What's the real story on SOURCE v/s DRAIN trimmers in JFET circuits?

[stm]

Well, if that's correct, you have just reduced two handles down to one handle.

Now we just have to figure out how to grab hold of it.

Exactly!

And I have found it.  Well, in fact I found two things, really:

1) I found the real story on SOURCE v/s DRAIN trimmers in JFET circuits.  The preferred biasing method is different depending if your source resistor is bypassed by a capacitor or not.

2) I found how to bias a JFET amplifier with a SINGLE trimmer without significant effect on gain, input dynamic range, output impedance, and harmonic content.  This works for both the bypassed and unbypassed cases.

[alanlan]

Sebastian,
I know this thread is about source vs. drain trimmer but it can be done without a trimmer.  AN102 works.  I've tried it on a batch of J201s as reported here:

http://www.diystompboxes.com/smfforum/index.php?topic=69885.msg561231#msg561231

[stm]

Sebastian,
I know this thread is about source vs. drain trimmer but it can be done without a trimmer.  AN102 works.  I've tried it on a batch of J201s as reported here:

http://www.diystompboxes.com/smfforum/index.php?topic=69885.msg561231#msg561231

Hi alanlan,

Last night I gave it a stab and it does work indeed.  I tired it in the sim with three different J201's (with Vp 0.5 V, 0.8V and 1.2V, and IDSS varying accordingly as well).  I got the Vp and Idss values from actual J201's I had, just to be sure I wasn't using nonsense Vp-Idss pairs.

In the end I had very little gain without source bypass capacitor (around 3 dB), so I added a 100u source bypass capacitor and gain boosted a good deal (around 20 dB).  Using a series resistor with said cap allows to establish gain to any desired value in between.  I verified gain didnt vary significantly from device to device, and drain voltage spanned about 1 volt between the minimum and maximum Vp.  I think this is a great alternative when you want to reliably amplify a low level signal with a known gain, as you can totally avoid trimmers.  It is very likely that you would need to set proper values for Rs and Rd depending on each batch, as JFETs from different batches and/or manufacturers could vary a good deal, as pointed before by Eb7+9, but it still works pretty good for what it is.

Regarding "our" use for stompboxes, there were some chareacteristics of the resulting circuit that I didn't like, though:

1) The need for bypassed source resistor added way too much 2nd harmonic distortion for my taste when signals get large.
2) The output dynamic range was greatly reduced (bottom peak at the output clips around 3 to 4 volts above ground, instead of 1 to 1.5V above).
3) Depending on if the device had large or small Vp, the tops or bottoms of the waveform clipped earlier/later, so getting into the clipping region will not be the same.

Based on the above, I prefer using a single trimmer "wisely" for stompbox use.  Indeed, using one trimmer is not bad at all.  In several three-stage JFET circuits you can find some interesting sweet spots by tweaking the biasing of each stage by ear anyway.  This is particularly true now that I've found how to do it for unbypassed source circuit while keeping gain and input dynamic range constant.

[Gus]

http://www.sdiy.org/oid/mics/Oktava-MK219.gif


http://www.sdiy.org/oid/mics/Oktava-MK-319.gif
R3, R4 fixed bias, R7, R8 self bias

R7 and R8 look to be selected for the fet for gain control and bias when you open the microphone you will see two numbers written on the transformer case they are the R 7 and R8 values.  It is a combo of fixed and self bias.  The microphone is powered by 48VDC with two 6.8K matched to pins 2 and 3(google phantom power)

Another thing to look at is the use of a negative supply so you can use a higher value source R to make it a quasi constant current bias, you can find this in books.  You can partial(cap and gain control resistor) or  fully bypass(cap) the source to ground.  This is what is used in the early 80's fender harvard FET preamp.  So make say a +- 15VDC supply and have fun.  It could also be a say +20 and -10 or ......

or a
higher voltage with a
MIX of fixed and self bias.  The fet "feels" only the difference between the fixed and self bias part this allows a greater voltage from source to ground for a quasi constant current HOWEVER one would need to be mindful of the way the circuit powers up.

[stm]

http://www.sdiy.org/oid/mics/Oktava-MK219.gif
http://www.sdiy.org/oid/mics/Oktava-MK-319.gif
R3, R4 fixed bias, R7, R8 self bias

R7 and R8 look to be selected for the fet for gain control and bias when you open the microphone you will see two numbers written on the transformer case they are the R 7 and R8 values.  It is a combo of fixed and self bias.  The microphone is powered by 48VDC with two 6.8K matched to pins 2 and 3(google phantom power)

Another thing to look at is the use of a negative supply so you can use a higher value source R to make it a quasi constant current bias, you can find this in books.  You can partial(cap and gain control resistor) or  fully bypass(cap) the source to ground.  This is what is used in the early 80's fender harvard FET preamp.  So make say a +- 15VDC supply and have fun.  It could also be a say +20 and -10 or ......

or a higher voltage with a MIX of fixed and self bias.  The fet "feels" only the difference between the fixed and self bias part this allows a greater voltage from source to ground for a quasi constant current HOWEVER one would need to be mindful of the way the circuit powers up.

Thanks for pointing out these schems.  They are good examples of how to use a JFET as a practical preamp.  I find interesting the use of a 50% split source resistor with just one half bypassed: some more gain but not too much.  I wonder if this hides some special property like more stable gain and/or biasing point, reduce harmonic distortion, etc.

I remember having seen a Carvin guitar amp with a JFET preamp input stage.  I don't have it hand, so I don't recall the exact model, but I believe it did use positive and negative supplies as well.  IIRC, it also had CD4049's or the like for the distortion stages.

IME being bound to a single +9V supply prevents from getting the full potential from a given circuit.  Yes, there are tricks to resort to single supply biasing, but low voltage v/s high voltage supplies provide additional benefits like what you mentioned: a more constant-current behaviour for biasing circuits; it also allows opamps to have better output capability and less distortion, apart from increased headroom and dynamic range.  All of these has audible effects.

Regarding the Oktava MK-319 schem, it should be noted that R2 and R5 are kiloohms instead of megaohms. 
EDIT: Gus confirmed the megaohm values as normal for use in mic circuits, so the line above should be disregarded!

[Dragonfly]

Dragonfly
If it is the one in the members section.  I noticed it it looks like right out of AN102  a mix of self bias and fixed bias,  Google MK219 oktava schematic you might like this gain stage.  Look for the harvard schematic.

  When you increase voltage make sure to note what happens with greater drain to gate voltage differences.

Its the one in the Marston article...probably the same as AN102 (id need to look, but it makes sense)...seems like a nice starting point for stable bias on a wide range of Jfets.

I'll check the Oktava schemo ...thanks !

BTW...for those who couldnt locate it, here's the schematic I posted (from Marston/AN102)

[Gus]

R2 and R5 are 680megohms and in some condenser microphones up to 3Gig.  You don't want to discharge the charge on the capsule because of the way a condenser microphone like this works.  The values of R7 and R8 are different in different 219s and 319s and don't have to be in a 50/50 ratio like the one that was traced.  I QUESS they measure the fet and use a look up table to select the gain and total bias value.

[stm]

R2 and R5 are 680megohms and in some condenser microphones up to 3Gig.  You don't want to discharge the charge on the capsule because of the way a condenser microphone like this works.

Oh boy, it's true that one learns something new everyday.  I added a comment in my post above so it won't cause confusion.

The values of R7 and R8 are different in different 219s and 319s and don't have to be in a 50/50 ratio like the one that was traced.  I QUESS they measure the fet and use a look up table to select the gain and total bias value.

I see, it makes sense.

[stm]

Cool schem, Andy.  Is your circuit intended for J201's or another type of JFET?

[Dragonfly]

Cool schem, Andy.  Is your circuit intended for J201's or another type of JFET?

Its straight from the Marston article. I believe it specifies 2N3819's, though I posted it mainly as an example of a possible biasing scheme that could be implemented for a wide variety of fets with a bit of value switching. i'm sure it could be adapted to different voltages and a wide variety of fets with a some value changes.


http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=17356

[stm]

Cool schem, Andy.  Is your circuit intended for J201's or another type of JFET?

Its straight from the Marston article. I believe it specifies 2N3819's, though I posted it mainly as an example of a possible biasing scheme that could be implemented for a wide variety of fets with a bit of value switching. i'm sure it could be adapted to different voltages and a wide variety of fets with a some value changes.

http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=17356

Now it makes sense, since with J201s I ended with a gate voltage around 1 volt or so, instead of nearly 2.8V.  Tonight I'll post the values I used to make a J201 bias with this method.

Cheers.

[Dragonfly]

Cool schem, Andy.  Is your circuit intended for J201's or another type of JFET?

Its straight from the Marston article. I believe it specifies 2N3819's, though I posted it mainly as an example of a possible biasing scheme that could be implemented for a wide variety of fets with a bit of value switching. i'm sure it could be adapted to different voltages and a wide variety of fets with a some value changes.

http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=17356

Now it makes sense, since with J201s I ended with a gate voltage around 1 volt or so, instead of nearly 2.8V.  Tonight I'll post the values I used to make a J201 bias with this method.

Cheers.

cool - its a simple enough circuit that it may work well for some people.

[earthtonesaudio]

I know everyone's been talking about the class-A common source configuration, but there are other ways of making gain from JFETs...

What about common gate?  If you use it as a current follower you can get voltage gain which is independent of device characteristics.  Then you just slap some source followers on the input and output, which gives you current gain while simultaneously providing good input/output impedances.  If you're willing to add the complexity of two more transistors, you can get power gain without phase inversion which is tolerant of device variations.

I have this on the breadboard right now, and it's proving to be surprisingly easy to tweak, even without coupling capacitors.  Right now it's two MOSFETs followed by one JFET, but I'm working on all-JFET and BJT/Darlington versions next. 

I don't know if anyone would be interested in exploring this more, but if so I can start another thread with some pretty pictures and stuff.

...After re-drawing this circuit in a more conventional way I discovered it's actually just a long-tailed pair or differential input stage.  Nothing new or exciting, so never mind.   

[stm]

Here's my take on a practical trimless JFET preamp using Siliconix AN102 mixed biasing method.
Gallery access: http://www.aronnelson.com/gallery/main.php/v/STMs-Circuit-Ideas/Trimless+JFET+Preamp-STM-rev1_0.png.html



Schematic contains expected voltages and range of JFETs that are verified to work.
Gain is fixed at 6dB with the values as shown.  If more gain is desired, R7 can be reduced or replaced with a 10k pot.  R7=0 ohms produces a max gain of 20 dB.

[alanlan]

Great work Sebastian!

[slacker]

I breadboarded the TJP this afternoon and threw in a random selection of J201s without bothering to measure them first. They all worked fine and biased up as per the schematic, subsequently measuring them showed they were all within the range of verified values. Adding a pot as Sebastian suggested makes this into a nice little booster

I then tried a couple of 2N5458s with VP around -2.55v and IDSS around 6mA and they worked fine as well and to my ears didn't sound significantly different than the J201s. So it looks like this circuit should work for a variety of Jfets, it would be interesting to see how far from the specified values you can go.

[stm]

I breadboarded the TJP this afternoon and threw in a random selection of J201s without bothering to measure them first. They all worked fine and biased up as per the schematic, subsequently measuring them showed they were all within the range of verified values. Adding a pot as Sebastian suggested makes this into a nice little booster

I then tried a couple of 2N5458s with VP around -2.55v and IDSS around 6mA and they worked fine as well and to my ears didn't sound significantly different than the J201s. So it looks like this circuit should work for a variety of Jfets, it would be interesting to see how far from the specified values you can go.

Great, thanks for reporting.  I'll take a look tonight to the effect of using 2N5458's in the same circuit.

[slacker]

No trouble, I'm not qualified to enter into the technical side of the discussion so getting the breadboard out is the least I can do.

Forget what I said earlier about it working with 2N5458s though, turns out I'd mistakenly used a 2k7 resistor in place of the 8k2 for R5. Using the correct value my 2N5458s won't work.

[stm]

No trouble, I'm not qualified to enter into the technical side of the discussion so getting the breadboard out is the least I can do.

I like to address engineering problems via three frontlines:

1) Math (for analysis)
2) Simulation (for design)
3) Breadboard (for testing and verification)

Each method has advantages and disadvantages (too long to detail here), but the three together allow for maximum understanding and best results IME.  I usually concentrate the most effort in 1 and 2, so any practical result is very welcome as it validates the other two!

Forget what I said earlier about it working with 2N5458s though, turns out I'd mistakenly used a 2k7 resistor in place of the 8k2 for R5. Using the correct value my 2N5458s won't work.

No problem.  A quick check showed that you would need to drop the drain resistor down to 3k9 and reduce the gain resistor (in series with the 1u cap) down to 1k5 to maintain 6 dB of gain.  Of course this is just a hack on the article biasing method, as it is most likely that the gate voltage should be adjusted as well for best results.  This was with a 2N5458 with Vp=-2.3V and Idss=5.6mA (reasonably close to your device).

[slacker]

I use mostly methods 2 and 3 although I find the more I learn the more I use the simulator especially for trying to understand new circuits and ideas.

I should have pointed out before that using the correct resistor values the circuit worked as expected with the J201s.