[Cauhtemoc]

Quote:

Originally Posted by Crowbar Unfortunately those THD analyzers don't show the individual harmonics, which is critical since different harmonics have a very different perceptual weighting.

To do what jcx said above, you would use the filters and of a THD analyzer to attenuate the fundamental frequency, then you amplify the residual with a low noise, low distortion opamp, and feed it into a sound card. That way, even a relatively cheap sound card like the Juli@ can be used to detect very low THD.

What normally makes a low distortion THD analyzer hard to build is that it need to be very precise, as you're looking to attenuate the fundamental by some 120 dB or more, while leaving the second harmonic largely intact. This calls for a very, very narrow filter, which is hard to make correctly.

You should not have this problem. A sound card like the Juli@ can already measure down to around -100 dB THD, so you only need to attenuate the fundamental by some 40 dB (and then amplify the residual with the same amount), in order to measure a THD of -140 dB. Compared to a 120 dB deep notch filter, a 40 dB one is a piece of cake to make.

 

[jcx]

another feature of the indirect IMD method is that you also bypass the need for ultra low distortion source

the DAC ouputs of the soundcard can each have one of the high frequency sines, then add them together with 2 good quality resistors

since the I/V stage should have a low impedance input you don't even need to buffer the Resistor summed sines from the 2 soundcard DAC channels

even poor DAC output is fine since you'll be looking for the IMD frequency difference in the I/V output, harmonic distortion of the individual DAC channels doesn't affect that difference product which only occurs in the I/V after the currents from the resistors are added together


yet another trick is if the I/V is a classic V feedback op amp style stage then you can just (amplify&) measure the V at the I/V input

if your feedback components are trustworthy - metal film or bulk foil R, polystyrene C - then the V at the input only contains the classic negative feedback "error" = distortion + input signal/loop gain

the input signal/loop gain term should be very small so it isn't generally required to filter it out

 

[Crowbar]

The I/V is current feedback and embeds a filter. There's no useful voltage at the input since the input impedance is fraction of milliohm and not really constant over the audio band.

 

[hachetman]

hey, first of all, this isnt my mod, but I can forward your questionsl, though NOT COMMERCIAL thing for sure.

so here is a modded e-mu 1212m with seperate supply, and components.
http://img147.imagevenue.com/img.php..._122_985lo.jpg
http://img147.imagevenue.com/img.php...122_1049lo.jpg
http://img223.imagevenue.com/img.php..._122_207lo.jpg
http://img147.imagevenue.com/img.php..._122_814lo.jpg