learner asks: how to correctly bias a jfet?

[vitruviano61]

I am a Category 2: Learner

From the "What to do if it doesn't work" checklist:

1.What does it do, not do, and sound like?

My circuit has a very fizzy sound, and low output. I am using a 100k multi-turn (15) trimpot, and cannot to 4.5v at the drain without creating a horrible squeal. I'm using a DMM, red connected to drain, black connected to ground. Is this correct? I have tried several 2N5457 transistors, and several 100k trimpots.

2.Name of the circuit = ROG Fetzer Valve

3.Source of the circuit (URL of schematic or project) = www.runoffgroove.com
4.Any modifications to the circuit? N
5.Any parts substitutions? 2N5457
6.Positive ground to negative ground conversion? Y or N
(not sure what this means)

7.Turn your meter ...: What is the out of circuit battery voltage? => 8.82

Now insert the battery into the clip.... g:

Voltage at the circuit board end of the red battery lead = 8.82
Voltage at the circuit board end of the black battery lead = 0

Q1
C = 1.81
B = 0
E = 1.08

For testing purposes I have connected the circuit to an extremely simple LM386 op-amp, using the following pins:
1. n/a
2. input (from volume pot at drain, as per schematic)
3. ground
4. ground
5. 220uF to 8ohm speaker (+)
6. 9v+
7. n/a
8. n/a

Help? Advise? Flames? All welcome!

best,
--v

[PenPen]

Ok, first off the pins on a fet are named differently. The Collector is called the Drain, the Base is the Gate, and the Emitter is the Source.

Hook your meter up to the drain of your jfet to read its voltage. Now, turn the trim pot until you read approximately half of your battery voltage on the drain. This would be roughly 4.5V on the drain.

The actual resistance value for this drain is always different for each fet, hence the use of a trim pot to adjust the voltage until its half of your power supply.

[George Giblet]

When you vary the bias pot can you change the voltage on the JFET source?  If no then you may have the JFET pins mixed up

Try putting a 10ohm in series with 47nF from the output of thre LM386 to ground as per say,

http://www.runoffgroove.com/littlegem.html

[vitruviano61]

Ok, first off the pins on a fet are named differently. The Collector is called the Drain, the Base is the Gate, and the Emitter is the Source.

Hook your meter up to the drain of your jfet to read its voltage. Now, turn the trim pot until you read approximately half of your battery voltage on the drain. This would be roughly 4.5V on the drain.

The actual resistance value for this drain is always different for each fet, hence the use of a trim pot to adjust the voltage until its half of your power supply.

Yes, thanks for the info PenPen, this is exactly what I've been trying to do, with poor results.

I appreciate your response, though -- thanks!
--p

When you vary the bias pot can you change the voltage on the JFET source?  If no then you may have the JFET pins mixed up

Try putting a 10ohm in series with 47nF from the output of thre LM386 to ground as per say,

http://www.runoffgroove.com/littlegem.html

I'll try this right away... thanks George!

best,
--v

[vitruviano61]

When you vary the bias pot can you change the voltage on the JFET source?  If no then you may have the JFET pins mixed up

Try putting a 10ohm in series with 47nF from the output of thre LM386 to ground as per say,

http://www.runoffgroove.com/littlegem.html

George -- thanks. Yes, I am able to vary the voltage on the JFET source with the trimpot. I'm quite certain that the pins are in the correct order, as I've verified this with the Motorola datasheet for the JFET I'm using.

I've added the 10ohm resistor in series with the 47nF capacitor to the ground, and the frequency changes slightly, but is still unusable. I'll try building out the LM386 as per the Ruby Fetzer schematic and see if that eliminates the problem.

Still learning... and thanks so much for your help!
best,
--v

[Jered]

  Test your 1M resistor and make sure its 1M. If so, make sure its connected to the gate and to ground. If your only using two pins of your trim pot, make sure its pins 1 & 2, or 2 & 3. Not 1 & 3.
  Jered

[vitruviano61]

  Test your 1M resistor and make sure its 1M. If so, make sure its connected to the gate and to ground. If your only using two pins of your trim pot, make sure its pins 1 & 2, or 2 & 3. Not 1 & 3.
  Jered

Thanks Jered -- I've tested them all.

I think the problem is perhaps because I discovered I was using a 10K volume pot, as shown in the Ruby schematic, rather than the 100K volume pot in the Fetzer+Ruby schematic.

Since I don't own a 100K volume pot today, and it's a holiday here, I've substituted a 47k pot, which helps a bit. I can bias the JFET up above 2v now at least.

As an experiment, I've tried placing a 47k resistor between the drain and my 47k volume pot, which allows me to bias the JFET a bit higher still.

I have a lot to learn about pots today. Going to re-read "The Secret Life of Pots" now.

Thanks to all who have been kind enough to offer help!
best,
--v

[Jay Doyle]

I think the problem is perhaps because I discovered I was using a 10K volume pot, as shown in the Ruby schematic, rather than the 100K volume pot in the Fetzer+Ruby schematic.

OK, first off I think you need to split the circuit you are designing to JUST the Fetzer, it is hard to help because I can't see the whole schematic. But a 10k volume pot after the Fetzer would definitely load down the JFET stage but WOULD NOT mess with the bias.

Since I don't own a 100K volume pot today, and it's a holiday here, I've substituted a 47k pot, which helps a bit. I can bias the JFET up above 2v now at least.

This should up the output a bit, but you will probably want to find that 100k eventually.

As an experiment, I've tried placing a 47k resistor between the drain and my 47k volume pot, which allows me to bias the JFET a bit higher still.

This is telling. Double check the 22n cap off of the drain of the JFET to make sure that it is good. adding a resistor either before or after the cap to the pot should NOT alter the bias in any way. The cap should block all DC so that only the trim on the drain has any affect on the bias.

As a best guess from what I can tell by what you have written, the 22n cap is either leaky or a short and therefore your DC bias on the drain is then bleeding off to the pot to ground. You need to make sure that cap between the drain and the volume pot is good; the ONLY thing that should adjust the bias voltage on the drain is the trim pot.

Let me know if this makes sense, I think that you are almost there.

Good Luck!

Jay Doyle

[George Giblet]

>I've added the 10ohm resistor in series with the 47nF capacitor to the ground, and the frequency changes slightly, but is still unusable.

As far as the oscillation goes, it sounds like the amplifier is interracting with the preamp. If you look at the link I posted before the circuit has a  100uF cap across the supply.  Adding that might help but it must be located very near to the LM386.

The other possibility is the way your power and ground connections are done. 

If the wiring on the power or ground goes in the order:  power supply -> preamp power/ground wires -> LM386 power/ground wires then that can cause oscillation.  If goes power supply -> LM386  power/ground wires -> preamp power/ground wires you are probably safe.

Another solution is to put a 100ohm resistor in series with the power rail on the preamp then add a 47uF cap on the preamp side side the resistor to ground.

[vitruviano61]

Joy!

I'm now playing it through a 4x12 cab, and I can finally see what all the Fetzer+Ruby fuss is about -- sounds very nice to my ears.

To recap, hopefully for the benefit of others, here are the remaining steps I took:

I think the problem is perhaps because I discovered I was using a 10K volume pot, as shown in the Ruby schematic, rather than the 100K volume pot in the Fetzer+Ruby schematic.

OK, first off I think you need to split the circuit you are designing to JUST the Fetzer, it is hard to help because I can't see the whole schematic. But a 10k volume pot after the Fetzer would definitely load down the JFET stage but WOULD NOT mess with the bias.

Since I don't own a 100K volume pot today, and it's a holiday here, I've substituted a 47k pot, which helps a bit. I can bias the JFET up above 2v now at least.

This should up the output a bit, but you will probably want to find that 100k eventually.

As an experiment, I've tried placing a 47k resistor between the drain and my 47k volume pot, which allows me to bias the JFET a bit higher still.

This is telling. Double check the 22n cap off of the drain of the JFET to make sure that it is good. adding a resistor either before or after the cap to the pot should NOT alter the bias in any way. The cap should block all DC so that only the trim on the drain has any affect on the bias.

As a best guess from what I can tell by what you have written, the 22n cap is either leaky or a short and therefore your DC bias on the drain is then bleeding off to the pot to ground. You need to make sure that cap between the drain and the volume pot is good; the ONLY thing that should adjust the bias voltage on the drain is the trim pot.

Let me know if this makes sense, I think that you are almost there.

Good Luck!

Jay Doyle

This made sense, and so I followed the steps, and was particularly encouraged by Jay's feeling that I was "almost there". I don't know how to test for a bad 22n capacitor, but the volume pot wasn't changing the bias on the JFET drain, so...

>I've added the 10ohm resistor in series with the 47nF capacitor to the ground, and the frequency changes slightly, but is still unusable.

As far as the oscillation goes, it sounds like the amplifier is interracting with the preamp. If you look at the link I posted before the circuit has a  100uF cap across the supply.  Adding that might help but it must be located very near to the LM386.

The other possibility is the way your power and ground connections are done. 

If the wiring on the power or ground goes in the order:  power supply -> preamp power/ground wires -> LM386 power/ground wires then that can cause oscillation.  If goes power supply -> LM386  power/ground wires -> preamp power/ground wires you are probably safe.

Another solution is to put a 100ohm resistor in series with the power rail on the preamp then add a 47uF cap on the preamp side side the resistor to ground.

1. I moved the 100uF cap across the power supply very much closer to the LM386.
2. I rewired everything to look like this: power supply -> LM386 power/ground wires -> preamp power/ground wires.
3. I added a 100ohm resistor in series with the power rail on the preamp, then added a 47uF cap to ground on the preamp side of the resistor.

It wasn't until I completed step 3 that the horrible noises stopped coming from the little test speaker and I was able to connect a fretted instrument of wood and wire to the input jack and fill my little shed with sonic wonder, with a properly biased drain on a Fetzer+Ruby.

I hope this helps some other poor devil.

And a most heartfelt thanks to all who replied to my question -- I used all of your info along the way.

Lastly, man, how do you guys know all this stuff? I'm reading all I can, but it still seems like so much black art/science and witchcraft to me, a Category 2 Learner.

best,
--v

[PenPen]

Huh. So it appears that particular jfet needed more than 100k resistance in the voltage divider to bias properly. Never heard that one before, and I've been using nothing but j201s. The cap I'm assuming was smoothing out the occilations.

And as far as knowing this stuff, I am wrong as often as I am right, I'm only a Cat. 5 Learner I guess. Is there some chart that ranks us? 
Just keep reading, stuff eventually starts to click and make sense, and then other things you didn't understand before make sense, and it snowballs.

[Jay Doyle]

Glad to hear that it worked out! Though, in all honesty, I have no idea how better filtering and power layout helped you to bias the FET! But...

I have read, reread and reread again so many different things until they make sense to me. Eventually, it WILL make sense.