I am trying to use the spectra plus analysis software to test the distortion of my soundcard(soundblaster live). I know it is not a good card but the tests done athttp://www.pcavtech.com/soundcards/ct4620/index.htm show that .01% THD+N using the loop back test can be achieved. However, when I send a 1kHz sine tone(-0.6dB) I get about .09% THD+N and when I test my signal generator(1kHz) I get over 2% THD+N. These seem too high, could these errors be caused by other components in the computer? I have been working at it for over a day now and can not seem to get the correct results. Any help would be greatly appreciated!
My function generator is very very old Mark I. I didn't realize that 2% THD was not out of the question for old equipment. On that note, can you use your sound card as a signal generator sending it through the unit under test back into the SC's input to analyze it. If so, I wouldn't need a new signal generator right? I hope....
I may be recieving a Lynx One Sound Card next week so that will hopefully double as my signal generator and analyzer.
I may be recieving a Lynx One Sound Card next week so that will hopefully double as my signal generator and analyzer.
A long time ago I had a Tek SG501 on my workbench. It was the companion to the 501 distortion analiser. Wish I still had that sertup, it was simple to use and very precise. I recall that the oscillator’s sine wave output was good .001% thd.
I also had this HP Function generator that did a lot of nice sweeps and odd waveforms. Besides being 3ru high, the other trade off for all that neat functionality was that the sine wave was not all that clean.
But jeessh 2% is ruff. Unless the unit is failing I would suspect that there may be a gain stage issue in your test setup. Other than that, all I can think is that it may be an adapter or cable issue... Sounds like you’ve been driving at that issue for some time now.
Having messed with the software for all of 10 or 15 minutes now, I would say that your suggestion of using the sound card output to feed signal to the device under test, out of the dive under test, and back into the soundcard for analysis, is perfectly sound (sorry for the pun). The only caveat I can think of is that the PC is fast enough to perform the function generator and analysis functions in unison. I don’t think that that should require too much horsepower.
I like the software, couldnt find the price. Can I ask what it costs?
Performing the loop test provides you with the base line test system performance. Breaking the loop and inserting the device under test will show the value of the test apparatus plus the device under test. In theory, you then subtract the loop test value to get the value for the device under test.
A few years ago I had a nice talk with Sam Berkow, who writes the SMART software distributed by JBL. Sam is a very sharp acoustician. He’s also not at all to remember talking to me, I’m not. The bulk of our conversation was about using PC sound cards test purposes. No reason not too. If there is something ugly in the cards design, it’ll show up in the loop test.
I tried looking up the Mark I. I'm afraid that I'm not familiar with it, and couldn’t find out much from googling. Was that HP, Tek...?
I'm also not sure that I’ve read your posts closely enough. Hope I'm not telling you and bunch of stuff you already know.
-Dave
I also had this HP Function generator that did a lot of nice sweeps and odd waveforms. Besides being 3ru high, the other trade off for all that neat functionality was that the sine wave was not all that clean.
But jeessh 2% is ruff. Unless the unit is failing I would suspect that there may be a gain stage issue in your test setup. Other than that, all I can think is that it may be an adapter or cable issue... Sounds like you’ve been driving at that issue for some time now.
Having messed with the software for all of 10 or 15 minutes now, I would say that your suggestion of using the sound card output to feed signal to the device under test, out of the dive under test, and back into the soundcard for analysis, is perfectly sound (sorry for the pun). The only caveat I can think of is that the PC is fast enough to perform the function generator and analysis functions in unison. I don’t think that that should require too much horsepower.
I like the software, couldnt find the price. Can I ask what it costs?
Performing the loop test provides you with the base line test system performance. Breaking the loop and inserting the device under test will show the value of the test apparatus plus the device under test. In theory, you then subtract the loop test value to get the value for the device under test.
A few years ago I had a nice talk with Sam Berkow, who writes the SMART software distributed by JBL. Sam is a very sharp acoustician. He’s also not at all to remember talking to me, I’m not. The bulk of our conversation was about using PC sound cards test purposes. No reason not too. If there is something ugly in the cards design, it’ll show up in the loop test.
I tried looking up the Mark I. I'm afraid that I'm not familiar with it, and couldn’t find out much from googling. Was that HP, Tek...?
I'm also not sure that I’ve read your posts closely enough. Hope I'm not telling you and bunch of stuff you already know.
-Dave
If it's like most function generators, it's actually a triangle wave generator with a differentiator to make the triangles into squares and a sine-wave converter which more-or-less crudely curves the side of the triangle to simulate a sine wave. And often that converter has a trimpot which you can use to minimize the distortion, though it will never be very low.
Integrators and differentiators
Absolutely. Function generators usually produce about 1% THD and a "Norman arch" can generally be seen at the peaks of the sine wave. However, they have their uses. The response flatness when sweeping the frequency of a function generator from end to end of its range is much better than a typical Wien bridge oscillator (which struggles to better 0.1dB). You almost always need a frequency counter with a function generator because their frequency calibration is poor, whereas the frequency calibration of valve Wien bridge oscillators is generally very good.
Absolutely. Function generators usually produce about 1% THD and a "Norman arch" can generally be seen at the peaks of the sine wave. However, they have their uses. The response flatness when sweeping the frequency of a function generator from end to end of its range is much better than a typical Wien bridge oscillator (which struggles to better 0.1dB). You almost always need a frequency counter with a function generator because their frequency calibration is poor, whereas the frequency calibration of valve Wien bridge oscillators is generally very good.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.