Where is "GND" with multiple DC power supplies?

[mmagliaro]

Imagine a DC power supply (could be a track throttle, or even just some general purpose accessory DC power).
AC comes in from somewhere, goes through a bridge rectifier, and out comes DC.  We get + and - out of the rectifier,
but the "-" isn't GROUND.  It isn't grounded to anywhere.  It is just a reference point.  We might call the two output
lines "0" and "12", but really, we just have a reference point and then 12 volts above that.

Now, imagine a SECOND DC supply. 

And now, we tie the two "-" terminals together from the two supplies.   And we connect the "+" outputs  to various loads.
What exactly happens since the two negatives were just floating reference points?  Obviously, they are forced to be at the same absolute point above true ground because they are now tied together.  But what happens if one happened to be 0.1v above ground and the other was 0.3v (before we connected them together)?  Doesn't some current actually flow from the one that was floating higher, because it is being pulled down by its connection to the first one?

Am I just over-worrying about this?

[DKS]

Assuming the "+" of the two supplies are not tied together, the "-" remains the reference point; whatever the first power supply is powering, that's the voltage they see, and likewise for the other power supply.

What you have is a classic "common return" arrangement. You can connect as many different power supplies together, and it makes no difference to the outputs, or the things they power. You can even connect the "+" to the common return (along with all of the other "-"s) and now you have a source of negative voltage, relative to the common return.

Even stranger, you can connect one terminal of an AC power supply to the common return, and it will behave just as it would ordinarily; whatever it powers will see the AC, regardless of what other supplies are connected in common.

[peteski]

Like DKS mentioned, if the power supplies are not grounded to the AC ground prong (if they use a grounded power cord) then all the outputs are floating.  That also assumes there is no current leakage internally in the AC transformer.

Just like if you were to connected single side of 2 car batteries or 9v batteries together.  Since each is a source of independent voltage if single lead of each battery is connected to the other, there is no voltage differential.  To have a differential, there has to be another path for the current to make the electrons flow.

Yes, you can connect a positive output of one power supply to the negative output other one and use them as a dual polarity (positive/negative voltage) power supply. The connection between them then becomes the "common" point (zero volts).

The nomenclature commonly used makes things confusing because the common return can be called "ground", "earth" or "common", even if it is  not connected to earth ground in the AC outlet.

The above is true for the standard AC transformer-based power supplies with 120V primary and isolated low voltage secondary.  With the modern switching power supplies there might be some path from the low-voltage secondary to the 120V AC side of the supply.  This is especially true with the switching power supplies that use grounded AC power cords.  If trying to connect multiple switching power supplies together I would be careful and at least take some measurements of any voltage or current between the outputs before attempting to connect them into a common mode supply.

[wvgca]

if the grounds are tied together , either at the output or a grounded plug input, they can be referred to as 'common ground' , if they are not tied together in any way, they are simply called 'floating ground'  ..and yes , on floating grounds it quite common to tie a positive to a negative  ..

[mmagliaro]

Guys, I get that all the outputs are floating.   But they are floating above points that were at different potentials above absolute ground.  So if power supply 1's negative terminal was 0.1v above ground, and 2's was 0.3v above ground, and I tie them together, they are now at the same potential above true ground.  But what value?   Of course, I've done this for years (connected the negatives together as a common).  But I never thought about what happens if there is big difference between the floating negatives.   

[wvgca]

if you tie them together, and they are floating, not tied to earth ground in any way ....does it matter what potential they float at ??

[DKS]

Guys, I get that all the outputs are floating.   But they are floating above points that were at different potentials above absolute ground.  So if power supply 1's negative terminal was 0.1v above ground, and 2's was 0.3v above ground, and I tie them together, they are now at the same potential above true ground.  But what value?   Of course, I've done this for years (connected the negatives together as a common).  But I never thought about what happens if there is big difference between the floating negatives.

In a common return scheme, there is no "absolute ground." Everything is "floating," and a common return system is not affected by differences between the potential at any supply terminal and "absolute ground." I'm assuming you're definition of "absolute ground" is household ground, or "earth." The common return of a group of power supplies would not, and indeed should not, be tied to household ground, nor should either output terminal of a supply be tied to household ground internally. This could otherwise cause grief. It's also why one should be cautious of using a switching power supply in such an arrangement, as Peteski mentioned.

if you tie them together, and they are floating, not tied to earth ground in any way ....does it matter what potential they float at ??

No, it does not. As noted earlier, the potential of any one terminal tied to a common return can be anything, from -50 VDC to +0.02 VDC to 6 VAC relative to the other terminal, or relative to "absolute ground." It makes no difference.

Am I just over-worrying about this?

Yes.

[wvgca]

they shouldn't float a great deal, most [but not all ] newer supplies have some sort of tie in back to the neutral, which is tied to the earth ground at the main distribution panel ..

[peteski]

Guys, I get that all the outputs are floating.   But they are floating above points that were at different potentials above absolute ground.  So if power supply 1's negative terminal was 0.1v above ground, and 2's was 0.3v above ground, and I tie them together, they are now at the same potential above true ground.  But what value?   Of course, I've done this for years (connected the negatives together as a common).  But I never thought about what happens if there is big difference between the floating negatives.

What exactly do you consider "absolute ground"?  As I explained with using individual batteries, there is no absolute anything. Isolated power supplies are just like those batteries - there is no common connection or voltage/current between them until you connect single terminal of each battery together. Then that becomes the common, or a ground if that is what you want to call it.  Any voltages difference or currents will only begin to flow once you you connect an electrical load between that common and one of the other terminals of either battery.

Now if you were to also connect the second terminal of each battery together then unless they produce absolutely identical voltages with the same polarity, there will be some current flow  between them.  But that is not really done. I realize that you already know all that, but you also have to wrap your head around the fact that there is no "true ground" with floating voltage sources.  If some "true ground" existed (which would be at some different potential from the common of those 2 floating voltage sources), how would you measure that? 

[mmagliaro]

Okay.  Let's consider "true" or "absolute" ground to be the earth outside my house with a copper rod shoved into it.
I take my two power supplies and measure between the "-" and earth ground on each one.  On the supply 1, it's "-" is 0.1v above earth, and supply 2's minus terminal is 0.3v above earth.
We put a load on each one.  Leave them completely independent for now.  Both loads have 12v across them (assume both supplies are 12v).

Now... tie the two "-" terminals together.   They are now the same.  What voltage are they at, relative to earth? 

[peteski]

Okay.  Let's consider "true" or "absolute" ground to be the earth outside my house with a copper rod shoved into it.
I take my two power supplies and measure between the "-" and earth ground on each one.  On the supply 1, it's "-" is 0.1v above earth, and supply 2's minus terminal is 0.3v above earth.
We put a load on each one.  Leave them completely independent for now.  Both loads have 12v across them (assume both supplies are 12v).

Now... tie the two "-" terminals together.   They are now the same.  What voltage are they at, relative to earth?

Unless you actually need to connect them to your house's "earth ground" what you mentioned is totally irrelevant.  If this was a real-world example as to why the there is a difference in measured voltages, there could be many reasons.  Do those supplies use grounded plug or just a standard 2-prong plug?

Did you try the measurements using both AC and DC volt range?  Did you try measuring the current instead of the voltage?  Just curious. 

EDIT: To answer your question, if you tie the two "-" terminals connected between the floating power supplies to earth ground, then they will be at the same potential as earth ground (regardless if there is any load on the output of the power supplies).

In a perfect world (or in a simulation) there should not be any measurable voltage or current present between earth ground and any of the outputs of a floating power supply, but in real world, there could be possibilities for that not to be true.

[mmagliaro]

No, when I tie the two "-" together, I do not connect them to earth ground.  Just to each other.  My curiosity is just  that before they are connected to each other, they are floating at different levels relative to earth.  So when connected to each other, isn't there going to be some small current flow before they settle at one common point?

[peteski]

No, when I tie the two "-" together, I do not connect them to earth ground.  Just to each other.  My curiosity is just  that before they are connected to each other, they are floating at different levels relative to earth.  So when connected to each other, isn't there going to be some small current flow before they settle at one common point?

Again, to bring up my floating 9V batteries explanation, there should be no voltage difference or current flowing if only one lead of each power source is connected.  "Floating"means they are totally separate and independent. there is no reference to anything (like earth ground).

Are you just theorizing, or you are using some real example?  If this is a real example then go ahead and connect an ammeter between the negative leads of the power supplies and measure the current.   If there is a current flowing than something somewhere in the power supplies is leaking current connecting completing the circuit, causing the current flow.  But the current will not "settle". It will always be flowing.  It is not like some capacitor where the current would stop flowing when the capacitor fully charges to the some lever as the voltage applied to it.

[DKS]

No, when I tie the two "-" together, I do not connect them to earth ground.  Just to each other.  My curiosity is just  that before they are connected to each other, they are floating at different levels relative to earth.  So when connected to each other, isn't there going to be some small current flow before they settle at one common point?

Okay, let's try this. Measure the potential from the "-" terminal (or the "+" terminal, for that matter) of any power supply relative to "absolute ground." If you see any potential at all--as in 0.1 V, 0.3 V, anything--then you have a problem. There is internal leakage in the power supply between its output terminals and its input, i.e. the wall plug. In this case, you should not use it in a common return scheme (indeed, I wouldn't use it for anything). There should never be a potential present between the output terminals and ground or anything else other than one another. Ideally, a power supply should behave just like a battery.