Distortion Analysis using an Oscilloscope

[gingertube]

I went to a Royal Australian Airforce equipment disposal sale the other day. The Hewlett Packard Distortion Analyser I had my eye on finally sold for AUS$2700. I bailed out of the auction at AUS$600.

I have a 400 MSPS Digital CRO (60 MHz analog bandwidth) with deep memory and FFT capability BUT due to the 8 bit digitiser that is ONLY good for distortion products down to about -45dB, hardly ideal.

I seem to recall an old paper discussing using Lissajous Figures on an oscilloscope to look at distortion. Might have been a Crowhurst paper.

Does anyone have a useful link to info on distortion analysis using Lissajous.

Thanks,
Ian

[ilimzn]

Your scope may be much more useful if you can null out the fundamental, that way you (theoretically) have -45dB down from the strongest harmonic, much more useful.
I'm using the same technique to increase resolution of a PC soundcard based setup - I use a HP332 distortion meter's output, which has the fundamental filtered out.

[Conrad Hoffman]

For tube circuits, where you're typically not dealing with the 0.0000017% THD produced by all good solid state amps, a pretty straight forward circuit should suffice. There's a distortion analyzer based on a twin-T notch filter that's been published in various circuit how-to manuals. I'll try to find a copy. Or just build a generic twin-T filter for the frequencies you want. IMO, looking at the output of that on the scope should tell you a lot. Another more involved path is to build the analyzer on Bob Cordell's site. It was published in Audio magazine way back in "the day". Very little will compete with it, but to build it today might be expensive if you have to buy the rotary switch parts new- a lot of wafers!

[pmillett]

If you're just interested in FFT spectra you can do well with a $150 PC sound card. Mine gives me better than -110dB floor if I'm careful.

I really do prefer looking at the distortion residual on a scope, though, especially if "tuning" something. To do that you need a distortion analyzer, or at least a very deep notch filter.

Pete

[Conrad Hoffman]

There was a simple passive twin-T notch filter published in EW years ago. There was also an FET based version with frequency range switching published in various places- maybe an old National Semiconductor FET app guide. Marston published an updated version using opamps in a test equipment circuit book that looks about like this.
I was never able to get the FET version to work, maybe wrong FETs, or maybe a circuit error. The opamp version should be easier, and range switching could be added.

[kevinkr]

Originally posted by pmillett
If you're just interested in FFT spectra you can do well with a $150 PC sound card. Mine gives me better than -110dB floor if I'm careful.

I really do prefer looking at the distortion residual on a scope, though, especially if "tuning" something. To do that you need a distortion analyzer, or at least a very deep notch filter.

Pete

I use Arta and Audio tester with a 24 bit M-Audio 2496 useful sfdr is about 130dB if you are very careful. With a few simple external accessories you can do all sort of measurements electronic circuitry and speaker systems.

On the generator side (sound card output) I have a small scaling amplifier so that I can produce output voltages of up to 4Vrms into 600 ohms.

On the analyzer side (sound card input) I have an attenuator to extend the dynamic range beyond that of the sound card, a buffer to provide high input impedance, and a clamp to prevent overdriving the sound card inputs.

I also have a homebrew mic amplifier and a beringer instrumentation microphone for doing acoustical measurements.