I meant that it would have more feedback under 200Hz, than if there 0.001 cap were used alone, without the 2.2M resistor
Out of context it makes no sense to me either.. the largest capacitor which didn't make the amp oscillate was 0.001, and I changed it to a shelf filter to increase gain under 200Hz, and improve performance.
Last edited:
Not surprising. Problems with high frequency stability/oscillation can often manifest as a low frequency problem. That's what I'd look for, given this information. Hollow state amps aren't like solid state where the finals poof if there's instability that leads to oscillation at frequencies significantly above the audio spectrum. If your low frequency time constants are appropriately staggered, then look to high frequency phase margin. You might need to reduce the level of gNFB or try different phase shifting capacitors in the gNFB loop and/or the first stage Zobel.
Hollow state designs, like the Williamson, that used heroic levels of gNFB required much attention to OPT design like the Partridge OPT used in the Williamson.
I apologize to the authors of the circuit, but this is a pretty poor design. Feedback amplifiers can't be simply imagined into existence through deep breathing exercises - they need a little more planning.
Lots of folks have offered valid and useful tourniquets, but this needs a chest seal kit. I'm sorry to sound grumpy, but there it is.
All good fortune,
Chris
Lots of folks have offered valid and useful tourniquets, but this needs a chest seal kit. I'm sorry to sound grumpy, but there it is.
All good fortune,
Chris
I find it quite disturbing that many good amplifiers (tube) still sounds as good with 1/4 of the power to both B+ and tube heaters, with almost imperceptible sound degradation...
This was pretty much the only hifi tube amp ever made in the USSR, and it was made pre information age. Who knows what sort of info the designers were working from, or how much time they had to design it.
I think it isn't bad sounding at all - my intention was to stabilize it a bit. What would you suggest, a whole new front end?
I know the design is a bit weird
The design is conventional, but the implementation is weak. DF96 has already suggested inserting a dominant low-frequency pole as part of the existing CR couplings. He likes it to be upstream, because that maximizes signal headroom within the driving stages. I prefer placing it at the location of clipping (the grids of the output valves for any decent Class 1 design) because that minimizes overload recovery time. You get to choose for yourself.
But, several low-frequency poles are built into any conventional vacuum valve amplifier. Allowing the output transformer, which walks all over the county in critically important ways, to set the dominant pole is foolish, to be kind.
All good fortune,
Chris
Another way to look at the feedback design is to note that 29dB stock feedback level is quite high.
Given your oscillator goes down to 15Hz, you could do a few spot frequency checks on how much feedback you have now with the modified coupling filter. If your filter started dropping loop gain at 250Hz, then feedback may be still substantial at 50Hz.
With less feedback at hum frequencies, any hum injected within the loop is not being suppressed as much as stock setup. The simple filtering of plate, screen and fixed bias supplies may now be the dominant cause - although fixed bias ripple could be easy to suppress, the law of diminishing returns may kick in.
Given your oscillator goes down to 15Hz, you could do a few spot frequency checks on how much feedback you have now with the modified coupling filter. If your filter started dropping loop gain at 250Hz, then feedback may be still substantial at 50Hz.
With less feedback at hum frequencies, any hum injected within the loop is not being suppressed as much as stock setup. The simple filtering of plate, screen and fixed bias supplies may now be the dominant cause - although fixed bias ripple could be easy to suppress, the law of diminishing returns may kick in.
The forward gain is lower but the amount of feedback is the same. Hence the loop gain is lower, which leads to better stability.kward said:
Last edited:
There was some raspy buzz in the audio, not regular 120Hz or 60Hz hum.. it was not present before the modification. Experience told me the signature of this sort of buzz (periodic spikes on the zero signal output waveform) can be a leaky power supply rectifier diode, and given I had done a bunch of full power testing, and Soviet diodes are horrible quality, I changed the Soviet made bridge rectifier for 4x1N4007, and the buzz is absolutely gone, the remaining hum is completely inaudible. This amplifier had the worst component quality ever.. pretty soon the tubes and transformers, and a few resistors will be all that remains! I still like it though, it's very different and it is powerful!
Now, with that out of the way, the amplifier sounds very nice. If I had to make a criticism I would say it sounds a bit strong on the high end, but it is subtle. This is most likely just me getting used to an amplifier with a decent damping factor, the one I am accustomed to listening to has a very poor damping factor, and I think it suppresses the high treble due to an interaction with my speaker system's impedance. This Russian amp sounds more like a transistor amplifier to me, than a typical tube amp.
The modification I did as I understand it substantially raises the frequency of the dominant pole, and the negative feedback takes care of fixing up the low end frequency response. The EF86 doesn't swing into distortion, and the shelf filter helps preserve the low frequency damping factor.
Now, with that out of the way, the amplifier sounds very nice. If I had to make a criticism I would say it sounds a bit strong on the high end, but it is subtle. This is most likely just me getting used to an amplifier with a decent damping factor, the one I am accustomed to listening to has a very poor damping factor, and I think it suppresses the high treble due to an interaction with my speaker system's impedance. This Russian amp sounds more like a transistor amplifier to me, than a typical tube amp.
The modification I did as I understand it substantially raises the frequency of the dominant pole, and the negative feedback takes care of fixing up the low end frequency response. The EF86 doesn't swing into distortion, and the shelf filter helps preserve the low frequency damping factor.
Last edited:
Was about to repy to the question of multi pole compensation but I see other members have covered it nicely. Re:- power supply-I have found it of benefit to bypass the smoothing caps with 0.1 microfarad poly prop or class y paper caps this helps reduce hf noise on the line some designs use soft recovery diodes as well to prevent diode swiching noise.This is just one reason some use thermionic rectifiers. Mind you it seems the origional design nust have been pretty ropey !
I had trouble with some "standard" diodes on a valve mixer I designed. The didoes caused large switching spikes. So I bought in some Schottky diodes and they worked a treat.
Lack of rail decoupling on any amp can cause havoc. I unplugged my soldering iron and the mains spike blew up my transistor amp. My fault for not adding 100uf and 100n on the rails.
Lack of rail decoupling on any amp can cause havoc. I unplugged my soldering iron and the mains spike blew up my transistor amp. My fault for not adding 100uf and 100n on the rails.
Priboy, Priboj, Priboi ..
I am late at the USSR tube amp party, so I had a look first :
Priboy gives 25 results in this forum,
Priboj 44 and finally
Priboi 124 results in total ..
so you want to know where to watch out.
In fact I observed the first and second wave of Прибой amps
flooding the local market but was successful to refuse any
repair of these for about 25 years now, because it was obvious
from the beginning that their construction has many problems.
However a situation may come up where you can not refuse any
longer (no, it is no lady) and so I had two of these on my workbench
last week and took a closer look at them.
The observations in this thread were considered to be useful, but I
found no apparent problem in the clipping behaviour. My two examples
from slightly different eras had a short hickup after overload, like a half
wave overshoot, albeit with small amplitude and I think this is rather
ordinary for many amplifiers.
I also noticed no tendency for hf oscillation in my set up.
For repair I recommend using the schematic in post 15
http://knob.planet.ee/fix/priboi/Priboi_Original.pdf
because the original drawing is not clear enough and the important
bias balance pot R14 can easily be overlooked.
People experience frequent blow out of the power tubes and cathode
and screen resistors. Multiple replacement is possible, but the tubes
are not exactly available next door. The pcb suffers from desoldering,
it is of poor quality, as most of the parts and chassis. Bias is safe at
300 mV at cathodes (i. e. 30 mA current).
Some people bias at twice that .. Avoid tubes of new manufacture.
I was told by the original first importer to use tubes with date code
of or before 4-92.
Generally I can not recommend the purchase of one of these amps
for the ordinary user. For experienced people on the other hand this
product is of no particular, not even educational, value. Rearranging
the circuitry or conversion to different (output) tubes as advised here
and there seems to be an uneconomical effort and waste of time.
I am late at the USSR tube amp party, so I had a look first :
Priboy gives 25 results in this forum,
Priboj 44 and finally
Priboi 124 results in total ..
so you want to know where to watch out.
In fact I observed the first and second wave of Прибой amps
flooding the local market but was successful to refuse any
repair of these for about 25 years now, because it was obvious
from the beginning that their construction has many problems.
However a situation may come up where you can not refuse any
longer (no, it is no lady) and so I had two of these on my workbench
last week and took a closer look at them.
The observations in this thread were considered to be useful, but I
found no apparent problem in the clipping behaviour. My two examples
from slightly different eras had a short hickup after overload, like a half
wave overshoot, albeit with small amplitude and I think this is rather
ordinary for many amplifiers.
I also noticed no tendency for hf oscillation in my set up.
For repair I recommend using the schematic in post 15
http://knob.planet.ee/fix/priboi/Priboi_Original.pdf
because the original drawing is not clear enough and the important
bias balance pot R14 can easily be overlooked.
People experience frequent blow out of the power tubes and cathode
and screen resistors. Multiple replacement is possible, but the tubes
are not exactly available next door. The pcb suffers from desoldering,
it is of poor quality, as most of the parts and chassis. Bias is safe at
300 mV at cathodes (i. e. 30 mA current).
Some people bias at twice that .. Avoid tubes of new manufacture.
I was told by the original first importer to use tubes with date code
of or before 4-92.
Generally I can not recommend the purchase of one of these amps
for the ordinary user. For experienced people on the other hand this
product is of no particular, not even educational, value. Rearranging
the circuitry or conversion to different (output) tubes as advised here
and there seems to be an uneconomical effort and waste of time.
I just noticed your message nearly two years late... for me it was more about fun working on something weird. And surprisingly after rebuild it had been very reliable, including the tubes. I have noticed though that if the tubes are operated at any angle except vertical during testing, they can short circuit. I have had several fail this way.