standards the negative leg in dc based systems?


Thread Starter

Mark Riche


I was wondering what are the preferences or standards for grounding or floating the negative leg in dc based systems ?


Mark Riche
Acres International Limited
79 Selkirk Avenue
Thompson, Manitoba
R8N 0M5
tel (204) 677-0711
fax (204) 677-0718
[email protected]
I have seen systems with and without grounding the "negative leg", or rather DC common leg. I prefer to ground the DC common. This ensures all DC supply have the same reference, ground.
Ontario Electrical code Section 10-100

(1)Two-wire direct current systems supplying interior wiring and operating at not more than 300v or not less than 50V between conductors shall be grounded, unless such system is used for supplying industrial equipment in limited areas
and the circuit is equipped with a ground detector.

that's what the good book says anyway. We pretty much always ground the neg. leg of DC here.

Patrick Allen

control systems technician
Atoma Technical Centre


Michael Griffin

Normal practice appears to be to ground the negative (or "common") side. It should be noted that a good many electronic devices have a limit on the difference between ground and a floating supply voltage.

In cases where you need to float it, check the electrical code to see if your power supply needs to use a "shielded" transformer. The shielded transformer ensures the secondary winding is isolated from the primary in the event the insulation breaks down. I know someone who ran into this problem once during an inspection on a (rather large) DC system.

Michael Griffin
London, Ont. Canada
Seems to me the option of grounding depends on whether the "negative leg" is isolated from the ac line or not. For example, if the dc system is derived directly from the ac line via a diode rectifier/capacitor filter circuit, the so called negative leg is connected to the ac source in a polarity alternating way. In this case, grounding the negative leg would cause a ground fault. A negative leg that is truly isolated does not have this concern. Makes sense?

Guy H. Looney


I have never heard one valid reason for connecting earth ground (GND, SHLD, frame ground, etc.) to 0VDC (common, negative, -VDC, etc.). I think it's a very bad idea, but unfortunately I'm a minority. Why is it done? Here's the #1 reason given to me when I ask: "...they're all grounds, right?". Nope, not really.

Here are a few of the reasons why I feel you should NOT connect GND & 0VDC:
1) There's no reason to!!!
2) Ground loops mean noise. I've seen a panel's GND circuit register a 46VAC potential when compared to the GND circuit of an outlet located
directly beside (but outside) the panel. Now this is a rare occurrence, but normally there is always some level of AC voltage on the GND circuit
(usually less than 1VAC). This is because a neutral is derived from the GND circuit. AC voltage is a noise generator when connected to 0VDC. Many circuits still operate at 5VDC and it doesn't take much noise to disturb them.
3) Debugging / troubleshooting a panel is harder. If you're trying to isolate a voltage signal, it's a heck of a lot harder if the GND & 0VDC are connected. I can elaborate if this doesn't make sense.
4) Short circuits / lighting strikes. What happens if a 460VAC, 100 HP motor fries? The voltage & current have to go somewhere. Anybody that doesn't think some, most, or all of it isn't dissipated through the GND leg needs to think again. Well if you've got the GND connected to the 0VDC signal can you think about what might happen? Lighting strikes are also another source of "contamination" on the GND circuit. Same argument holds about where some, most, or all of the voltage goes.
5) Communications problems. Ever had problems with a desktop communicating to a programmable device, but a laptop works just fine? How about one laptop working and another not? How about when a laptop is running on battery it's okay & when it's plugged into the wall it's not? Assuming the COM port is working & the software is okay, the problem is probably due to a
potential difference in the 0VDC on the computer and the programmable device or noise. Technically, this is due to either a ground loop or a problem with terminating resistors (RS485/422 hardware layer only). But the ground
loop wouldn't affect the communications if the GND & 0VDC weren't connected!!!!

I'm sure I've left a few out (potential differences between panels for example), but I think you get my point.

I'd LOVE to hear some support for connecting GND & 0VDC. Anyone????


Guy H. Looney
Motion Control Engineer

A.C.E. Systems, LLC
work: (615) 754-2378
fax: (615) 754-0098
cell: (615) 330-0044
[email protected]

Gary H. Lucas

I can offer you one quite good reason. If the DC circuit is not grounded then you will not detect a ground fault in your DC circuit if one occurs,
the fuse won't blow. This ground fault ( sharp edge cutting into a wire?) may wreak havoc with your control system because it creates another ground loop that very well may have circulating currents flowing in it. It may be VERY hard to detect, especially if it is intermittent in nature. Your concerns about ground loops are very good. When I see a piece of machinery
with a ground rod driven along side connected to a heavy ground wire I am sure I am looking at the work of people that just don't get. You know who
REALLY doesn't get it? Encoder manufactures that supply shielded cable tied to the encoder case! You install it and connect the shield wire hanging out and you've got yourself a nasty ground loop.

Gary H. Lucas
> I have never heard one valid reason for connecting earth ground (GND,
> SHLD, frame ground, etc.) to 0VDC (common, negative, -VDC, etc.). I think
> it's a very bad idea, but unfortunately I'm a minority. Why is it done?
> Here's the #1 reason given to me when I ask: "...they're all grounds,
> right?". Nope, not really.

Grounding is a complex issue froth with uncertainties. While it is certainly not a requirement to ground low voltage DC circuits, I can give you some good reasons why you might want to:

1. Safety - Should a higher voltage circuit come into contact with an
ungrounded DC instrumentation circuit, an instrument technician may be
in for a rude shock(literally) and possibly a fatal one.

2. It's been my experience that large integrated(and for that matter
generally) instrumentation systems work better if they have a common

3. It's is difficult to maintain true floating in a system unless you are
using a battery or DC-DC converters.

4. If you use intrinsic safety and use passive barriers, you are required
to ground.

5. While not really a reason, one might wonder why all major DCS
manufacturer's require that their DC side be grounded.

It is certainly true that grounding to earth does bring its share of problems but in a well designed system they are minimal.

Bill Mostia
William(Bill) L. Mostia, Jr. PE
Independent I & E Consultant
WLM Engineering Co.
P.O. Box 1129
These opinions are my own and are offered on the basis of Caveat Emptor.

Johan Bengtsson

Someone was into it in another post but anyway...
Some devices, for example profibus devices (all RS485 devices?) don't allow more than a certain voltage difference between 0V and the RT/RX wires. This means all devices on the same wire must have some common enough DC voltage. One way to do this would be to ground them, ok the voltage will not be the same for all units but that is not required either as long as it is close anough.

If my memory doesn't fail completely the allowed difference in the case of a profibus unit is something around 40-50V.

That doesn't mean I don't agree to your statements, this is just an example of why you might want to do it. (Yes I realise it could be solved in other ways too).

/Johan Bengtsson

P&L, Innovation in training
Box 252, S-281 23 H{ssleholm SWEDEN
Tel: +46 451 49 460, Fax: +46 451 89 833
E-mail: [email protected]
Do you need education in the area of automation?