6V6 PP AB1 circuit

[ccschua]

Hi,

I have the following circuit from one of the commercial tube amp. It uses 1x12AX7 and 2x12AU7 as pre and the 4x 6v6GT in push pull AB1 with self bias.

I trace the circuit point by point but I cant reach the capacitor. Disregarding the power circuit for now,

1. This is Direct DC couple between 12AX7 and 12AU7. I wonder what is the connection of R8 and C3. I have no way to trace what is connected at R8 and C3 ! It seems to me R8 and C3 should be connected to B- ?

2. Isnt wise not to use direct DC couple since any fluctuation in supply will affect the sound ?

3. I identify the signal components are C5, C6, C7 and replace them with Mundorf Supreme. Besides these, what are the critical signal components ?

4. Should I change R17, R 18 and the feedback resistors to something better like Riken or Audionote tantalum ? The Negative feedback is in 2 modes, High or Low. At high, it is 680R and at Low, it is 150K, is this correct ?

5. For the negative feedback, no capacitor was used ? Is it better to have capacitor and what value and type ?

5. Should I put a grid stopper before 12AX7 with a value let say 6.8k arbitralily?

[Frank Berry]

R8 and C3 may be for power supply hum reduction. What are the values of R8 and C3?
A 1k grid stopper on the 12AX7 won't hurt but if it wasn't in the original design, it may not be necessary.

[ccschua]

I dont know the connection of R8 and C3 because I cant trace it. The typical 12AX7 anode voltage can be around 150V. How can the150V pass to the grid of 12AU7 ?

[Wavebourn]

It is an oscillation harnessing thingie. It is probably 10K and 33 pF?
When feedback is deep the amp has wide frequency band, but starting from some frequency phase shift is 180 degrees, and if a gain through a feedback loop is more than 1 on this frequency the amp starts oscillate generating radio waves.
What does that R-C Zobel network, it lowers an amplification factor on frequencies below that one on which phase shift is 180 degrees, so where phase shift is 180 degrees an amplification factor is below 1, so oscillations don't occur.

The amp is good, should work fine.

[ccschua]

It is an oscillation harnessing thingie

I dont quite get what u say.

I found changing C5 from stock cap to Mundorf supreme really helps alot in the sound. I wonder if changing the resistors will help to improve the sound.

I am using Mullard CV series tubes while the 6v6gt is tungsol reissue. Guitar acoustics really rocks.

[Wavebourn]

Originally posted by ccschua


I dont quite get what u say.

I found changing C5 from stock cap to Mundorf supreme really helps alot in the sound. I wonder if changing the resistors will help to improve the sound.

I am using Mullard CV series tubes while the 6v6gt is tungsol reissue. Guitar acoustics really rocks.

It the cap was dead changing it you've stopped oscillations. The RC network compensates the amp on high frequencies that are unaudible, otherwise it oscillates on high frequencies with feedback applied.

If you put much bigger value of the cap you could roll-off highs on audible frequencies.

[ccschua]

Originally posted by Frank Berry
R8 and C3 may be for power supply hum reduction. What are the values of R8 and C3?
A 1k grid stopper on the 12AX7 won't hurt but if it wasn't in the original design, it may not be necessary.

It looks to me R8 and C3 serve to add Bias to Grid of V3A ? I think I will measure the DC condition and if necessary, dismantle the board for the capacitors connection.

[gingertube]

Hey guys - listen to Wavebourn, he is giving you the real answer.

R8 and C3 are a absolutely typical "zobel" or "step" network to stabilise the amplifier when global feedback has been applied - as in the case with this amplifier. Don't let R6 fool you. Remember that the power supply is effectively at signal (AC) ground, it is tied there by those big electrolytic bypass caps (C10 in this case). So as far as AC or signal is concerned that R8 and C3 could just as easily been tied from the 12AX7 anode to 0V. It makes no difference to their operation.

As you go higher in frequency the output transformer introduces phase shift. At ultrasonic frequencies the phase shift through the output transformer can be enough that the negative feedback (from the output tranny secondary) becomes positive feedback. If the amplifier gain at this frequency is more than 1 then it oscillates. R8 and C3 reduce the gain at high frequencies (while leaving the phase response largely unmodified) such that at that frequency where the feedback becomes positive the gain has been reduced to less than 1 and so the amplifier remains stable rather than oscillating.

Semi technical bit (leaving aside the vector algebra and talking a bit loosely): - Skip if of no interest
Gain of an amplifier (open loop or no feedback) is found by multiplying the gains of each stage together.
Gain of that input stage is proportional to the anode load - 100K in this case.
As frequency goes up the impedance of C3 drops and starts to reduce the gain - when Xc (the capacitor impedance) is approximately 90K then R8 + C3 = 100K and the effective anode load is then 2 x 100K in parallel or 50K and the gain will be halved. As frequency continues to go up Xc continues to go down and gain continues to be further reduced until Xc becomes insignificant and the 10K (R8) dominates the anode load.
However significant phase shift is not introduced until Xc falls to 10K (the R8 value) or below. When Xc = 10K there will 45 degrees phase shift. This happens at a frequency 9 times higher than the 1/2 gain point.
So a zobel or step network basically reduces high frequency gain while leaving the phase untouched until very much higher frequency - just what is required for stabilising the amplifier when global feedback is applied.

The values of R8 and C3 Wavebourn suggests (10K and 33pF) are typical and in the right "ball park". For correct zobel opperation teh resistor (R8) needs to be approximately 1/10th of the value of the load resistor (100K). C3 sets the frequency at which gain starts to roll off. If the amp still oscillates make it bigger.

You asked, "how can +150V go to the grid of the 12AU7?". For a typical bias of say -10V from grid to cathode of a 12AU7 that means that the cathode is sitting at +160V. From that you can work out that there is 7.3mA current flowing through that 22K cathode resistor on the 12AU7, or 3.65mA per side of the 12AU7 differential driver (160V divide by 22K = 0.0073 Amps).

Cheers,
Ian

[ccschua]

Thks for your reply. It makes ssense to me now, cause I dont really see such network in the old 'typical' tube amp.

As for the global feedback, the adjustment allows for 680ohm or 150K. I 'normally' see a capcitor of 100p to 1000p added in paralle to the resistor. what is that use ? improve high frequency ?

[gingertube]

The capacitor across the feedback resistor also helps with stability.

As we have said, at higher frequencies the output transformer delays the signal a bit, what we call a phase lag.
The capacitor across the feedback resistor is generally called a phase lag compensation capacitor. It allows high frequencies to bypass the resistor a little, speeding up the high frequency component of the feedback from the secondary of the output transformer (that is, working against what the transformer has already done by delaying the high frequencies). Generally, this capacitor is the last thing you set in a design. It is set by experiment. Once the amp is all working and is stable with feedback applied, you look at a 5 or 10kHz square wave into a dummy resistor load (instead of a speaker) using an oscilloscope. You then add capacitance across the feedback resistor (this phase lag compensation cap) to get rid of any major overshoot or ringing on the top and bottom of the square wave.

Hope this makes sense.

Cheers,
Ian