With a scope that is set to 20us/division, then:
A 10kHz square wave will be 5 divisions long.
That should tell that it is 20us/division, or if it is 20us across all 10 divisions.
A 10kHz square wave will be 5 divisions long.
That should tell that it is 20us/division, or if it is 20us across all 10 divisions.
Unfortunately I had only analog control over that sig gen, a knob took me from 3k-65k. And I hope it's merely overheated from a long day, it was not delivering a stable signal last I checked.
Um I thought energized OPTs with no load were a bad thing. Like, "Go buy a new OPT"
Three things about unloaded output transformers.
1. Only use a Very Small signal to the amplifier input.
2. Quickly and immediately watch for oscillation, if there is . . .
immediately re-connect the load,
immediately remove the signal,
immediately power off the amp.
3. If you have about 5 hands to do number 2, you might need a new output transformer.
No matter how fast you are, there is a risk. A full blown maximum voltage oscillation can be very bad.
Certain kinds of amplifiers are unconditionally stable with no load.
Example:
Most well designed 2 stage single ended 45, 2A3, 300B with no negative feedback will be stable with no load.
The same often is true for single ended Triode Wired pentodes and Triode Wired beam power tubes (no global negative feedback).
However, the signal level at the amplifier input must be small to keep the output stage voltage swing small.
My balanced amplifier, with 2 stages, and Triode Wired beam power tubes is stable with no load.
(I use a CD player with balanced XLR outputs that barely can drive it into clipping).
For larger signal testing:
I use a generator with balanced output, and I make sure that the 8 Ohm load resistor is connected to the amplifier output.
1. Only use a Very Small signal to the amplifier input.
2. Quickly and immediately watch for oscillation, if there is . . .
immediately re-connect the load,
immediately remove the signal,
immediately power off the amp.
3. If you have about 5 hands to do number 2, you might need a new output transformer.
No matter how fast you are, there is a risk. A full blown maximum voltage oscillation can be very bad.
Certain kinds of amplifiers are unconditionally stable with no load.
Example:
Most well designed 2 stage single ended 45, 2A3, 300B with no negative feedback will be stable with no load.
The same often is true for single ended Triode Wired pentodes and Triode Wired beam power tubes (no global negative feedback).
However, the signal level at the amplifier input must be small to keep the output stage voltage swing small.
My balanced amplifier, with 2 stages, and Triode Wired beam power tubes is stable with no load.
(I use a CD player with balanced XLR outputs that barely can drive it into clipping).
For larger signal testing:
I use a generator with balanced output, and I make sure that the 8 Ohm load resistor is connected to the amplifier output.
Last edited:
I killed a small 5w 8k SE OPT in a Champ guitar circuit because I'd wired to the wrong side of a switching jack. Thing was toast by the time I figured out what I'd done. Figured that was the norm.
Not necessarily. 6A3 makes a good point, very small voltage input and just long enough to get a glimpse of the scope. He's also on point saying the some amps are stable without a load, and a feedback amp can be as well if it's tuned properly. My Williamson amps are, but that's only thanks to Dave Gillespie's tutelage. ;-)
I did briefly energize this amp with one speaker disconnected earlier today, it ran with signal through the other channel for ~15 seconds. Didn't pop.
Amplifiers can be designed to protect themselves in case of operating without a load. Not everything is, and some of the protection mechanisms can change the sound (Which is why many don’t).
The bigger stuff I build will run without a load, but not indefinitely at full power because it will overdissipate the the screens when the plate currents go too low.
A zobel network goes a LONG way - because it becomes a snubber. You get high voltages but nowhere near as high. Freewheeling diodes are a full fail safe, other than screen dissipation. Their action is audible when they do conduct - but normally they don’t.
The bigger stuff I build will run without a load, but not indefinitely at full power because it will overdissipate the the screens when the plate currents go too low.
A zobel network goes a LONG way - because it becomes a snubber. You get high voltages but nowhere near as high. Freewheeling diodes are a full fail safe, other than screen dissipation. Their action is audible when they do conduct - but normally they don’t.