A tube amp has the potential to oscillate if there is no input connected or a load on its output. I am aware of this but many people are not.
What type of protection circuit should I add to prevent the potential of oscillation with...
Case A. No speaker connected.
Case B. No input connected.
The amplifier is a KT88 push pull parallel unit. Case B would be easy if RCA shorting jacks existed (if they do, feel free to point me to a source).
What type of protection circuit should I add to prevent the potential of oscillation with...
Case A. No speaker connected.
Case B. No input connected.
The amplifier is a KT88 push pull parallel unit. Case B would be easy if RCA shorting jacks existed (if they do, feel free to point me to a source).
The reason many people are not aware of this is that it is not generally true. Valve amps can be damaged by being run with no output load but it is not oscillation but hard limiting which creates the problem. If you can't guarantee a load will always be present you will have to add a Zobel network.
Oscillation with no input connected is probably a sign of poor layout, and possibly poor design with marginal stability. Don't add a 'protection circuit'; correct the design.
Oscillation with no input connected is probably a sign of poor layout, and possibly poor design with marginal stability. Don't add a 'protection circuit'; correct the design.
I bet it is, but you're not looking in the right place ! If it buzzes, check unscreened input leads and make sure ones lab bench light doesn't have a flourcesent tube. A well designed amp should be able to accept hot input switching between an o/c and a dummy input load should be reflected in a slight noise level difference. The exception is when those nasty RCA input plugs are used.
I'm just finalising my powerful 250+250W amp; no way would I allow such problems !
richy
One way to get buzzing when no input is connected is to have a huge input capacitor (too high in value, possibly audiophile boutique component) which acts as one plate of a capacitor. The other plate is every nearby conductor, including mains supply, rectifiers etc.
Try a smaller capacitor. Mount it nearer the chassis. Put a few pieces of grounded wire over the top of the cap to act as a crude electrostatic screen.
Try a smaller capacitor. Mount it nearer the chassis. Put a few pieces of grounded wire over the top of the cap to act as a crude electrostatic screen.
You shouldn't have this problem. If you do, then you need to fix your open loop design before you add the NFB. (One project I did showed a weak oscillation at ~100KHz with the input unloaded. A grid stopper fixed that problem.)
The only project that oscillated when running with an unloaded input was the audio deck for a TRF longwave receiver that had a couple of op-amps and 120dbv of audio gain. Without the CW BPF switched in, it oscillated at 2500Hz. That stopped with an input load. Given the extreme gain, that (and microphonics) were unavoidable. (And one reason why superhets replaced TRF's: gain distribution between IF and AF.)
My KT88 PPP amplifier resonates when I connect a 4 ohm speaker to the output. This was with the input shorted or a source connected. The frequency was about 310kHz. Its way up there. The 8 ohm dummy load presented no oscillation.
I wonder if that was coming from the speaker winding resonating?
As soon as I can I will try it with other speakers and see what happens.
I wonder if that was coming from the speaker winding resonating?
As soon as I can I will try it with other speakers and see what happens.
Last edited:
It's becoming popular here on DIYAudio to install quenching diodes from the output valves' anodes to ground. This seems to be enough to protect the output transformer from misadventure, and has no important ill effects. Especially with pentode outputs, DIY amps have been known to oscillate without load, and some commercial ones too.
All good fortune,
Chris
All good fortune,
Chris
Right. But plug/socket pair must be made such a way that ground is connected before input, or/and both devices are properly grounded.
I did such thing to guitar amps that I repaired. No more complains on damaged output transformers after that.
How many secondary tap-offs? It's highly doubtful that it's speaker self resonance that's doing it. One common way to connect gNFB to the OPT secondary is to make the connection to the highest secondary tap-off, not necessarily the same one to which you connect the speaker. It's something I avoid with my own projects when using stock OPTs that include a variety of secondary tap-offs. It's a convenience for the end user, not having to move the gNFB pick-off point when connecting speeks.
Nyquist Criterion: "The number of unstable closed loop poles is equal to the number of open loop unstable poles plus the number of encirclements of the point: -1 + j0". Of course, you avoid open loop unstable poles unless it's your intention to make an oscillator.
That makes for stability problems since the unused portion of the secondary can have parasitic reactances that are unloaded. Also, connecting speeks with different impedances can change the magnitude and/or the phase of the feedback signal. That can lead to hitting that nasty -1 + j0 point, or coming too close to it (which gives what can be problematic increases in gain over a narrow range of frequencies).
Do spark gaps of appropriate trigger voltage do the job equally well? I would also like to protect my OPTs, but if possible without giving up the 'no semis' design idea...
Greetings,
Andreas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.