Signal noise cDAQ-9172

[Margaret F.]

Hello Richard,

Please try adding bias resistors. Any resistor value within the 10k-100k range is fine; however, higher resistance allows for less error. By exciting the sensors, I meant are you having to provide any external voltage. Please let me know the outcome of using the bias resistors. Thanks!

[dthess]

Using a bias resistor is a good suggestion.  It takes care of floating grounds which are differences in normal ground voltage between the detector and the ADC.

If data bunching and the bias resistor aren't enough, one of these algorithms might help:

http://www.chem.uoa.gr/Applets/AppletSmooth/Appl_Smooth2.html

A google search on "Data Smoothing Algorithms" for sensors leads to many white papers

Message Edited by dthess on 01-20-2009 11:54 AM

[rmichels]

Thanks for the advice, both of you.

I still haven't had a chance to try using bias resistors, but we did have an interesting day today with this.  We started removing the signals 1 at a time to see if any one of the devices was causing the problem.  Every time we moved anything (whether removing or reattaching) the input signals, our values would move back to what we expected for about 30 seconds and then go back to the weird, unexpected signals (0.4V expected, moves down to 0.39 or 0.38V).  All but 1 of our devices were in the 9205 voltage input module at this point.  We found that when we removed either our CO or CO2 signal, all of our other signals would go back up to the expected values and stay there.  The weird thing was that even if we connected either CO or CO2 to the 9203 current module, ALL signals in the 9205 voltage module are affected.  So a signal entering the current module somehow affects the voltage module signals, but we found that the signals in the current module were unaffected.

So we moved all of the signals 1 by 1 into the current module.  This module only has single ended connections, which is why we didn't like to use it before.  When we turn on the fans in the facility we get quite a bit of noise.  But with the signal smoothing I've thrown into labview I'm hoping this is acceptable.

I'll try bias resistors on Thurs.  But if that doesn't work, it looks like we just can't use our 9205 module.  Maybe it's defective?   Any thoughts?

Again,  Thanks for your time spent on this.  If I knew where you were, it would be coffee all around.

I'll keep you posted.

Richard

[Margaret F.]

Hi Richard,

I am still convinced that this is a grounding issue. Please give the bias resistors a try and let me know the outcome. You may have been seeing clean signals when you attached and unattached because you touching the system was acting as a ground. Go ahead with the bias resistors and let's see what happens. Have a great day!

[nuumamonjaa]

hello

I have the same problem with my cDAQ9172. I am using a NI9221 to acquire data and when i want to observe a 1V continuous signal, i get a noise of about 0.3V. (RSE configuation)

I don't know why but when i plug the 9221 to a NI USB-9162 (hi speed usb carrier) with the same acquisition configuration, the noise is reduced to 0.1V

if you managed to solve your noise issue could you held me with mine.

thanks

[nuumamonjaa]

problem solved lol

the designers of the cDAQ-9172 forgot (?) to ground the cDAQ with the other devices plugged into it. I only had to wire a metalic pat of the cDAQ to the ground of the NI9221 and thats it, no more noise ... (dear NI, it's bad to make trainees develop your devices ^^)

[rmichels]

I actually never solved the problem. I talked to a support guy for probably 10 hours or so over a few weeks and finally I found if I just didn't use the 9205 voltage  module and used the 9203 current module instead, the crazy voltage drop didn't happen and if instead of taking 1 sample every second, I took a couple hundred samples and averaged them, I could reduce the noise to an acceptable level.  I tried grounding the module.  We thought that worked for about 10 minutes.  It looked good, but then the voltages all just dropped again and I threw in the towel.  For a master's thesis I couldn't afford to lose more time.  This wasted a couple months.  Plus when the service guy couldn't figure it out, I figured there was nothing I could do having absolutely no background in electronics. 

Glad to hear your system works alright.

Richard

[KyleBryson]

Hi nuumamonjaa,

The NI 9221 is electrically isolated; the reasons for isolation are discussed here and in several other articles. The manuals for the cDAQ-9172 and NI 9221 provide guidance on properly grounding your data acquisition system.

Regards,
Kyle

[stilly32]

A couple things: The 9221 is an isolated module - the AI front end is isolated from the chassis ground by design so (for example) that you can connect  a source with a high common mode voltage and still measure it without frying your front end. By connecting COM to the chassis ground, you are essentially providing a ground point for the module and removing isolation from the chassis. If your signal source has a ground point, I would recommend using that as COM.

As for why you would see more noise in the chassis vs the carrrier - I'm not really sure. One thing that is recommended is to make sure the 9172 ground is connected to your measurement system ground. I have seen this reduce line noise in measurements by quite a bit.  

Hope this clears things up, 

Andrew S

Message Edited by stilly32 on 04-10-2009 09:26 AM

[D_Zuo]

Hello Andrew,

Is it possible for you to clarify the issue about isolation? So if I connet the ground for 9172 to the COM of my input module (9205), my input module is no longer isolated from 9172. I understand this. What do you mean by saying that the AI might be fried if we have a high voltage source? And what do you mean by saying the 9172 should be grounded together with the measurement system? My sensor needs excitation and I am supplying it with NI 9264. Should I connect the grounding point of 9172 and the COM of 9264? I am seeing spikes in my acceleration measurement signal (signal ended system). I have been wondering if it was due to incorrect grounding.

Thanks.

Delong