Oscillation Disappears with Dummy Load

[dorrisant]

Why would an oscillation happen when on a speaker load but disappear when run on a dummy load?

I am probing the output jack, same settings, just changing from speaker load to attenuator. I've tried 3... resistive load, Weber Mass and Ultimate. All three will kill the oscillation. I can run it through a Friedman Mic-No-Mo and still see the osc on my scope. I have to have the amp almost dimed to get it to happen, so it is very annoying. (Note to self: I really need to build an isolation box.). I removed the NFB and it goes away, so I will play with that value (8k2) and see if it goes away. Probably needs to be a bigger value.

Anyway, thanks for reading! Let me know if you have an answer.

[R.G.]

Do different speakers change the oscillation, or does any speaker make it oscillate? Only some speakers?

Edit: this comes from the thought that something about the speaker load is not duplicated by a dummy load of any kind. Speakers get very inductive at high frequencies, and inductive loading is hard on any amplifier with feedback. So the question in my mind is "how much of a speaker not modelled by a dummy load is sufficient to make it oscillate?"

[pdf64]

It may be due to the complex impedance of a speaker load, eg several phase flips.
A low value cap across the series feedback resistor can sometimes help.
Rolling off the HF inside the loop can help too.
A schematic is required really to comment further.

[dorrisant]

It happens with any speaker. I raised the NFB value from 8k2 to 15k and it went away. I watched it shrink in steps on a decade box.

[bepone]

What is the osc. frequency?
Oscillations are showing how good the amp is.

Measuring with the resisstance load is often wrong because R is not triggering the oscillations..

[neskor]

Once I had similar problem, it was coupling between reverb tank and speaker magnet when dimed.

[Stevem]

Oscillation is good in a way because it means you have a heathy level of overall gain taking place from input to output.
That said, the fact that NFB when implemented right will lower gain within the circuit that it surrounds until that amount of db reduction that the feedback voltage provides drops off to a insufficient level.
This kinda tells me that something odd is going on in your amp.

The major pain in the ass when you have a oscillation issue is getting to the root cause of it and not just unknowingly doing a bandaid fix.

Can you post up gut shots of the amp and even a hand drawn schematic if need be?

[dorrisant]

The amp was a 6V6 ODS copy from Ceriatone. A customer built it but couldn't get the osc dialed out. It was only at high master volumes (4 o'clock or higher) that the osc would start. I lowered the volume and chopsticked the lead dress to watch the osc jump or shrink. If the osc would stop, I would turn up the master a bit more to make it start again. Then I would chopstick other leads to see if the osc would drop in amplitude even more. I ended up re-routing two wires (pi plate and NFB lead) and just moving one wire (other pi plate) down against the chassis to minimize it as much as lead dress would allow. Then I went up from 8k2 in decade steps with the amp totally dimed. It was barely present at 12k and completely gone at 15k, so that is what I went with.

Sorry I didn't get any pics, amp is back with owner now. Odd thing was that he was using this as a Small special but with high plate values similar to #183. He has the OD switching disabled and no tube in slot #2 for increased headroom. He noticed right away that the amp was a bit more lively, probably due to the decrease in NFB. He though it was just because the osculation was gone. The thing is, most times I see an oscillation I also see a sharp rise in the AC mains current that will coincide with the osc appearing and disappearing. This one was not doing that at all. Current remained constant, osc on or off. Also, the frequency was way above audible by itself, but would cause a cracking sound that trailed after held notes. Not sure what the frwq was exactly, I'm using my backup scope as my other one is down and it has the frequency counter.

All is good now and if I don't hear from the owner again, it is probably because it is fine now. That the thing about the repair business... Usually no news is good news.

[Stevem]

On many amps that are oscillating I find that it’s related to the length of the PI wires.

Since you can’t just move that tube closer to the outputs the best thing to do ( and especially when laying out a new build ) is to move the PI output coupling caps right to the grid of the output tubes.

Many times I can even do away with that tone sucking grid resistor or atleast get it down to 1200 to 820 ohms.

This works because any tubes input grid is a high impedance and acts like a antenna.
The output of a tube is a much lower impedance and as such far less likely to cause a oscillation problem.