[SOLVED] please explain non 0 Volt grounds

[embed68k]

I am in the process of converting a uATX power supply into a lab workbench. I will add a 10W, 10ohm resistor as a load, but it works without a load( I don't know how long it would continue to work without a load). Anyway, it has been suggested from diy sites that I can get 3.3, 5, 12, and -12 Vdc coming off the rails. But if I wanted a 7 or 24 VDC source I could connect the positive to 12V and the negative to 5 volt. Or 12V postive -12V negative to get 24V. How is this possible and what are the current sourcing effects of doing this. In the power supply the rails at different voltages have different current sourcing abilities, so am I to assume I just take the lower of the two current ratings or would this combine them to the max the fuse/power circuitry allows. I'm very confused about non 0 V grounds, which means I don't understand OP amps either, or anything else that takes negative voltages. Can anybody help clarify with formulas(non calculus or set theory) or analogies how this is possible. I can probably assume this combining voltage mess only works with voltage sources that share a common ground? I loaded some power sources and a multimeter up in MultiSim and confimed what the diy sites were telling me, I just don't know if it is safe or reliable to do in real life, especially the current ratings.

 

[Element_115]

Just think about if you had two 1.5V batteries.

If you put them in series and connected them to a load across them what would
the load see? 3V (of voltage potential ).

Now at the "GRND" / "Return" Reference of the
first circuit (the bottom of the two batteries in series) you can attach the positive
terminal of another 1.5V battery and the negative terminal of the new 1.5V battery
you tie it to Earth Ground ! There will be no change in the first circuit. But if you
measure the new circuit with "Earth ground as your reference you will see lager
values for you voltage.

Now if you were to get a 2 output Power Supply (set each channel to 5V) and tie
the Neg terminal of Output CH1 and the Positive of Terminal of output CH2 to
Earth GRND (The green banana jack typically). Then use a multimeter with the
black (Negative) probe on the Power supply's Earth GRND, and use the red
(Positive) Probe to measure the Output of the power supply you will see +5V on
O/P CH1 and -5V on O/P CH2! Now use the Multimeter and put Black Probe on the
O/P CH2 and the Red Probe on O/P 1 and you will red +10V.

Also look up ground loops this will give you more insight to "GRND REF"
As for current you need to make sure all channels can handle the current you
expect in you circuit.

 

[embed68k]

I'm planning on only having one ground path for all voltages. I simply don't have room on the small power supply to break out a ground for all voltages. There are plenty of ground wires, but most of them will be clipped off at the board. So I am correct in assuming that using the two rails at different current sourcing abilities I can expect the lower of the two? And also connecting 12+ to 5+ to get 7+ is a normal property of electronics and is just as safe and reliable as connecting a 12V to a variable voltage regulator? I won't be attempting to hook up huge loads that draw tons of current. I will use no more than 1 or 2 amps from this supply at any one time. Explaining it as voltage potential does help my understanding slightly. I work at a automotive battery plant. If I were to hook up a 12V auto battery and a 8V golf cart battery(I promise I won't really) There is a voltage potential of 4V DC and would last as long as the cranking amps of the lower battery? We ship the batteries overvolt, but just assume 12.00 and 8.00 values instead of 12.60 and 8.46V. It seems like there could be some kind of water pipe analogy, but I just can't seem to put my finger on it.