diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Equipment & Tools (http://www.diyaudio.com/forums/equipment-tools/)
-   -   Instrumentation amplifiers (http://www.diyaudio.com/forums/equipment-tools/220741-instrumentation-amplifiers.html)

dirkwright 30th September 2012 08:20 PM

Instrumentation amplifiers
 
I was wondering if anyone had thought about using a high CMRR instrumentation amplifier like the AD8429:
AD8429 datasheet and product info | 1 nV/?Hz Low Noise Instrumentation Amplifier | Instrumentation Amplifiers | Analog Devices

as part of a THD test facility. What I mean is to feed an oscillator input to one input of the in amp, feed the output of a DUT to the other input of the in amp, and thus the output would be the distortion which then could be fed to a FFT analyser. Seems like a simple idea to me.

jan.didden 30th September 2012 08:31 PM

Yes it's simple, if you can adjust the two levels to be exactly the same (for a -100dB distortion resolution, for instance, you need to match levels to 1:100,000 which is about 0.0001dB) and also compensate for the additional phase delay for the signal going through the DUT ;)

Conrad Hoffman 30th September 2012 09:01 PM

As above, plus unless you build the rest of the THD analyzer with auto tuning and phase nulling, you'll probably degrade the FFT output as opposed to improving it!

I do like differential tests where you compare the input and output of the DUT, say a power amp, playing music into a load, and the instrumentation amp is perfect for that.

dirkwright 30th September 2012 10:43 PM

Quote:

Originally Posted by janneman (Post 3185280)
Yes it's simple, if you can adjust the two levels to be exactly the same (for a -100dB distortion resolution, for instance, you need to match levels to 1:100,000 which is about 0.0001dB) and also compensate for the additional phase delay for the signal going through the DUT ;)

Thanks. I had actually realized that I have to match levels, but didn't mention it. I was thinking of providing some kind of tuning indicator that shows when the signals are exactly equal. I had not done the math though, so a 0.0001dB match would probably be really hard to make unless the error was amplified by 100dB or so.

The issues of phase shift I had not considered and would be an obvious problem.

I like this idea though because it would cancel out the distortion components from the osciallator.

dirkwright 30th September 2012 10:46 PM

Quote:

Originally Posted by Conrad Hoffman (Post 3185333)
As above, plus unless you build the rest of the THD analyzer with auto tuning and phase nulling, you'll probably degrade the FFT output as opposed to improving it!

I do like differential tests where you compare the input and output of the DUT, say a power amp, playing music into a load, and the instrumentation amp is perfect for that.

Yeah, it's all still a thought bubble in my head right now. I don't know about correcting for phase shifts, so I have some reading to do. Thanks for your reply.

dirkwright 30th September 2012 11:32 PM

I guess you'd need an oscillator with a really stable output as well, but that's probably true for all measurements.

coluke 1st October 2012 08:49 AM

Quote:

Originally Posted by dirkwright (Post 3185262)
I was wondering if anyone had thought about using a high CMRR instrumentation amplifier like the AD8429:
AD8429 datasheet and product info | 1 nV/?Hz Low Noise Instrumentation Amplifier | Instrumentation Amplifiers | Analog Devices

as part of a THD test facility.

Yess :) - Analog Devices AN-245 : Instrumentation amplifiers solve unusual design problems, by Scott Wurcer and Walt Jung - page 10-15. It works very well, and if you're only interested in extending the dynamic range of a spectrum analyzer, then you can relax quite a bit limits on null depth - Bob Cordell's THD magnifier is based on the very same principle.

L.

dirkwright 1st October 2012 12:24 PM

Quote:

Originally Posted by coluke (Post 3185942)
Yess :) - Analog Devices AN-245 : Instrumentation amplifiers solve unusual design problems, by Scott Wurcer and Walt Jung - page 10-15. It works very well, and if you're only interested in extending the dynamic range of a spectrum analyzer, then you can relax quite a bit limits on null depth - Bob Cordell's THD magnifier is based on the very same principle.

L.

Thanks! It looks like I've reinvented the wheel, which is a good thing actually. I have some reading to do!

I've never looked at Bob Cordell's "THD magnifier", so thanks for the tip.

Samuel Groner 1st October 2012 02:17 PM

Quote:

0.0001 dB.
As you're feeding a FFT analyzer that good level matching is not necessary. 60 dB suppression should be enough. But keep in mind that the instrumentation amplifier has its own distortion contribution in such an application as it sees large common-mode swing.

Samuel

dirkwright 2nd October 2012 05:50 PM

I ordered Linear Audio number 0 today. I need the schematic for Bob's DM. I want to see if I can add some features to it.


All times are GMT. The time now is 08:55 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2