Understanding grounds

Total hypothetical here. I'm trying to teach myself and there's something here I really fundamentally don't understand.

Let's say I have two devices with separate power supplies on separate circuits. Let's say they are 9V and 12V DC power supplies. Let's say I measure voltage between the grounds on the two devices and the one with the 9V power supply is +2V relative to the one on the 12V power supply, for whatever imaginary reason.

Now let's say I want these two devices to talk to one another over a serial connection. So I connect the two grounds. What happens?

Does current flow from one to the other through my new connection? If I were to go back and measure from positive to this new common ground on the two individual power supplies, will they still say 9V and 12V? Or does everything magically adjust itself to the new reference point? Or is it something that falls out in the math?

I just really don't understand how to imagine this.

OK I know it's kinda strange but as long as the two power supplies are each galvanicley separated and DO NOT share a ground with earth it should be fine to hook the grounds together. They will be 0 reference to the highs 9 and 12 no changes in voltage on the highs but you may still have a voltage between the ground and earth ground. If you have doubts about it hook a low voltage light between the grounds and see if it lights if it does you have a problem don't continue with it.

say I measure voltage between the grounds on the two devices and the one with the 9V power supply is +2V relative to the one on the 12V power supply, for whatever imaginary reason.

Everything depends on that reason.
It the are seprated then everything works. And that measurement did not give a real result.

If they are not seprated then lots and lots of current can flow and potentially go bang.

Delta_G:
Let's say I measure voltage between the grounds on the two devices and the one with the 9V power supply is +2V relative to the one on the 12V power supply, for whatever imaginary reason.

That's not correct. In reality the grounds are "floating", they're not referenced to anything.

(Or they should be if everything is working properly)

The test to see if they are floating is to connect them together with a 1K resistor. If you still measure the same differance then they are not isolated and you should not connect them together.

Grumpy_Mike:
The test to see if they are floating is to connect them together with a 1K resistor. If you still measure the same differance then they are not isolated and you should not connect them together.

So I'm getting that the moral o the story is, "It depends on what happens when you connect them".

So I'm getting that the moral o the story is, "It depends on what happens when you connect them".

No, exactly the opposite in fact.

What happens when you connect them depends on what they were to begin with.

The test to see if they are floating is to connect them together with a 1K resistor. If you still measure the same difference then they are not isolated and you should not connect them together.

+1

For AC circuits up to 120VAC, I use a large 225W 1K resistor to determine the degree of isolation between 2 contact points. The resistor is connected across the voltmeter (multimeter) terminals and any voltage measurement is also the milliamps, where 1V = 1mA.

Excellent for determining the level of isolation / leakage / safety level. Note that 0V reading may imply a full isolation OR a complete short between the 2 points. If there's no change in voltage with resistor connected or disconnected, then there's no isolation between the points.