Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Sharing of DC commons 2

Status
Not open for further replies.

MemsElec

Electrical
Sep 3, 2003
3
I am designing a control system with a Power supply providing +15V, -15V and +24V DC. My question is in regards to the use of the 24V common. Since I only have 1 power supply providing power for all my DC signal, I use the same common for every function that requires use of the 24V common. I use it for a PNP common of a digital output, I use it for the PNP common of a digital input. for power common whenever a component requires it and for a 24V relay common to close the relay, as well as a common for a some Solid State relays that I use. My question basically comes down to this. Is it acceptable to use this same common for everything listed, or should I have some some of filters, cap, inductors or whatever, between each of the different uses. Each component draws in the mA range so current consumption is not a issue. though added up could mean a total of maybe 2.5A. I appreciate any help you could provide.
 
Replies continue below

Recommended for you

Hi,
For 24VDC control circuits the way you described it is the common practice, provided that they are all isolated from primary power.
About +/-15VDC it is not so clear, since you did not mention what it is for.
If there is some analog signal involved, you might need to separate it's common, thought not necessarily.

 
So you are saying that its ok to share the 25V common for every purpose without the requirement of any filtering between the different components. The Power Supply itself is Isolated from the AC primary power.
+/-15VDC is mainly used to supply power to other components. The analog signals run independantly back to the analog monitor, but the analog commons are shared, which I think, causes some ofset between some of the lower voltage signals.

thanks again
 
Ideally you would want to seperate any analog and digital signals. This includes commons. If they both use the same source this is difficult (of course). I would not try to isolate these commons but I would route wires/traces such that all inductive loads(relay) are seperated as much as possible from any digital. The same can be said for analog and digital. They will have to be tied together somewhere eventually unless you want to isolate them entirely. I suggest you set up a seperate ground plane for analog and digital if you want to reduce the chances of having strange things happen.
 
As general rule, different circuits should run separate power supply returns back to the source.

This will tend to eliminate any cross-coupled effects caused by induced voltages in the return lines.

TTFN
 
The comments already posted are to the mark, and may already have answered your question.

If not, to help the analysis, decide where you want your zero reference to be. Especially the zero reference for your digital and analog inputs. This is usually either at the power supply or on a pcb that it connects to. The impedance of your return will add either add offset voltage (think noise)to the zero reference relative to the board, or the board will be running at reduced voltage relative to the power supply.

The latter is usually an easier design problem, you can ignore it, increase the voltage of the power supply, or use remote sensing. Adding bypass caps at your 'zero' voltage point reduces the high frequency impedance. I would guess from your application that having a foot or two of cable between the power supply and the rest of the circuitry would be an advantage. This makes the pcb or power distribution block in your controller a good choice, and bypassing should be done across the power block. Most off-the-shelf controllers will already have adequate bypassing and a place to connect digital 'ground'.

The impedance of the lines will be radically different at dc and at 10 or 100 mHz, which has to be considered.

Since your digital inputs have finite noise tolerance, any noise added by drops across the power supply leads will reduce your noise margins. Heavy loads with fast edges are especially important to control, and you should run these signals separately back to your 'zero point' wherever it happens to be, rather than running them back through a shared common.


If your system is widely distributed (more than a few feet between components) then separate returns using twisted pairs will be even more advantageous.

DspDad
 
Thanks guys for all your help.
just to finish off what I am doing. The DC Power supply using twisted is 2 1/2 ft away from a pcb that distributes the signals. the relays are mounted on the PCB. Once the signal exit the PCB the wiring Also goes out in twisted pair to all the rest of the components (microelectromechanial electronic solenois valves, Solid State Relays for heating etc..)

Thanks again,
MemsElec
 
Dont forget to use some suppression (diode, diode-resistor, RC) across the coils if operate/release time is not too critical. This will improve the overall stability of the design.
 

Skillful use and interpretation of differential-mode oscilloscope measurements at the prototype stage and in later real-world conditions can be worth their weight in gold.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor