1
\$\begingroup\$

A single ended earth grounded data-acq. board on the left in the below illustration has inputs coupled by BNC cables coming from force transducer amplifiers A1, A2, A3, A4 as shown on the right:

enter image description here

If no BNC is hooked up, when I measure the voltage between any BNC connector terminal to earth I read around 100VAC by a multimeter. But by using low impedance setting of the multimeter I verified that this is stray voltage aka ghost voltage.

Since I measure this stray voltage between BNC inner pin to earth and BNC outer pin to earth, I guess this stray voltage is in phase both terminals of the BNC.

There is no direct ground loops in the system i.e the amplifiers'/sources' grounds are not earthed. During data acquisition sometimes I see noise coming and going for some seconds sometimes, and I was suspicious of capacitive coupling.

My questions are:

1-) The data-acquisition max input voltage is +-10VDC. Would this stray voltage be problem in this case? Is it exceeding maximum common mode range? Because the moment I hook up the reading are okay. But Im not sure is this stray voltage still have side effects.

2-) Is this configuration is more susceptible to error in readings than in differential ended inputs? How would diff ended. connection eliminate the noise caused by stray voltages?

\$\endgroup\$
11
  • \$\begingroup\$ If your multimeter reads 80-100 VAC, then you have a real problem. The multimeter needs current flow of around 18 uA (assuming 10M Ohm input resistance) so it's not just a 'ghost voltage' I'd be looking very carefully at your PSU. \$\endgroup\$
    – Jack Creasey
    Commented Mar 6, 2018 at 16:47
  • 1
    \$\begingroup\$ All power supplies I have this stray voltages between their terminal and the earth around 90VAC. It is coming from the power supplies. \$\endgroup\$
    – floppy380
    Commented Mar 6, 2018 at 16:51
  • \$\begingroup\$ If you need low noise, you need the correct type of power supply (and probably a linear output stage), correctly earthed and with low capacitive coupling between input and output. \$\endgroup\$
    – Jack Creasey
    Commented Mar 6, 2018 at 18:05
  • \$\begingroup\$ Do you know any power supply type which has almost zero coupling like battery? I bought medical approved ones still they have this issue. I tried linear I tried SMPS. Cannot get rid of this demon. \$\endgroup\$
    – floppy380
    Commented Mar 6, 2018 at 18:08
  • \$\begingroup\$ Look for CF rated power supplies such as TDK-Lambda produce, these have superb isolation. You have to be careful in selecting supplies because some have mixed capability such as this BEAR unit. 24 V CF rated and 12 V BF rated. bearpwr.com/pr_medical-dualCF-BF.shtml \$\endgroup\$
    – Jack Creasey
    Commented Mar 6, 2018 at 18:31

1 Answer 1

Reset to default
1
\$\begingroup\$

This question and answer explains how the output of many two-pronged AC (earthless) SMPSs use a capacitor internally to reduce excessive output switching noise because of internal transformer capacitive coupling.: -

enter image description here

The side effect is that both DC output terminals (open circuit) are "wobbled" around at the AC power frequency. It is harmless to humans but might cause a problem to some inputs of some DAQ boards.

The data-acquisition max input voltage is +-10VDC. Would this stray voltage be problem in this case? Is it exceeding maximum common mode range?

You'd need to look at the data sheet of the DAQ card to see if it might be a problem.

Is this configuration is more susceptible to error in readings than in differential ended inputs?

Yes it is, The use of an unbalanced input means that there will be more current flowing to earth from the interfering common-mode signal than what is flowing down the signal line and this might create a differential interfering signal of low-ish size.

How would diff ended. connection eliminate the noise caused by stray voltages?

The input impedance to ground presented by both differential inputs is equal so, what was mentioned previously about imbalanced currents isn't applicable HOWEVER, if those impedances are high and no external capacitances are applied to suppress the common-mode voltage you could easily swamp the common-mode input range of the DAQ's differential input.

\$\endgroup\$
4
  • \$\begingroup\$ @ Andy What is the current thru the 2,200 picoFarad cap? 100 uA? and its lop-sided. \$\endgroup\$
    – analogsystemsrf
    Commented Mar 6, 2018 at 17:35
  • \$\begingroup\$ @analogsystemsrf the AC power current depends on what the DC output terminals are connected to. \$\endgroup\$
    – Andy aka
    Commented Mar 6, 2018 at 17:42
  • \$\begingroup\$ @Andyaka Since the noise appears in all channels Im almost sure it is this kind of common mode capacitive interference. But what could be the reason it happens sometimes for some seconds(in all channels) not even periodic. Noise increases for couple of seconds and goes away. What parameters could affect the amount of capacitive coupling changing randomly? Can it be a change due to the earth currents or SMPS filter or switching? I really wonder why this happens randomly. Thanks. \$\endgroup\$
    – floppy380
    Commented Mar 6, 2018 at 18:12
  • \$\begingroup\$ @user1234 I have no idea without a system scheme. Loading effects on the secondaries, the nature of the load and its impedance to earth, the switching turning off then on cyclically due to low loads being present. There could be many reasons. Sorry about the late reply. \$\endgroup\$
    – Andy aka
    Commented Mar 7, 2018 at 10:30

Not the answer you're looking for? Browse other questions tagged or ask your own question.