Audio Measurements

stephie's place

Moderator: sbench

Post Reply
hotvalve
Posts: 3
Joined: Fri Dec 21, 2007 9:16 pm

Audio Measurements

Post by hotvalve »

Greetings, and Happy New Year. :D

I hope you won't mind me asking, what do you guys use for measuring/plotting frequnecy response and measuring distortion?

I have an old Fluke 8060 DMM which had a dB function for measuring frequnecy response, but this has now given up the ghost, must be 20 years old.

I was wondering if you chaps use software and a PC?

Thinking about software got me thinking about PC distortion analysis.
What do you chaps use?

Your response will be greatly valued.
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

A lot of different techniques and equipment. Most of the distortion curves I do that show the distortion separated by component in Excel spreadsheets come from "specialized" equipment:
1. Sources are highly modified "Heathkit" oscillators with passive LC tuned circuits that net me somewhere around .001 to .0005% residual distortion.
2. Preselected passive LC notch on the receive end to remove about 40dB of fundamental followed by about 40dB of low noise gain. This drives H-P 30x series (300, 302, 302A) wave analyzers. Ancient equipment but the system allows me to see down in excess of 100dB.
3. Ballentine true RMS voltmeter is primary source of RMS level output. I keep it calibrated with my Epley standard cell (1.046+ volt absolute reference).
hotvalve
Posts: 3
Joined: Fri Dec 21, 2007 9:16 pm

Post by hotvalve »

Mr. Bench; thankyou so much for your reply.
Your opinions I value more than most. :D

Wow! .001 to .0005% from a Heathkit? Thinking of it, I'm sure I read something about this on the 'Fun With Tubes' site; I'll have to have a look.
Out of interest, which model numbers are yours?

I have (briefly) looked into Wave Analysers. Over here in the UK the Marconi units are more common than the HP ones. I may pursue this further.

I was thinking software may well be the way forward, but now I'm not so sure.

Perhaps when you have the time, you could expand on some of your measurement techniques on your (already excellent) website.

Many thanks for your reply and best regards,
Rob

PS. Is your Ballentine meter one of these?
Image
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Have to go look at the #. IIRC, it is a 320A. 1mv to something like 300v out to better than 1MHz. True RMS for crest factors of up to 20:1. Old instrument, though: 19" rack mountable.

My approach is rather different for test equipment: most of the "guts" inside the instruments has almost nothing to do with the original circuitry from the factory. Also, since I end up with specific test frequencies for distortion, generating a low distortion waveform is easy. Filtering works well so long as the reactive components are reasonably linear and very stable.

Similar on the measuring end. By filtering out the fundamental low distortion and passively, the input dynamic range of "stock" equipment gets much better, typically ending up being "noise" limited, rather than distortion limited.

Others take different approaches. The traditional approach is modern equipment with high dynamic range.
hotvalve
Posts: 3
Joined: Fri Dec 21, 2007 9:16 pm

Post by hotvalve »

Sounds like a very useful instrument.
Out of interest, what was the model of Heathkit oscillator you modified?

Many thanks for sharing some details of your methods, much appreciated, please feel free to share more if you have the time!

Are there any details of your notch filter on your site?

Thanks again for all your help.
cda75
Posts: 10
Joined: Sat Jun 12, 2010 9:11 pm

Post by cda75 »

sbench wrote: 1. Sources are highly modified "Heathkit" oscillators with passive LC tuned circuits that net me somewhere around .001 to .0005% residual distortion.
Very interesting, Stephie.
The IG-18 has been modified by many.
Is this the model you have?
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

One of several. Two Heathkit models, an old Eico model that I've modified to cover up to 1MHz, and a little hand held unit I designed and built that uses an Analog Devices 9834 DDS chip as its base and controlled by a PIC. Also several "TIMS" (transmission impairment test set) used primarily for telecom applications, but useful for noise and IM testing on audio.

Stephie
cda75
Posts: 10
Joined: Sat Jun 12, 2010 9:11 pm

Post by cda75 »

The HH unit looks great, saw the other thread.
Lots of broken nails whilst constructing, I'll bet.

Image
Post Reply