c3g self oscillation problem - a newbie needs help!

[tinitus]

Originally Posted by nlinus

I'm not sure I understand that about "filament fragility" and "cool for a couple of minutes before removal".

meaning it does not like to be shaken(or sturred), especially when on and hot, and that its best to handle carefully, in general, I suppose
or in other words, not the best tube for a road touring guitar amp

[nlinus]

Aah, I get it. Luckily, I plan to have mine stationary in the living room

I have now promised myself I should not do any more work this evening, but I came to some interesting conclusions before I finished up.

Have a look at the schematics for the PSU (first post). The voltage drop seems to occur over the two 22 ohm resistors in parallel. The drop is about 500mV, and if I measure after the 10H inductor, the drop is of equal size.
The drop could only be caused by a sudden change in load, and since the drop is of equal size before and after the inductor (95 ohms) I guess the only thing that could cause this is the two capacitors in series?!

The capacitors are these and the resistors in parallel with those are 3w wire wound. The 22 ohm resistors are also wire wound.

[euro21]

C3g, C3m, C3o are low noise (German post) telecom (long distance telephone transmission, cable repeaters -under the ocean-) tubes. C3g inner -cathode- structure mostly similar than old direct heated triodes, therefore very fragile, when hot or warm. Be patient, wait few minutes before move it.

C3g is the least of microphonic tube, but some of them produce a little noise about four-six kilohertz narrow spectrum.

They are very linear tubes, my collection (few dozen) produce 0.33-0.86 % THD at 240Vpp!

But... likes power, below 20mA THD is higher than ordinary.

Nlinus, I think, this configuration is underestimated, max. swing only 100Vpp, and THD is a bit high (simulation: 0.85% at 92Vpp).

[euro21]

Sim:

[nlinus]

90vpp would be enough for me i guess since the power tubes (KT88) are biased at ~40v. Maybe there would be other benefits by increasing the load resistances...

[euro21]

If you increase anode and cathode resistor to 4K7, and increase anode current to 17mA (1V bias), at 92Vpp output distortion is 0.17%.

[kevinkr]

Yes, you've misunderstood something.. The 470K is the grid (bias) resistor, the grid stopper appears in series between the grid and everything else that is external. The grid stopper effectively reduces the Q of external wiring inductance and provides isolation between that and the grid's stray and miller capacitance, it also creates an HF pole which can be intentionally chosen to roll off HF gain above a certain frequency to reduce susceptibility to external RFI in an input stage as an example.

[nlinus]

Well then...
I must have gotten the terminology wrong. Anyway, I have a grid stopper resistor connected directly at the tube.

The voltage fluctuations is driving me nuts. Could it maybe be caused by mechanical movement/vibrations in the transformer? Or bad capacitors (although they are brand new)? Well, I'll continue with this tomorrow...

[kevinkr]

Not quite sure where you are in terms of the build. I assume you have not closed the feedback loop so far and those blips every 330ms on the supply rail aren't because the amplifier has a subsonic oscillation which it very well could. Open global feedback loop if present if this has no effect then try removing one tube (input tube) and see if the blips go away - if they do then it is feedback through the supply and you will need to revise the design to provide more decoupling between the various amplifier stages..

Also at this point it seems worthwhile if you could post a schematic of the entire amplifier.. Snapshots of your notebook pages will be just fine for this purpose.

[nlinus]

Actually only the PI and driver stages are wired, but this phenomenon appears even with all three tubes removed from their sockets.

As I mentioned earlier: The voltage drops are of equal magnitude before and after the inductor - which causes me to believe that the cause of these drops should be located before it. Have to check more tomorrow...

The transformer and inductor are made by a Polish company named Indel. I bought them here (the inductor is here). These are very cheap in comparison to other brands, but I thought I'd try. Right now, I'm not really sure if these could have something to do with my issues :s