Distortion analyser circuit

[steevo]

Has anyone got a good distortion analyser circuit/PCB layout to build

Thanks
Steevo

 

[tschrama]

build mine on pref board... similar to Rod Elliots design.. work fine down to -100dB.. below that will require some significant extra effort..

 

[Conrad Hoffman]

It's a pretty major project, but the analyzer on Bob Cordell's site is hard to beat.

 

[MRupp]

Cyril Bateman has a series of articles on capacitor distortion that include a distortion analyser circuit, fixed frequency but extremely low distortion threshold: http://uk.geocities.com/cyrilb2@btinternet.com/downloads_4.html

 

[unclejed613]

there are two types of distortion analyzer you can build. first is a notch filter type. it's simple, but fixed frequency. it also requires a super low distortion oscillator. the second is a subtraction type where you use an instrumentation amp to subtract the amp output from the oscillator signal. it's a bit more complex, but the distortion requirements for the oscillator aren't as stringent, and it works at any frequency.

there's actually a third option.... use the sound card on your PC as the oscillator, and as a spectrum analyzer. you need to build a very good attenuator for your sound card, but it's a lot simpler than a complete distortion analyzer. a good freeware program for this is Spectrum Lab by DL4YHF. it can be used with full duplex sound cards with up to 196k sampling rates. then if you want to translate db measurements into percentage, i'm sure you can find a chart of this on the web.

 

[tschrama]

http://sound.westhost.com/project52.htm

 

[gaetan8888]

Hello

Renardson made a good web page about a subtraction type distortion analyser;

http://www.angelfire.com/ab3/mjramp/simple2.html


Here is a more detailed version article about it;

http://www.angelfire.com/ab3/mjramp/distortiontest.pdf


Gaetan

 

[jackinnj]

the 1kHz or 2kHz distortion is going to tell you a lot about the DUT -- so if you are in hurry build the Cordell oscillator (the basis for the Tektronix SG505) for 2kHz, run it off batteries -- it will yield THD% in the arena of the Audio Precision family. It's a lot less fussy than Bateman's, and you won't break the bank purchasing the opamps.

for $49 you can purchase the Texas Instruments ADS1271EVM -- it's a USB based -- you'll probably need an MMB0 mother board as well. The two are sold as a package for $149 from TI.

I don't know if you ever see them in the UK, but the later Heathkit THD analyzers were pretty good. I used a Heath IM analyzer and it worked as well as my Tektronix AA501.

Here's a GIF of the ADS1271's output.

http://www.tech-diy.com/TestEquipment/THD Analyzers/ADS1271/ADS1271.gif

 

[unclejed613]

i have one im thinking about building one that will use CFB op amps to improve the accuracy of high frequency measurements. regular VFB op amps begin introducing errors above their LF corner. (sorry for my lack of proper punctuation, but for some reason, every time i try to enter an apostrophe, i get a find window at the bottom of the screen)

 

[jackinnj]

i have one im thinking about building one that will use CFB op amps to improve the accuracy of high frequency measurements. regular VFB op amps begin introducing errors above their LF corner. (sorry for my lack of proper punctuation, but for some reason, every time i try to enter an apostrophe, i get a find window at the bottom of the screen)

Current feedback opamps are not very forgiving when it comes to layout. I found after several iterations of the Linear Tech Super Oscillator that it worked best when the two gain blocks were symmetrical in layout, and copper was poured all around. Not a great project for someone just starting out.

But what an oscillator:

An externally hosted image should be here but it was not working when we last tested it.

 

[jackinnj]

In case you're wondering, here are the distortion residuals for an AP2022 -- the display is averaged 32X and the sweep time is 50 seconds:

An externally hosted image should be here but it was not working when we last tested it.

The fundamental is 10kHz -- if you take the sqrt of the sum of the squares it is within a gnat's eyelash of the THD level shown on the AP.

 

[unclejed613]

are you operating that analyzer in linear mode? shouldn't your readout be in db/div and not uV/div? just wondering....
i usually use an analyzer in log mode.

 

[jackinnj]

Yes, it is in linear mode -- there's too much information in log mode.

I love this HP3577 -- and have been programming via the HP-IB of late.