certain frequency artifacts appear in data

[myra]

Thanks for the help - I'm really stumped by this noise. 

I've used a function generator a few times now to input a known frequency and have always seen at least the 4 and 8k signals on top of it.  Afterwards, I've generally tested the same function generator signal with my oscilloscope, where these two noise frequencies do not show up...  These measurements were always taken in the differential configuration though - so would you recommend that I instead make an RSE measurement on a known signal from the function generator?  I've used differential measurements to prevent charging from bias currents, but I don't know what access currents are.  Thanks for the link, I've read it many times before but am still trying to understand how to balance the entire measurement!  Guess I should read it a few more times. 

I'm not sure if I can perform a loop back test on my DAQ board because I can't really produce an analog signal from my AO channels (at least that's my understanding, hopefully I'm wrong).   I thought it could only produce a DC voltage.  If you think it would be useful to wire this up, I'll give it a try!

I was wondering what the source of the noise could be.  It seems like the inside of a computer would be a really noisy place for a card to be, right?  Maybe the DAQ board is picking up on some kind of clock signals?

Thanks so much, Myra

[Corby_B]

Hello Myra,

Thanks for the post back! I think you are indeed right that these frequencies are a bit mysterious. I would think that if these were coming from an on board clock of some kind that they would be higher in frequency. I think an RSE measurement would be a great test on your board. This way we could see if the problem is still there in a different configuration. Another option would be to add the bias resistors to your + and - terminals. Try any value from 10 K ohm to 100 K ohm although the RSE test from a function generator would be a better test. Also your board does have analog output capabilities but as you mentioned it does not have hardware timed AO. So you could do a software update output but the max frequency would be as fast as your OS can go witch is usually on the order of milliseconds, so maybe 1000 Hz. Let us know how the RSE measurements go from the function generator of known frequency. Let us know how the test goes!

Cheers!

Corby_B

http://www.ni.com/support

[ASakhno]

Hi,

I have the simalar problem. I use NI 9215 (with Ni-9163/Enet) to measure 50Hz current signal, I have try different sample rates from 1k to 100k and have the same trouble, some periodic signal 4kHz on my 50Hz signal, my scheme and result are on the attachments. I try shemes with 1MOhm resistor between "com" and "AI0-" and without it, resuls are equal. I measure this signal with another daq device - Dataq company and with oscilloscope and there is no any 4kHz noise on my signal. What wrong in my scheme, I couldn't use such signal in my application. Please answer, if such problem was solved and what I should do.

 Alexandr

[Aaron_W.]

Alexandr,

Thank you for posting on the NI Forums. This spur that you are seeing at 4 KHz could be from a couple of things. What sampling rate are you using to measure the 50 Hz signal? Is your system in a noisy environment where it could be getting a 4 KHz signal from some piece of equipment in the surrounding area? This should help us find out the problem that is occurring with your system. 

[ASakhno]

Hellow, Aaron.
Thank You for Your reply.I have wrote in my post that this 4k frequency signal is registrated by Ni9215 on all sample rates from 1k to 100k, usually I use 50k sample rate. I use NI9215 on the high voltage substation, but there is no 4kHz sources in surrounding area. I try to mesure known 50Hz signal by NI9215 and everything was OK, there were no 4kHz signal. So NI9215 is OK and there is no real 4kHz signal because other daq system and oscilloscope does not registrate it. I dont have such problem with such scheme and DATAQ, Dataforth systems. I think that the problem is in the scheme of NI9215 connection, maybe something wrong, I try different schemes that find on NI web but results were the same - 4kHz signal with amplitude near 250mV on the main 50Hz signal. You can see it on my previos post attachments.Please, reply as soon as possible because I have very little time. Wating for Your soon answer.

Alexandr

[Aaron_W.]

Alexandr,

I would recommend going into Measurement and Automation Explorer and using the test panels for that device. If you use channel DevX/_aignd_vs_aignd you should see an output that is almost exactly 0 Volts. If you are still seeing the 4 KHz signal here there could be a problem with the device or noise is getting into the system from some location. If you are seeing this 4 KHz signal in all ranges this would indicate that the noise is because of a problem with the card or noise in the system. 

Also if I understand correctly you have input a 'clean' 50 Hz signal into the 9215 and you saw a clean square wave with the 9215? At the same time you have also measured the actual signal with an oscilloscope and you don't see the noise with that device?  This has got me confused because if there was a problem with the device itself you would see the 4 KHz noise when you input the 'clean' 50 Hz signal. 

I know you are talking with Jordan F. on another post here and he actually sits next to me and we have been discussing the issue. We will work together to see what solutions we can come up with for you. 

[ASakhno]

Hellow, Aaron.

>"I would recommend going into Measurement and Automation Explorer and using the test panels for that device. If you use channel DevX/_aignd_vs_aignd you should see an output that is almost >exactly 0 Volts. If you are still seeing the 4 KHz signal here there could be a problem with the device or noise is getting into the system from some location. If you are seeing this 4 KHz signal in all >ranges this would indicate that the noise is because of a problem with the card or noise in the system."

if I understand correctly, I should measure without actual signal? I tried this, and there were no 4kHz signal, only "white noise" with level near 3uV

>"Also if I understand correctly you have input a 'clean' 50 Hz signal into the 9215 and you saw a clean square wave with the 9215? At the same time you have also measured the actual signal with an >oscilloscope and you don't see the noise with that device?"

Yes, You are right

It seems like, everything is OK with Ni9215, everything is OK with signal, but when I connect it together, there apears some mistake.
Thank You very match for Your reply. I'am wating for Your advices.

Alexandr.

[ChrisR.]

Perhaps it's not really 4kHz noise, but something that aliases to 4kHz.  The 9215 has a bandwidth beyond 400kHz.  (http://www.ni.com/pdf/manuals/373779e.pdf)  If there's 404kHz noise on the signal, for example, and you sample at 100kHz, or 50kHz, or 20kHz,... the 404kHz noise will appear to be at 4kHz.  Try sampling around 99kHz instead of 100kHz.   If the noise no longer appears to be at 4kHz, then it is aliased from a higher frequency.  What is the bandwidth of the Dataq device?

Chris 

[Aaron_W.]

Alexandr,

I would recommend trying what Chris suggested and let us know how that works. 

[ASakhno]

Hellow, Chris.
Thank You very match, for Your reply. I will try to mesure on 99 kHz sample rate tommorow, and let You know a result. Indeed I have tried only 100kHz, 50kHz, 30kHz, 10 kH, 1kHz sample rates. Maybe in this was my fault.

Thank You very match.

Alexandr.