How to Get Different Voltages From Single Power Supply

[bostonman]

I have various bench top power supplies, some single output, some with multiple.

The board I want to power needs 12V DC in, and two pins 0-5V DC each (so maybe 2.5V on each pin just for testing purposes - but ideally adjustable between 0V and 5V).

To simplify my question, which is something that I've been confused at for a while, how do I connect the outputs from the power supply to the circuits since the grounds on the power supply are separate?

Most likely I can't connect the grounds together, the circuits can't be powered by two separate grounds, so I'm curious whether I'm overlooking something or I just need a single supply and one (or more - depending on my needs) voltage regulators?

[r6502]

Hi bostonman,

The lab power supplys that I knew do have galvanical isolated outputs. this makes it possible to connect them in any order, you want.

eg: bipolar supply needed:
output 1 negative connected to output 2 positive and acts as GND / common
Output 1 positive connection acts as positive rail
Output 2 negative connection acts as negative rail

If you need an additional rail eg. use as input signal for your circuit, you could use a 3rd output (or another power supply) and connect negative output to common if positive signal is needed or connect positive output to common if a negative signal is needed.

Some power supplies do have a button, where you could turn on / off all rails at the same time, very practical.

Guido

[Terry Bites]

Isolated DC-DC converter to the rescue!
See this article on adjustment
https://www.analog.com/en/resources/technical-articles/digital-adjustment-of-dcdc-converter-output-voltage-in-portable-applications.html

[bostonman]

output 1 negative connected to output 2 positive and acts as GND / common
Output 1 positive connection acts as positive rail
Output 2 negative connection acts as negative rail

I was talking positive voltages, however, this may become a need in the future and/or was a  need in the past.

My power supplies have isolated outputs. Unless I'm thinking of this incorrectly, if as example something has a four-pin input connector, +5 and GND, and +12 and GND, I can't connect channel 1 from my power supply to +5V and ground, and channel 2 from my power supply to +12 and ground, because I'll have two separate grounds. If the grounds are tied together at the power supply end, then it will induce groudning issues, correct?

As for the DC-DC converter, that was my assumption on how to solve the need for separate voltages because the converter will be powered by a single channel on the power supply. i.e. a single ground return.

[ebastler]

No problem at all with that. All the negative (typically black) terminals from your power supply need to be connected together, and connected to your prototype's Ground. Then the various positive (typically red) terminals on the power supply carry the different positive voltages relative to Ground, and can be connected to the prototype accordingly.

No DC-DC converter is required if your power supply has enough outputs for the different voltages you need.

[bostonman]

All the negative (typically black) terminals from your power supply need to be connected together

This is my confusion.

Would this cause a difference in voltage potentials? Some of the current from the (in my example) +12V channel would flow into the +5V channel and vice versa. Then my circuit would have different potentials.

[Ice-Tea]

Not sure if I fully understand where your confusion comes from but picture this: you start with two completely separated circuits, one running on 5V and the other on 12V.

If you now connect the grounds from these two circuits nothing really changes. You're just making sure both circuits are referenced to the same potential, that's all..

[ebastler]

All the negative (typically black) terminals from your power supply need to be connected together

This is my confusion.

Would this cause a difference in voltage potentials? Some of the current from the (in my example) +12V channel would flow into the +5V channel and vice versa. Then my circuit would have different potentials.

No, it will not cause any crosstalk between the different supply voltages. Imagine each voltage source as a reservoir of water (charge) raised to a different height. Now you provide separate hoses from each reservoir into a common kiddie pool at ground (in both mechanical and electrical terms) level.

Water (current) will flow from each reservoir to the ground level, driven by a pressure (voltage) defined by the height difference between the reservoir and ground.  That's what you want. There will be no unwanted interactions between your sources.

[bostonman]

Years ago I blew a transistor in a power supply at an old job, but I don't remember what I did to blow it. As a result, I've never done any (what I considered to be) alternate connections; especially with my personal equipment.

My needs over the years have been utilizing the 'series' option on a power supply to get more voltage, never used the 'parallel' option though. So I'm inexperienced trying to power a circuit with a single ground that has two or three different voltages from three different channels, but all grounds tied together.

One of my power supplies I believe is three channels which will be perfect for my needs (as stated, I need 12V DC, and two separate 5V adjustable from 0 to 5), and I'll connect all three ground terminals together at the power supply end.

[ebastler]

One of my power supplies I believe is three channels which will be perfect for my needs (as stated, I need 12V DC, and two separate 5V adjustable from 0 to 5), and I'll connect all three ground terminals together at the power supply end.

Yup, that should do the trick.

You can do the same with multiple power supplies too, by the way, if you need even more different voltages. As long as each of them provides a voltage which is floating relative to mains (and any "lab power supply" will!), you can wire all their grounds together.

[bostonman]

floating relative to mains

I assume you mean isolated as in going through a transformer?

You can do the same with multiple power supplies too

This was a surprise as I would have assumed this would certainly cause ground loop issues.

Maybe I'm not understanding still, and apologize, but the visual still causes confusion. Let's assume my circuit needs 5 and 12V DC, for purposes of discussion 1A each, and a single ground (maybe 12V to power a display and 5V for logic). Now if my two-channel power supply has the grounds tied together, why can't the 1A that goes through the 12V circuit for the display see a lower resistance through way of the 5V ground and the full 1A from the 12V and full 1A from the 5V go through this channel thus sending 2A through channel 2 possibly blowing the transistor (let's assume the bench top supply can only supply 1A each channel).

In my head I'm visualizing the possibility of all the current flowing into a single channel whereas in a normal configuration, let's say two totally different circuits, the current is flowing in a loop within each of the bench top output channels.

[ebastler]

Hmm -- I can't quite understand where you see the potential problem. Looking at the attached simplistic diagram with two voltage sources which supply two different parts of a circuit (just represented as resistors here): Where would the undesirable current path be?

[bostonman]

I'm thinking internally channels 1 and 2 are isolated from each other and the current that should flow back into channel 1 now has some, or all, flowing into channel 2, then the transistor(s) will see greater than 1A flowing through it.

[ebastler]

I'm thinking internally channels 1 and 2 are isolated from each other and the current that should flow back into channel 1 now has some, or all, flowing into channel 2, then the transistor(s) will see greater than 1A flowing through it.

Please draw a fully connected line in the circuit diagram which shows how some of the current that flows through R2 manages to accidentally flow through source V1. 

That is impossible in my book. All the current flowing through R1 will also flow through V1 (only), and all current flowing through R2 will flow through V2 (only).

[bostonman]

Obviously my thinking is incorrect.

The external circuitry and tied grounds make sense. The transistors internal to the power supply in my mind should have separate grounds. Tying them together (again, in my mind) can cause current from one channel to flow through the other.

I'd assume each channel/transistor is fed by the secondary on the transformer by separate windings. Maybe, as usual, I'm thinking too deep and confusing myself rather than keeping it simple.