An interesting continuation of this would be to expand on your discussion of distortion by using a spectrum analyzer to show the effects at different operating points.
The thing that really knocked my socks off is what appeared at 15:18 . When I decided to learn RPN, I naturally looked towards HP. The 32S made its appearance then and the dots matrix numbers quickly won me over. The simple clean keyboard was magnificent. I grew to love this sooooo much that I had to get a backup. THEN the display failed and HP "repaired" it by giving me a brand new 32SII . Much more cluttered keyboard and see through pages in the manual. This is when the love story ended. I had good times squeezing the most out of it. My most complex program calculated the miles per gallon and estimated miles left in the tank plus it also told me when it was time to change my engine oil. One answer was before the decimal point, one answer was after the decimal point and the expansion from 4DP to 11DP said it was time to buy engine oil. On, I learnt a lot about JFETS here also.
HP repaired my HP67C in the early 80s by giving me a brand new one in a box. I had no idea that was going to happen. They called to tell me it was ready to go. I went to pick it up and was very confused when they handed me a brand new calculator. The repair rate was a flat $99 no matter the repair. I still have it although it needs some TLC. I wonder what would happen this time. Likely no warehouse full of HP67Cs out there, huh.
Wow. This was perfect for me. I'm a tube amp guy so this was cool. These things work like tubes, with the obvious differences of course. Looking forward to watching more of your videos.
Also a very easy circuit to do impedance matching. Common use is to allow a very high impedance microphone, like D-104, to be used with modern day transceivers that use lower impedance microphones. I used a 10M ohm resistor from the input to ground to accommodate the D-104 microphone.
A quick question. If you know. Can you tell me the difference between an amplified D-104 typically used for cb but also ham and one made in the 1950's long before the 9 volt battery amp was ever though of other than lack of the amp? Thanks.
@@W1RMD The original D1 04 used an extremely high impedance Rochelle salt crystal. The newer ones use a lower quality element and use the amplifier as impedance matching circuit
@@bobkozlarekwa2sqq59 I've noticed that I HAVE to run the amp in this mike with this D TEN FOUR. Even though I have a 1963 Johnson Invader 2000 transmitter. I've tried to bypass the amp, but it doesn't sound good. I have to use the amp and keep it turned all the way down to get a good sounding audio. It's too bad they've turned this (the D104 product) into a "CB" mike over the years and I understand that you're dealing with 5 watts on a CB so it needs all the help it can get , but that's not an issue with a large transmitter. That I know of, you can no longer get the correct element, or the mike anymore. 73's
I wish my Electroncis professor at the uni had this practical and down to earth approach to circuits, instead, he would flood us poor students with tons of equations and formulas - and as much as I loved maths I hated the completely theoretical approach. That made me decide to take a software career even if I loved electronics more. Well, I'm 52 now and picking it up after a 30 years hiatus and channels like yours are bringing the love back, even stronger actually! Thanks for your great videos.
The math in most engineering science did not start with the chalk board but with experimentation. Teaching should start with more practical applications and work the math backward from there. More intuitive and easier to remember.
I've been designing circuits for over 15 years and this is very similar to a first order design I do to solve a problem. Of course sometimes you need to calculate then simulate and understand 100%. But sometimes you just need something you can trust as a foundation and go from there. Quite refreshing to see Im not alone in how I approach things.
You've explained well the negative potential between G and D! The best part of the game is calculation and next check values and oh yeah the pleasure of a positive feedback. My favorite moment is when you move the camera and... bang‼️... the circuit is already playing on the breadboard 😂👍❤️
Great video! I agree about not having to run all the numbers on everything. One of the important things in engineering is to not over-design, or spend too much time on things. You're operating in the safe range of the device, it's used as part of a hobby, and you have the gain you want over the frequency range you want, by measurement. If it were for a production run of a thousand, then sure, do some math. Even for production use as part of a test setup, you'd use gain figures by measurement, not calculation, because the only thing that matters there is the actual value, not the theoretical.
There are depletion mode MOSFETs which act aomewhat like JFETs, at least in part of their characteristic. They are fairly rare and not much used, but they do exist. One nice property of those is that, as opposed to a JFET, you can allow the gate to swing to a positive voltage, not only negative. Because it's not a junction, and it won't become directly biased if it gets positive. As for choosing JFETs for an application - in my type of applications (voltage-controlled current sources mostly) I start by looking at JFETs with an Idss (drain current when Vgs=0) in the range that I am interested in. Idss also happens to be the maximum current that the JFET will conduct. Everything flows from that point on.
"I chose this particular JFET transistor because I have a lot of them, and it's a very very very common one." or was, until it went very very very EOL. Central is still selling them, but naturally it's at their "where else are you gonna go?" prices.
@@glasslinger those would be the offering from Central, yes. Hard to justify paying that when we aren't all that far removed from paying buck and a half *for* 10. Better off finding a substitute IMO.
@@glasslingerso far, I've managed to maintain my frivolous affinity for RN-series resistors without expressing any sneering contempt for those poverty-stricken Xicon users. After all, they do the same job for less money. It's a great video from a great instructor covering general use JFETs. Market forces have dictated that this particular JFET is no longer available at jellybean prices. The content is still applicable to other parts at a lower price point, or indeed for those who are determined to pay the premium for the named part. None of these strike me as controversial statements or an invitation for financial advice.
So happy I recently found your channel. You’re very well spoken and are now one of my primary sources for my DIY projects :) keep the awesome content coming. (Especially excited about your new analog synth series)
There's a REALLY big difference between designing a hobby circuit and designing a mass production circuit that will be built in million piece quantities. So many people don't seem to understand the difference. I'm glad you do and happy I found your channel!
It’s been over 30 years since I learned how JFETs work and I’ve never used them since, so I had forgotten how they work and why I’d need them. Thanks for the refresher course and also what type of circuit you would use them in.
AFAIK it's similar to the BF245 fet. One thing about them is that they are very good at high frequencies. I built a small (boradcast band) FM transmitter from just 3 of those, they are a joy to work with.
Really nice way of explaining J-FETs Thank you. Since I never worked with J-FETs before I didn't know that they behave this way. This was a nice lesson for me. Thank you once again.
One interesting use for these are packaged as 2-lead devices and called current regulating diodes. They're sorted by their nominal current into part numbers.
JFETs have a great similarity with vacuum tubes in the sense that they conduct at 0 bias and need negative bias to stop conducting. They are very suitable as current sources (limiters)
I've used MMBFJ112 (50 mA) as overcurrent protection in a circuit. A resistor in series with the JFET and gate connected before the resistor. So when current increases, gate voltage increases and finally cuts off current.
I serviced vintage Fluke lab grade 6 1/2 digit 8502 multimeters and these 1989's jewels used tens and tens of discrete JFETs mainly as analog switches. Today, JFETs are practically no more used except as input devices inside high input impedance op-amps, as "start-up" circuits inside analog chips and inside miniature Electret microphones. They have been heavily replaced by MOSFETS which are waaaay easier to implant as analog switches since their gates are capacitively coupled to the channel instead of being P-N coupled to it so they may be positively and negatively biased without any problem. Digi-Key stock over 12 000 different models of MOSFT's over only 235 JFET's, demonstrating the huge difference in popularity. But, surprisingly, i did not know it, they are some kilovolt-level high-power SiC JFETs available. They definitely worth a look if you need depletion devices instead of enhancement FETs.
There isn't a "linear range" for a JFET. The flat lines in Figure 2 (at 13:05 et seq) are not equally spaced anywhere. What the graphs shows is that Id is independent of Vds when Vds > 2.5V. The spacing between the lines is simply what you could read off Figure 3, which shows Id vs Vgs at Vds=15V. It's pretty well-known that Id = Idss.( 1 - Vgs/Vgs(off) )² which is a parabolic relationship between Id and Vgs, not a linear one. Note that figures 2 and 3 show characteristics for a sample which happens to have Vgs(off) = -1.2V. The datasheet also supplies figures 4 & 5 for a sample that has Vgs(off) = -3.5V, and figures 6 & 7 for a sample that has Vgs(off) = -5.8V. The point is that there are no _typical_ Id vs Vgs characteristics for the 2N5457; they depend entirely on the Vgs(off) of the sample you have, and Onsemi only specify that it may be somewhere between -0.5V and -6V. If your supply voltage and your load resistor is fixed, your source resistor will have to be determined by trial-and-error to set the drain voltage somewhere around the midpoint between the supply voltage and the quiescent voltage at the source in order to maximise the output swing. If someone else uses a 270R source resistor, then they might get a sample of 2N5457 whose Vgs is -3.5V, and will find that with just -0.3V on the gate, they will get a source current of more than 1.1mA which will drive the circuit into saturation. They would have to set Vgs= -2.3V, to get Id=0.6mA, giving Rs=3.9K instead. For the Vgs(off)= - 5.8V sample, they would need around -4.5V to get Id=0.5mA which sets Rs=9K, which is unlikely to give any gain at all. Frankly, I'd pick a somewhat smaller load resistor and/or a higher supply voltage to allow for the wide variation of expected Vgs(off). The way in which you get linear amplification is by choosing a source resistor that is significantly larger than the reciprocal of the forward transfer admittance (Yfs). For the 2N5457, the Yfs is between 1 and 3 mS when Vgs=0, which equates to an equivalent intrinsic source resistance of between IK and 330R. So with the sample you have, that's going to be difficult. Calculating the gain is "kinda" not an ugly calculation. It's just a matter of estimating Yfs for the sample you have. Then the gain is Rd / (Rs + 1/Yfs). very similar to the way we calculate the gain of a common emitter stage. For the datasheet and your chosen operating point, I estimate Yfs should be around 2mS, so 1/Yfs = 500R. Your gain should therefore be about 10K / (270R + 500R) = 13. It's hard to tell from your scope, but I think your input is 100mVpp and the output is around 1.2Vpp, giving an actual gain of 12. My calculations are certainly good enough for me to get a decent ball-park figure for the gain. They also show the real difficulty of designing for JFET amplifiers that have such a wide range of their most critical parameter, Vgs(off). Just copying somebody else's circuit simply won't work for common-source JFET amps, and I'd be willing to bet that over 75% of people who just copied your circuit would end up with a transistor in saturation.
@Rex Schneider What is your point? Is it your tech? I can’t understand why you are such a prick unless it threatens your livelihood. May your knowledge die with you as a proper dickdead.
The gain of a FET amplifier is typically a trade-off between linearity, gain and output swing. At first you have to make sure, that you operate in a linear region. Therefore you have to work with relatively high bias currents. You need to define one and calculate your gate voltage to get it. From this gate voltage and the bias current you get your source resistance. In order to get maximum output swing you have to calculate your drain resistor to set the drain working point right at half of your operating voltage by use of the bias current. But if you use a higher drain resistance you can get more gain by sacrifice of some output swing. But if you work at 12 V and you need only 1 volt output swing, then you will set your drain resistance a bit higher to get more gain for a given transconductance. You will then end up with a lower operating point. But it doen't really matter so much.
Main difference between 2N5457 and J310 is current handling. J310 will pass tens of mA with 0v on gate, they can get very hot when passing this maximum current (Idss) so if you want to measure it, be quick. JFET characteristics can vary widely between ones of the same type, so using a trimpot for your source resistor, set to your estimated value, then adjust for best results.
For JFETs, the simplest way to evaluate gain is to look trans-conductance (mA/V) from the datasheet and multiply it to the load resistance. The source resistor creates negative feedback and you may bypass it with a cap to max the gain.
A JFET and a small BJT combo make for a great active test probe. Very high input impedance so you have minimal impact on the circuit under test. You can add a 1N60 diode for demodulating RF if you wish. More than enough gain to feed into a small amp and speaker. Run the whole lot off a 9V battery.
I went looking for JFETs a couple years ago to try my hand at making an active patch antenna. The only JFET I could find that seemed suitable was the NTE312 which I think is one of those replacement parts. Coupled with a BC559 PNP transistor, it surprisingly works pretty well for SWL. While just empirically experimenting I found the PNP second stage seemed to have less IM than an NPN. Not sure why but it would seem to be a linearity issue.
Hi IMSAI Guy. You didn't mention that these devices operate in the "DEPLETION MODE"!!!! Its right there at the top of the data sheet. Most MOSFETs operate in the "ENHANCEMENT MODE". Basically: Enhancement mode has enhanced pinch off at 0 volts and conducts only with greater voltage, (Most MOSFETS). Depletion mode does conduct at 0 volts an requires negative voltage to achieve pinch off, (All J-FETS). (Of course, there are P channel versions where all the polarities are reversed.) I said "Most" MOSFETs because there are some MOSFETs that do operate in the depletion mode. redrok
Hi, yes, you are correct. Imsai guy failed to explain the working of the depletion mode of the device he has taken up for study; or rather he is not aware of it, I believe. Someone below has said that the gate is protruding into the channel in case of jfet while in the MOSFET it's having insulation; he wants to attribute the difference in channel conduction this way...!! De VU2RZA
You can actually get current above Idss by putting a positive voltage on the gate. However, the gate will become a forward biased diode and limit the positive voltage.
I agree with you to a point. Yes, you can "experiment" with electronic components to a point. But unless you actually understand what the spec sheets are saying, you will earn a lot of blown parts in your efforts. Is that a bad thing? Maybe not if the parts are very cheap, but knowing some math can save a lot of blown parts as well. And once you understand OHM's law well, designing circuits gets interesting when you see the circuits you design, work just like the math you used.
Can you please elaborate on what allows you to ignore the 10K resistor when calculating the current for the bottom resistor? Is it because at -.2v, the JFET is nearly closed? (and had you picked a different operating point, where the JFET was partially open, might it still be a consideration?).
By picking a current of 0.7mA, it implies that the 10K load resistor is going to drop 7V across itself, setting the dc bias point of the drain at +5V with a 12V supply. I'd maybe have chosen a lower current, say 0.5mA (using an 820R resistor), and moved the bias point up to +7V to give a bit more headroom for the output swing - but we'd lose some gain and with outputs no more than 2Vpp or so, it's not critical.
Don't really know of why JFET's were originally developed other than the radio receiver front end/RF amplifier applications. One military application was those sensitive listening probes used in the Vietnam war. The gate has an arrow on the circuit symbol. Another common JFET is the 2N3819. The 2N3820 is a P channel JFET.
FETs are the solid state version of a vacuum triode and have been predicted decades before the bjt transistor was invented. But it took longer to actually produce one. They behave very much like there glass brothers&sisters, but on lower voltage range and unfortunately also on the input/gate side. Unfortunately for the area where according to hard core people, vacuum can never be replaced: guitars. There output signal is no problem for tubes, but for JFETs it‘s often too high. You need to use a high pinch off (negative) voltage type to not completely clip, but just use the „warm“ transfer characteristic.
How to choose a jfet? By input voltage, as an ac signal needs to stay in the „linear range“ between cut off voltage and 0V. If you have a 2..3Vpp guitar signal, a -4V cut off is needed for not hard limiting that signal. On the other side, the small cut off V jfets produce higher gain.
Getting a high gain from a single transistor is not always the best way. Sometimes its better to get a small gain with a good clean output, then add another similar stage to increase it further. I suppose it all depends on what your needs and limitations are.
The best thing is to think about a JFET as being virtually the same as a valve triode, including how you bias it. This is what so many people get wrong.
@@fromgermany271 Yes but why would you design a guitar preamp using a JFET with such a relatively low impedance source, JFETs are much better suited to RF amplifiers in the 100s of MHz range amongst other things because of their really high input impedance and lower RF noise.
So the 2N5457 is not so common these days so I put the circuit into LTSpice (I know "the map is not the terrain") and ran through a few different devices starting with a J201, which gave about the same results. Some other devices did poorly in this exact circuit, many others did well. Now I understand the concern that variations in the device may produce wildly different results, but that is the cardinal rule in much of electronics to "design defensively" and not count on a particular value for any device's performance when ranges are specified by the manufacturer. For example 100 may be a good place to start when working calculations involving a particular BJT's beta, the smart move is to set the beta externally based on the range given. Then, as needed, do what you have to with the next stage once you get the first stage working. The example given in the video seemed good one to me: The "goal" of the stage was having a JFET front end because of some intrinsic property of the device, in this case high impedance. So once you pick your device and get that working, go to the "toolbox" and grab something standard to handle the variations. In the example an op-amp was that tool. Do what you need to do in that stage to accommodate what the JFET may present. Often (almost always) this is an iterative process where many cycles are required to get a solid working design. And please note that I am in no way a professional in circuit design. I have been playing with various things for around 45 years but never had to earn a living with any of it.
It really does seem that way. Reading through classic designs from e.g. Pass Labs, very often any JFET mentioned is no longer made. I’m not sure why, I wouldn’t think demand for audio gear has changed over the years
Thanks for the video! It helps to better to understand how JFETs work. Will a 2N5457 work well with a 25mm condenser capsule? (I have B version of it) Or is it better to use 2SK170 BL (in the manual they say - Low Noise Audio Amplifier Applications)?
I would go with the sk170 for noise. you will also need a good 1 gig-ohm resistor for the gate. there is an odd jfet that has a built in gate resistor, used in the really cheap and small mic capsules 2SK596
I wonder, has anyone in here ever been able to get FETs from... ahem, unofficial sellers, like Amazon vendors, ebay or banggood? I have a feeling one would receive fet-marked BJTs...
I’ve ordered both signal and power MOSFETs from no-name sellers on Amazon and they work just fine. I’ve also gotten samples from the manufacturers for free, but I’m sure it’s more difficult to create accounts today without talking to sales people.
@@NavinF The 'free stuff' ... or samples, from manufacturers... was ruined by free-loaders egged on by TH-cam influencers to take advantage. Now if you want to play, you pay. No more boxes full of free samples in the basement: never intended to be used. I rant. Yes. But, it's made it much harder for countless students, and small businesses who actually used, and appreciated these samples.
I order from Ali Express quite a bit... mainly for the jelly bean components where the details don't matter so much. E.g., I have hundreds of LM358s, and they've all worked just fine. If I remember correctly, I think IMSAI Guy did a video on some counterfeits, so it's DEFINITELY a risk... but I haven't been bit hard yet!
@@NavinF In general... Nope. The 'get stuff free by mail' was was a popular way to make money by showing people how to get 'stuff for free, by mail.' TH-cam 'influence' really put the nail through the heart of free samples.
Just wondering if you could get the 12V vacuum tubes that work with a B+ of not higher than 12V and less than 120V would that not do the same job as a JFET.
How does V=IR (ohms law) calculate the value of the 270 ohm resistor? Don't you need to factor in the impedance of the transistor and resistances of both the 4.7M and 10K resistors?
It's still not 100% clear to me how the self-biasing works. I see that the 270 ohm resistor is involved, but how it magically biases the gate to -0.2 volts, I don't understand. I would have assumed that the AC coupling to the gate through the capacitor meant that the input would swing up and down around the 0 volt point (which I think it does relative to ground, but somehow doesn't relative to the voltage at the source?)
Another way to explain this is: initially when power is applied, the Vgs voltage is 0V (no current flow), and therefore the JFET is turned on fully. Then the current through the 270ohm resistor will rise up to a point when the voltage across it becomes positive with respect to ground. Since the Gate voltage is tied to ground through the 4.4M resistor, that means it’s DC voltage is always 0V, but the Source voltage goes up. When the Source voltage gets to ~0.2V the JFET gets to its steady state because the voltage at the Gate is negative 0.2V with respect to the Source. If the DC current tries to increase the voltage at the Source goes up, but that means the Vgs is more negative, turning off the FET a bit more. So it self regulates the Source voltage depending on the value of the Source resistor. I think that’s what is meant by self-biasing.
I would add that because there is zero DC current flow through the 4.4M resistor it is at the same DC potential at both ends> Therefore the gate is tied (DC) to ground. The AC signal sees 4.4M resistance as signal current tries to flow though the 4.4M.
Imagine a 12 meter tall building, with you and your friend standing in front of it. From ground level, you both say the building is 12 m tall (+12 m). Only you Go up a floor, say 3 meters. To you, the ground is lower by 3 meters (-3 m). When you are standing at 3 m (which is your reference), you can say, the building is 9 meters above you (+9 m) and your friend is 3 meters below (-3 m). For your friend who is standing on the ground, the building is still +12 meters and you are standing at +3 meters. Your name is source and your friend name is gate. Source must stay at a negative voltage to gates. Regards.
The 2N7000 is fine as long as you don't plan on a continuous drain current in excess of 200mA. Also if you're driving the gate with only 4.5V, its Rds(on) could be as much as 5R, so you may need to be careful of power dissipation.
Without the bypass capacitor, it's low frequency response is limited by the 0.01μF input capacitor and the 4M4 gate resistor, giving a roll-off at 3.6Hz. Adding a 10μF bypass capacitor adds a dominant low frequency roll-off around 30Hz (as well as loads more distortion and a bit more gain). The high frequency response is anybody's guess. Onsemi claims it's designed for audio applications, but there's not much shown on the datasheet to limit its bandwidth apart from the 3pF max reverse transfer capacitance. That would likely limit the hf to audio applications when used with a high input impedance and reasonable gain.
It will cause the gate-source junction to become forward biased and so the gate will conduct driving a current from gate to source. It's hard to predict what that current would be without knowing the rest of the circuit, source impedance, etc. but it's not a viable way of increasing Id above Idss for any JFET. Some samples of 2N5457 might have an Idss as high as 9mA and you can always pick a different JFET with a higher typical Idss anyway, so it may just be a case of trying several until you get one that works for your application.
There is no enhancement JFET. It only „pinch off“, just like a vacuum triode, but while the tube allows (limited) positiv gate voltage, a JFET does not really. It then degrades to diode with very low backward current.
Could one of these transistors be used to simulate an electret microphone? I want to pass audio from a headphone amplifier (radio, tape deck, DAC, etc) into the microphone (headset) input on a cell phone.
@@IMSAIGuy 10K across the output and capacitor from the wiper to the input? Isn't the mic supposed to look like a 1K resistor that varies with the sound pressure?
Still not seeing how using a discrete jFET has any advantage over using a TL082. The jFET-input op amp gives you both the high input impedance and the high gain. Only possible reason I can think of for using this is, again, if you're going to bias this circuit closer to 0.5mA to get That Sweet, Sweet Even Harmonic Distortion(TM) commonly associated with triodes.
@@petersage5157 while the JFET I call out is just one I had, companies like Rode and Neumann use a JFET and 1G ohm gate resistor, this is low enough noise to send to a mic preamp. look at the datasheet for a INA217. this preamp has typical specs for mic preamp. it is much lower than the TL082.
Thanks for the video. It was a good intro in JFET. I'm not sure the pedantic value of confusing people with a drawing more depicting a MOSFET than a JFET. The gate(s) of JFET definitely protrudes into the body of the device. This explains why all the channel current will flow into the gate, if the JFET is biased incorrectly. The MOSFET gate is insulated from the body of the device. A gate current cannot flow in a MOSFET, in this way.
@@markanderson8066 A rectangle (drawn for the body of the JFET, with a parallel line adjacent to one of rectangles longer sides (for the gate of the JFET) is not the JFET schematic symbol (if that's what you mean.) The gate of the JFET, in any pictorial representation of the physical construction of a JFET, ought to show the gate material, of the JFET, protruding into the body of the JFET material. Check any reference on the physical construction of a JFET.
The calculated value of resistor was supposed to be 286 ohm not 285, you have to round the values sir. Actual calculation of JFET amplifier is pretty verbose and mathematically nerdy. Yet, they way you did it is pretty good
So why would these be chosen for the front end of a receiver. I remember when I bought my FT-847 They made a thing of the front end containing Jfets. Is it because of the high impedance and low voltage shift on the gate to make a stable linear amplifier?
Maybe it is as you point out (depends on the amplifier design). The big advantage of JFET is they are lower noise than BJT, and most MOSFET. It is a big deal if you can trim the noise in the front end: you'll get a better signal to noise ratio, and a 'better' radio. I think IMSAI Guy did something on this topic in an earlier video.
Lower noise and high linearity (thus low distortion and high dynamic range) is the thing. You can get a noise figure around a couple dB with a well-designed JFET LNA. As to the physics reasons they have those attributes... I wish I knew!
With BJTs, their best noise performance tends to come with low impedance sources, but with JFETs the opposite is true. The 2N5457 datasheet show typical noise figures of 14dB with 3K source resistance, falling to around 1dB at 1M source resistance. Horses for courses.
Why do you imagine the negative terminal is called the source and the positive is called the drain? Because current flows from negative to positive not the opposite.
Conventional current flows from positive to negative. If you're concerned about the actual charge carriers (heaven knows why?), then in an n-channel FET, the charge carriers flow from negative to positive; while in a p-channel FET, the charge carriers flow from positive to negative. I think it best to just stick to conventional current and avoid any worries about sub-atomic particles and quantum mechanics.
Please, DO NOT EXPECT HIGH (almost infinite) INPUT IMPEDANCE on the GATE of FET amplifier. This is completely wrong. They have a large (almost infinite) input resistance (because of insulation of the gate), but the internal structure of the FET (gate is isolated) forms two capacitors (any insulation forms capacitor, this is how capacitors are made - two electrodes with a insulation between them), one between gate and source, another between gate and drain. That one between gate and source is just an capacitor, that project its capacitance into a imaginary part of input impedance depending on the frequency. It just sits there doing its regular job, because the voltage between gate and source is almost constant, so this capacitor just pulls your input impedance to the ground, nothing more. The second one is a little bit tricky, because the voltage between the drain and the gate is not constant, it's multiplied by the gain of that circuit, which means, that you have to catch this voltage change by the input current of that capacitor, which in fact multiplies its capacity by the gain of that amplifier. This is well known from the vacuum tubes as a Miller effect, but it is something common to almost everything that have at least some gain. So if we speak about low frequencies (high and higher frequencies is another story), then the gate of the FET amplifier has almost infinite resistance, but relatively low input capacitance (depends on frequency), which forms pure capacitive load - the input behaves like an almost perfect capacitor. And when it comes to AC, then the capacitor can load a lot.
That formula only works if Rs is much larger than 1/gm, which it isn't in this case. Here the gain can be calculated as RL/(1/gm+Rs), with gm given in the datasheet (listed as Yfs) as typically 3000 umohs the gets you an expected gain of 16x. Still more than we see here but there is a very big variation expected on gm (it's given as something between 1000 and 5000 umohs)
@@rjordans Thanks, you are right. I forgot all about 1/gm and probably ro. The degenerated gain Gm = (gmro)/(Rs+[1+gm+gmb)Rs]ro) which is approximately Gm =gm/(1+gmRs) and voltage gain then is gmRL/(1+gmRs) if gmRs >> 1 this would become RL/Rs if it is not then you should use gmRL/(1+gmRs) here gm is about say 0.001 then the gmRs is about 0.001*270 or 0.27 So Av= (0.001*10K)/(1+0.001*270) which is 7.8 for gm of 1000umohms , which is very close to 10 and using a actual gm value in the range of 1000 to 5000 we should get Av=10
@@rjordans I think the Yfs will be rather less than 3mS since that's quoted at Vgs=0. Yfs is the slope of the Id vs Vgs curve and that shows that Yfs will be lower as Vgs goes more negative. I took it to be more like 2mS, giving 1/Ys as 500R and yielding a gain of 13. The output on the scope seems close to 1.2Vpp, but it's difficult to estimate the input with any precision, although if a signal generator was used, there must have been a temptation to set 100mVpp. So I think the real gain was more like x12. Obviously with that biasing, the non-linear 1/Yfs dominates the source resistance, so expect a noticeable amount of 2nd harmonic distortion.
Hey. You might say you are no expert but the way you look at circuits really help us all technicians and engineers.
Tks so much for helping!!
Thanks for the easy approach to understand the JFET characteristics by showing us its circuit under operation !
Absolutely, a breadboard demo makes ALL the difference and I love your statements of how you selected the JFET.
An interesting continuation of this would be to expand on your discussion of distortion by using a spectrum analyzer to show the effects at different operating points.
Too much nitpicking in this video's comments.
I learned something; thanks for sharing.
You know you're listening to an optics guy when he says that's the "cyan trace." :-)
guilty
The thing that really knocked my socks off is what appeared at 15:18 . When I decided to learn RPN, I naturally looked towards HP. The 32S made its appearance then and the dots matrix numbers quickly won me over. The simple clean keyboard was magnificent. I grew to love this sooooo much that I had to get a backup. THEN the display failed and HP "repaired" it by giving me a brand new 32SII . Much more cluttered keyboard and see through pages in the manual. This is when the love story ended. I had good times squeezing the most out of it. My most complex program calculated the miles per gallon and estimated miles left in the tank plus it also told me when it was time to change my engine oil. One answer was before the decimal point, one answer was after the decimal point and the expansion from 4DP to 11DP said it was time to buy engine oil. On, I learnt a lot about JFETS here also.
HP repaired my HP67C in the early 80s by giving me a brand new one in a box. I had no idea that was going to happen. They called to tell me it was ready to go. I went to pick it up and was very confused when they handed me a brand new calculator. The repair rate was a flat $99 no matter the repair. I still have it although it needs some TLC. I wonder what would happen this time. Likely no warehouse full of HP67Cs out there, huh.
Wow. This was perfect for me. I'm a tube amp guy so this was cool. These things work like tubes, with the obvious differences of course. Looking forward to watching more of your videos.
Also a very easy circuit to do impedance matching. Common use is to allow a very high impedance microphone, like D-104, to be used with modern day transceivers that use lower impedance microphones. I used a 10M ohm resistor from the input to ground to accommodate the D-104 microphone.
A quick question. If you know. Can you tell me the difference between an amplified D-104 typically used for cb but also ham and one made in the 1950's long before the 9 volt battery amp was ever though of other than lack of the amp? Thanks.
@@W1RMD The original D1 04 used an extremely high impedance Rochelle salt crystal. The newer ones use a lower quality element and use the amplifier as impedance matching circuit
@@bobkozlarekwa2sqq59 I've noticed that I HAVE to run the amp in this mike with this D TEN FOUR. Even though I have a 1963 Johnson Invader 2000 transmitter. I've tried to bypass the amp, but it doesn't sound good. I have to use the amp and keep it turned all the way down to get a good sounding audio. It's too bad they've turned this (the D104 product) into a "CB" mike over the years and I understand that you're dealing with 5 watts on a CB so it needs all the help it can get , but that's not an issue with a large transmitter. That I know of, you can no longer get the correct element, or the mike anymore. 73's
I wish my Electroncis professor at the uni had this practical and down to earth approach to circuits, instead, he would flood us poor students with tons of equations and formulas - and as much as I loved maths I hated the completely theoretical approach. That made me decide to take a software career even if I loved electronics more. Well, I'm 52 now and picking it up after a 30 years hiatus and channels like yours are bringing the love back, even stronger actually! Thanks for your great videos.
The math in most engineering science did not start with the chalk board but with experimentation.
Teaching should start with more practical applications and work the math backward from there. More intuitive and easier to remember.
I love your channel. You teach this stuff in a way that helps me understand stuff I could never get my head around on my own. Thank you
I've been designing circuits for over 15 years and this is very similar to a first order design I do to solve a problem. Of course sometimes you need to calculate then simulate and understand 100%. But sometimes you just need something you can trust as a foundation and go from there. Quite refreshing to see Im not alone in how I approach things.
You've explained well the negative potential between G and D! The best part of the game is calculation and next check values and oh yeah the pleasure of a positive feedback. My favorite moment is when you move the camera and... bang‼️... the circuit is already playing on the breadboard 😂👍❤️
You taught me in very effective way that how to choose source resistor to self bias JFET amplifier, keep going 🎉
Great video! I agree about not having to run all the numbers on everything. One of the important things in engineering is to not over-design, or spend too much time on things. You're operating in the safe range of the device, it's used as part of a hobby, and you have the gain you want over the frequency range you want, by measurement. If it were for a production run of a thousand, then sure, do some math. Even for production use as part of a test setup, you'd use gain figures by measurement, not calculation, because the only thing that matters there is the actual value, not the theoretical.
There are depletion mode MOSFETs which act aomewhat like JFETs, at least in part of their characteristic. They are fairly rare and not much used, but they do exist. One nice property of those is that, as opposed to a JFET, you can allow the gate to swing to a positive voltage, not only negative. Because it's not a junction, and it won't become directly biased if it gets positive.
As for choosing JFETs for an application - in my type of applications (voltage-controlled current sources mostly) I start by looking at JFETs with an Idss (drain current when Vgs=0) in the range that I am interested in. Idss also happens to be the maximum current that the JFET will conduct. Everything flows from that point on.
"I chose this particular JFET transistor because I have a lot of them, and it's a very very very common one." or was, until it went very very very EOL.
Central is still selling them, but naturally it's at their "where else are you gonna go?" prices.
Yep, and the only ones I found were a lower current version.
Mouser. 30,000 in stock. Buck and a half each in 10's.
@@glasslinger those would be the offering from Central, yes. Hard to justify paying that when we aren't all that far removed from paying buck and a half *for* 10. Better off finding a substitute IMO.
@@grantlack If you are that strapped for cash you are not going to succeed. Rethink your overall program.
@@glasslingerso far, I've managed to maintain my frivolous affinity for RN-series resistors without expressing any sneering contempt for those poverty-stricken Xicon users. After all, they do the same job for less money.
It's a great video from a great instructor covering general use JFETs. Market forces have dictated that this particular JFET is no longer available at jellybean prices. The content is still applicable to other parts at a lower price point, or indeed for those who are determined to pay the premium for the named part. None of these strike me as controversial statements or an invitation for financial advice.
This is a great tutorial- very clear explanations. Look forward to more!
So happy I recently found your channel. You’re very well spoken and are now one of my primary sources for my DIY projects :) keep the awesome content coming. (Especially excited about your new analog synth series)
There's a REALLY big difference between designing a hobby circuit and designing a mass production circuit that will be built in million piece quantities.
So many people don't seem to understand the difference. I'm glad you do and happy I found your channel!
It’s been over 30 years since I learned how JFETs work and I’ve never used them since, so I had forgotten how they work and why I’d need them. Thanks for the refresher course and also what type of circuit you would use them in.
AFAIK it's similar to the BF245 fet. One thing about them is that they are very good at high frequencies. I built a small (boradcast band) FM transmitter from just 3 of those, they are a joy to work with.
BF245 is an excellent product. Used them in oscillators, and they worked like hell
Really nice way of explaining J-FETs Thank you. Since I never worked with J-FETs before I didn't know that they behave this way. This was a nice lesson for me. Thank you once again.
Glad it was helpful!
One interesting use for these are packaged as 2-lead devices and called current regulating diodes. They're sorted by their nominal current into part numbers.
JFETs have a great similarity with vacuum tubes in the sense that they conduct at 0 bias and need negative bias to stop conducting. They are very suitable as current sources (limiters)
Good point...
I've used MMBFJ112 (50 mA) as overcurrent protection in a circuit. A resistor in series with the JFET and gate connected before the resistor. So when current increases, gate voltage increases and finally cuts off current.
For sensitive laser diode can be perfect.
I serviced vintage Fluke lab grade 6 1/2 digit 8502 multimeters and these 1989's jewels used tens and tens of discrete JFETs mainly as analog switches.
Today, JFETs are practically no more used except as input devices inside high input impedance op-amps, as "start-up" circuits inside analog chips and inside miniature Electret microphones.
They have been heavily replaced by MOSFETS which are waaaay easier to implant as analog switches since their gates are capacitively coupled to the channel instead of being P-N coupled to it so they may be positively and negatively biased without any problem.
Digi-Key stock over 12 000 different models of MOSFT's over only 235 JFET's, demonstrating the huge difference in popularity.
But, surprisingly, i did not know it, they are some kilovolt-level high-power SiC JFETs available.
They definitely worth a look if you need depletion devices instead of enhancement FETs.
There isn't a "linear range" for a JFET. The flat lines in Figure 2 (at 13:05 et seq) are not equally spaced anywhere. What the graphs shows is that Id is independent of Vds when Vds > 2.5V. The spacing between the lines is simply what you could read off Figure 3, which shows Id vs Vgs at Vds=15V. It's pretty well-known that Id = Idss.( 1 - Vgs/Vgs(off) )² which is a parabolic relationship between Id and Vgs, not a linear one.
Note that figures 2 and 3 show characteristics for a sample which happens to have Vgs(off) = -1.2V. The datasheet also supplies figures 4 & 5 for a sample that has Vgs(off) = -3.5V, and figures 6 & 7 for a sample that has Vgs(off) = -5.8V. The point is that there are no _typical_ Id vs Vgs characteristics for the 2N5457; they depend entirely on the Vgs(off) of the sample you have, and Onsemi only specify that it may be somewhere between -0.5V and -6V.
If your supply voltage and your load resistor is fixed, your source resistor will have to be determined by trial-and-error to set the drain voltage somewhere around the midpoint between the supply voltage and the quiescent voltage at the source in order to maximise the output swing. If someone else uses a 270R source resistor, then they might get a sample of 2N5457 whose Vgs is -3.5V, and will find that with just -0.3V on the gate, they will get a source current of more than 1.1mA which will drive the circuit into saturation. They would have to set Vgs= -2.3V, to get Id=0.6mA, giving Rs=3.9K instead. For the Vgs(off)= - 5.8V sample, they would need around -4.5V to get Id=0.5mA which sets Rs=9K, which is unlikely to give any gain at all. Frankly, I'd pick a somewhat smaller load resistor and/or a higher supply voltage to allow for the wide variation of expected Vgs(off).
The way in which you get linear amplification is by choosing a source resistor that is significantly larger than the reciprocal of the forward transfer admittance (Yfs). For the 2N5457, the Yfs is between 1 and 3 mS when Vgs=0, which equates to an equivalent intrinsic source resistance of between IK and 330R. So with the sample you have, that's going to be difficult.
Calculating the gain is "kinda" not an ugly calculation. It's just a matter of estimating Yfs for the sample you have. Then the gain is Rd / (Rs + 1/Yfs). very similar to the way we calculate the gain of a common emitter stage. For the datasheet and your chosen operating point, I estimate Yfs should be around 2mS, so 1/Yfs = 500R. Your gain should therefore be about 10K / (270R + 500R) = 13. It's hard to tell from your scope, but I think your input is 100mVpp and the output is around 1.2Vpp, giving an actual gain of 12. My calculations are certainly good enough for me to get a decent ball-park figure for the gain. They also show the real difficulty of designing for JFET amplifiers that have such a wide range of their most critical parameter, Vgs(off). Just copying somebody else's circuit simply won't work for common-source JFET amps, and I'd be willing to bet that over 75% of people who just copied your circuit would end up with a transistor in saturation.
@Rex Schneider What is your point? Is it your tech? I can’t understand why you are such a prick unless it threatens your livelihood. May your knowledge die with you as a proper dickdead.
Make some instructional videos as well :)
Yeah this happened to me while using a fixed resistor from a schematic.. so much variability in parts and it's so difficult to get the sweet spot.
The bias variability is the reason JFETs are not used more.
The gain of a FET amplifier is typically a trade-off between linearity, gain and output swing. At first you have to make sure, that you operate in a linear region. Therefore you have to work with relatively high bias currents. You need to define one and calculate your gate voltage to get it. From this gate voltage and the bias current you get your source resistance.
In order to get maximum output swing you have to calculate your drain resistor to set the drain working point right at half of your operating voltage by use of the bias current. But if you use a higher drain resistance you can get more gain by sacrifice of some output swing.
But if you work at 12 V and you need only 1 volt output swing, then you will set your drain resistance a bit higher to get more gain for a given transconductance. You will then end up with a lower operating point. But it doen't really matter so much.
Fortunately some (6 string) applications are happy with the tube like non-linearities. 😂
This is how they ought to teach electronics. I remember my college back in the day making this as dry and boring as possible.
The FET that puts the FET in FET Compressor *1176
Excellent explanation!
Main difference between 2N5457 and J310 is current handling. J310 will pass tens of mA with 0v on gate, they can get very hot when passing this maximum current (Idss) so if you want to measure it, be quick. JFET characteristics can vary widely between ones of the same type, so using a trimpot for your source resistor, set to your estimated value, then adjust for best results.
Very refreshing tutorial...cheers.
very clear and enjoyable, thank you! Yes it would be great to see the Neumann schematic
the link is in the description
For JFETs, the simplest way to evaluate gain is to look trans-conductance (mA/V) from the datasheet and multiply it to the load resistance. The source resistor creates negative feedback and you may bypass it with a cap to max the gain.
A JFET and a small BJT combo make for a great active test probe. Very high input impedance so you have minimal impact on the circuit under test. You can add a 1N60 diode for demodulating RF if you wish. More than enough gain to feed into a small amp and speaker. Run the whole lot off a 9V battery.
You are a legend, sir!
A great refresher, thank you. I am a new subscriber.
You can find JFET's inside of electret microphones, not branded but working pretty well în radiofrequency ranges..!
many of those also have the gate resistor built into the device.
I went looking for JFETs a couple years ago to try my hand at making an active patch antenna. The only JFET I could find that seemed suitable was the NTE312 which I think is one of those replacement parts. Coupled with a BC559 PNP transistor, it surprisingly works pretty well for SWL. While just empirically experimenting I found the PNP second stage seemed to have less IM than an NPN. Not sure why but it would seem to be a linearity issue.
Much fun. Thanks a lot. It's appreciated.
Hi IMSAI Guy.
You didn't mention that these devices operate in the "DEPLETION MODE"!!!! Its right there at the top of the data sheet.
Most MOSFETs operate in the "ENHANCEMENT MODE".
Basically:
Enhancement mode has enhanced pinch off at 0 volts and conducts only with greater voltage, (Most MOSFETS).
Depletion mode does conduct at 0 volts an requires negative voltage to achieve pinch off, (All J-FETS).
(Of course, there are P channel versions where all the polarities are reversed.)
I said "Most" MOSFETs because there are some MOSFETs that do operate in the depletion mode.
redrok
Hi, yes, you are correct. Imsai guy failed to explain the working of the depletion mode of the device he has taken up for study; or rather he is not aware of it, I believe.
Someone below has said that the gate is protruding into the channel in case of jfet while in the MOSFET it's having insulation; he wants to attribute the difference in channel conduction this way...!!
De VU2RZA
Yes he did mention depletion mode early on. lol
Always an interesting video from IMSAI Guy. So when i understand correctly JFET resembles a electron vacuum tube the most ....
Great explanation.
Excellent video.
You can actually get current above Idss by putting a positive voltage on the gate. However, the gate will become a forward biased diode and limit the positive voltage.
I still have a bunch of MPF102 JFETs I bought at a hamfest decades ago...
I don't know why they quit making them, they were like the 2n2222 of the JFET family.
@@edwardneuman6061 They're literally the only JFET I could name off the top of my head! I didn't know they'd stopped making them - that's too bad.
very well explained
Great video! Appreciate it!
I agree with you to a point. Yes, you can "experiment" with electronic components to a point. But unless you actually understand what the spec sheets are saying, you will earn a lot of blown parts in your efforts. Is that a bad thing? Maybe not if the parts are very cheap, but knowing some math can save a lot of blown parts as well. And once you understand OHM's law well, designing circuits gets interesting when you see the circuits you design, work just like the math you used.
Seems just like a triode! No wonder these get used in solid-state adaptations of tube circuits
I'm subscribing just because of the IMSAI 8080 reference.
I have a playlist for 'IMSAI'
Thank you.
Can you please elaborate on what allows you to ignore the 10K resistor when calculating the current for the bottom resistor? Is it because at -.2v, the JFET is nearly closed? (and had you picked a different operating point, where the JFET was partially open, might it still be a consideration?).
By picking a current of 0.7mA, it implies that the 10K load resistor is going to drop 7V across itself, setting the dc bias point of the drain at +5V with a 12V supply. I'd maybe have chosen a lower current, say 0.5mA (using an 820R resistor), and moved the bias point up to +7V to give a bit more headroom for the output swing - but we'd lose some gain and with outputs no more than 2Vpp or so, it's not critical.
Don't really know of why JFET's were originally developed other than the radio receiver front end/RF amplifier applications. One military application was those sensitive listening probes used in the Vietnam war. The gate has an arrow on the circuit symbol. Another common JFET is the 2N3819. The 2N3820 is a P channel JFET.
FETs are the solid state version of a vacuum triode and have been predicted decades before the bjt transistor was invented. But it took longer to actually produce one. They behave very much like there glass brothers&sisters, but on lower voltage range and unfortunately also on the input/gate side. Unfortunately for the area where according to hard core people, vacuum can never be replaced: guitars. There output signal is no problem for tubes, but for JFETs it‘s often too high. You need to use a high pinch off (negative) voltage type to not completely clip, but just use the „warm“ transfer characteristic.
How to choose a jfet?
By input voltage, as an ac signal needs to stay in the „linear range“ between cut off voltage and 0V. If you have a 2..3Vpp guitar signal, a -4V cut off is needed for not hard limiting that signal. On the other side, the small cut off V jfets produce higher gain.
Getting a high gain from a single transistor is not always the best way. Sometimes its better to get a small gain with a good clean output, then add another similar stage to increase it further. I suppose it all depends on what your needs and limitations are.
The best thing is to think about a JFET as being virtually the same as a valve triode, including how you bias it.
This is what so many people get wrong.
Only with a much lower input swing before you clip your signal. No issue for HF, but tricky with 2..3Vpp output of a guitar pickup.
@@fromgermany271 Yes but why would you design a guitar preamp using a JFET with such a relatively low impedance source, JFETs are much better suited to RF amplifiers in the 100s of MHz range amongst other things because of their really high input impedance and lower RF noise.
So the 2N5457 is not so common these days so I put the circuit into LTSpice (I know "the map is not the terrain") and ran through a few different devices starting with a J201, which gave about the same results. Some other devices did poorly in this exact circuit, many others did well. Now I understand the concern that variations in the device may produce wildly different results, but that is the cardinal rule in much of electronics to "design defensively" and not count on a particular value for any device's performance when ranges are specified by the manufacturer. For example 100 may be a good place to start when working calculations involving a particular BJT's beta, the smart move is to set the beta externally based on the range given. Then, as needed, do what you have to with the next stage once you get the first stage working.
The example given in the video seemed good one to me: The "goal" of the stage was having a JFET front end because of some intrinsic property of the device, in this case high impedance. So once you pick your device and get that working, go to the "toolbox" and grab something standard to handle the variations. In the example an op-amp was that tool. Do what you need to do in that stage to accommodate what the JFET may present. Often (almost always) this is an iterative process where many cycles are required to get a solid working design.
And please note that I am in no way a professional in circuit design. I have been playing with various things for around 45 years but never had to earn a living with any of it.
i got 2n5485, 2SK30A, BF256 in To-92.
good video.
With the modern op amps now available no reason to use these except for RF applications.
almost every dynamic microphone uses this circuit. millions of them
@@IMSAIGuy Old stuff. Low end stuff.
sorry to disagree, but Rode, Neumann all make modern and expensive microphones with a JFET with 1G ohm gate resistor.
great content!
does anyone still make these or the 2n5458? JFETs in general seem hard to come by these days.
It really does seem that way. Reading through classic designs from e.g. Pass Labs, very often any JFET mentioned is no longer made. I’m not sure why, I wouldn’t think demand for audio gear has changed over the years
Well done. Thanx.
Great video, explanations, thank you 👌
Wonder what the output impedance would be ?
Some of my best friends are JFETS!
good video, there are also enhancement mode jfets
Thanks for the video! It helps to better to understand how JFETs work.
Will a 2N5457 work well with a 25mm condenser capsule? (I have B version of it)
Or is it better to use 2SK170 BL (in the manual they say - Low Noise Audio Amplifier Applications)?
I would go with the sk170 for noise. you will also need a good 1 gig-ohm resistor for the gate. there is an odd jfet that has a built in gate resistor, used in the really cheap and small mic capsules 2SK596
for a fancy one, Rode uses the IF4520 (SNJ450) SNJ450113
see my video: th-cam.com/video/8IS7rScFL2U/w-d-xo.html
I wonder, has anyone in here ever been able to get FETs from... ahem, unofficial sellers, like Amazon vendors, ebay or banggood?
I have a feeling one would receive fet-marked BJTs...
I’ve ordered both signal and power MOSFETs from no-name sellers on Amazon and they work just fine. I’ve also gotten samples from the manufacturers for free, but I’m sure it’s more difficult to create accounts today without talking to sales people.
@@NavinF The 'free stuff' ... or samples, from manufacturers... was ruined by free-loaders egged on by TH-cam influencers to take advantage. Now if you want to play, you pay. No more boxes full of free samples in the basement: never intended to be used. I rant. Yes. But, it's made it much harder for countless students, and small businesses who actually used, and appreciated these samples.
@@willthecat3861 Eh the heyday of free samples was over 15 years ago, well before TH-cam EEs became popular. Your rant is kinda late.
I order from Ali Express quite a bit... mainly for the jelly bean components where the details don't matter so much. E.g., I have hundreds of LM358s, and they've all worked just fine. If I remember correctly, I think IMSAI Guy did a video on some counterfeits, so it's DEFINITELY a risk... but I haven't been bit hard yet!
@@NavinF In general... Nope. The 'get stuff free by mail' was was a popular way to make money by showing people how to get 'stuff for free, by mail.' TH-cam 'influence' really put the nail through the heart of free samples.
Just wondering if you could get the 12V vacuum tubes that work with a B+ of not higher than 12V and less than 120V would that not do the same job as a JFET.
How does V=IR (ohms law) calculate the value of the 270 ohm resistor? Don't you need to factor in the impedance of the transistor and resistances of both the 4.7M and 10K resistors?
14:50
Nice video
So much like valves (tubes).
thanks
Clever.
Does the gate voltage have to be genuinely negative (i.e. below zero) or just negative with respect to the source voltage?
just with respect to
Enlightenment comes when you realize that there is no such thing as genuinely negative. All voltage is relative.
Awesome,👍👏💕🙏
It's still not 100% clear to me how the self-biasing works. I see that the 270 ohm resistor is involved, but how it magically biases the gate to -0.2 volts, I don't understand. I would have assumed that the AC coupling to the gate through the capacitor meant that the input would swing up and down around the 0 volt point (which I think it does relative to ground, but somehow doesn't relative to the voltage at the source?)
the source is at +0.2 V the gate is at 0.0 V so the gate is 0.2 volts LOWER than the source.
Another way to explain this is: initially when power is applied, the Vgs voltage is 0V (no current flow), and therefore the JFET is turned on fully. Then the current through the 270ohm resistor will rise up to a point when the voltage across it becomes positive with respect to ground. Since the Gate voltage is tied to ground through the 4.4M resistor, that means it’s DC voltage is always 0V, but the Source voltage goes up. When the Source voltage gets to ~0.2V the JFET gets to its steady state because the voltage at the Gate is negative 0.2V with respect to the Source. If the DC current tries to increase the voltage at the Source goes up, but that means the Vgs is more negative, turning off the FET a bit more. So it self regulates the Source voltage depending on the value of the Source resistor. I think that’s what is meant by self-biasing.
I would add that because there is zero DC current flow through the 4.4M resistor it is at the same DC potential at both ends> Therefore the gate is tied (DC) to ground. The AC signal sees 4.4M resistance as signal current tries to flow though the 4.4M.
Imagine a 12 meter tall building, with you and your friend standing in front of it.
From ground level, you both say the building is 12 m tall (+12 m).
Only you Go up a floor, say 3 meters. To you, the ground is lower by 3 meters (-3 m).
When you are standing at 3 m (which is your reference), you can say, the building is 9 meters above you (+9 m) and your friend is 3 meters below (-3 m).
For your friend who is standing on the ground, the building is still +12 meters and you are standing at +3 meters.
Your name is source and your friend name is gate. Source must stay at a negative voltage to gates.
Regards.
where's the roadrunner & wiley-E cyote? the tale of 2 alphas = 1 Omega!!
Thanxs
How about a *2N7000* is a great device for driving relays when a *2N3904* won't do.
The 2N7000 is fine as long as you don't plan on a continuous drain current in excess of 200mA. Also if you're driving the gate with only 4.5V, its Rds(on) could be as much as 5R, so you may need to be careful of power dissipation.
I would like to see the frequency response of this amplifier.
Without the bypass capacitor, it's low frequency response is limited by the 0.01μF input capacitor and the 4M4 gate resistor, giving a roll-off at 3.6Hz. Adding a 10μF bypass capacitor adds a dominant low frequency roll-off around 30Hz (as well as loads more distortion and a bit more gain).
The high frequency response is anybody's guess. Onsemi claims it's designed for audio applications, but there's not much shown on the datasheet to limit its bandwidth apart from the 3pF max reverse transfer capacitance. That would likely limit the hf to audio applications when used with a high input impedance and reasonable gain.
Jungle FET
What happens if you put a positive voltage on the gate. Will that increase the D-S current above 1mA?
yes, but will damage the part I believe
It will cause the gate-source junction to become forward biased and so the gate will conduct driving a current from gate to source. It's hard to predict what that current would be without knowing the rest of the circuit, source impedance, etc. but it's not a viable way of increasing Id above Idss for any JFET. Some samples of 2N5457 might have an Idss as high as 9mA and you can always pick a different JFET with a higher typical Idss anyway, so it may just be a case of trying several until you get one that works for your application.
Depletion vs. enhancement mode JFET ?
There is no enhancement JFET. It only „pinch off“, just like a vacuum triode, but while the tube allows (limited) positiv gate voltage, a JFET does not really. It then degrades to diode with very low backward current.
@@fromgermany271Thank you..
Could one of these transistors be used to simulate an electret microphone? I want to pass audio from a headphone amplifier (radio, tape deck, DAC, etc) into the microphone (headset) input on a cell phone.
you just need a resistor and capacitor
@@IMSAIGuy Could you be just a little more specific?
@@GordieGii 10k pot to ground to act as volume, 10uf into input
@@IMSAIGuy 10K across the output and capacitor from the wiper to the input?
Isn't the mic supposed to look like a 1K resistor that varies with the sound pressure?
@@GordieGii you are over thinking the input. it just wants to see a voltage swing.
Still not seeing how using a discrete jFET has any advantage over using a TL082. The jFET-input op amp gives you both the high input impedance and the high gain. Only possible reason I can think of for using this is, again, if you're going to bias this circuit closer to 0.5mA to get That Sweet, Sweet Even Harmonic Distortion(TM) commonly associated with triodes.
low noise
@@IMSAIGuy TL082 noise is typically 25 nV/√Hz and 0.01pA/√Hz. How well does your 2N5457 preamp compare to this?
@@petersage5157 while the JFET I call out is just one I had, companies like Rode and Neumann use a JFET and 1G ohm gate resistor, this is low enough noise to send to a mic preamp. look at the datasheet for a INA217. this preamp has typical specs for mic preamp. it is much lower than the TL082.
1mA to light, so brightly, a LED? That is surely not a typical LED, right?
Note that this is a very original approach, at least, to me, it is the first time I see it presented that way.
GaN Blue led, yes they are common, very efficient.
Hey, where did you get this kind of indicator light?
it is just a simple blue LED
@@IMSAIGuy OH, I´m sorry but my question is the indicator light at the very beginning of your vid.
......tagged NORMAL ABNORMAL
@@juarezvivo-sc2qi oh, here it is: th-cam.com/video/NVg-UfeLSkw/w-d-xo.html
Cheers
Is that the same as the VN66AF MOSFET?
no
Thanks for the video. It was a good intro in JFET. I'm not sure the pedantic value of confusing people with a drawing more depicting a MOSFET than a JFET. The gate(s) of JFET definitely protrudes into the body of the device. This explains why all the channel current will flow into the gate, if the JFET is biased incorrectly. The MOSFET gate is insulated from the body of the device. A gate current cannot flow in a MOSFET, in this way.
This is the correct jfet symbol.
@@markanderson8066 A rectangle (drawn for the body of the JFET, with a parallel line adjacent to one of rectangles longer sides (for the gate of the JFET) is not the JFET schematic symbol (if that's what you mean.) The gate of the JFET, in any pictorial representation of the physical construction of a JFET, ought to show the gate material, of the JFET, protruding into the body of the JFET material. Check any reference on the physical construction of a JFET.
what about a 100pf across 270R resistor.
you need to calculate the reactance of 100pf at the frequency you are using.
What if I pull out 4.4 MOhm from gate?
then the gate will not have the correct bias. it will float and not function
The calculated value of resistor was supposed to be 286 ohm not 285, you have to round the values sir. Actual calculation of JFET amplifier is pretty verbose and mathematically nerdy. Yet, they way you did it is pretty good
thinks
So why would these be chosen for the front end of a receiver. I remember when I bought my FT-847 They made a thing of the front end containing Jfets. Is it because of the high impedance and low voltage shift on the gate to make a stable linear amplifier?
Maybe it is as you point out (depends on the amplifier design). The big advantage of JFET is they are lower noise than BJT, and most MOSFET. It is a big deal if you can trim the noise in the front end: you'll get a better signal to noise ratio, and a 'better' radio. I think IMSAI Guy did something on this topic in an earlier video.
@@willthecat3861 Yeah, Makes sense. :)
Lower noise and high linearity (thus low distortion and high dynamic range) is the thing. You can get a noise figure around a couple dB with a well-designed JFET LNA. As to the physics reasons they have those attributes... I wish I knew!
With BJTs, their best noise performance tends to come with low impedance sources, but with JFETs the opposite is true. The 2N5457 datasheet show typical noise figures of 14dB with 3K source resistance, falling to around 1dB at 1M source resistance. Horses for courses.
@@yakovdavidovich7943 I think it's mainly the contribution of noise _current_ into the BJTs, which is essentially absent from JFETs.
Why do you imagine the negative terminal is called the source and the positive is called the drain? Because current flows from negative to positive not the opposite.
Conventional current flows from positive to negative. If you're concerned about the actual charge carriers (heaven knows why?), then in an n-channel FET, the charge carriers flow from negative to positive; while in a p-channel FET, the charge carriers flow from positive to negative.
I think it best to just stick to conventional current and avoid any worries about sub-atomic particles and quantum mechanics.
@@RexxSchneider See that just confuses people. I always think in terms of electron flow..
@@edwardneuman6061 Why? Doesn't that just confuse people?
Please, DO NOT EXPECT HIGH (almost infinite) INPUT IMPEDANCE on the GATE of FET amplifier. This is completely wrong. They have a large (almost infinite) input resistance (because of insulation of the gate), but the internal structure of the FET (gate is isolated) forms two capacitors (any insulation forms capacitor, this is how capacitors are made - two electrodes with a insulation between them), one between gate and source, another between gate and drain. That one between gate and source is just an capacitor, that project its capacitance into a imaginary part of input impedance depending on the frequency. It just sits there doing its regular job, because the voltage between gate and source is almost constant, so this capacitor just pulls your input impedance to the ground, nothing more. The second one is a little bit tricky, because the voltage between the drain and the gate is not constant, it's multiplied by the gain of that circuit, which means, that you have to catch this voltage change by the input current of that capacitor, which in fact multiplies its capacity by the gain of that amplifier. This is well known from the vacuum tubes as a Miller effect, but it is something common to almost everything that have at least some gain. So if we speak about low frequencies (high and higher frequencies is another story), then the gate of the FET amplifier has almost infinite resistance, but relatively low input capacitance (depends on frequency), which forms pure capacitive load - the input behaves like an almost perfect capacitor. And when it comes to AC, then the capacitor can load a lot.
You are confusing a jfet with a mosfet.
Your gain should be RL/Rs , 10K/270 which is about 37, It was surprising that you got gain of 10! What gives.
That formula only works if Rs is much larger than 1/gm, which it isn't in this case. Here the gain can be calculated as RL/(1/gm+Rs), with gm given in the datasheet (listed as Yfs) as typically 3000 umohs the gets you an expected gain of 16x. Still more than we see here but there is a very big variation expected on gm (it's given as something between 1000 and 5000 umohs)
@@rjordans Thanks, you are right. I forgot all about 1/gm and probably ro. The degenerated gain Gm = (gmro)/(Rs+[1+gm+gmb)Rs]ro) which is approximately Gm =gm/(1+gmRs) and voltage gain then is gmRL/(1+gmRs) if gmRs >> 1 this would become RL/Rs if it is not then you should use gmRL/(1+gmRs)
here gm is about say 0.001 then the gmRs is about 0.001*270 or 0.27
So Av= (0.001*10K)/(1+0.001*270) which is 7.8 for gm of 1000umohms , which is very close to 10 and using a actual gm value in the range of 1000 to 5000 we should get Av=10
@@rjordans I think the Yfs will be rather less than 3mS since that's quoted at Vgs=0. Yfs is the slope of the Id vs Vgs curve and that shows that Yfs will be lower as Vgs goes more negative. I took it to be more like 2mS, giving 1/Ys as 500R and yielding a gain of 13. The output on the scope seems close to 1.2Vpp, but it's difficult to estimate the input with any precision, although if a signal generator was used, there must have been a temptation to set 100mVpp. So I think the real gain was more like x12.
Obviously with that biasing, the non-linear 1/Yfs dominates the source resistance, so expect a noticeable amount of 2nd harmonic distortion.
Nooooo. Don't copy someone's else's value, that's unscientific. Use Monte Carlo analysis instead! ;)