• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

ST35 Cathode Feedback


2011-06-10 3:25 am
Hi all,

I'm thinking of building an ST35 from scratch as my first tube amp. I'm currently going over the circuit trying to understand it all, and am learning a lot about how tubes work. They no longer seem like black magic to me!

Anyway, I've been looking at the cathode feedback from the preamp tube, a 7247 in the original circuit, or half each of a 12ax7/12au7 in the DIYtube circuit.

It goes from the high mu cathode of the 7247 then has a 18pf capacitor which leads to the primary winding of the OPT/pin 9 of one EL84, and parallel 27pf capacitor and 27k resistor leading to the 16ohm tap on the OPT on the original. The DIYtube circuit changes the primary side capacitor to 15pf, and if using a 8ohm output tap transformer they change the parallel resistor to 20.3K but leave the capacitor at 27pf. The modern reproduction OPT's have an 8K primary.

Original schematic: Dynaco ST35 Tube Amplifier Schematic and Manual

DIYtube: http://www.diytube.com/st35/ST35_REVD_2009.pdf on page 11.

First question, this forms what people call an NFB, right?

Second more important issue, I'm not understanding the values used on the secondary winding. They form a high pass filter, but when I model the frequency response from 1hz through to 150,000hz the resistance barely drops by the 150,000hz, way way less than 1%. In audible ranges it is functionally flat. When I model it with 20.3K the resistance is flat for all intents and purposes way past 150Khz. What is the point of even having the 27pf capacitor? I looked at some other EL84 designs and the frequency response curves are are completely different for the capacitor/resistor combinations they are use, the resistance is much lower and the capacitance much higher, actually having an effect at audible frequencies.

Third question, I think I understand the purpose and function of the parallel R and C connected to the OPT secondary in theory (not in practice with the values here though!), but I'm not understanding the 17/15pf cap on the primary side, I haven't seen any other EL84 based schematics that have an equivalent. Can someone also explain to me what it is doing? I'm not sure whether it is a bypass for pin 9 of the EL84 to the OPT secondary, or is a low pass filter combined with the primary impedance for for the 7247 high mu cathode, or is it something else entirely?



2010-02-16 9:00 am
1. Yes, it is the global negative feedback loop.

2. 27pF capacitor in parallel with 27k feedback resistor is there to compensate the phase shift that takes place in output transformer at high frequency.
Due to GNFB the gain of the amplifier tends to increase at high frequency and this can be compensated with this parallel capacitor.
The value needed is depending a lot on the characteristics of the OPT.

3. This capacitor - from the UL-tap to GNFB-line - serves the same purpose. It is part of high frequency negative feedback, but now local (inside the output stage) not global (from output to input).
These two separate high frequency compensations ensure the stability and linearity better than the typical "one capacitor" compensation at the main GNFB-line.
This construction was typical to most Dynaco amplifiers.


2010-02-16 9:00 am
No, it is not that easy.
The values of the NFB-resistors can be accurately simulated, but it is quite difficult to model or simulate the real effects of the capacitors used in NFB-circuits.
The reason is that the characteristics of output transformer determine the values of these capacitors.
If we had very accurate transformer models, then the simulation could be possible.

In general, when the amount of GNFB is reasonable, below some 10...12 dB, the need for capacitive compensation in NFB-circuits is small or it does not exist at all.
Contrary, in case of "Williamson"-class GNFB (above 20 dB) the overal performance and stability of the amplifier depends a lot on the used OPT and GNFB-loop features.

When I fine tune the new amplifier, I first adjust the the optimum amount of GNFB (with the resistance in NFB-loop) and then optimize the transient response with the parallel capacitance.
One way to do this is to use 10 kHz square wave as a test signal and optimise the slew rate and ringing.
This process is usually a search for a good compromise.
The key-component is the OPT and it's features.


2011-06-10 3:25 am
Thanks artosalo, that helps, I think I'm starting to get it now.

Unfortunately I don't have a 'scope (I've been thinking of getting one though), so I guess adjustment would just have to be done an a trial and error basis with a trim pot for R, and alligator clips for switching out C.

I'm going to be using a reproduction Z565 for an OPT, here is a test of one (pages 4-6 are a repro, 1-3 are an original) for power and distortion at 3 different output levels, is that helpful at all? How do I determine what range of C values to try?


Too bad trimmer caps aren't available in 500v ratings...


2010-02-16 9:00 am
If I were without test equipment, I would begin without capacitors first and with bigger resistor that suggested.
Then the stability problems which are typical with (too ) high GNFB can be avoided.

From the documents of Z565 can not directly have any information about required GNFB components.

This type of ceramic trimmer capacitors cope with high voltages:
JFD Precision VC56C Piston Trimmer Capacitor 5 PF ? 50 PF Lot of 2 | eBay

Last edited: