cRIO 9022 Supply C terminal and chassis internally connected, noise problem

[askay]

Hi all,

I have a question, I have been using cRIO 9022 with 9205 AI module. I am using differential mode for Analog readings.(from torque sensor)  In my test bech, as soon as DC motor driver is enabled (driving a 10kW brushless AC motor), my torque reading are badly effected. I am getting heart beat type once every 1 or 2 seconds,  25Hz 0.01V p-p amp ripple for a duration of 10msec. Normally, noise level of the signal before driver is enabled is less then 5mV p-p, no additional periodic low freq. ripple at all. 

Now I am trying to find what causes this. One thing I figured out is that 9022 chasis ground is internally connected to the ground of the cRIO supply, i.e. 24 VDC GND. If DC motor driver is giving harmonics to the AC mains and AC ground, my DC ground, which I thought it is isolated from AC ground is now connected AC ground. cRIO is hooked up to cabinet and cabinet is grounded.

I am planning to acquire same data after isolating cRIO ground from AC ground by removing it from cabinet.

One more thing to mention, 8 channels of 9205 is compiled to use, but only one of them (torque feedback) is physically connected, others are either not connected. I am planning to tie unused channels to ground, since 9205 is muxed and not ch to ch isolated.

I dont have on site measurement with me for the moment but I have painted what I saw in the attachment.

9022  has the C terminals are internally connected to the controller chassis, how about new gen. cRIOs ? when I read the new gen user manuals, I havent seen any.

thank you

sean 

[ikelli]

Hi Sean,

This type of ghosting can occur when the empty channels are left unwired.  Have you tried switching the channel you are measuring from to determine if it is specifically that channel or a similar result?

[alandry]

Hi,

I know this topic is a few months old, but did you figure out a solution? We are having the exact same problem using an AC-AC motor driver, with ripples occuring in data every ~2 seconds once the motor starts turning (just powering on the drive doesn't do this, the motor has to turn). I'm using a 9066 cRIO and also just figured out that the grounding lug of the chassis is internally connected to the C terminal of the power supply. I'm using the same 24V  power supply to power pressure sensors, encoders, etc., which means all my sensors are grounded. Channels are set to either RSE or DIFF, with the negative terminal of the power supply connected to AI GND of my NI 9205 module. Does this create a ground loop? Should I instead use a separate, isolated power supply for my sensors and use AI SENSE to measure the negative therminal of this power supply?

Thanks

Alex

[askay]

Hi Alex,

We had exact same problem: solution to the problem was:

- measure  differentially,

- remember to hook up the com pin, coms of the cRIO c modudes were internally short, conneting one of them to the system gnd shoud suffice. You should play around this.

- once the motor drivers enabled, the motors in your setup will effect your measurements, although it looks like a low freq. ripple on your analog measurements,i.e. one every second,  it is indeed high frequency switching in your motor driver which causes this result. our attempt was to use analog anti-alising filter, I was closing the loop at 5khz, loop bandwith was 500Hz,  and my analog filter cut-off freq. was 2500hz which is the nyquist frequency. My ripple is now very low, that is less than 2mv p-p at +/-10V scale.

hope this helps

good luck

Sener

[alandry]

Thanks.

For sensors that only have one COM reference for supply and signal, wouldn't DIFF or RSE measurements be the same?

I have 3 AC-DC power supplies (24V, 12V and 5V), all the negatives are connected together and to the COM of my analog input (9205), analog output (9263)and digital input/output modules (9401). I have a thermocouple module (9213), but I haven't connected the COM and kept it isolated. What I realised is that since my 24V power supply for the cRIO is also used for sensors, and since the cRIO internally grounds the negative terminal to the grounding lug (which I connected to earth ground), basically all my channels are now referenced to earth ground. Is this OK? Should I isolate the cRIO power supply and use a separate one for my sensors? I feel like I'm measuring a lot of noise but I'm not sure what is "acceptable". Analog inputs (pressure) show about +- 10 mV of noise. Thermocouples show about +- 0.1 degC. Is this normal?

I will try filtering the data.

Thanks for the help,

Alex

[askay]

Hello, 

I believe, seperating crio power supply from sensor supplies won't help, if they all used for powering sensors and controller in the same setup, let them share same gnd (negative terminals), tie them together, including crio supply negative and com ports.   since crio supply negative terminal was internally short to it chasis, your cabinet gnd is now at earth gnd level, considering your cabinet is grounded (which has to be)

 If the crio controller and high power (drive) elements are is the same cabinet, then your drive, power supplies and controllers sharing the same gnd. If there is a way to seperate high power and low power away from each other, do it. If it cannot be avoided, make sure your grounding resistance is proper, any grounding failure might damage your controller. 

regards