ni cdaq 9184 earth ground

[s²]

Hello,

i wonder what is, exactly, the utility of connecting the cDaq chassis to the earth ground, does it improve the quality of the signal measured?

I ask this question because i have difference between cDAQ measurement and ni usb 6218 BNC measurement for the same signal (ground referenced in both case).

The measure is quite good with the ni usb system and and "very bad" with the cDAQ.

Thanks

[jlrapter]

Hi:

I have similar problem with the NI-9184. I have see with an osciloscope that when I connect the power supply of this cDAQ, strong noise (50 Hz) is present in all the system and of course the measurements are very bad. There is some expert from NI that could give me some help. Thanks.

[s²]

Hello,

Thanks, i will check that with them.

[Now_With_Underscores]

The 50Hz noise you're seeing is a ground loop, which wikipedia explains much better than I, but in short, there's a current flowing between two grounds at a different potential along the ground shield of your cabling... thus, you see the 50Hz (or 60Hz in the states, etc).  Tying all grounds to the same point (star) will eliminate ground loop noise, unless of course one of your systems is sinking current to ground (bad).

So what's the ground lug for in the first place?  If for some reason mains voltage were to get into your NI product, the exposed metal surfaces could present a shock hazard.  Grounding the lug means that instead of leaving your metal case 'hot', the ground will pull the mains voltage down to where you're standing (the earth), and hopefully some upstream circuit will pop.  Another reason is to make sure your NI product's ground potential is the same as whatever you're measuring, which is particularly important with single ended measurements.

[s²]

Thanks for yout reply, i will check this!

[jlrapter]

Dear to all friends:

First, I know what is ground loop and the causes. The point is that I discovered that the power supply, which in my case NI supplier sold me, was working very bad. In fact, when I mesured the output of this adapter with the oscilloscope I saw only the DC component, but when I connected a resistive load I started to see the 50Hz noise. In this way, I connected another adapter as power supply for the CompacDAQ 9184 and the noise in the data aquisition disappeared. Obviously, I am not saying that NI sells bad quality adapter with this cDAQ model, but if your problem is similar of mine, so there is a problem here!!!!

Bye.

PD.- Let me know what is about your problem. Thanks.

[Now_With_Underscores]

Hey Jlrapter,

To make sure i understand you correctly, you connected your scope directly to an NI power brick and are seeing 50Hz noise when you load the supply with (presumably) a ceramic (1/5/10Ω?) resistor?

The power supply that ships with the 9184 (NI PN 190415B-01) is an off-line converter, which rectifies the incoming line voltage to high voltage DC, then uses a 63kHz isolated chopper to vary the duty cycle, and thus produce the desired 12V output.  For this reason - and several others - I'm cautious to conclude from your testing that our 50Hz AC line frequency makes it through our converter to the 12V output, hence my suspicion that the noise is coming from elsewhere, probably due to a ground loop. 

Of course, we need to check our assumptions as to whether or not 50Hz noise on the 12V line is actually an issue.  The first thing our 12V rail touches inside the 9184 is a common mode choke, and then we dump into a high frequency buck SMPS, which modulates the input voltage to the working 5V rail used inside the chassis.  Yes, there is noise on the 5V rail due to the chopper, but it's very high frequency and cares very little about a 50Hz input.  Of course, if your actual noise is due to a ground loop, the wiring issue is likely past all the supplies, sitting right at the wiring to wherever you're connecting to your DUT, hence my recommendation to check for ground loops.

Ground loops are difficult, particularly when taking single ended measurements.  The ground loop could be as simple as the earth grounded negative scope lead on a single ended probe.  Whenever I find a ground loop, the first thing I do is plug everything electrically connected to my DUT in any way directly into the same power strip.  This prevents current sunk to ground (instead of neutral) from affecting the system, unless something in your system is actually connected incorrectly and sinking to ground. If the noise persists, the next thing I do is unplug all my DUTs and measure just the measurement equipment.  Note that a differential scope probe can be useful if you're unsure whether or not the noise is from the scope's ground lead.

Of course, I have very little insight into your system, so it's possible there is no groung loop, but that would be my strongest suspicion until I proved otherwise.

Note, if you'll post up your schematic and which NI C series modules you're using, I bet somebody will comment further.

[jlrapter]

Hi Jhon:

Now I am very confused. Today I repeated the measurement with the oscilloscope, power supply in question and Power Resistance of 22Ohm. The noise was not present, but this time oscilloscope and power supply were connected at the same power strip and I don't remember how they were connected the last time (around  one month ago).

In any case, the schematics of connections are attached follow. I ilustrated in not thecnical schematics the connections before and after that I found "the practical solution" changing the power supply of c-DAQ 9184. As you can see, by the application I can not connect all equipments at the same power strip but looking by a solution, I noticed that effectively the power supply NI P/N: 190415B-01 has isolated de CC output voltage from earth AC lead. Even I also noted that the lead of negative input bias voltage has connectivity with c-DAQ chasis. I think that in any moment there was a ground loop. But I could do a misstake in this, if you look a ground loop, Could you show me where it is? Thanks.

                

Also, I am using NI-9205 aquisition module and I getting signals in NRSE. But I did several tests with NRSE, RSE and Differential configurations and in all cases I was getting very hard 50 Hz noise. Also, another thing is that the DUTs block very high impedances in the output signal channels but I dont think that this could be the problem.

I was forgetting, as I resolved the 50 Hz noise I dismantling the system until future use but as general practice for testing I did several measurements over only the system without DUTs and the results in Data aquired showed that the noise was also present but only the positive part of sinusoidal 50 Hz due to the buffer circuit is only bias with positive voltage. From several measurements I could detect also that noise was higher in Channel 1 and lower in Channel 8 (Last channel) in my system input references obviously.

Thanks in advance by your observations.

Ing Jorge H. A.

Student UNICAL