VTA MK3 Oscillation. NFB?

I built 4x VTA MK3s (quad amping) from scratch a few months ago. I posted a thread when I had a phase issue I couldn’t figure out… turns out out of the 4 transformers, 2 had the secondaries taps reversed, the colour coding was wrong… that’s why it took me forever to figure it out.

Anyway, since then the amps have been working ok until I did some tube rolling. If I use a 12BH7 or E80CC, I started to get an oscillation. While monitoring the bias, I notice that if I turn the volume up on my preamp past a certain point, the bias goes up by about 10% and the amp oscillates. If I turn down the bias, the “pitch” of the oscillation lowers and then disappears.

So it’s like an input sensitivity. The 12BH7 and E80CC have a bit more gain but even with a 12AU7, if the volume is loud enough, the amps will oscillate.

I realized I neglected something about the NFB and I wonder if this could be the issue: the schematic takes the NFB (6.8K/680pf) to the 16ohms tap. The OTs I use has a single 6ohms secondary tap (and the speakers are 4ohms).

I soldered a 8k2 in parallel with the 6k8/680pf network to bring it down to 3K9 ish but now as soon as the the tubes light up, the oscillation appears right away with no signal at the input...

you really need an oscope and a sig gen to get the NFB correct,
but try 10K and 220pf instead of the orginal
the resistor controls the gain, the capacitor controls the high frequencys and square wave response

Thanks Roy! I only changed the resistor to 10K and it works fine with 12AU7s and 12BH7s. Unfortunately, still seem to oscillate a bit with E80CC although I'm not sure... it sounds more like a hum?

Otherwise, would lowering the capacitor to 220pF brighten the sound a bit? Unfortunately, I wouldn't know how to use a scope... Thanks!

I don't have a MK3 amp here to try anything, so it's all trial and error without an amp, scope, and sig gen

Regarding the E80CC, I think I'm getting the issue because it's biased too hot. If I'm reading the schematic correctly, the cathode resistor is 780r (680+100), right? If I change R6 to something like 680, the 1K36 should be good as far as bias. I'm just not sure I'm reading the schematic properly...Or should I change R10 to 1K2 instead of R6?

not sure why you're having trouble with an E80CC tube unless it has some defect, as it has the same operational characteristics as a 12AU7 except
that it uses twice as much filament current. R6 should remain 680 ohm and R20 should remain 20K unless you are measuring voltages at plate and cathode that are more than
10% different than nominal in the VTA circuit (140v plate, 0.47v cathode)
R10 affects the percentage of NFB and shouldn't be a factor.
There are tons of great quality 12AU7 tubes out their, so why mess with something that works?

I just have a few of them and I really like how they sound. In previous builds where I used them (not MK3s), I remembered having to use a higher cathode value, like 1k2, 1k5 otherwise it would distort… it has a higher mu than 12au7 (17) vs 25 with the 80. I’ll check the voltages.

I’ve just been spending time tube rolling and coming up with pretty interesting results, using 4 amps really expands what can be done between balancing highs and lows. I’ll post the results of that in another thread once I’ve figured out the 80 thing.

Hi MHProd,

You say, "The OTs I use has a single 6ohms secondary tap (and the speakers are 4ohms)." So, you are not using the original Dynaco transformer design. This could have something to do with your problem.  The Dynaco transformers are very high quality and less prone to ringing.  With the original Dynaco transformers in my upgraded driver board MkIIIs the 10kHz square wave response was very respectable.  The same amp fitted with the Edcor "equivalent" transformers rang / oscillated severely on a 10kHz square wave. Not only did I have to play with the global negative feedback I had to add some local damping upstream to solve the issue.  Note also that not all oscillation can be heard.  You may think that you have solved the problem but then start experiencing tweeter failures at higher volumes.  So, as mentioned in this thread, an oscilloscope (and audio frequency generator) are essential in troubleshooting this kind of problem. Otherwise you are pretty much shooting in the dark.  

Jack

Hey just saw this.

Maybe helpful. Not “our” amps, but shows how nfb can be tuned.

https://youtu.be/LD10RvhKNP4

Rec wrote:Hey just saw this.

Maybe helpful. Not “our” amps, but shows how nfb can be tuned.

https://youtu.be/LD10RvhKNP4

Another thing to do is to make sure the amp still does well of a 0.5uF cap is paralelled
to the resistive load. Reactive load is more difficult and must be taken care of as
many speakers has caps ( and coils) .