A Measurement Pre-Amplifier to use with my Soundcard for Audio and Noise Measurements

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi All,

I have finally completed this build and decided it'll be nice to share with others on this forum, in case this will be of interest/use to anyone.
I have been looking into making some audio frequency measurements in a more reliable and simple way than I was using in the past. I also wanted something that isn't too expensive, and preferably DIY, this is for hobby use after all. The solution was to build a pre-amplifier with the required ranges and input protection to accompany my sound-card. This will accompany my modified EMU 0404 USB.

The result is the following:
DSC_6595c.jpg


I was looking for something that will fit my needs, which meant using both SE and BAL sources and loads, as I sometimes measure balanced amplifiers (such as the B22 I own).
Additionally, I wanted something that would go down to very low input ranges so that I could measure the noise + ripple of low noise regulators. Due to the high input impedance requirements (100K per input) this isn't as low noise as it could have been without such a requirement, but at 8nv/rt(Hz) on the low voltage (2mV/20mV) ranges, this is more than good enough for most applications.
I have originally looked for a pre-made solution to my needs, and came across Pete Millett's excellent sound-card interface which was also posted here at DIYaudio almost a decade ago. Unfortunately, it didn't meet all of my needs, so despite the extra time it took to design something instead of ordering pre-made boards and panels, I went down this path. I would like to thank Pete for sharing his project, its was without a doubt helpful for me as I'm sure it was for many others.

I have posted about it with all of the information I though relevant on my blog:
Audio Measurement Pre-Amplifier – Part 1 – Motivation and Circuit Design
Audio Measurement Pre-Amplifier – Part 2 – Board Layout and Assembly
Audio Measurement Pre-Amplifier – Part 3 – Initial Setup and Measurement Results
Audio Measurement Pre-Amplifier – Part 4 – Casing the Pre-Amplifier
I don't think it makes sense to repeat all of it here, especially considering the fact I have shared there not just to final results but also different experiments and modification I did along the way to find the best solution/fix mistakes I did earlier in board layout :)

I'm only able to measure it to the capabilities of my sound-card + few other accessories around it, as I have no high performance audio analyzer. Therefore it is probably better than I can measure with this setup, but even with these limitation it seems quite capable, and more than adequate for my needs.

DSC_6463-1.jpg


THD at 1KHz 2.0Vrms (at the 2V range) is <0.0001%.
Input noise density on 2mV/20mV ranges is 8nV/rt(Hz) measured (7nV/rt(Hz) calculated, so this is in good agreement).

This was by no means designed to compete with a high-cost audio analyzer, and it could have been optimized/designed differently according to other needs. I have designed it to meet my needs, and I'm happy to say it meets them completely. It actually performs as I've expected, so overall I think this was a successful (and obviously fun) project. More importantly, I can now use it for my builds/repairs which will be the best part about building it.
I'm far more limited by the EMU 0404 USB sound-card than I am with this pre-amplifier. However, since I can't justify spending a hefty sum of money on a high performance sound-card (or audio analyzer) for limited hobby use, I guess I will stick with the EMU for now. If you are familiar with something that is both cheap, and can out perform the 0404 USB by a noteworthy margin, please, do share :)
 
Beautifully done!


The input attenuator could be split into two parts, one a low impedance power attenuator for use with power amplifiers (high impedance not a requirement, just power handling),
and an active attenuator / buffer / amp as the first stage. It should be possible with cunning to keep ~100k series resistances from before the first active device.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.