Tube Amps / Music Electronics
|For current discussions, please visit Music Electronics Forum.|
|10/12/1999 1:53 PM|
||Re: OT protection tests|
Thanks for running this test and reporting the results. I, for one, am glad to see that the 260 ohm resistor did the job of limiting the reflected voltage to a safe level, at least during a steady state "high load" condition.
If you get another chance to run a test, you might want to try running your existing setup, but with a speaker load connected. With the amp running hard, pull the plug on the speaker cabinet to see what happens with the reflected voltage at the initiation of the inductive spike. I wonder if the voltage will climb high enough to jump your spark gaps? It may only be one tiny, short duration spark, so it might be a good idea to perform the test in low ambient lighting in order to not miss winessing an arc. This test would simulate actual operating conditions we would most likely encounter with guitar amps. (How about those undetected loose speaker jacks on a shredder's Marshall cabinet, or an accidental pull caused by stepping on the cable mess on stage?)
A tiny arc jumping an external safety gap is much better than a pierced insulating layer on an expensive output transformer primary. Even a single shorted turn causes enough inefficiency to raise the transformer's normal operating temperature, accelerating degradation of the remaining good insulation.
|And now, a word from our sponsors:
|10/12/1999 4:57 PM|
I'll try pulling the plug out and see what happens. I won't be able to use a real speaker load though, the noise would be unbearable. I don't know what difference not having a voice coil inductance, mechanical resonance, etc, will make to the results.
|10/13/1999 5:20 AM|
Alright Stephen! Hard core proving!
The difference might be that the amp will already be putting out more AC voltage into a speaker then it could into a DCR dummy load, with everything else being the same.
This of course would depend on the frequency and the true reactance of the speaker!
As the field breaks down in the speaker (from being disconnected abruptly) would the counter EMF flyback to the secondary on one side and then on to the primary?
Maybe that was dumb...
|10/13/1999 2:43 PM|
More results from the secret tone lab. Doc asked me to try winding it up with the load connected, and then removing the load to simulate a dodgy connection.
The test rig was the same as before, with a safety spark gap between each OT primary leg and the centre tap, and a 510 ohm resistor across the OT secondary. I connected a 16-ohm dummy load, and drove the amp to rated power with an 80 Hz sine wave, so it was just beginning to clip. I then broke the connection to the dummy load.
Again, nothing exciting happened. The output waveform became almost square, and the p-p voltage got half as big again. I didn't observe any huge spikes or ringing. I tried disconnecting and reconnecting the load several times, and got more or less the same results. Even with the lights out I saw no sparks jumping the gaps.
Now it's time for you to have a go.
|10/13/1999 7:06 PM|
Thanks again for your lab work. I'm sure you felt a little shaky, and wondered if you were sacrificing a perfectly good amplifier. We commend you for taking the bull by the horns.
Your results are conclusive enough for me, proving that the extra "safety" resistor wired permanently across the output transformer secondary is effective in limiting harmful inductive spikes that can be created from the collapse of the magnetic field upon open-circuiting the normal load. It seems that no other protective devices are really necessary, at least for guarding against this particular failure mode.
Anyone building, servicing, or improving an amp can add one inexpensive resistor and gain a large margin of safety for their output transformer. As I mentioned in an earlier thread, this was common practice in Fisher and Scott circuits, and someone else noticed it in some others, from our golden Hifi era. So it's demonstrated prior art. Giving the technique some cute name and taking out a patent can't keep the rest of us from using it. Y'all know what I mean.
|10/13/1999 9:11 PM|
|Gus||Lets watch and see if someone takes...|
Lets watch and see if someone takes out a patent, like people who try to patent electron flow in tubes. I know what you mean Doc.
|10/14/1999 5:22 PM|
|Geoff Van Brunt
||Re: OT protection tests|
I hate to rain on a parade, but there is one more test that may be required to see if this "mod" really does the trick. I remember reading somewhere (I think it was Lord Valve on AGA), that a common destroyer of 2 Ohm OT's in old Fenders was the use of a 16 Ohm cabinet. This is of course only an 8x mismatch, not the nearly infinite mismatch of an open circuit, but there is the issue of back EMF, and the complex load that a real speaker presents to the amplifier.
When the field collapses in the OT so does the field in the speaker and the cone starts to return to its normal position creating a current that is forced back into the transformer. AFAIK this counter-current is kept in check by the damping factor of the amplifier. But when a mismatched load is presented to the amp, the DF cannot keep it in check, and large voltages may present themselves on the secondary, and even larger ones on the primary.
There is also the issue of the complex load the speaker presents. It does different things at different frequencies and combinations thereof. Steven used a 80Hz signal, a lot of which would be rolled off by the circuitry of the amp itself, so that signal at the secondary, would be much lower than a higher frequency. Although he did drive the signal into hard clipping, the harmonics created would be less than the original signal. I think a 1 kHz signal would have been a better choice. However it is still not a complex signal like a guitar, which would be an even better choice.
I'm no guru when it comes to power amp theory (or electronics for that matter), so forgive me if I've made some mistakes, in the above. Even if the theory is not quite sound, I'm one of those "seeing is believing" type of guys, and I like to see test results from real world use of the device under test.
The problem with using a real guitar signal however, is that it is hard to measure peak voltages at the secondary, as the guitar signal can't be locked onto with a scope. A peak detector would be required. A dual - op amp setup as a buffer and high impedance load with a cap in between would work, but it would have to handle some pretty high voltages, as the power amp is already putting out @ 40V RMS into a 16 Ohm load, which is almost 57V p-p. Of course it should be lower if there is an impedance mismatch as in our test. And there is the issue of loud volumes associated with the speaker being used.
Any idea's, comments or suggestions?
|<<First Page||<Prev||Page 2 of 4||Next>||Last Page>>|