LOLIN S2 Pico – A compact ESP32-S2 board with an OLED display
LOLIN S2 Pico – A compact ESP32-S2 board with an OLED display
If you're into small MCU boards with an integrated display, you're in luck as LOLIN launched the S2 Pico board with ESP32-S2 and an OLED display about at the time same as LILYGO T-Display RP2040 board we covered yesterday.
Wemos/LOLIN S2 Pico board offers WiFi connectivity, a 128×32 OLED display, USB Type-C port for power and programming, as well as the usual GPIO headers in a compact 50×23 mm form factor.
$10 Raspberry Pi Pico Alternative Comes With LCD Screen
Tom's HARDWARE: $10 Raspberry Pi Pico Alternative Comes With LCD Screen
Microcontroller boards based on the RP2040 chipset, the same SoC that powers the $4 Raspberry Pi Pico are becoming very popular among makers. Newer boards are popping up with extra features appearing on them, such as this 1.14 inch color display, the $10 LILYGO T-Display spotted by CNX Software.
It's not the first such board, of course, with the Arducam Pico4ML pulling a similar trick - and with a resolution of 240×135 pixels it's hardly HD - but it comes in at just under $10 with the RP2040 board attached. Along with the ST7789V SPI controller needed to run the screen, you get all the usual Pico accoutrements such as the dual-core Cortex M0+ processor, 2 x UART , 2 x SPI and 2 x I2C connections, along with a generous 4MB of flash storage. Power and data connectivity is via USB-C, a good choice of connector as it is now becoming the norm on maker boards. Where the LILYGO T-Display falls short is the GPIO. The board looks to be wider and a different pin layout to the traditional Raspberry Pi Pico, so creative hacking is required to connect accessories designed for the Pico.
Sadly, a major drawback is that it cannot run Windows 11.
Related Stories
The Telegraph reports that Raspberry Pi Trading has offloaded stakes to Lansdowne Partners and the Ezrah Charitable Trust in a move that values the operation at around $500m. Most manufacturing is able to be done in the UK, and last year's sales amounted to 7.1m units for a profit of £11.4m.
Lansdowne Partners' presence in the list of investors is less surprising than Ezrah Charitable Trust. The latter was founded by former Goldman Sachs vice-president and Farallon Capital Management partner David Cohen in 2016 to focus "on the poorest of the poor, especially in Africa" – an indicator that it may be the work of the not-for-profit Raspberry Pi Foundation that was of interest.
The Register mentions that the foundation's 2020 financials show an income of over £95.8m, up nearly double from the £49.5m it reported in 2019.
Raspberry Pi Trading makes the hardware, the magazines, the peripherals, and so on. The Raspberry Pi Foundation runs the charitable programs.
Previously:
(2021) Two New Microcontroller Boards Released with Built In Displays
(2021) Raspberry Pi Begins Selling its RP2040 Microcontroller for $1
(2020) Raspberry Pi 4 Gets 8 GB RAM Model, Also 64-bit OS and USB Boot (Both in Beta)
and more.
(Score: 4, Interesting) by JoeMerchant on Thursday September 16 2021, @12:51PM (9 children)
The $10 Pico w/display is tempting, but it lacks connectivity.
What are the good similar packages for a PiZero W?
I mean, here's an e-ink: https://www.amazon.com/2-13inch-Display-HAT-Two-Color-Raspberry/dp/B07Z1WYRQH/ref=sr_1_8?dchild=1&keywords=Pi+Zero+HAT&qid=1631796445&sr=8-8 [amazon.com]
And color e-ink: https://www.amazon.com/Waveshare-2-13inch-HAT-Three-color-Communicating/dp/B075FR81WL/ref=sr_1_13?dchild=1&keywords=Pi+Zero+HAT&qid=1631796445&sr=8-13 [amazon.com]
I just like the idea of the pre-attached package (one less thing to screw up / have to mess with) - I'd like it even better with a case solution already worked out.
🌻🌻 [google.com]
(Score: 2) by looorg on Thursday September 16 2021, @01:00PM
If these become popular for something, floppy emulators or something, I'm sure someone will create and put up a decent case at Thingiverse (or some similar site) for easy printing.
(Score: 2) by Freeman on Thursday September 16 2021, @02:55PM
Those seem to be the only pre-attached options for the Pico. I kind of doubt there will be a whole lot of pre-attached package options. The more utilitarian approach of the Pico doesn't lend itself to plug and play options. It's an interesting little device, though. It's cheap enough, maybe someone will make a product using Picos?
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 2) by Marand on Thursday September 16 2021, @05:00PM (6 children)
It's different hardware, but you can find tiny ESP32-based chips with small displays pretty readily. I got one of these [amazon.com] a while back, for example, though I still haven't done anything with it. I just wanted an ESP32 board and figured "fuck it, get one with a tiny display for fun".
(Score: 2) by JoeMerchant on Thursday September 16 2021, @06:35PM (5 children)
You've got WiFi on that, too... did you ever setup a toolchain? If so, which one?
That's (one of) the attractive things about the Pi Zero, to me. Raspbian is such a close relative of Ubuntu I can develop on my "normal" desktop, then setup the same toolchain on the Zero and duplicate the results there, complete with USB keyboard, mouse, HDMI out, wifi network, etc. Downside of that approach is the boot time - if that ever got to be an issue you can bare-metal the Pi Zero for faster startup, but I've never felt the need to go there.
🌻🌻 [google.com]
(Score: 2) by Marand on Friday September 17 2021, @11:30AM (4 children)
Yep, that was why I got it. Wanted an ESP32 because it's pretty well known/supported for odd stuff like NodeMCU, figured "fuck it, get wifi+lcd one".
Nope, got side-tracked and never really did anything with it beyond making sure it wasn't a dud. I was wanting to test some higher-level language options on it like Lua (or one of the various compiles-to-Lua languages out there) using NodeMCU, Ferret [ferret-lang.org] (a lisp that outputs C++11 for use with microcontrollers), or even setting up OCaml directly, but never got far with it. The OCaml route didn't pan out because the info for setting up a toolchain for it was really outdated, and before I could try out Ferret or NodeMCU I got distracted by the Pi Zero W, which I got at the same time along with a small LCD HAT that has a 4-way dpad and a couple other buttons.
The boot time on the Pi Zero isn't great, though you can make it a little better by cutting out some of the cruft manually, but it doesn't matter because the convenience and familiarity is hard to beat, it's just another Linux environment with everything I expect to have. Which is why I ended up messing with it more instead of going further down the ESP32 rabbit hole.
The main issue I have with it (aside from the boot time) is that compile times suck if you're compiling on it directly, but that's stupidly easy to work around if you use Linux elsewhere as well: mount an RPi OS image via loopback, set the mounted directory up as a chroot, and use qemu-user-static [debian.org] to do binary translation from ARM to amd64. There's still some overhead but being able to compile on an 8c/16t desktop with 64GB of RAM is still way faster than doing it on the Pi Zero (or even another, faster Pi). With a small shell script to automate the steps I could compile an OCaml program, rsync it to the Pi Zero, and execute it on the Pi Zero in less time than it would take for some interpreters to warm up and launch on it directly.
Of course, I've since gotten a Pi 400, which has made me lazy. The chroot's still faster but most of the time the pi400's "good enough" so I just compile stuff on it first, at least for now; it'll be back to the chroot whenever I get around to switch from 32-bit to 64-bit RPi OS (formerly known as Raspbian). Probably should have just gotten a regular RPi4 since I mostly interact with it using Barrier or over ssh, but the integrated keyboard was amusing and I like the idea that I have a mostly self-contained little PC I can use in odd situations that may come up.
(Score: 2) by JoeMerchant on Friday September 17 2021, @03:09PM (3 children)
My usual setup is a git/ssh server then Qt Creator on both the Ubuntu "powerful" desktop and the Pi4 or Zero W running Raspbian - often running a VNC server on the Pi so I still use my Ubuntu system's KVM, but it is nice having the option to direct plug KVM into the Pi hardware when things go pear shaped.
Development happens on the fast system, then I will do a clone/pull on the Pi and compile native in Raspbian while I am doing something else so I don't notice how long that takes.
My favorite interface for the Pis lately is some kind of webserver, often using Stefan Frings template: http://stefanfrings.de/qtwebapp/index-en.html [stefanfrings.de] but anything from Apache to other webserver frameworks do well also. The main thing is easy access/control from a phone browser.
The Pi4 is nice and powerful, but more often than not the Zero W is still overpowered for the task at hand, and in a lot of cases it is great to have "single purpose" widgets that just do what they do, instead of cramming all the features into a single (point of failure) home PC.
🌻🌻 [google.com]
(Score: 2) by Marand on Monday September 20 2021, @09:28AM (2 children)
Same basic idea, except I hate waiting so I moved the compilation step to an ARM chroot on the amd64 desktop. Not nearly as big a deal now that I have the Pi400 to use as an intermediate step, but it was a great trick. Good for setting up new Raspbian images, too, because I can do bulk installs via my desktop's faster storage and wired internet connection, then write it to sdcard when done.
Yeah, once it's running the Zero W is honestly overkill for a lot of things, which is why I wanted to mess with ESP32 chips. Which are kind of overkill for many things as well, but kind of at a sweet spot of price/perf, so not worth worrying about. But laziness wins in the end, so I'd rather use the slow-to-boot Pi Zero for something unless it's going to be turned off/on frequently and would need the faster boot time.
I'll admit to being guilty of the latter. Always-on desktop with 64GB of RAM and more CPU cores than I typically need for normal use, so I'll sometimes just throw a VM at a problem. Trying to branch out a bit more, though, especially for anything I want to use a display with, since I'm limited on the # of display outputs available otherwise. My primary GPU is already maxed out at at 4 outputs (it has 5, but can only use 4 simultaneously), so if I want to use another display for something it has to be through the secondary GPU, which is used by VMs via GPU passthrough.
(Side note: I currently use a Windows VM for some games that way, but I'm considering trying the upcoming SteamOS 3.0 in one for Linux-native gaming instead of using the host OS for that. Cleaner separation of games vs. other things, hopefully fewer issues as a result. A lot of native games don't like my weird system configuration.)
For example, I got a fairly massive plasma television for free because a business was replacing them, and rather than try mounting that ridiculously fucking heavy thing for use as a TV, I repurposed it into a secondary "desk". It's kind of fun to display different wallpapers, especially subtle ones, on it; makes it look like different types of surfaces at a glance. Or even subtly moving videos, or put useful information it. I started driving it from an rpi instead of trying to just use it as an extra desktop display, made it a bit more flexible in use that way.
(Score: 2) by JoeMerchant on Monday September 20 2021, @11:06AM (1 child)
Plasma display desk, sounds like a good plate warmer.
🌻🌻 [google.com]
(Score: 2) by Marand on Tuesday September 21 2021, @01:50AM
It has a lot of fans and does a good job of cooling itself for the most part, so I don't have to worry about having stuff on it like a normal desk...though I suspect it'd make chocolate a bit soft if left on it while running.
Adds a fair bit of heat to the room when it's on, though, so for goof-off use I mostly only run it in the fall and winter, maybe early spring. During hotter times I only turn it on when I have a specific use for it.
(Score: 2) by looorg on Thursday September 16 2021, @12:58PM (5 children)
I just quickly scanned this at the moment to see how much power the display requires but didn't really notice in any of the summaries, but I could be blind or missed it. The reason for it is that most things I have built or assembled that has some little microcontroller (raspberry or other) doesn't actually need or have much use for a display. So I don't want to be wasting power for something that isn't used. That said there are versions without that are still apparently a lot cheaper so you'll just get this one when you need that kind of display. It's quite an uptick in price tho from just around $4 to $10. Hopefully if you get some you can just remove/disconnect or desolder the display if not needed without the entire board just going crazy if it's not there. But you never know.
It's just that a lot or most things that these little controllers are used for, by me, just end up in another larger box and there just isn't a need for a display. It will just be a totally worthless power drain that in the worst case just sits there lighting up the case from the inside (if not turned off or disconnected). Perhaps it will be nice initially when you program it or as a diagnostic tool but beyond that not so much. In such cases just having a few LEDs or something is just as good and doesn't need full display (if you can ever call something that is 1.14" as an actual full display).
For the things I have made that uses a display I don't think full colour have ever been useful. Some kind of single or double row monochrome thing is more then enough and good enough. I would much prefer it to have a display I connect when needed, if needed or wanted and not something that comes pre-soldered and not possible to remove (that said nothing is impossible with force, violence and enough heat or desoldering).
(Score: 2) by Rich on Thursday September 16 2021, @09:50PM (4 children)
Think about the time saved for mounting two distinct components and wiring between them.
If you need a display, you can just grab one of these. It's not that your generic 2x16 is "stupid". It still has the HD44780 controller chip. And you end up with 8 data wires if you can live with the stupid 4-bit banging, or 12 if you want 8-bit parallel. Or stuff another PCF8574 I2C controller in front of it. Pricewise, we're now at where that LOLIN S2 is, so we can as well take the whole integrated package courtesy of some entrepreneurial residents of Guangdong. The colour is sort-of free. They just can't be bothered to make black and white displays again, I guess.
I can imagine quite a lot of situations where a little display is helpful. I've never seen a decent charger for NiMH batteries, one could easily be rigged up and indicate charge current and status for batteries. The RPI2040 packs mighty PWM abilities, it could show stepper and encoder positions for closed loop motor control (one of your debugging use cases, but incredibly helpful, if there's a complex stack above it that only speaks G-Code). Or, extremely simple, for anything that needs more than on-off settings, show an encoder position.
What I'm not fond of, is the incredibly small size of those screens. I really wouldn't want to display more than, say, a parameter name, and it's value at the same time, because otherwise one would need a magnification glass. Having colour might be actually useful, because colour cues can be used to assist with the screen information.
(Score: 2) by looorg on Thursday September 16 2021, @10:30PM (3 children)
I think I rather have a fairly long display display. I'm fairly partial to the 2x16 (or larger). Still 2x20 or 4x20 or 2x40 or whatever are options they are usually to large for the item built so the screens become larger then the rest of the little device so it becomes it bit odd. So somehow I found that 2x16 is sort of sweet spot (2x20 is ok to). It can usually put out enough text there without having to resort to like scrolling and other annoying display tricks. That said I think there could be uses for this to. I was mostly just concerned with I wondered how much power it would require. It might be a trivial amount and not matter but it could also be an issue. Still it might be really useful to just display some graphics instead of text or in combo with text. Or as you note perhaps just have the entire tiny little screen fill up with specific colours for very quick visual diagnostics instead of having to read lines of code. Having full colour then instead of just one or two gives a lot of options.
(Score: 2) by looorg on Thursday September 16 2021, @10:39PM
Considering the entire board appears to require or make do with 5v it shouldn't be any issues I guess.
(Score: 2) by Rich on Friday September 17 2021, @07:03AM (1 child)
I have to stand corrected on the colour thing. There are indeed a good number black/white OLEDs on offer. I gave one such datasheet (multi-interface display only, no MCU, 1-inch 128x32 black/white) a quick glance, so FYI the power consumption is around a quarter microamp for the controller (1.8 to 3.3V), and the display itself draws a tad over 1mA per 10% illumination (from a 7V source). The one I saw on the datasheet could DC/DC convert the 7V (display draw 7.25V at 50% typ 5.1mA, with converter typ 15.5mA from supposedly 2.8V).
Lifetime on this 10 year old datasheet was given as 10000 hours typ, a bit over a year, so one wouldn't want to have it always-on anyway. That's the good thing about the good ole' alphanumeric LCDs: these still work great after 40 years. The datasheet I saw was from 2011, so they might have gotten a little better, but I think the above numbers still give a rough idea what to expect.
(Score: 2) by looorg on Friday September 17 2021, @09:24AM
Good catch. I didn't even considering the life time, as noted previous displays more or less lasts forever. If this thing can't be turned off at will or disconnected you have a very set limit on it's lifetime and usability. Seems somewhat less and less interesting as a board.