Is Grounding in Circuits Truly Relative?

[ajl254]

Hi,
I have a question about grounding. Everything's relative right? Take two power supplies, each at 15+ DC and link one to the other (in series). Connecting to the circuit you will have a zero point (ground), a +15 VDC, and a -15 VDC. Reconfiguring the circuit and calling -15 VDC ground you will now have 0-30 VDC power source.

Now take a different case. Two power supplies; +10 VDC and +5 VDC. Link them together in the same fasion (series). You now have a few options (without drop resistors for alterations)): 0-15 VDC -15-0 VDC, -5-+10 VDC, and so on.

A particular IC (none specific, seems to happen to most); This IC needs a Vin of +5 VDC and a 'ground' point to operate. Theoretically one could have Vin as +10 Vdc and 'ground' +5 VDC, but i.f.f. any branches of circuits involved with that IC, and their child branches use +5 VDC as 'ground'. If for instance an entire circuit is powered this way, and the true ground (0 VDC) never sees the circuit, then the circuit always sees potential in reference to the 5 and 10 V potentials. This (or at least I think, maybe I'm wrong; correct me if so) means that the IC/circuit is seeing 0-5 VDC the same as a true ground and +5 VDC power source.

Setting up and attempting multiple configurations and scenarios; I have seen that this is not true.

What is wrong with this? It's obviously wrong or it would work; the problem is I don't understand why. If it matters: we are putting power to a DSPIC 33F6014A. Other IC's being used are Murata NMT0572SC.

Any help would be greatly appreciated.

 

[berkeman]

Hi,
I have a question about grounding. Everything's relative right? Take two power supplies, each at 15+ DC and link one to the other (in series). Connecting to the circuit you will have a zero point (ground), a +15 VDC, and a -15 VDC. Reconfiguring the circuit and calling -15 VDC ground you will now have 0-30 VDC power source.

Now take a different case. Two power supplies; +10 VDC and +5 VDC. Link them together in the same fasion (series). You now have a few options (without drop resistors for alterations)): 0-15 VDC -15-0 VDC, -5-+10 VDC, and so on.

A particular IC (none specific, seems to happen to most); This IC needs a Vin of +5 VDC and a 'ground' point to operate. Theoretically one could have Vin as +10 Vdc and 'ground' +5 VDC, but i.f.f. any branches of circuits involved with that IC, and their child branches use +5 VDC as 'ground'. If for instance an entire circuit is powered this way, and the true ground (0 VDC) never sees the circuit, then the circuit always sees potential in reference to the 5 and 10 V potentials. This (or at least I think, maybe I'm wrong; correct me if so) means that the IC/circuit is seeing 0-5 VDC the same as a true ground and +5 VDC power source.

Setting up and attempting multiple configurations and scenarios; I have seen that this is not true.

What is wrong with this? It's obviously wrong or it would work; the problem is I don't understand why. If it matters: we are putting power to a DSPIC 33F6014A. Other IC's being used are Murata NMT0572SC.

Any help would be greatly appreciated.

Welcome to the PF.

What you have said is basically correct. Why do you say that when you try it it dosn't seem to be true? What did you try exactly?

 

[ajl254]

We’ve tried multiple configurations. I work for the navy in research and development. A colleague of mine is making a circuit and it is beginning to become slightly complex. I was working with him and thought of this solution (simpler circuit and lower cost); however, he said he had tried that already and it did not work. We then had a discussion, working out circuits and theory and why this should work, but isn’t. We then decided to cut the fat and put basic theory to the test. Using the stated PIC and the programming board we connected +9 VDC to ground, and +18 VDC to Vsource. All the grounds in the circuit are connected and isolated completely from Earth ground; the circuits ‘ground’ is +9 VDC. The circuitry can’t ‘know’ that, it only sees references. Connecting it to a comp to program it yielded a ‘no voltage’ failure (connection to comp does not tie ground in, only has 4 wires, + - Tx and Rx). I should state that the circuit has electrolytic caps in it, but still same principle, as long as the + side of cap is at higher potential then no problem. We’re stumped as to why it won’t work.

 

[Antiphon]

You should post schematics. That will break loose some good advice.

 

[berkeman]

We’ve tried multiple configurations. I work for the navy in research and development. A colleague of mine is making a circuit and it is beginning to become slightly complex. I was working with him and thought of this solution (simpler circuit and lower cost); however, he said he had tried that already and it did not work. We then had a discussion, working out circuits and theory and why this should work, but isn’t. We then decided to cut the fat and put basic theory to the test. Using the stated PIC and the programming board we connected +9 VDC to ground, and +18 VDC to Vsource. All the grounds in the circuit are connected and isolated completely from Earth ground; the circuits ‘ground’ is +9 VDC. The circuitry can’t ‘know’ that, it only sees references. Connecting it to a comp to program it yielded a ‘no voltage’ failure (connection to comp does not tie ground in, only has 4 wires, + - Tx and Rx). I should state that the circuit has electrolytic caps in it, but still same principle, as long as the + side of cap is at higher potential then no problem. We’re stumped as to why it won’t work.

When you measure the voltages with a DVM before and after connection to the programmer, what do you see? Are you sure that all the power supplies are floating from Earth ground?

 

[Okefenokee]

I think he's asking why a chip cannot be powered by connecting a +5V source to the chip's ground and a +10V source to the chip's Vin. That's a really good question.

I would think that it would work. Keep in mind though, that even though you are absolutely correct that it's the voltage difference that matters in a circuit, the environment will usually be at zero volts. If you tie the chip's ground pin to 5V, field lines will extend out from the ground line inside the chip to the nearly 0V environment. Maybe the field lines create some kind of Hall effect that prevents the chip from running. Also, any pins that would have been floating with a normal power hookup will certainly be relatively negative.

 

[berkeman]

I think he's asking why a chip cannot be powered by connecting a +5V source to the chip's ground and a +10V source to the chip's Vin. That's a really good question.

Oh, that's different. I didn't pick that up. No, that shouldn't work. power supplies can only source current out of their + output, they cannot sink current via their + output.

 

[sophiecentaur]

It would work as long as the 'lower' of the two supplies is already sourcing enough into other circuits. (Or if it is a shunt regulator)
But why does this IC need to be operated between 'unusual power rails"? Not only is it difficult to power it but will you not have issues with DC level changes.
Or was it just a whatif type question?