I get more accuarate results measuring DC(Volts) by scanning muliple channels than I do with measuring each channel seperately; unless I sample for a full second.. Why?

[Datadude]

Using a PCI MIO 16XE 50 I'm scanning 16 channels. I decided to measure each channel idependantly as a comparison. I found the scanning method, via Acquire Waveforms, more accurate. The DAQ user manual suggests this would not be the case.

[markwysong]

What is meant by "more accurate"? And what do you mean by sampling for a full second? Are you averaging the signals over a second?

Are you measuring the signals going directly into your MIO card?

I've had similar instances of this type, so I'm trying to find out exactly what you are seeing.

Mark

[Datadude]

I collect 1000 samples at a sample rate of 1000s/s. I then take the mean. I'm measuring a dc output connected to my breakout box. My dc output is of a lowpass filter with a 32k resistor and a 47nF cap. I am measuring every 100 mS as my oven temperature ramps up. I subtract each measurement by its previous and don't expect to see greater than 500 to 600uV of change. I see 1 to 5mV of change when I measure each channel independantly. I must admit that it doesn't happen all the time. I couldn't get it to happen yesterday when I was experimenting with different techniques of acquiring single channel data. If you want to discuss further let me know. Regards.

[Datadude]

The 1000 samples at 1000s/s was the condition for the single channel collection experiment I tried. What I described with respect to the collection of data in the oven was 16 channels,1000 samples, at a scan rate of 1250s/s.

[markwysong]

Thanks for the info. In your 16 channel acquisition, are you getting 1000 samples at 1250 samples per second...is that continuous? Do you average those?

Mark

[Datadude]

I'm using the Acquire Waveforms.vi so it is continuous.
I store all the data in an array and take the mean of each indivual channel.

[Filipe A.]

Hello;

We need more information on how you are measuring the channels separately, and what exactly do you mean by sampling for a full second?
Filipe

[Datadude]

I did some experimenting and here's what I found. I have 16 devices inside an oven with wires running out and connecting to my breakout box. With the oven On or Off I get the desired results with a multi channel (Acquire Waveforms.vi) approach. With the oven off I get desired results with the one channel at a time (Acquire Waveform) approach. With the oven On I get the 1 to 5mV of noise. The amount of samples for both methods was 1000. The Scan Rate for multi channel = 1250s/s. The sample rate for the single channel was 20ks/s. The oven bieng on results in noise on my devices. For some reason the multi channel approach reduces it.