2nd stage distortion

[otto]

We know theory about 1st stage triode. Without bypass cap cathodic voltage swings together the grid, with bypass cap it doesn't, and so on ...
What happens in the 2nd stage where amplitude in input is far bigger?
Let's consider 2 different solutions.
One is without bypass cap (jcm800 mono-channel, soldano atomic16 ...). In this case cathodic voltage swings together with the grid, being however the input signal of high amplitude it happens that cathodic voltage reaches (and goes below?) zero voltage (ground).
The other case is with bypass cap (ax84 high octane, engl thunder ...). In this way grid voltage, having an high amplitude, goes over the cathode voltage.
In both cases we go out the field explained on books about little swings of 1st stage and we will have a distortion I think that is not the cut off one at the opposite site of the transfer curve.
Can anybody explain the difference in sound between the 2 solutions?

[MWJB]

Phew, long story short...

No cap = less gain, looser sounding, maybe better fidelity.

With cap = increase in gain/distortion, tighter sound, cap value affects frequency response.

Basically the same as with stage 1. There may be a nominal difference in the effect of the cap in the 2nd stage, as compared to the first stage, but it's not as marked as your question suggests.

Most practical approach is to try with & without 2nd stage bypass cap, then if you like it with the cap, try 2 or 3 values and decide by ear.

[otto]

amplitude

Thanks, but what about the fact that in the 2nd stage the amplitude of the input signal (with gain pot full up) is bigger than the difference in idle voltage between grid and cathode?

[MWJB]

Well, you asked about the effect on the sound. Can you describe the audible symptoms of the problem you are having?

[otto]

Well, you asked about the effect on the sound. Can you describe the audible symptoms of the problem you are having?

No problem. I am modifying the crunch channel of my amp and I have considered 4 diagrams as reference: jcm800 mono-channel, soldano atomic16, ax84 high octane, engl thunder 50. The first 2 are without cap on the 2nd stage cathode, the last 2 have it.
I've studied the circuits with superspice and I've designed mine. I am going to build it but this morning I had the doubt described. What is (in the 2nd stage) the effect of a signal bigger than biasing?

[Angelo Skouras]

Otto my friend

Remember that a triode, if we forget the grid for a moment, is a diode, where only a positive current can be allowed. A cathode, can only emit electrons, and this is true for all kinds of valves.

This way, if the voltage of grid (with respect the cathode) comes more negative than a limit (dependent on your design), then the valve stops conducting until the voltage comes back inside the limits. This way the signal, after the amplification loses a part of its waveform, and the negative pick is like if is "cut with a pair of scissors".

Something identical happens when the voltage of grid comes more positive with respect the cathode, because the grid "seems like an anode" and receives electrons, witch have to be sinked by the circuitry that drives signal to the valve. Anyway, in that instance, input resistance of the valve falls dramatically, and unless special designed, the driver cannot supply the grid current, and the positive pick of the waveform is lost the same way.

Both defects result in distortion.This phenomenon is labeled as overloading. A sine wave input in an overloaded valve would give a square wave output. In a guitar amp this would give a sound like fuzz, or overdrive, and that is how these guit. effects work. If fidelity is paramount, then overloading is forbidden.

An amplifier without a cathode cap is overloaded more difficultly, but don't forget that this increases very much output resistance, causing problems in hight frequency response due to next stage input capacitance.

If you overload a triode with a cathode capacitor, the capacitor will be charged with the cathode voltage of the overload, and will discharge very slowly (seconds), making a momentary overload a worse problem...

I recommend you avoid overloading and unbypassed cathodes. Try loadlines with less gain and triodes with less gain. The 12sn7gt/6sn7gt have low gain, and are the best for low distortion. If a stage cannot avoid overloads, then a voltage reference (zenner or so) can substitude the capacitor/resistor combination. (At output stages, were anode load is a primary, you MUST bias with a resistor, for current limiting reasons).

MWJB said only true words to you.

Without an osciloscope you 'll be blind. Try Morgan Jone's "Valve amplifiers"

Good luck!!!!

[drewl]

What he said-
I usually put another .68uf in parallel with the 820ohm in the 2nd stage of my Marshalls and clones. Bumps the grind up just a bit, tightens the bottom end, anything over 100uf is too flabby for my tastes.

[otto]

Otto my friend

...

Something identical happens when the voltage of grid comes more positive with respect the cathode, because the grid "seems like an anode" and receives electrons, witch have to be sinked by the circuitry that drives signal to the valve. Anyway, in that instance, input resistance of the valve falls dramatically, and unless special designed, the driver cannot supply the grid current, and the positive pick of the waveform is lost the same way.

...

Good luck!!!!

Thanks Angelo, you have been very kind to sum up all this.
Distortion is not bad to me because I am making a crunch channel. I want to understand if the one obtained when the voltage of grid comes more positive with respect the cathode is good or not.
From your answer I understand that it is very similar (as sound sensation) to the clipping obtained with the grid more negative than the cut off.

[Angelo Skouras]

Otto my friend

Whatever you want we are here to help. You are realy welcome.

It is not a matter of good or bad. It is a matter of what you want to do. For me it is "bad" because I design high fidelity amplifiers (at least at time), and distortion must be avoided whatever the cost. Since you design a circuit that distorts, it is a matter of personal taste.
I advice you to listen (or play) music through it and let your own ear make the choice.

I let you know that if you drive a valve with a source able to provide more current, grid current flows when the grid comes positive with respect the cathode and the valve enters class A2. This way a far bigger part of the positive pick survives, until the grid "wants" more current than driver source can provide.

If the grid of the valve is not designed for such a current, will soon be destroyed, and valve will be dead. For any reason, reed very good the specifications of the device you use, and consider the current abilities of the previous stage.

If you work only with low signal valves, there is no possibility of grid current, because they cannot drive grid current to each other. Anyway, be very careful!

[otto]

If the grid of the valve is not designed for such a current, will soon be destroyed, and valve will be dead. For any reason, reed very good the specifications of the device you use, and consider the current abilities of the previous stage.

Thanks Angelo. I am modifying a Marshall amp with 12ax7 tubes and I don't make innovations respect reference circuits so I hope I won't make disasters.
Maybe I will give a little more positive swing than a similar circuit. How can I understand if the grid of the valve is designed or not for such extra current?
Do you mean that in the previous stage I must add this current and verify that watts required are enough?
I am speaking of a positive voltage peak in the grid bigger of 50% respect a reference circuit with a similar solution

[Steve Conner]

Sounds like you are talking about blocking distortion. This is a nasty effect where the sound cuts out completely for a fraction of a second after each loud note. It's caused by the coupling and bypass capacitors being charged up to an abnormal state when the tube is overdriven heavily enough to draw grid current. The stage ends up biased into cutoff and useless as a linear amplifier of small signals, until the capacitors have discharged back to normal.

In later stages of high-gain amps, you prevent this by using smaller coupling capacitors than usual, also smaller cathode bypass capacitors, or no cathode bypass capacitors at all. The idea is to allow the capacitors to charge and discharge faster so that they can follow the envelope of the signal, hence the stage never blocks.

Attenuating the signal a little between stages can help too, so the tubes aren't driven so heavily into grid current. Ray Ivers also proposed a zener clamp to limit the capacitor charge. But I believe that driving a bit of grid current and letting the bias point jump around actually adds more life to the sound. I think it's a major factor in the tube vs. transistor sound thing.

[otto]

Thanks Steve, as written in the previous post I'm going to give to the 2nd stage tube grid a positive voltage peak 50% bigger than a reference circuit (I am speaking about jcm800, Caswell mod, and things like these if you know them ...). Do you think that if it is too much I will hear a blocking distortion but no danger for tubes?

[Steve Conner]

No, there is no danger that I know of. The average DC grid current is limited by the grid leak resistor which is typically 1 Meg.

[Steve Conner]

The 2nd stage of what? If you mean the classic tube amp circuit, then the volume pot/tone stack act as the grid leak ofthe 2nd stage, and have an effective resistance of several hundred kOhms, so again there is no problem.

Use the "Manage Attachments" button to insert images. You have to scroll down to see it.

[otto]

You are right, I've realized this and I've deleted my question. But too late.
Thanks