4 to 20 ma, 2 wire EMC coupling noise

[wreckage]

I am using Daqcard 6036E with SCB-68 and shielded cables to read 4 to 20 ma signal from a 10,000 psi transducer. 2 wire connection with 15 VDC supply. 250 precision sense resistor gives me 1 to 5 v measurement. If I hold transducer in free air, I get a clean signal. If I touch the transducer case to the unit under test, I get about 30 mV noise at 4 or 5 Hz. Process connection is still at zero. Transmitter manufacturer says EMC circuits provide capacative coupling between transducer electronics and case. Nothing they can do about it. All cables are twisted pair shielded instrument cables with shields connected only @ SCB-68 end. My DAQ device is set for current measurement using RSE. I could filter the signal but I need t
o analyze frequencies in the same range as the noise.

Any fast response transducers on the market which will isolate the noise?

[Ian_Ren]

Hi, wreckage

I guess you do not have any anti-aliasing signal conditioning before SCB-68.

The 4 or 5kHz noise you saw may actually be the noise at a much higher frequency. I have some experiences from engine cell projects that an MIO card would measure high noise level, a DSA card (eg, NI 4472, which has anti-aliasing filter) returned very clean signal.

Unless you are sure that there is no high frequency noise above half of your sampling frequency (Nyquist frequency), you probably need to use some kind of anti-aliasing signal conditioning.

Also, have you tried differential instead of RSE?

Regards

Ian Ren

Roush Industries

[wreckage]

The noise frequency I am seeing is 4 or 5 Hz not 4 or 5 Khz and I was sampling @ 1000 Hz (every millisecond) so it's not an aliasing problem. I have not tried differential. I have a 4 to 20 ma signal coming to the SCB-68 and use a 250 resistor on the solder pads as per NI instructions. The only time I see the noise is when the transducer case is grounded (via pressure connection) to the unit under test or to structural steel in the building. I believe EMC circuit has a capacitor from current amplifier to case which becomes current path for ground loop.

Any thoughts much appreciated.

[RIO-Todd]

Greetings,

With the EMC capacitor giving us a current path for ground loops, using the differential configuration will be optimal here. This will prevent any coupling of noise from the shield to be coupled onto the ground of the DAQ device (since you won't be connecting to AIGND when using differential).

Therefore, differential should help us a lot here. Let me know if moving to differential doesn't help out.

Regards,

Todd D.
NI Applications Engineer

[wreckage]

To go differential, I connect 4 to 20 signal to pin 68 on the SCB-68 and move 250 ohm resistor from pad B to pad E. To complete the current loop, I have to connect terminal 34 to something. The only option I see is power supply ground which completes the ground loop. One way or another the current has to "exit" the system through power supply ground which is same as AIGND (I think). Remember the transducer is 2 wire, power and current signal. There is no other current path between transducer and SCB-68.

I'll try differential input but I'm unsure where to connect terminal 34 without completing the ground loop.

Thx all for your comments

[RIO-Todd]

Hey again,

So the current output of the transducer doesn't have two leads? Is there a ground from which the current is being supplied on the transducer? Usually, with a transducer (either voltage or current) you will have a positive and negative terminal. If the transducer is getting power from another source, then you could try that power supply's ground for the pin 34 connection.

Measuring Current Using the Analog Input of a DAQ
Board?
http://digital.ni.com/public.nsf/3efedde4322fef19862567740067f3cc/82508cd693197ea68625629700677b70?OpenDocument

Regards,

Todd D.
NI Applications Engineer

[wreckage]

Only 2 wires on the transducer, +15 VDC supply and signal. No ground. The noise has to be coming down the signal wire so a differential setup won't reject it. I am convinced that EMC circuits in the transducer couple noise into the signal. The laptop and transducer power supplies plug into same 110 wall plug. I have terminal #24 (AIGND) connected to transducer power supply common to complete the current loop. Cable shield terminates to chassis of SCB-68. Using a SHC68-68-EP shielded cable from SCB-68 to Daqcard 6036E so I assume shield terminates at laptop power supply common.

[RIO-Todd]

Have you tried a differential configuration with pin 34 (channel 8...CH-) connected to the power supply common?

If you think the noise is being coupled to the signal, it sounds like you're stuck if hardware filtering isn't an option. You could try to use a software filter (which would be more of an ideal filter) to filter out the noise. You might be able to use a narrow band, software filter and keep your signal level intact.

Regards,

Todd D.