I've found it less interesting to pump square waves into solid state amps because they produce them so damn well. 
For tube amp tuning and feedback loop tweaking, I do this:
I start out with 1 KHz square waves at a couple watts. (I use 15 watt 8 ohm resistors for loads). If everything looks good there, I go down to 20 Hz and check my coupling caps and phase shift there. Hint: With a tuber amp 20 Hz square waves are rarely square. If I'm satisfied with the low frequency response, I head up to 10 KHz and check the feedback loop compensation cap and any other compensation caps in the forward loop. I look at the transitions and the ringing behavior.
Unless it's a super powerful tube amp, I then switch to sine waves and run it up to clipping.
While I've got the generator out, I also check the bandwidth. then I usually pull out the spectrum analyzer and look at distortion spectra. But I like to make sure the amp looks good with sine and square waves first. It's a lot faster to sweep the function generator than to wait for the spectrum analyzer.
Sheldon
For tube amp tuning and feedback loop tweaking, I do this:
I start out with 1 KHz square waves at a couple watts. (I use 15 watt 8 ohm resistors for loads). If everything looks good there, I go down to 20 Hz and check my coupling caps and phase shift there. Hint: With a tuber amp 20 Hz square waves are rarely square. If I'm satisfied with the low frequency response, I head up to 10 KHz and check the feedback loop compensation cap and any other compensation caps in the forward loop. I look at the transitions and the ringing behavior.
Unless it's a super powerful tube amp, I then switch to sine waves and run it up to clipping.
While I've got the generator out, I also check the bandwidth. then I usually pull out the spectrum analyzer and look at distortion spectra. But I like to make sure the amp looks good with sine and square waves first. It's a lot faster to sweep the function generator than to wait for the spectrum analyzer.
Sheldon
I agree with Sheldon, I typically test with a 1kHz sine wave at 1 watt into 8 ohms. Initially, I use a light bulb connected in series with the amp being tested to limit the current (a Variac also works well at this point).
If the amp looks stable, the I plug it directly into the AC outlet, still at 1W and 1kHz. Then I perform a sweep from 10Hz to 100kHz, all at 1W, monitoring the THD. If all looks good, I then check the full power output, first at 8 ohms then at 4 ohms, and possibly lower depending upon the amp. (If you wish to stress the amp, you can run it at 1/3 power for 1 hour with a 1kHz sine wave and then measure it's THD at full power across the audio spectrum using sine waves, but you should know the amp is reasonably stable before doing this.)
After all this, then I look at the square wave, and then usually just at 10kHz into 8 ohms, to check for ringing and slew response. If you want to further check stability, you can parallel the 8 ohm load with a 0.47 to 2.0 uF film cap to see its effect upon the square wave (ie, ringing, overshoot, etc). Sometimes at this point, adding 5 to 20pF (or maybe more, again depending upon the design), in parallel to the feedback resistor, can help clean up this square wave response.
I have tested some amps to full power with a 20kHz square wave, but only after I'm sure all is stable. I wouldn't recommend going out further than that as you risk blowing your output devices (or burning up the RC filter on the output).
Regards, Robert
If the amp looks stable, the I plug it directly into the AC outlet, still at 1W and 1kHz. Then I perform a sweep from 10Hz to 100kHz, all at 1W, monitoring the THD. If all looks good, I then check the full power output, first at 8 ohms then at 4 ohms, and possibly lower depending upon the amp. (If you wish to stress the amp, you can run it at 1/3 power for 1 hour with a 1kHz sine wave and then measure it's THD at full power across the audio spectrum using sine waves, but you should know the amp is reasonably stable before doing this.)
After all this, then I look at the square wave, and then usually just at 10kHz into 8 ohms, to check for ringing and slew response. If you want to further check stability, you can parallel the 8 ohm load with a 0.47 to 2.0 uF film cap to see its effect upon the square wave (ie, ringing, overshoot, etc). Sometimes at this point, adding 5 to 20pF (or maybe more, again depending upon the design), in parallel to the feedback resistor, can help clean up this square wave response.
I have tested some amps to full power with a 20kHz square wave, but only after I'm sure all is stable. I wouldn't recommend going out further than that as you risk blowing your output devices (or burning up the RC filter on the output).
Regards, Robert
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.