Analog Devices ACS software with AD7606-18

[cncjerry]

Hello, I am interested in trading notes with anyone using the subject software, specifically if used with the AD7606-18 (or -16) DAS chip and support board.

I've had great results using the 8 channel board, collecting samples, over-sampling, etc.  It has exceptionally low noise, wide bandwidth up to 1MSpS but then there are applications that don't make sense.

For instance, I have an SSM2019 single-ended amplifier being used to measure power supply noise with my HP3562 displaying the FFT.  The amp has 60db of gain.  When displayed on my HP3562, the noise level is across ~-80dB level (including 60dB of gain, actually  ~-140).  You would expect the AD7606 with at least that much dynamic range to see -80 and on other outputs it does.   For instance if I run an AC 1Khz signal into the amp, I get the expected output. But for some reason the very low frequencies near DB (<1hz) are displayed on the ACE software FFT very high (0db) and masking the lower numbers with a display of -300dB when I try to profile a power supply DC output.  I tried additional filtering to take that lower stuff out with the same results.

I realize this is confusing without pictures, plots etc. and I'll try to post some up if I get my GPIB running with the 3562.

Thanks

Jerry

[moffy]

Could it be related to 1/f noise which is well known?

[cncjerry]

What I found was that the ssm2019 for some reason related to input bias I assume, was causing the output to flip to either the negative rail or positive depending on the input unused.  This caused the input caps to charge on the DAS board overloading the input.  So not having a ton of experience with the ssm2019 in single ended mode. I switched now to the circuit from Jim William's AN83 as it has equal or better NF compared to the ssm2019, works, and I'm more familiar with the layout. 

What I am doing is measuring power supply noise.  I am building a standalone unit with battery supply using the dewalt 20v batteries.  I'll put a few diodes inline to drop the voltage to something a little lower than 18v, and go from there.  A pair of 20v batteries each having 4AH will last for many runs.

By the way, for the money, this 8 channel, 18 bit, 1MspS DAS with over sampling has very nice dynamic range.  The ACE software is lame so I use it to capture data and then process it with Octave or Matlab.  I plan to setup so the channels can be used in a simple average instead of over sampling.  The board is cheap at $64 but the support board isn't. 

Jerry