Fetzer Valve does sounds horrible when overdriven!

[sajy_ho]

Hi guys, I'm trying to build a practice amp for a friend. The schematic is three cascaded Fetzer stages followed by a TDA2040 as power amp:

Everything is alright as long as I play clean (gain less than 50%), but when I overdrive the third Fetzer stage I get this horrible clipping sound:https://soundcloud.com/sajy_ho/voice-001
I've built Fetzers before but they always sound awesome when overdriven. I remember I heared this bad clipping sound out of my Tube Reamer long time ago.
So what do you think? What could be the problem?

[thehallofshields]

Well... If you've certain the problem is happening at Q3, then perhaps your Gain control is throwing off the Input Impedance of Q3? Does the sound get worse as you turn the Gain Control down?

[sajy_ho]

Well... If you've certain the problem is happening at Q3, then perhaps your Gain control is throwing off the Input Impedance of Q3? Does the sound get worse as you turn the Gain Control down?

Thanks for the reply, My problem begins when I turn the gain pot up to overdrive. In the sound sample I posted first the gain is down and there is no problem at all, but then I turn the gain up to start overdrive and this Cracken-type clipping comes into the sound.

[ggedamed]

Maybe Q3 doesn't like being fed with DC through the Gain pot. Try a big resistor (2..5M) from the gate of Q3 to the ground.

[sajy_ho]

Maybe Q3 doesn't like being fed with DC through the Gain pot. Try a big resistor (2..5M) from the gate of Q3 to the ground.

Thanks man, I'm trying to understand your point,but I'm not seeing any DC going to Q3 gate. Anyway you are right about big gate resistor, maybe Q3 gate needs to be higher of the ground. I'll try it and I'll report.
Thanks again

[Seljer]

Yeah, C5 would block any DC before it gets to Q3, so it's happily referenced to ground via the gain pot and R16.

For good measure, I'd add some more filtering on the Vdd rail near the preamp sections.


What are the DC voltages on the three transistors?

[sajy_ho]

Yeah, C5 would block any DC before it gets to Q3, so it's happily referenced to ground via the gain pot and R16.

For good measure, I'd add some more filtering on the Vdd rail near the preamp sections.


What are the DC voltages on the three transistors?

I tried replacing the gain pot with a 1M pot, but still that sound!
Here are my measured voltages:
Q1;  Vd1: 7.93    Vs1: 0.36
Q2;  Vd2: 7.94    Vs2: 0.38
Q3;  Vd3: 7.95    Vs3: 0.38
I adjusted all the voltages using Fetzer Calculator.
Maybe the problem is the Q3's load, becouse it's directly coupling to the opamp buffer, what do you think of that?

[sajy_ho]

Also added extra filtering cap right behind Q3, also didn't work!

[Seljer]

The gate voltages are all 0V right?

The opamps buffer is practically no load at all to the last FET stage, the opamps input impedance is practically infinite so you're left with only the 1Meg resistor that sets the bias. At a glance, I'd say roughly the same load is presented upon Q1 with that 820k resistor and it doesn't appear to be acting up.


Maybe some kind of high frequency oscillators are the issue?  I'd hook up an audio probe before and after Q3 to verify it really is the stage that is distorting.

[sajy_ho]

The gate voltages are all 0V right?

The opamps buffer is practically no load at all to the last FET stage, the opamps input impedance is practically infinite so you're left with only the 1Meg resistor that sets the bias. At a glance, I'd say roughly the same load is presented upon Q1 with that 820k resistor and it doesn't appear to be acting up.


Maybe some kind of high frequency oscillators are the issue?  I'd hook up an audio probe before and after Q3 to verify it really is the stage that is distorting.

Thanks man, I splited the cap before the opamp and added 100k load resistor for Q3, also didn't work.
Could it be the opamp distorting? I've heared opamp overdrive before but this sound is not like overdriving an opamp. I don't know what else to do, I'm giving up on this project.

[anotherjim]

Try a stronger Vref - 100k supply splitter is a little weak. As you have 12volt, try 47k or 33k.

[Pojo]

I'll go out on a limb and suggest moving the gain pot to between the Q1 and Q2 stages and lowering the value of C3....substantially. Probably down to 1.5nF or possibly further down into the pF range. I've found with overdriven cascaded jfets that you really have to limit the low end to get good sounding clipping. This will sacrifice the fullness of clean tones of course so might be good to have a switch that bypasses a stage and put a bigger cap in parallel with C3.

[sajy_ho]

I'll go out on a limb and suggest moving the gain pot to between the Q1 and Q2 stages and lowering the value of C3....substantially. Probably down to 1.5nF or possibly further down into the pF range. I've found with overdriven cascaded jfets that you really have to limit the low end to get good sounding clipping. This will sacrifice the fullness of clean tones of course so might be good to have a switch that bypasses a stage and put a bigger cap in parallel with C3.

You are right man, becouse it doesn't make that horrible sound on the higher notes. My friend that I'm building the amp for him is a pedal guy and I don't think he uses amp's overdrive at all, so I'm thinking of omitting the third stage completely for a clean amp.
Anyway thanks for all your help.

[ggedamed]

Maybe Q3 doesn't like being fed with DC through the Gain pot. Try a big resistor (2..5M) from the gate of Q3 to the ground.

Thanks man, I'm trying to understand your point,but I'm not seeing any DC going to Q3 gate. Anyway you are right about big gate resistor, maybe Q3 gate needs to be higher of the ground. I'll try it and I'll report.
Thanks again

As I see it, feeding GND through the pot is feeding DC (as in Direct Current). I tend to avoid biasing through pots and you'll see that almost every cascaded JFET stages circuit is isolating the pots with caps. But it seems it's a different problem here. Have you tried to audio probe the circuit? You did.
I would try increasing R9, seems like blocking distortion.

[Eddododo]

If i remember correctly, isnt the pinch-off voltage an independent threshold for clipping the input, regardless of Vcc, Vbias, etc?

could clipping the input stage be contributing to nastiness (independent of the output voltage)

[PRR]

> As I see it, feeding GND through the pot is feeding DC

Yes; but the Current into the Gate of a JFET is incredibly small.

Vacuum tube Grid is similar. We often used pot-wiper as grid-leak, without general problem.

However a cap and separate grid-leak (gate leak) does tend to de-bias the stage in heavy overload, which *may* sound better (or different) than brute-force overload.

IMHO, *three* of *any* high-gain stage in cascode is liable to go beyond distortion into garble.

The usual trick is to insert voltage dividers between gain stages, so the overall gain is not so high, and each stage is just over-loaded, not OVER-OVER!-loaded.