#1186 2N5457 JFET Applications

Episode 1186 chip of the day

[urlkrueger]

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.

[jonpattison]

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!

[TechProbe]

Thanks for the easy approach to understand the JFET characteristics by showing us its circuit under operation !

[rickwise]

You know you're listening to an optics guy when he says that's the "cyan trace." :-)

[MarioGonzalez-ogzf]

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!!

[amps]

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.

[homemdosaco]

Too much nitpicking in this video's comments.
I learned something; thanks for sharing.

[bobkozlarekwasqq]

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.

[elye]

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 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.

[grantlack]

"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.

[veum]

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.

[stamasd]

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.

[stefanopassiglia]

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.

[markcummings]

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.

[davebullard]

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

[RexxSchneider]

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.

[majordabalert]

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)

[floretion]

This is a great tutorial- very clear explanations. Look forward to more!

[hamzahkitech]

You taught me in very effective way that how to choose source resistor to self bias JFET amplifier, keep going 🎉

[yanava]

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.