# Why the measured voltage increases when I change ground reference?

The following image shows my connections:

The grounds of all devices are connected to a common PCB. When I measure the 1.2V referenced to the ground on my PCB, it reads 1.201V which is great. But when referenced to the common GND PCB, it increases by 23mV, I expected it to be lower than 1.201V (because of bigger ground loop, therefore more voltage drop) not higher, how come?!

Edit #1 : Notes

• I have only ONE voltmeter used for both measurements, sorry it's not clear in the image.
• "My PCB" is supplied by 64V from the shown power supply, then it's converted to 12V then to 1.2V

Edit #2 : More Notes

• These measurements are the same even before I soldered the rest of the board. I solder power modules first of all, even before a load exists I got these results!

• The "common PCB" is a one layer board that is only used as a common ground. It has no ICs or passive components. Only copper pour and holes for crocodiles to be attached.

Because there is current flowing thru the ground wire between your PCB and the common ground point. The wire has a small resistance. The current time this resistance causes a offset voltage. For example, if there were 1 A flowing in the wire, then it would only need to have 23 mΩ resistance to exhibit the 23 mV offset.

Another more subtle possibility is that you have high frequencies in the system that are confusing the meter electronics. Above some frequency, probably somewhere in the 10s of kHz to 100s of kHz range, some meters without adequate filtering may start seeing AC as DC. Old purely mechanical meters are pretty much immune to this, but modern meters with active electronic front ends provide various ways for high frequency AC to be treated non-symmetrically and therefore to add a offset to the DC measurement. That all said, this is rather less likely than the much simpler explanation of a offset voltage accross the ground wire due its resistance times current thru it.

• In the comment he said that the two PCBs are supplied from the same source; this should bring the ground lines to the same potential... – clabacchio Mar 5 '12 at 16:44
• @clabacchio: That's not what the diagram implies, so it's hard to believe everything the OP is saying. Note that the power supply lines are completely missing from the diagram. We don't know there isn't current thru the ground wire, and there probably is. Basically I don't trust this OP to believe there isn't whether he says so or not. – Olin Lathrop Mar 5 '12 at 16:46
• @OlinLathrop, Kindly check Edit #1 and #2. These measurements are the same even before I soldered the rest of the board. I solder power modules first of all, even before a load exists! – Abdella Mar 5 '12 at 16:56
• @clabacchio, the "common PCB" is a one layer board that is only used as a common ground. It has no ICs or passive components. Only copper pour and holes for crocodiles to be attached. – Abdella Mar 5 '12 at 16:58

Probably there is a ground loop between the various instruments and boards, because some of them may be earth-referenced, or referenced to a different potential than others; this creates currents on ground paths, and if they are significantly big (this seems to be the case) it can result in voltage drops.

You should disconnect your instruments and measure the voltage between their ground connectors, to check if there is any difference.

• Currents on ground paths should result in voltage drops as you mentioned. What I'm in now is that the voltage Increases when I measure reference to the common ground. I was expecting it to be less than the 1.201V I measured referenced to GND on My PCB. – Abdella Mar 5 '12 at 16:29
• @Abdella this means that My PCB is sinking current for a drop of 23 mV over the ground wire, and probably this current flows in the second Voltmeter; how is My PCBsupplied? – clabacchio Mar 5 '12 at 16:33
• I have only one voltmeter, I should have cleared it in the drawing. I'll edit the question to clear this. "My PCB" is supplied by 64V from the shown power supply, then it's converted to 12V then to 1.2V. – Abdella Mar 5 '12 at 16:36
• @Abdella it's the same if you disconnect the scope and the PC? – clabacchio Mar 5 '12 at 16:42
• Yes, same measurement, just tried it now! – Abdella Mar 5 '12 at 16:54