2

When I plug a data-acquisition PCI card to a PC, I'm trying to picture the ground connections. Some following confusion arises:

  1. Is the PC motherboard's ground and PCI's ground common?
  2. Does a PC have analog and digital grounds? Ore one common ground?
  3. If so, is this PC common connected to the chassis earth as well?
Magic Smoke
  • 772
  • 3
  • 13
user16307
  • 12,049
  • 52
  • 184
  • 320
  • Is there any reason why you can't borrow a meter and measure the resistance of the connection between these points? – Andy aka Oct 26 '15 at 13:14
  • The reason Im asking recently Im using a daqboard which measures the offset more than the absolute accuray in the specs. . I find out for zero current the reading(through a shunt resistor) is around -14mV. So the absolute error shouldnt exceed 3mV but it is infact -15mV after the auto calibration. So I contacted to the manufacturer and here I got the following reply: "Maybe you are making a single-ended measurement? In that case, maybe there is an offset between the PCI-DAS6036's ground (that comes from the PC power supply ) and your power supply." – user16307 Oct 26 '15 at 13:20
  • And were you making a single ended measurement and are you able to make a differential measurement? – Andy aka Oct 26 '15 at 13:22
  • im making 16 channel single ended measurements. can it be that one of those channel inputs' ground causing a problem? But then all channels should have the common ground inside the PCI in single ended. Im confused – user16307 Oct 26 '15 at 13:26
  • In my experience, when trying to measure small levels accurately using computers and "NI type" DAQ stuff, differential measurements are the way to go. – Andy aka Oct 26 '15 at 13:34
  • i see, but did you understand what he means? can it be the reason? he wrote " PCI-DAS6036's ground (that comes from the PC power supply )". Is that true? – user16307 Oct 26 '15 at 13:35
  • 1
    I totally understand it - if you are remotely measuring a voltage and that voltage is "supposedly" at the same ground potential as local ground then it's quite common to see several milli volts of difference between the two grounds. Very common indeed. – Andy aka Oct 26 '15 at 13:39
  • So the PC power supply ground(not the earth, not the neutral) is connected to the motherboard ground and therefore to the PCI's ground and therefore to the single ended input's ground. Is that right what I understand? – user16307 Oct 26 '15 at 13:44
  • I can't tell you what the internal earthing regime is. I can only point out that a few milli volts can easily be the difference between points that are normally regarded as the same voltage. – Andy aka Oct 26 '15 at 13:47
  • even they re common grounded, there might be some difference? – user16307 Oct 26 '15 at 13:48
  • so it seems the absolute specs only valid for diff. ended measurements, which is not mentioned in their manual. – user16307 Oct 26 '15 at 13:52
  • Try making a single ended measurement of the "other" ground point to see what it measures. This should then convince you are maybe point to another problem. It could be that there is an AC voltage between the two points and maybe averaging a few readings this gets better. – Andy aka Oct 26 '15 at 14:01
  • the "other" ground point ? what do u mean? – user16307 Oct 26 '15 at 14:10
  • Well, you haven't shown a picture of your system but you have talked about different ground points so the thing you are measuring that has the error, if you measured on the grounded side of that thing you would get an idea what the errors are without measuring the "wanted" part of the signal. It just gives you an idea what the problem is. Then try earthing that input back at the DAQ card and see what you get. – Andy aka Oct 26 '15 at 14:15
  • Im going to make drawing and link to u as another quation. – user16307 Oct 26 '15 at 14:33
  • http://postimg.org/image/44uwx2z51/ sorry for the primitive drawing. bu the i wanted to show the basic system. – user16307 Oct 26 '15 at 14:50
  • the potential difference between A and B is measured by the daq with an offset more than specs. with or without current offset doesn't change. so if i plug off the power supply of the regulator there is still offset. all channels' grounds connected to pci's ground hence to the PC's psu's ground hence to the regulator's psu's ground. so everything seems have the same reference point. but there is still offset. – user16307 Oct 26 '15 at 14:53

1 Answers1

2

Yes in a normal desktop PC mains ground is connected to both the chassis and to the ground pins on the power supply output connectors. In turn the motherboard connects both chassis ground and the ground from the power supply to the ground of the expansion slots.

Whether your aquisition card connects the ground of the PCI slot to the ground of it's input connector is a question for NI but I expect it does.

Peter Green
  • 22,193
  • 1
  • 39
  • 82
  • ok but what do u mean by "PC mains ground"? input of a PC psu has line neutral and earth; and output has +dc and ground. "PC mains ground" refers to which one`? – user16307 Oct 26 '15 at 14:11
  • mains ground=mains earth (the former term is usually used in america, the latter in the UK, dunno about other countries) – Peter Green Oct 26 '15 at 14:26