I made further measurements and it really seems that the spikes and oscillations are happening exactly at the same times the rectifier stops or starts conducting (=when there's rapid change in transformer current).
So it seems that I would need some kind of snubber in parallel with mains transformer secondaries. Anybody has experience about what kind of snubber would remove this kind of problem?
So it seems that I would need some kind of snubber in parallel with mains transformer secondaries. Anybody has experience about what kind of snubber would remove this kind of problem?
It is totally clear where the buzz is coming from. Switching noise is magnetically coupled from mains transformer to the output transformer. I made some metal shielding that I put around output transformer and when I take it off, I can clearly hear difference in volume of the buzz. And I can see difference in amplitude of the spikes on the anode voltage.
The question is just how to fix this. I tried 10nF capacitor and 1k resistor in series as a snubber in mains secondary, but I didn't notice any difference.
Maybe I should consider putting this to pentode mode instead of ultralinear. I guess pentode would be more tolerant to the anode voltage disturbances. And maybe I could then fix rest the NFB.
The best way to reduce magnetic coupling is to add some distance. Sometimes changing orientation can help, but not always easy with a big transformer. Are you sure it is the transformer, and not the wiring? Toroids don't couple very much to things around them.
Is the heater supply clean? You can get rectifier spikes from the HT supply getting into the heater wiring and then getting into the audio circuitry. I once had this with an elevated heater supply; no amount of decoupling stopped it but grounding the heater CT killed it completely. You do have a DC reference for the heaters?
Is the heater supply clean? You can get rectifier spikes from the HT supply getting into the heater wiring and then getting into the audio circuitry. I once had this with an elevated heater supply; no amount of decoupling stopped it but grounding the heater CT killed it completely. You do have a DC reference for the heaters?
I'm pretty sure it is the transformer magnetic coupling because I can see the spikes in anode pin if I remove the tubes from the amplifier. And also because I was able to reduce the buzz and spikes seen in scope when I put some metallic shielding around the output transformers.
The spike is from a poor rectifier and filter setup. For example, if you use the smallest current rating UF type rectifier diode, and I suggest without any cap in parallel with a diode, and layout the diodes and first filter cap to minimise what noise gets past the first filter cap in to your amp then that should be a good start. If the PT has substantial leakage inductance, then a snubber, as per bell-ringer threads may assist further.
The rectifier turn-off spike has gotten past your main filter cap and is capacitively coupling through the transformer primary winding to secondary winding interface, so that you see a spike on the speaker when you have removed the output stage valves.
Your original scope plot showed low frequency hum as well - that ingress path is not yet identifiable I suggest.
The rectifier turn-off spike has gotten past your main filter cap and is capacitively coupling through the transformer primary winding to secondary winding interface, so that you see a spike on the speaker when you have removed the output stage valves.
Your original scope plot showed low frequency hum as well - that ingress path is not yet identifiable I suggest.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.