Seperation between Signal Ground and Frame Ground


Thread Starter

Mark Meng

Hi, All,

Some control devices (like Japanese PLC) seperate 24VDC 0V (Signal Ground) from PE (Frame Ground), but some European products link these two grounds together inside the device. Can anyone tell me what is the pros and cons of these two different
grounding approaches? In practice, if we have these two kinds of devices in the same system, should we link the two grounds together everywhere or still seperate until linking together at only one point?

Thanks a lot!

Mark Meng
> Some control devices (like Japanese PLC) seperate 24VDC 0V (Signal Ground) from PE (Frame Ground), but some European products link these two grounds together inside the device. <

By coincidence this came up for me yet again a couple of weeks ago. In summary it's probably best to link the grounds with a resistance. Some studio audio stuff uses 47K, some lab instruments as low as 100ohm - I prefer around 1.5K.

If you just link grounds at more than one place magnetic fields (which are around in any industrial plant) will induce currents in the loop formed - you can easily get a few amps and a thus a few volts from wiring resistance in the ground system. The result is false signals.

If you don't link the grounds at all leakage will mean quite high voltages can appear between plant - maybe they will function, maybe components will be damaged, operators will get shocks, you will get false signals when you plug things and so on.

I figure that about 1.5K will limit currents so as to get mV signals, which should not be a problem in most plant, plus will trip 30mA ground fault breakers in safety situations when the two grounds separate by more than 45V or so.

This is all very qualitative - you will get situations where interference and safety considerations mean other measures - extra shields and so on - will be needed. Safety grounds (frame) absolutely need low resistance to the "real" ground and simply should not be used as signal grounds, and local codes may vary as to loops in them.

Best of luck

David Boyes

Anthony Kerstens

It depends on your panel construction. The power feed to your panel should be grounded with a ground conductor that goes back to the grounding of the power source. This is the physical earth and is the electrical "safety" ground. That means the installer has scraped paint on the bonding surface, and did a megger or high-pot test to verify the integrity (if you're lucky :).

Signal ground is used as the zero volt potential on your 24VDC power circuits. In practical circumstances, the PE (panel) is your only grounding point. That said, what I normally do when building a panel is to have copper bars mounted in the panel as ground terminal points. I usually have separate bars for power systems PE and signal ground connections.

However, as long as everything in the panel is properly bonded anyway, the only pro of this is neatness and nice presentation. The con is the extra cost. However, the cost is minimal so I do it anyway. I like neat and pretty panels.

Anthony Kerstens P.Eng.
I have a similar question regarding grounding. I hired a CE consultant to advise us on the design of a machine going to Europe. I have always grounded the neutral of any transformer used to step down the native primary voltage for control power and he told me this can not be done if it is going to pass for CE
certification. We discussed this for a very long time and I still don't agree with him, but am doing it to pass his inspection for CE
certification. My argument against this is if a two wire device is used and it has no physical ground and one of the lines develops a short to the case that this creates a hazardous condition. I looked in the NEC and it specifies that systems may be wired like this if a ground fault system is used or only qualified people work on the
system. Can someone explain what I am not understanding about this?

Dale Witman

Daniel Boudreault

Hello Dale,

I am not familiar with the CE requirements, but we have sent stuff over to Europe with the Neutral grounded... And no one caught us on that... I guess it doesn't really make it right...

Neet thing about not grounding the Neutral is that it makes secondary of the transformer float...and you can touch either of the wire, and not get electrocuted...but don't touch both at the same time...

Dan B.

Santosh S Tangsali

Hi list
For robust system ,I would prefer three seperate grounds one each
for 1) system 2) barriers 3) panel .
for 1 & 2 clean earth is used , in this case the ground resistance is less than 1 ohms. for panels dirty earth system is used , in this case the ground resistance is less than 10 ohms.

all these earthing system are isolated from each other to ensure that there are no ground current loops because of large difference of fault current .

clean earth is formed by connecting 3 to 6 nos earth pits in parallel to ensure low resistance.

dirty earth is usually connected to the plant earth grid , and is a single pit. pit to pit connection is done by galvanised iron flat bars.

although such a design is comparatively costly it is well proven.

one has to ensure that the pits dont go dry in summer . this will increase the ground resistance.

Anthony Kerstens

After doing a little research........

Grounding systems are not designed to carry current in normal circumstances, whereas the neutral is. Also, under NEC the ground and neutral may only be tied together at one point. If the plant where this panel is being installed already has the neutral and ground connected at the main power distribution, then they are also likely providing separate conductors (green/yellow and white) for ground and neutral.

>From that, I might assume that if CE disallows the grounding of the neutral in your panel or transformer, it would then require the tying of the neutral and ground at one single location in the plant.

By all means, ground your panel and properly bond the conduits and trays, but leave the neutral separate since it may only be tied at one point, CE and NEC. Tying them together at multiple points
may result in ground loops.

Anthony Kerstens P.Eng.
Opinions ? no. Real world OK.
Panel ground (call it electrical ground) obeys
the electrical code (here in North America). This is one business.
Signal ground is your business, i.e your Instrumentation ground. As it has nothing to do with Electrical but rather to signal you need your own grounding, and well installed.
The best example is the practice in fairly new plants, like nuclear.
Signal ground and Electrical grounds meet only
in the graphite field outside plant (wet, deep ...) where the earth potentential is considered
nill (zero).
Ground loops and inverse current flows are the two main ennemies in signal measurement.
You can always minimize by pencil and paper and
graphing your signal ground system and tracking the electronic flow.
Too good connections ? Never (galvanic effect).

Mark, there is a fourth ground on insulating fluids: the piping ground between flanges.