http://diy.smallbearelec.com/HowTos/BreadboardBareAss/BreadboardBareAss.htm
This article explains some basics of how a JFET (Junction Field Effect Transistor) works and how to measure the parameters that make it work as an amplifier. I will walk you through breadboarding a two-stage "Bare-Ass" clean boost with tone control, an extremely useful utility for stage or recording studio. There are other sources for much of the material that I present here, but I sometimes found their language and level of detail hard-to-follow. I will focus on useful basics and practicalities, and I'll do my best to make the theory accessible to other non-engineers.
If you have never used a solderless breadboard, please refer to the intro article on breadboarding before you continue. That How-To covers a lot of basic information and techniques; I presume that you have been through it, have done the demos, understand how the tool works and have started learning to use your multimeter. If you are already familiar with breadboarding, note that the level of detail in this article is meant to guide complete beginners, so please be a little patient.
(Score: 0) by Anonymous Coward on Sunday August 04 2019, @01:17PM (27 children)
I usually use op amps cause I lack knowledge but tutorials like this makes me want to buy some transistors
(Score: 0) by Anonymous Coward on Sunday August 04 2019, @01:33PM (26 children)
I forgot to ask our resident analog EE, who write really well (Rupert Pupnick), are JFET plain old PnP and npn transistors?
(Score: -1, Redundant) by Anonymous Coward on Sunday August 04 2019, @01:49PM (1 child)
Hmm, is knowledge of how to find information lost on some?
And yet, the above, simple, obvious, method of learning seems lost on some.
(Score: 2, Informative) by Anonymous Coward on Sunday August 04 2019, @06:02PM
I can read Wikipedia and use google like anyone else. You totally missed the point of asking that question here instead of asking goggle.
Here you have acces to knowledgeable experienced peoples (Morpheus, VLM, Robert and probably others) providing great answers that focus on the useful for an hobbyist informations. It's much better than the dry pseudoacademic rendering of Wikipedia.
(Score: 5, Informative) by morpheus on Sunday August 04 2019, @01:58PM (3 children)
JFETs (stands for Junction Field Effect Transistor) do have a pn-junction but they differ npn (or pnp) transistors in two significant ways: 1) JFETs have a single np (or pn) junction, whereas npn and pnp transistors (usually referred to as BJT for Bipolar Junction Transistors) have two; 2) in a normal application this junction is biased in reverse (that is the voltage is applied in the opposite direction to the one in which the junction will conduct. On the device level, the current through the source-drain channel is controlled by the voltage applied to the gate-source junction and no current is flowing through the gs-junction itself. This is JFET's main attraction over the BJT as for the latter you need the current flowing through the emitter-base junction, thus wasting input power. On the other hand, the be junction is usually small while the gs one is large which increases its capacitance so JFETs have higher losses at higher frequencies (in general, although there are some really elegant tricks that can mitigate these problems).
(Score: 0) by Anonymous Coward on Sunday August 04 2019, @02:51PM
Thanks, it's appreciated and much better than a wiki article.
(Score: 2) by Rupert Pupnick on Sunday August 04 2019, @03:17PM
Couldn’t have said it better myself.
(Score: 3, Interesting) by RS3 on Monday August 05 2019, @04:13AM
The bias depends on whether the FET is an enhancement mode or depletion mode.
The GaAsFET (Gallium Arsenide FET) revolutionized UHF / microwave electronics, and ushered in the MMIC- monolithic microwave integrated circuit. https://www.macom.com/technologies/gaas [macom.com]
AFAIK, GaAsFETs are used in cell phone RF circuits. https://www.mitsubishielectric-mesh.com/download/catalog/weibo/HighFrequencyDevices_e_201311.pdf [mitsubishielectric-mesh.com]
(Score: 0) by Anonymous Coward on Sunday August 04 2019, @03:15PM
P-channel, N-channel
(Score: 4, Informative) by VLM on Sunday August 04 2019, @04:25PM (17 children)
morpheus's answer is technically correct and pretty awesome, so I'll go in a different direction.
bipolar pnp npn are oldest tech and eat a little current to control a huge current, ratio ("beta") in the hundred-ish range (typically).
Eating current costs... current... which is energy means automatically can never be above 90-something percent efficient, big heat sinks, etc.
A really cool idea for a better transistor would be a gadget using capacitive gates and stuff that uses low/no current to control a huge current. We go these, the modern MOSFET. Of course there are efficiency issues where AC signal hitting a capacitive gate means it reacts "like a resistor" so 100% efficiency eludes us, although its not bad, much better than bipolar transistors.
There was sort of an intermediate idea that peaked around 1970 and would be retro to younger boomers or older gen-X where making gate capacitors was a PITA so given that a reverse biased varactor diode IS a decent capacitor, kinda make a psuedo - MOSFET and skip all that "real" capacitor stuff and use what boils down to a varactor diode instead to make a "fake" capacitor gate. Note a lot of hand waving here and I know its not the best analogy. Note that in 2019 its easier to make a mosfet than a jfet due to decades of mfgr experience, so ironically jfets are no longer "cheap simple mosfets" you simply use actual higher performance mosfets.
As near as I can tell, jfets are pretty much dead in industry and hobby stuff. The venerable MPF102 that was in ten thousand ham radio schematics around 1980 is discontinued although similar substitutions are available. Its a MOSFET world in 2019.
Figure, maybe, gate current of a jfet is a billionth that of a bipolar, and gate current of a mosfet is maybe a thousandth of a jfet.
Insert some hand wavy stuff about surface waves and 1/f noise here, but for some construction differences a jfet is lower noise than a mosfet and the system sometimes could be lower noise than bipolar (not so much today, LOL). In 1980 a MPF102 RF amp stage could blow the doors off a 1960 designed vacuum tube input stage, at least at higher frequencies. Now a days theres not much interest in putting a quarter century old preamp design on your half century old ham radio vacuum tube radio. JFET input opamps were a big performance gain over bipolar input opamps, in the 80s you could get TL081 jfet opamps at Radio Shack and IIRC they were pin compatible with legacy 741 opamps other than vastly improved small signal performance, so it was a ham radio thing to do, to take a filter or other project designed for a high noise 741 in 1970 and hot rod it in 1985 using a TL081 jfet input op amp to get better high freq performance and higher input impedance and lower noise (usually).
Again due to process improvements and obscure tradeoffs, its possible in 2019 (for low input impedance applications) to have a lower noise bipolar input stage than any fet input stage for an opamp.
The whole ramble above summarized to jfets are kinda gen-x/boomer retro nostalgia because in 1980 they were "better" than bipolar transistor designs and "better" (and often cheaper) than a mosfet design. But the EE world has moved on and it hasn't been 1980 in awhile so the famous MPF102 is a discontinued transistor and the design criteria for 1980 just don't work anymore in 2019 so its kinda a hobbyist topic, like tunnel diodes or, frankly, vacuum tubes.
The link is a cool post and strikes me as the kind of thing I've have done at a smaller scale in the 80s using a radio shack 200 in 1 experimenter kit or in the very early 90s in an actual EE classroom lab.
(Score: 2) by Rupert Pupnick on Sunday August 04 2019, @05:40PM (1 child)
You had the 200 in 1 Experimenter Kit???
(Score: 2) by VLM on Monday August 05 2019, @12:39PM
Yeah the futuristic one with the front panel and a plastic lid. Its kinda a life long hobby, and I recall the microprocessor project kit, some kind of "solar experimenters kit", the 50-in-1 when I was elementary school age, something like a 150-in-1 when I was older, and finally the pretty nice 200-in-1.
Something rarely mentioned was if you owned more than one kit, you could take the interesting but mostly useless audio amplifier in one experimenters kit, add a length of wire, and connect it to an AM radio in the other experimenters kit, and similar larger scale experiments.
(Score: 2) by crb3 on Sunday August 04 2019, @09:05PM (1 child)
> Again due to process improvements and obscure tradeoffs, its possible in 2019 (for low input impedance applications) to have a lower noise bipolar input stage than any fet input stage for an opamp.
Noise isn't the only consideration, though. Input bias currents matter when you're buffering a timing or filter cap. You can get very-low-I_in bipolar opamps, but there are tradeoffs to be studied in the datasheets and they're not jellybean (common, cheap and multiple-source) parts like the PJFET-input TL0xx are. You'll still see a lot of TL072/TL082 in modular analog synthesis and effects pedals because they're good enough. Admittedly those are niche markets, but they're not dying, in fact they're rebounding.
(Score: 2) by VLM on Monday August 05 2019, @12:41PM
Totally yeah. There's also 1/f noise vs higher freq noise and gain bandwidth product and all kinds of fun tradeoffs.
And, in 2019, still plenty of circuit designs rolling around where nothing special is needed so spec a five cent 741 thats half a century old.
(Score: 2) by morpheus on Sunday August 04 2019, @09:35PM (9 children)
I would have to agree that JFETs have become rather niche devices (I guess not in small part due to the difficulties involved in putting CMOS and bipolar devices on the same die) but some discrete JFETs are still made (MMBFJ175LT1G is an SMT JFET made by ON Semi with a pretty good 1nA cutoff current, tough to achieve with a MOSFET) and there are a few low-noise opamps with JFET input stages. In general though, JFETs do make me nostalgic for the 80s ...
(Score: 2) by RS3 on Monday August 05 2019, @04:19AM (8 children)
Brings up a question: I've seen lots of audio power amps with power MOSFET output stages. In fact I have one. But you don't see power JFETs, and I don't see small MOSFETs used in small-signal stuff much. I assume they're not super linear and would introduce distortion. Any thoughts / knowledge on this?
(Score: 1, Interesting) by Anonymous Coward on Monday August 05 2019, @05:03AM (1 child)
BJTs are much more linear than MOSFETs.
However, BJTs use more power than MOSFETs and, unlike MOSFETs, BJTs can start to consume MORE current as they get hotter, leading to thermal runaway and *poof* of component.
I could go on, but suffice it to say that analog circuit designers have used either type to make some amazing amplifiers--they just need to use different tricks to get the performance they want.
Except for digital circuits. There, the almost perfect on-off switching characteristics of MOSFETs made BJTs obsolete early on.
(Score: 2) by RS3 on Monday August 05 2019, @05:42AM
It's pretty easy to design in current-limiting and temperature compensation (I'm an EE).
(Score: 2) by VLM on Monday August 05 2019, @12:46PM (5 children)
RF mixers with delicious large signal performance and low (-er) spurious products. But yeah other than that oddity, yeah.
I don't see many low power class-D mosfet RF designs although its an interesting idea. Probably a complete waste of time, but then again ham radio is all about fun and there's probably someone out there using a class-D amp as an oscillator buffer amp simply for bragging rights, LOL.
(Score: 2) by RS3 on Tuesday August 06 2019, @03:59AM (4 children)
Yes, absolutely, especially the cool 2 gate mixer MOSFETs!
I should have qualified my question: by "small signal" I meant lower frequencies, instrumentation, audio, medical, etc. RF is its own beast. I don't think I've ever tried to design an amp stage with a MOSFET. But why not? I know that very high-end "condenser" microphones all have a MOSFET input stage...
(Score: 2) by VLM on Tuesday August 06 2019, @12:47PM (3 children)
Ridiculously high output impedance matches well with ridiculously high input impedance, FETs and condenser mics are a marriage made in heaven...
(Score: 2) by RS3 on Tuesday August 06 2019, @08:00PM (2 children)
I think you meant ridiculously low output impedance. (?)
(Score: 2) by VLM on Wednesday August 07 2019, @04:29PM (1 child)
Its all in the point of view of the ins and outs. The condenser mic is just a floopy capacitor so its high impedance and the preamp is turning that into a low impedance for later audio processing, which FETs are really good at.
(Score: 2) by RS3 on Wednesday August 07 2019, @10:25PM
I figured that's what you meant. It's not always easy to verbally describe this stuff, hence we use diagrams and specs. In another "discussion" I got slammed and downmodded 3 times over pedantry. Someday I'll study psychology of the ego.
I modded a guy's "Blue" condenser microphone and he loves it. Someone had knocked it over and it was intermittent, so I fixed something in it, but while I had it apart I noticed quite a bit of wire leading from the capsule to the FET. 1 G ohm bias resistors, BTW. So when you would move the mic, the wire would move, thus changing capacitance, and I = C dV / dt and all, and you get noise. So I slid a Teflon sleeve over it, which reduced the capacitance (I think!), limited the motion and resulting handling noise, and he loves the mic. (Grammy-winning recording engineer.)
I still can't get over the 1 G ohm resistors. Someday it would be interesting to know how they make such a thing. I think some mics have 2 and 5 G ohm bias resistors.
I worked for a short time for a small company that makes small A-D converter modules. Some have almost infinite input impedance. When one was on test, someone entering the room would wildly vary the output. IIRC it only had a common FET op-amp, but the construction and cleaning was key to the impedance- if not cleaned and rinsed and dried exactly correctly, the impedance was far low.
(Score: 2) by RS3 on Monday August 05 2019, @04:28AM
For _really_ big stuff we use the IGBT https://en.wikipedia.org/wiki/Insulated-gate_bipolar_transistor [wikipedia.org]
And for really small, quiet stuff, a BJT op-amp INA163 http://www.ti.com/product/INA163 [ti.com] might be of interest. Caution: it suffers from latchup- smoke will happen. It was originally designed by Burr-Blown, so... There may be other op-amps of similar specs without the latchup problem, I'm not sure.
(Score: 0) by Anonymous Coward on Tuesday August 06 2019, @05:30PM (1 child)
Do you give any credence to the statements made by audiophiles who claim tube-based amplifiers are inherently better (warmer, more natural, etc., etc.)? Can't you get better cleaner amplification with modern FETs than tubes? Is it worth the thousands of dollars to pick up 1960s-era equipment? I've also seen that argument [pbs.org] made as well. Do you think analog components sound better, or just different (or is it just some sort of placebo effect)?
(Score: 2) by RS3 on Tuesday August 06 2019, @08:23PM
Believe it or not, you're mixing 2 different things. Cleaner is not necessarily warmer, more natural, or pleasing. This gets into a complex area that I'm not expert in, but I know some who are.
They talk about even vs. odd harmonic distortion, whether a system is "musical", etc.
You might correctly assume that ridiculously low distortion should yield the best overall listening pleasure. But, what if the recording / mix / mastering was done through systems that introduced some degree of distortion (especially the speakers). The mixers / masterers might have compensated in a way that makes the recording sound good on certain playback systems, and maybe harsh on other systems.
I'm still trying to figure out what people like so much about tube systems. One is "power compression"- there's a natural compression that happens due to the inherent resistance in tubes, transformers, etc., so that the louder an input peak is, due to the resistance you don't quite get the full peak at the output.
EVH famously used a Variac on his tube amps- quieted them down, but added some distortion and power compression.
I have much more to write on the subject, but little time. In general I've learned that some people have super hearing and can hear things which are almost immeasurable, or, the people who define the measurement criteria don't consider how intricate some people's hearing is.
(Score: 2) by TheGratefulNet on Sunday August 04 2019, @07:55PM
jfet is more like a tube.
pnp and npn are 'junction' types.
quite different in how they work. they both can be amps or switches and both have 3 leads, yes.
"It is now safe to switch off your computer."
(Score: 3, Informative) by crb3 on Sunday August 04 2019, @03:16PM
More specific is this link: http://www.runoffgroove.com/fetzervalve.html [runoffgroove.com] ...which discusses how to make a JFET distort like a vacuum tube as its main topic. Down near the bottom of the page is the design of a compact JFET test jig (I built one on a scrap of copperclad to plug into my DMM) and measurements obtained with that jig from various JFETs commonly used in guitar-pedal building.
(Score: 0) by Anonymous Coward on Sunday August 04 2019, @04:07PM (8 children)
was a turning point for humanity. Like fire, it changed much of what we can do; and understand. Without it: bla bla bla....
With it we have come to understand molecular energy and reactions which we never would with vacuum tubes and chemistry. The simiconductor revolution triggered a change in how we looked at electricity.
All hail the mighty transistor!
and one day my first tattoo will be a transistor symbol... now which one... NPN or PNP....
Moss
(Score: 2) by crb3 on Sunday August 04 2019, @05:18PM (7 children)
Why not both -- https://en.wikipedia.org/wiki/Sziklai_pair [wikipedia.org]
(Score: 2) by c0lo on Monday August 05 2019, @03:36AM (2 children)
Ummm... which one [electronics-tutorials.ws] - NPN-like or PNP-like?
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 3, Funny) by Rupert Pupnick on Monday August 05 2019, @12:32PM (1 child)
If you have a preference for ground referenced signals and positive supply rails, I recommend an NPN tattoo. Show it off with pride and maybe you’ll get lucky!
(Score: 3, Funny) by c0lo on Monday August 05 2019, @12:54PM
Eh, too late for that, my luck ran out.
Anything-tatoo will look weird on an old liver-spotted wrinkled skin, and I'm definitely not gonna expose my bear-belly with pride (grin)
Also, speaking of "lucky", there's a good number of jokes out there about guys with tattoos on their penis [google.com]. Haven't seen one with a NPN Sziklai pair though.
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by c0lo on Monday August 05 2019, @03:53AM (3 children)
Why not both - https://www.electronics-tutorials.ws/amplifier/class-ab-amplifier.html [electronics-tutorials.ws] ?
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by VLM on Monday August 05 2019, @12:48PM (2 children)
In today's political climate he probably doesn't want to appear partially biased all the time (sorry about the awful EE joke)
(Score: 2) by c0lo on Monday August 05 2019, @12:57PM (1 child)
The configuration works well with a well balanced biases that cancel one another, no joke (grin)
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by RS3 on Tuesday August 06 2019, @08:26PM
He could always go for self-bias.