
M/K1^Q;.Q>P%O!ZJ(DV*TFH;;GL\8:T2%"2;8M;CA9#+SS;7%O3G(ON%$8FZQ
M1P?9(T2C!G&6]8&F^`)?<7CX#ER^T3LC=+36BG?`25!.1D,/O'NNR@'\00TO
M?8X+QF4#2#(^Z:I=PB9EK&5&AU,&4&O6E,=Q+ZL29^JK]*&"U[LC>/.#VSD-
M#CO"?,<1=&I]:(IFLG/OP0G!Q`>>)GR5:[0'!4^;HRQA91X;.Q@`\7V!.0$2
MHGK:HK.";2W*'K%D=?Y9*W32C!!3V8?W;Y]\H'EPA*DD7:UBG=RA>UT?TPE3
M1W2"(NS%.*+B^IU)K\VP8ZY!N*7[JT+N\S9;,1JXCLI_-9OR\JU-=1UK%4AP
M`BM![SUAJ/[87<+QNZ"8\5P?34OW$TY+BOC3I/@L$1_6M:6>:I<^70=V.;FN
MOW!&8N<SRH7S'E&Z0M'7/G'^C`HF_M;#;`@=&0HB(W44W2%.3"P^1;[W'&/8
ML$^HFG:^W3^2;%4*H_&'85PZ;.=&I!,%(T/DI`)T7'.DWY[/*GTA.LXJAMM+
M<:[2KKIEX"Q#^?FD[@F>8=G'1=JN!F(K),[.'XQ]<O<&XW,+V#'Y[&2>(39$
MB*/VOGSTWCT9WOF9Y3'.!8Y/)N4G\'7G(YL['QZ2U&(]21`=`^,SQO;QT3L]
M=?1J^SU!QCH$^P0;%YY/E:!CEF%\'A=_3[$>_H+^73H5``-&:WE.DVJ/]ZCG
M/PY&>M09\[V2P!\J2*IQTT:\?@>5VB)#_"OFCO[;"/EWSMG&2`@EYC_8_=*=
M.!,GG#C3B1F/E^Y-H,:H%\#@.IQ\@XWU(D)R]E5MB<3:7O47=8WCVW!R>QM>
MW2IQ;\/YK6LX8B(*BK:>S>G\1?'\M6*H>I4<1[*>PJ0(5`%F7WH4YRET;8+Y
MVQ_;I27*H4\:A&JPL4^,$U32JJ5:XUF9(N6H6)/(K7FK:Z_@JM[EDF"A0:@4
MC1AE?.G^,)V]N42S:MN$4-I0WH;S)$I<)V&+C5K*8X7IK^7,@^M6J`)QI#?K
M;*5)4(!ZJ)1,%`BT(3%FAA718HUJERC*/)*EKH+UVLIE2%:TMP&4Y-A;<8R^
M*HI4/B5K"C1]2)YZ&8O5\I1[<M7Q,Z*.3.AP0!]:KP*"Z'YU@JC`#5@2K12H
MK!#@B,-J->R@UYZVI<#>1NW(-IZY<3%(2EYV;VB&SV"%M>80Y)OB"&>%U&,4
MY6F)4C6$<ST0L3*77!G]]\][*WM($)3O9KN6!]:N><P.%<C5)C7MV[JF5+O5
MA)\_[N\)+2Z*[Q_8M>'JW;>AKLGNS@WIGQ_=UD*=:*A-8X>&`1I"*4:\O6T:
M'S!6:E#K,$IBUPZB#8DA<"=1YCX86R9C*5D1E\W$BQ"F[M,CY:_0%#=.VF%0
MSGQ@.E#[)D-D-QH,Q[JF@',C%6S>FG'RC4$C2H4K297IOY1"DF2K?;!`,3SY
M\534@"A!>`G"6(05DQ(:JWTTBA.I%=@E<)*N+8HQP%K=#D/SE\0Q!(`UZ*G#
M0=4&WGT7C2YQVTNL=;^:P'_E!?J!#U69USO@0S8;Q7>8&F9O55**I6.(,*ZS
MO+XT)%W&+-W__?<_C<-9BU@^I^=D)/<H6BE*8^LW2@%_*`'5B6.X`#^8G+IY
M9W:ICQS_D*9'X!`^`0=];G@KG#".:Q!>'T+3DR$4&#]_=0C-3HC0U:M$Z/)D
M"(47LU>)T)N3(31]E?B\/1T^KS1/7YT,H=DKS=/S$R+T.O-T,#D91)=_K$1M
MD:N*S1\JO`ZV%M3>JMI)06%3150E1#585Q#=45.E=,>:`J4:U@V"-3;S@\GD
M1SVJD2.ZA6!8'\14=[6`.V<\0[1=17)5%IM7;G*6XU%(6UNQ4H*(365GD;3H
M<LU9F:Z]JN]D;['/)+?%#L]^;Z)NO^\ZR]F>K'7=*KBJ^A]FK-:JBVE='.YI
M6RJ3<=/SL?&HE]M4=J[7JV(/2U4[>ZJ4SG67J+=WTW18>GNW>P!4W(_OX=8K
M#/51:WKOTS.[W6,`M8DZ0"MJ%PD+I`,,W>;/H6[P**9:X89Y4.6&8UCIAF5`
M[88^K/@`B_8OG!5K9!QU+XHXI@0G/4[)[OI\LH^A2@)M0K`:7P\TO-CJ"X[D
MADAUM)BF>&_?JVZ^(MZ\I>OM=3N>VAJ[""X'NF,3_:^G^]7T\@_TS=X>[IN9
ML!ULFTU-%C\639-JS@EGE<ONB"#:I<_2Z7J$YNKD@7[A>51_R#=,[U)'RFC1
M?]6BPW,;$!]^YPQ^PEKP6TG5#D*\U3IT'^J891Q8P\F9`[,=-17^MB>=,;),
M@G@T#I!3_AA2KEB\?4E)#]_253<D;]1!<K16T#1/"!Y\_O!\V@W=3IO'=-T6
M_]%JFEOD[U?)(VJ&[BU_74B-A(-(^J*)88PCSV9'.3*H%+R$3H,/K";F*'OF
M8W-\%)D;D;'F"(\W1WCVL[BY6YJT3^(Z-DB2*)K/OR\;C`ERL\MN:^`B&&@.
M-(0>^SZN/V",/CWJ,%EC%+_T+<<HZ"LT9^$0G&$P'J/C`J/&Z$52U<N!U&V3
MU'V-%:/Q\="=*:U\M]#->CI,1*H@CYZ*:(:1*/ESQ.M^\7V2$KOOW9+>>0-O
MI%BVK5.Z2NAE3F15GX^U4=B7`_R=MP*K2WC7#+Z4B[R:4..<8D_5=:J0[5CI
M7S<3MX?G@.L-'!,[P^8`"GH/H'I4[U<!-YM>!%?V>-V&;@B5C)PH8!@G[?>N
M\)$?1\3&R-)DV$%OXF(O459?'GNP0(.+/;CGA-#`3#E\Q%:9:\5X#"?C@\6@
M/VBRBI`A(;5;;MOW/I52,`2/$]IK^X5-HW7S6N<>O95W#(_I@(AAUDCY*6?T
M`$>5!R"Y:BY_5P\#2J4L8+SSSK)^QZW-%ZV/MAYLO6[J6<:U7JK=P=+O_T;_
M^O]02P,$%```"```+KV>3P```````````````!@```!#;VYF:6=U<F%T:6]N
M<S(O;65N=6)A<B]02P,$%```"```+KV>3P```````````````!H```!#;VYF
M:6=U<F%T:6]N<S(O=&]O;'!A;F5L+U!+`P04```(```NO9Y/````````````
M````&````$-O;F9I9W5R871I;VYS,B]T;V]L8F%R+U!+`P04```(```NO9Y/
M````````````````&````$-O;F9I9W5R871I;VYS,B]F;&]A=&5R+U!+`P04
M```(```NO9Y/````````````````'P```$-O;F9I9W5R871I;VYS,B]I;6%G
M97,O0FET;6%P<R]02P,$%```"```+KV>3P```````````````!H```!#;VYF
M:6=U<F%T:6]N<S(O<W1A='5S8F%R+U!+`P04```(```NO9Y/````````````
M````&@```$-O;F9I9W5R871I;VYS,B]P;W!U<&UE;G4O4$L#!!0```@``"Z]
MGD\````````````````<````0V]N9FEG=7)A=&EO;G,R+W!R;V=R97-S8F%R
M+U!+`P04```("``NO9Y/L7R8&A,!``#>`P``%0```$U%5$$M24Y&+VUA;FEF
M97-T+GAM;*U3VV[#(`Q][U=$O`>V/DVH:1\J[0NZ#V#$29'`(&RJYN^71.IE
MFC*M6M]\L'W.P9C-[AQ\=8),+F(C7N6+J`!M;!WVC?@XO-=O8K==;8)!UP&Q
MO@35V(=TA8TH&74TY$BC"4":K8X)L(VV!$#6W^OUK'1%=P;68KNJ;GJ=\U"/
M_7FX57?%^SH9/C9"+9'<C@.TSM0\)&B$2<D[:W@L4R=LY6Q8WON4?3;IZ"P)
M]8B/P[&$3S3.D^)+*!/V"SY<,#VH*?^0"@'S^#`DQV$N,#.<64WIAX@#L'DZ
MJ8W(TT2?S4L\>'C^#/81.]>7/&\'K96Q%CR,,&9E2\Z_7^5_6G]<5RHX69#%
M27O/,(EOU(\_NOT"4$L!`A0`%```"`@`+KV>3XV1C0B\%```RP\!``L`````
M`````````````````&-O;G1E;G0N>&UL4$L!`A0`%```"`@`+KV>3]4T<J.X
M"0``XS\```H`````````````````Y10``'-T>6QE<RYX;6Q02P$"%``4```(
M```NO9Y/````````````````&`````````````````#%'@``0V]N9FEG=7)A
M=&EO;G,R+VUE;G5B87(O4$L!`A0`%```"```+KV>3P```````````````!H`
M````````````````^QX``$-O;F9I9W5R871I;VYS,B]T;V]L<&%N96PO4$L!
M`A0`%```"```+KV>3P```````````````!@`````````````````,Q\``$-O
M;F9I9W5R871I;VYS,B]T;V]L8F%R+U!+`0(4`!0```@``"Z]GD\`````````
M```````8`````````````````&D?``!#;VYF:6=U<F%T:6]N<S(O9FQO871E
M<B]02P$"%``4```(```NO9Y/````````````````'P````````````````"?
M'P``0V]N9FEG=7)A=&EO;G,R+VEM86=E<R]":71M87!S+U!+`0(4`!0```@`
M`"Z]GD\````````````````:`````````````````-P?``!#;VYF:6=U<F%T
M:6]N<S(O<W1A='5S8F%R+U!+`0(4`!0```@``"Z]GD\````````````````:
M`````````````````!0@``!#;VYF:6=U<F%T:6]N<S(O<&]P=7!M96YU+U!+
M`0(4`!0```@``"Z]GD\````````````````<`````````````````$P@``!#
M;VYF:6=U<F%T:6]N<S(O<')O9W)E<W-B87(O4$L!`A0`%```"`@`+KV>3[%\
MF!H3`0``W@,``!4`````````````````AB```$U%5$$M24Y&+VUA;FEF97-T
:+GAM;%!+!08`````"P`+`/4"``#,(0``````
`
end
Spreadsheet to calculate tube lengths. Choose chassis options as Booleans. Set length, width and angle of front window where zero degrees is horizontal.
begin 644 tube0-0-2.csv.gz
M'XL(`"^."EX"`WU6WW.:2AA]]Z]@>.BH61M8EE7OC'=&-)ED)FE:L6GGOA'=
M*%.%#&+3]J^_Y]L%`F+SH,#'V?/].M^RCVH;KW;*NLZBO;*6QR=EW:EDDV\/
MC+$._19I^LQ<W#Q&B;[.XT/TM%-K:[I:J<.!.;!=_<JSR`I5E%N+]+5NVJ99
M;DU_[90V$N$RS:-=X85Y#ID_1X<#GE5FS=+]2Y3E>Y7DB"E-<L;=OR,6*LJ,
MLV03)ZK^CDV"\2!P'?PHZ&_Q&M[<H8][S6M]BY-U^FI-DPUB$U5P5(%%M(Z/
M!\8_^J5EB2K]2'2V'T4!O4FS^`^8D,PL'5P]/\>K6#N>/83=Q71^._T4=@-7
M]'I4/)7E\:H%#6\_M:#++(XHJ!-HX'J7X9?%LBL)=!/MGBV3E'[#8?MR1.ZH
M4,U,L5[%FVV^K5M'1*#(C&?N]GD_X$Y)^O;";[#6[:*JXCQ-L[*7$U/N<5]S
M%E7"\B2/\]_,@*RK:+4M[\/C4TYB8,O?+ZK#J65L,O7<?N"Y[/JX*U5B+8Y)
MHK*#AO#1!5HZP+4?#`F.Z#W.[M)7!!G&:U6A)_PBD#IC@GF`>:S>M"P]'+3D
M`?4N`@%T7WO@A!?`BS8^?(E692@N7`N*YJ+;2+V'I43AX^HS&J`BD8ZLZ"7>
M26:*^'!$@3L3]R+PBX808@C$L$#,XVB3)C0V4;91@`XT%!A^J5VA&-[H%'RO
MH.-]#6VPB-(;GV+#?;3;`>J@$'TJ7%]0+`AX*ARDZ;!*PF&>'7/4;%2':FD&
M'%7FJ"5'>AQ"Q6)T4[A,IUCS9BAXM0ZHHGJX-M93C7E9J)\D[ZU29QA*S^@;
M%PT&O!$>J^FU"B/(T,PV!_7516(F<<T!5B'8UY<7!!!$JQ^G'-0\6?'00/%&
M#)""\-]*:+2DEY8MY:7\!,T=EF"QD.>7F&:9:2`HI")*J=04:\+5*1<R+UQ@
M2\0J:$:,&.VA[R_B_2XIJ%>EI].JTJO>3@4Z*%K*.N,?A=54-"FD,!^%]AT3
MRKEU)E/,_/#-F0^I^&X[:;,1M)S:D_OI]RX2'L`["NST;")!IWS>JD&+XR\R
MTW'`XGNLMO,8;5K_J2SMG$Y&5^<AQJA:>:=)0.>+,R0/B>K8$S&@X!T6#`<!
M"L[M5CA0`!]J)BC-]\\P+5_3SD3T:RHM(]*[*?6QO"N[@O[Z\AS5-E-*ZX(/
M@F&OS5G&4_#@P1^>X;E.C]CVJ*.`MDNLKT9>)C4(UA^=XXE_JHY[,GWEIC*@
M/O4;3YH,6O7']?V].!;HW:4^F20^?ZQG4D*GTGE_D9G-,]&,JDA&510204FW
M33B+GN*D%85T312@E/S]1<5VWMZ37"J#6W1&HL2R\5'4D]#V7@ZN]$P$I"/Q
M[KKS9:"YH;;B7Z</L4J__9%M)6ZV1`E!2ODWN/'X=L+4D\M8X[@Y5X=XDT1Y
MG":=2?CUOHO3QC]3"7DQ_3C#XXP>[8EM?W^T[0]=*1SL[?C0.D[O@VU?PG9[
MW0W\?QUFVW/;QK]M][1-3,AVT[!)C;MKV#R-"YMK->ZQM-FU1![3W7&O&/+>
M[7!B-4?UPEC8'E:KXTN,,[DQHU^0MTN5QA%V,I/TA`[V/]]VT7H7YZ;9D%].
AA[QCCOEANHO7369C.B6NL=&/^(C+UUS_`Z[JZF]-#```
`
end
from the I-wanna-hold-your-hand-rail dept.
Bela Kosoian was at the Montmorency Metro station in 2009 when a police officer told her to respect a pictogram with the instruction, "Hold the handrail."
The situation grew tense when she refused to comply and declined to identify herself when asked.
She was arrested, detained for 30 minutes and finally let go with two tickets: one for $100 for disobeying a pictogram and a $320 fine for obstructing the work of an inspector.
She was acquitted of the two infractions in Montreal municipal court in 2012 and subsequently filed a $45,000 lawsuit against Montreal's transit authority, the City of Laval and one of the officers, Fabio Camacho.
Her suit was rejected by Quebec court in 2015 and by the Quebec Court of Appeal in 2017, which said Kosoian was the "author of her own misfortune."
The Supreme Court has ordered $20,000 be paid to Kosoian in damages. The Société de transport de Montréal, which operates the Metro, and Camacho will each be liable for half the amount.
The judges wrote in their decision that Kosoian was "entitled to refuse to obey an unlawful order and therefore committed no fault" in the case.
"A well‑informed person whose rights are infringed must be able to respond — within reason — without being held civilly liable," the court concluded.
Also running in the This-is-not-a-library dept, so go somewhere else to read this! Who loves ya, baby!
From the A-Rose?-by-any-other-name dept.
The town of Asbestos, Quebec, Canada has decided a change of name is due:
"There is really a negative perception around asbestos," Mayor Hugues Grimard told CBC News, referring to the widely banned mineral that was mined there for more than a century.
"We have lost businesses that don't want to establish themselves here because of the name."
The name-change discussion has been brewing for some time, he said. Over the last few years, officials have tried to improve the town's image with a new slogan and colours.
But, try as they might, the name is loaded with negative connotations, the mayor said.
"It's really the perception that you have to change," Grimard said. "
"Others suggested a range of replacement names, from the serious to the humorous.
Suggestions included Ville des Trois-Lacs, Nobestos, Amianteville, Asbestos 2.0 and Poumontousse (a play on the French words for lung and cough)."
https://www.cbc.ca/news/canada/montreal/asbestos-quebec-change-name-1.5375703
MY suggestion? Asbestos McAsbestosFace.
What do you think? Put this to 4chan for suggestions? Nah... SoylentChan!
VTOL
I've previously mentioned that an honorary/great uncle led the team that developed aerospace grade titanium which led to the development of the X-1 super-sonic aircraft. Friend 1 has a similar family connection with VTOL aircraft and from this spent many early years in San Diego and Tel Aviv.
My friend's father was a world leading expert in the construction of seamless tubular steel and was invited to the Pentagon to discuss the feasibility of vectored thrust aircraft. My friend's story may be highly embellished but involved being invited to a meeting for purposes unknown, passing three sets of armed guards on the way to the sub-sub-sub basement and finally being brought to a room with a projector which was swept for bugs before a man hand-cuffed to a brief-case signed over a carousel of slides. VTOL was a common trope in science-fiction and my friend's father had no expertise apart from steel manufacturing. However, he caused quite a commotion when the first slide was projected and he said "I *knew* you were working on VTOL!" Supposedly, this led to an investigation over possible leaks. After this was resolved, my friend's father developed predecessors to the Harrier Jump Jet before retiring in San Diego.
These events set the path of my friend's life. After such an exotic childhood, it was an obvious choice for my friend to join the air force. However, my friend had a very disjoint education and didn't have the technical knowledge to develop aircraft. Instead, my friend worked in logistics. My friend was qualified to drive a truck, off-road, lights-out in a war zone. However, much of the time was spent tinkering with the stock room computer and, eventually, sharing a desk with Nick Pope, the UK military's most famous UFO researcher.
UFOs
A scientific investigation of UFO phenomena can easily cause a person to lose faith in humanity. Firstly, you'll be disappointed by the number of people who report the Moon as a UFO. Secondly, you'll be disappointed about the number of UFO reports which surround civilian airports. Unfortunately, that is the life of a military UFO researcher. Is it the Moon? Is it a commercial flight? Is it a military flight? Is it a weather balloon? If no then place in next pile. This procedure eliminates more than 95% of reports. This procedure may create a few false negatives but that doesn't matter because the remainder is inexplicable and can be correlated with the initially rejected reports.
I have no idea about the remainder and my friend was very hesitant to be drawn about the issue. Indeed, what follows could be described as a mix of Babylon 5 and Barberella. My friend had a theory that human gestation only works if a soul associates with a fetus. So far, this is broadly compatible with the theories in the book: DMT - The Spirit Molecule by psychiatrist Rick Strassman. Indeed, if there are sentient aliens then perhaps a similar association between fetus and soul would be required. However, my friend has the notion that there is a shortage of souls elsewhere in the galaxy; possibly due the human population explosion. Therefore, aliens who wish to procreate do so in the vicinity of Earth where there is a relative abundance of souls. Actually, this sounds more like the work of Douglas Adams or L. Ron Howard.
Crop circles are supposedly alien spacecraft landing imprints and/or messages from Gaia, the collective unconscious or the future. (As I've noted with BitCoin conspiracy theories, time-travel is awkward to debunk.) Like UFO sightings, most crop circle formations can be trivially dismissed. Some of the crop circles are obviously made by drunken amateurs. Others are known to be constructed by people using string, wooden planks and geometry known to the ancient Greeks. Despite this, any crop circle attracts a gathering of hippies with dowsing rods and crystal woo-woo. Worryingly, they *always* find vortex energy and some claim that the fakes were made by the unwitting agents of Gaia or somesuch bunkum. To complete this freak show, a guy call Sørenson used to fly over crop circle formations in a micro-lite and then make compilation videos of crop circle formations. (I presume this has been replaced with drone footage.) An annual illustrated calendar was also popular. (I know too much about this madness due to working in Internet cafés and hydroponics.)
Graffiti
A few crop circle formations are inexplicable. The ones which follow tractor treads are invariably constructible with string. 2012 in binary is obviously made by people. To mis-quote Douglas Adams, the widely propagated interpretation is a dead give-away. 216 circles in a hexagonal radial pattern is barely possible without error; obviously not a first attempt. However, less than 5% cannot be dismissed easily. A friend describes it as rural graffiti. In the case of urban graffiti, elaborate designs are produced overnight. Many of them have a traceable origin and some of them span multiple canvases in an artistic manner. In either the case of crop circles or urban graffiti, can we trace the origin of all examples? No.
Should we trace the origin of all examples? Friend 2 suggests that we shouldn't but it'll happen anyhow and it'll lead to "bad mojo". I should possibly explain that this friend was a qualified electro-mechanical engineer who survived the King's Cross Fire and then quit in disgust due to multiple safety lapses after 31 fatalities. This included a front page exposé in a national newspaper about tins of paint thinner stored under escalators. The flammable paint thinner was covered in lint and the hastily instigated smoking ban was frequently breached. My friend was offered counseling but it was a career limiting move because counseling wasn't taken seriously in 1987. Instead, my friend dropped out, became a hippy, learned juggling, learned fire-breathing, repaired fairground rides and learned video editing. Despite my expertise with C and Unix, my friend was quite ahead of me with Raspberry Pi expertise and remains ahead of me with Kodi.
Anyhow, my friend has a hard to explain theory that an observed, willed or otherwise cajoled universe is constrained in unforeseeable ways. To quote Douglas Adams: "There is a theory which states that if ever anyone discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable. There is another theory which states that this has already happened." While this is humorous, there is, apparently, a more morbid side. Call it the Overton window, the Collective Id or morphic resonance but the rich stimulus of, for example, ubiquitous cameras has a price. A fluffy cat playing a theramin? Cute. Dashcam meteorite? Holy crap! Dashcam fatal aircraft crash? Now we're getting into engineering/snuff porn. This is "bad mojo". But it doesn't end there. The dystopian social credit systems rely heavily on CCTV and facial recognition. This is the downside of everything being observed and collated. Anything of note will be logged, tracked and shared without regard to privacy or dignity unless we make a conscious effort to leave matters unspoken and uninvestigated. To quote The Prisoner: "I will not be pushed, filed, stamped, indexed, briefed, debriefed, or numbered! My life is my own!"
Men In Black
Friend 3 has a more quixotic theory which I assumed was a ghost story or similar. My friend is mad and in poor health but suitably responsible to handle payroll and medical data as a DBA in the UK's NHS. In the course of this work, my friend would often drive at night in remote locations. My friend said that the landscape changed. Not in a get-off-my-lawn, everything-used-to-be-fields kinda way. The waterworks used to be by the main road. Now it is 1/2 mile away. In a year or so, it'll be closer. The pylons are in different positions. Again. The roads are wider/narrower. Just to spook me, my friend told me this near Halloween, while we were in a bar by a canal, major trunk road and chain of pylons. However, the best was yet to come. Near the bar was a World War 2 pill-box turret. Apparently, there are two lines of these turrets across the UK and they were intended to impede a possible Nazi land invasion. After the war, they were left in place. However, my friend claims that the pill-box turret (by an intersection of water, road and electricity) didn't exist during World War 2 and instead appeared spontaneously many decades after the war. My friend claims there is no official record of the turret. Investigation is likely to be inconclusive because war records would be sketchy, classified (as a matter of national, strategic importance), de-centralized to prevent compromise or retro-active by heritage groups.
My friend claims to have seen and spoken to MIBs [Men In Black] around the waterworks and also claims to have been driving along a road while it spontaneously widened and the vegetation changed. Some of this can be attributed to driving while tired. Regardless, my friend appears to believe that the landscape has been changing for centuries and that towns in a crescent shape all have Wiccan names. (Crescent shape disputed. Town names inconclusive.) I was certainly spooked. I subsequently had a dream that I was in a team of MIBs working behind a hologram screen under a road bridge sequestering pollution. This could be a composite of films I've watched. I've also had flying dreams around pylons and I was trying to induce similar before I used text-to-speech to read huge quantities of information.
Rendering Errors?
I'm quite relaxed about these matters but I've been prompted to look again. The first instance was a letter box in an inexplicable place and the post office dis-avows record of its placement. The second instance occurred while walking around London. There is a lovely road called Tufnell Park Road. It is a quintessential London street with an eclectic mix of houses and, at one of the junctions, there is a beautiful stone church. I wondered if it was possible to re-create a street like this using fractals or generative techniques. This is a wholly unoriginal idea inspired by little more than a screen-saver. XWindows has some very ornate screen-savers. One simulates galaxy collisions and, from watching it, the simulation quite obviously relies upon a large hidden mass to make the simulation work. I note that shortly after this screen-saver became widespread, the concept of super-massive black-holes at the center of galaxies *also* became widespread. Another screen-saver sketches lines and it makes patterns which look somewhat like street maps. An early iteration of this screen-saver only drew straight lines. A more current version adds curves. While thinking about this screen-saver, I walked along the other side of the main road (Holloway Road) and wandered along Tollington Park. This street is like a fractal version of Tufnell Park Road. The terrain is different. The angle of the road is different. The trees are different. However, it has a complimentary mix of houses and a different style of stone church at a differently angled junction. Streets are fractal and these two streets are pleasing examples.
I like the two miles or so along Tufnell Park Road, Tollington Way, Tollington Park and Upper Tollington. Ignoring the crack dealers and occasional murder, it is relatively quiet and relatively affluent. Along Upper Tollington, a railway line has been converted into a pedestrian walkway. Because it has become a literal rat-run, the neighborhood cats pick a random spot to survey and hunt rats leaving the walkway. One of these cats has a particular dislike of rain. It can often be found in a dry spot where it makes the most mournful pleading noises when it gets attention from anyone. Overall, it is pleasant to walk along this road and get lost in thoughts. It was therefore a surprise to pass through this road on several occasions without noticing a very unusual artifact near the junction of Turle Road. It is not a turret or an inexplicable post box. However, it is definitely in a similar class. There is an inexplicable brick archway near the West corner of Tollington Park and Turle Road. It is particularly unnerving to walk past this site on multiple occasions and only notice it while thinking of such matters. Was it there on each occasion? Probably. Am I certain? I would like to think so but I cannot guarantee it. Ahhh. I'm being afflicted by the madness of ley-lines and crystal dowsing woo-woo - except I've arrived at such notions via screen-savers and the fringe ideas of three people who were trusted with critical infrastructure.
Rationally, the arch was outside of my filter bubble or reality tunnel. This was abruptly corrected through thoughts on a related topic. An alternative explanation is that we live in a reality with far less fidelity than is generally acknowledged. I hope it is a lapse of perception. Perhaps it is like the time when I wanted to know the whereabouts of my local bus stop - and then realized that people who occasionally loitered outside my place were waiting for a bus. Resource discovery boot-strap. Unknown unknowns. You don't know what you don't know - and there is no easy way to correct it. It isn't always wilful ignorance but there is finite will and infinite ignorance.
The Mandela Effect
I conclude this tour of Fortean Fortina with the Mandela effect. It is possible for the collective memory of the general public the err and a good example is Nelson Mandela's 27 years as a political prisoner. When Nelson Mandela was released from prison, a common reaction was "Nelson Mandela? I thought he died in the 1980s!" And, of course, this led to the inevitable conspiracy theory that dead Nelson Mandela had been replaced with an actor. The unfortunate truth may be that people stopped caring and assumed that he'd died. This may be the root of the Mandela effect. If there is something you don't care about, there's a Mandela effect for you.
Don't care about sculpture? There's a Mandela effect regarding Rodin's Thinker. Don't care about spelling? Then you won't care if someone uses a MacIntosh in McDonald's to read about Berenstain Bears. Don't care about celebrities? Then you won't care who's dead, who's alive or care about celebrity sibling age gaps. Don't care about geography? Oh, this one is vast. There's so much to not care about because of the current/historical names and relative positions of so many places. Many kooks cite the Mandela effect applied to geography as evidence of parallel universes. I cite it as evidence that people care less about their out-group.
I'm an open skeptic but testing objective reality/my sanity gets to the crux of existential philosophy. René Descartes wrote: "I think therefore I am." However, subsequent Occidental philosophy discards causality and history. So, how can I test that I'm not segueing through/interacting with parallel universes when I cannot trust records or memory? More concisely, if you devise an objective test, how can I distinguish your truthful transmission of the correct solution from a false memory of the solution? For the duration of empirical experiment? And when acting upon conclusions? I cannot. It is not an artifact of neurons or atomic decay. It is an information universe with bit rot where nothing is certain. Everything is built on a foundation of sand but some foundations are packed more firmly than others. If we are in a simulated universe then nothing within the universe can ever be smart enough to act without error. If it is dualist then it has a particularly hooky user interface. Can I prove it? Not really but I'm fairly certain. Under such circumstances, what more could you ask?
Taking ideas from this mix (and also the worst Neal Stephenson book that I've ever read), I would like to make a computer game which lazily evaluates the game world as players move through it. In the trivial case, start with the top level of an octree, an undifferentiated sphere and some statistics, such as a total allotment of gold and the sphere being 30% land; divided into mountains, jungles, farms, towns and cities. Place one or more players onto the sphere and sub-divide the octree until everything in the player's field of view has sub-pixel resolution. From here, rendering is similar to the REYES algorithm. When a player moves, perform incremental sub-divisions, as required.
Somewhere in the process of sub-division, terrain will be approximated, streets will be apportioned, buildings will be outlined and everything down to the personality of the neighborhood cats will be assigned. Sometimes, this process will lead to an anomalous street, a stray archway or a homeless cat. To prevent the volume of game data growing without bound, leaf nodes may be reclaimed as a player passes an area of detail. This may occur using LRU or random replacement algorithms. Indeed, this is a source of inexplicable, transient quirks while being consistent with the bit rot of an information universe.
Programming such a simulation requires considerable effort and perhaps I describe nothing more than the ultimate goal of open-ended game design. However, I suggest goals for a very secondary objective. If particle physics is implemented with some precision, it may be possible to nest simulations. In the same way that it is possible to nest virtual machines, the inner ones have progressively less processing power and memory. With or without a seed and algorithm which approximates consensus reality, it remains possible to simulate one universe and nest approximations of it with decreasing fidelity.
See Also