|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.140.192.106
I made a simple passive preamp with a 100K 2W AB pot to make fine tuning the output signal possible while testing.
Test signal was 1KHz sine.
At 1W: Left Channel THD = .63% THD + N = .66%
At 1W: Right Channel THD = .66% THD + N = .69%
At 2W: Left Channel THD = 1.16% THD + N = 1.16%
At 2W: Right Channel THD = 1.12% THD + N = 1.12%
At 3W: Left Channel THD = 1.74% THD + N = 1.74%
At 3W: Right Channel THD = 1.58% THD + N = 1.58%
At the 3W power level, the left channel 3rd harmonic was 3dB above the 2nd harmonic. The right channel 3rd harmonic was about 1 dB above the 2nd harmonic.
I have another SE amp that uses global feedback, so I will test it to see if the 3rd harmonic is that pronounced as in this circuit.
Follow Ups:
with the passive preamp hooked up and your signal going into the passive preamp and then to the power amp? If that's the case,that isn't going to give you can honest reading even if it is passive.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
"with the passive preamp hooked up and your signal going into the passive preamp and then to the power amp? If that's the case,that isn't going to give you can honest reading even if it is passive."A shunt pot might work better to present a fixed impedance input.
Got another idea I can try out?
I'd like to fine tune the test signal coming out of the soundcard.
My THD measurements and the computer model were pretty close.
Edit: OK, I tried a shunt pot with 47K resistor connected to the 100K pot.
That gave me about the same distortion measurements as the 100K pot or no pot connected to the amp.
Edit 2: tried adjusting ARTA signal generator level and soundcard output level.
That got output voltage across load resistor close to target. Once again, distortion measurements seemed to be about the same as before.
Since the computer model and my measurements at the output of the amp were close, I will call it good enough unless someone else has any ideas.
Edits: 09/30/16 09/30/16
I'd like to fine tune the test signal coming out of the soundcard.
Which soundcard are you using? I can't believe the money they get for the Asus 124db soundcard.
Since you are dealing with very low power in this amp,different anomalies in the output transformers alone could show different specifications in the final test. What is the typical distortion rating for this type of an amp? Try loading the input of amplifier with no shunt pot hooked up and leave the 100k pot wide open and then do the test. Lets not anything influence the input signal whatsoever.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
My RH test amp has NO built in attenuation at the input, just a fixed grid resistor.
Not sure if you checked out the photo of the test amp or schematic I had posted previously.
The 100K pot was connected between the soundcard and amplifier input.
The pot was just lowering the input signal a bit in order to get an exact voltage over the load resistor so I could test at a specific power output.
This does not affect the distortion readings to any degree.
I have also removed the pot and just used the ARTA internal signal generator in combination with the soundcard level output to get a more precise output test signal level.
The soundcard I have been using for a few years is the ESI Juli@ and it has low distortion on the loopback tests. More than good enough for non-pro amp benchmarking.
I have read that the original RH amp benchmarks at 1% THD @ 5W, that was a simulation. Not sure if a real circuit was measured or not.
I took my left channel up to 4W and the THD% was 2.37% with the 3rd harmonic being dominant.
Hope this helps clarify things.
Having the third harmonic dominate in a SE amp suggests some "distortion cancellation" is happening. you should check the distortion at the grid of the output tube at 3W out and see if it is the driver stage that is causing this.
If the distortion is vanishingly low at the driver then the other place to look is the output transformer and a SE output running too much current will start generating even harmonics as you hit saturation.
dave
With the power output of the amp @ 3W , driver stage distortion:
Left Channel : THD & THD+N = 13.59% 3rd harmonic down 8.5dB from 2nd
Right Channel: THD & THD+N = 12.38% 3rd harmonic down 9.6dB from 2nd
In the non-feedback SET circuits I have built in the past, driver stage distortion would be very low compared this circuit.
Comments?
Why are those numbers so high? Even without GNFB to the cathode, the driver's output shouldn't be that distorted. And if it was, those numbers would show up at the output stage. Seems like something is wrong with the measurement.
--------------------------
Buy Chinese. Bury freedom.
Edits: 09/29/16
I thought that too but then not being well versed in the evils of feedback I built the circuit in spice and the numbers came very close to what dt667 measured.
here is the fft of the driver and the output @ 2.8W out.
and here is the numerical rundown of the distortion
If I remove the feedback and decrease the input value to get the same power out we get the expected low distortion behavior from the driver but the overall circuit gets worse.
To keep things complete, the other important thing that changes with the feedback is the output impedance. As built, the Zout is around 2 ohms and if the feedback is removed the Zout jumps to just over 100 ohms.
dave
I was just about to test the driver stage without the feedback resistor in place when I noticed your post.
In the past, I had tested driver stages in zero feedback SET circuits with my Virtins PC soundcard probes and got valid results.
Thanks for all the circuit modeling work.
Very informative.
dt 667
My fault, I wasn't thinking about the fact that the correction signal is connected to the anode of the driver. Dave Slagle's SPICE results make this clear; driver distortion can't be measured unless Rfb is removed. You're measuring a large percentage of the correction, not driver tube output per se.
--------------------------
Buy Chinese. Bury freedom.
Edits: 09/29/16
Asking for a friend.
Here were my earlier findings.
I'll need to investigate further and report back.
Thanks for the comments.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: