DIY Distortion Analyzer/Analysis?

Hi Guys

As someone mentioned way back, it would be much simpler to build a few fixed-frequency oscillators and possibly some matching filters for analysis. When you target <1ppm performance things are difficult enough for producing a low-distortion signal and then measuring the outcome precisely, so why not make things simpler by avoiding continuously-variable frequency?

For me, a 1kHz and a 20kHz frequency are fine. 1kHz will generally show you what THD is like at the most sensitive part of our hearing, where 20kHz will show you how linear the amp is and how susceptible it might be to TIM.

Have fun
 
update:

Ok I received the whiz bang sound interface. I ran self test in rightmark and here are the results. IMO not too good but it maybe my very old xp machine usb
 

Attachments

  • IMG_0998.jpg
    IMG_0998.jpg
    986.4 KB · Views: 424
Disabled Account
Joined 2012
Last edited:
I've been looking at a few solutions for doing testing of my audio gear at home. My "play" laptop, an Acer Aspire One is the test mule. With the inputs shorted, the laptop sound card noise levels are about -70db -- not good enough for measuring anything. The first USB sound card I purchased, an M-Audio Transit USB, measures about -90db. Better, but still not great. The newest sound card I purchased, a Behringer UCA-202, measures around -110db with one weird -100db peak at 600hz. That I can live with. Both of these sound cards cost me less than $20. On the test bench this weekend is a Technics SH-9010 equalizer. I just recapped it and I'd like to list it on eBay. I'm curious to see what the S/N of this vintage piece of audio equipment will end up being. The original S/N was spec'd at 90db. I'm almost finished building my Pete Millett so I can start measuring my diyaudio amplifier projects.
 

Attachments

  • LAPTOP_soundcard.jpg
    LAPTOP_soundcard.jpg
    264.2 KB · Views: 505
  • TRANSIT.jpg
    TRANSIT.jpg
    271 KB · Views: 484
  • UCA202.jpg
    UCA202.jpg
    263.5 KB · Views: 498
Last edited:
Member
Joined 2004
Paid Member
Make sure you are talking apples to apples wehn measuring noise. The spectragrams are misleading when talking numbers. SNR is a measurement from either max level (just before clipping) or a reference level to the integrated noise over a defined band. Many times its a-weighted noise which means a very specific range of noise is what is measured and the range below 300 and above 6K is filtered out.

I think RMAA can make those tests pretty automatically. And its free. For line level stuff its pretty useful. Not so for power amps or phono/mike preamps without some special stuff in the middle.