Audio Measurement ADC input stage

I'm considering building an ADC or using a stock sound card as an input in order to take advantage of computer-based measurement tools. How have others solved the input protection problem?

(While "always use an attenuator" is an option, I would prefer to have both a "protected" input as well as a bare "hi quality" one)

What about the circuit below, shamelessly adapted from the Bitscope plans? The bandwidth of interest would be 10-100kHz, and 103dB S/N with very low distortion would be considered minimal. I am not sure that the 82pF cap serves any purpose at these frequqncies. I've seen designs with zeners but I would think these to be far too noisy. Are there other alternatives? I am afraid to go active in what will be my measurement set.


  • untitled-1.gif
    2 KB · Views: 289
I have not been able to find any proof of LED versus normal diodes (Not Zeners), but would like to hear if any papers exists on this matter.

I think using LED as OVP is a bad idear. They are not as fast as ~1n4148 or schottky (only .2V drop) or fast recovery diodes.

I would go for the 1n4148 because of the lower capacitance across it.
For the cap 82pF across 10k: Depending on the size of the inputcapacitance of your ADC i would think it only has a little impact above 50 - 100KHz. Try run without first.