Hi,

I have a question to a subject for that I do not have a lot of knowledge. It is related to filtering between Protective Earth (PE) and Ground (GND) for a quite sensitive application.

Let me begin with some short notes on the application itself. Basically it is a device for lab applications like scales, microscopes or other sensitive stuff. Now I experienced the case that some users (worldwide) experience the appearance of mains humming (50Hz/60Hz - depending on country). But as this effect is only observed rarely, I did not put too much attention on it. This will change now

Please excuse my lack of knowledge. I studied physics and my knowledge on power electronics is very rudimentary

I suspect that for these users, they operate the device in an environment where the PE line might be pretty dirty. So I took a closer look at the power supply, which is a relative normal comercial AC/DC 12V power supply. The routing of this power supply is shown below.



So here the PE line is used for shielding. Moreover there is an EMI filter at the input of the AC/DC converter. Apart from that PE is directly connected to 0V (GND) on the DC side (secondary ground?).

For me it seems that this setup is not very practical as I suppose that if you have a "dirty" PE line, this noise should directly couple into my device? But what kind of filter do I apply here? I tried to teach myself about the filter types and measures how to deal with noise between PE and GND (Emi filter, common mode choke...). But all these examples were dealing with filters on the AC side. How would you deal in such a situation? Is it even possible that this flaw in the power supply (is it even a flaw?) can lead to mains humming in the device? Can you recommend literature or other sources on this subject?

Thank you very much in advance for your help!

 

In many cases it is useful to disconnect the negative output from the "protective earth"connection. That is because the PE circuit may indeed include noise. And the 12volt circuit is not a shock hazard that needs a PE connection.
Keep in mind that the PE connection is to prevent a shock hazard. So the PE connection might add noise, or it might not.

 

Hi,
I guess, the "Filter" in square is just a 2-5 nF capacitor. In some cases, a small ferrite choke is in series with the capacitor. This technique is to minimize RF interference, which can lead to humming , too.

A good book:
https://www.amazon.com/Electronic-System-Design-Interference-Techniques/dp/0132521237

Take in mind, it's not a simple task to win noises and humming, in most cases

 

Really, the type of filter that you require ON THE OUTPUT SIDE depends both on what sort of noise is present and how demanding the requirements for "silent DC" are.
With a switch-mode power supply, the filtering on the mains side is primarily to keep supply generated out of the mains wiring. The filter in the green-wire (PE) connection may be a benefit in both directions. If you are able to separate the supply output negative from the PE connection that will probably be a benefit, PRESUMING that the supply provides adequate isolation between the mains and the output side. Usually adequate output filtering is provided by capacitors connected across the output. Not knowing the requirements of the application makes any reasonable guess difficult.

 

Hey,

Sorry for the late respone and thank you very much for your support. I will try to clarify the situation a bit more. But beforehand let me adress some of your tips.

@MisterBill2: Unfortunately I cannot disconnect secondary GND and PE as in this case 50Hz humming is even more present. I am also irritated by that. Additionally the power supply is not very sophisticated. But what makes me wonder why in some cases we have noise in our system and at other locations there is not any. Especially in difficult environments with heavy machinery, the PE seems "dirty" which is then also visible in our electronis.

@Pyrex: You are right with the Filter on the primary side. Nevertheless I am not sure how this filter affects the connection between PE and secondary GND

So let me explain the setup a bit more in detail. We are supplying active vibration isolation for sensitive equipment. The power supply looks as follows:

View attachment PE_GND_Loop_extended.png

With this setup I tried to quantify the influence of "dirty" PE to the setup. Therefore I used a accelerometer, measuring the residual vibrations and determine the noise at 50Hz. For generating a "dirty" PE I used a soldering station in the same multiple socket adapter. I got the following results:

Noise amplitude at 50Hz (a.u.)
Vibration isolation active - 0.091
Vibration isolation active and soldering iron on - 0.122
all power cables detached from vibration isolation unit - 0.041

In order to improve this, I soldered an LRC Filter between PE and secondary GND. The new setup looks as follows:


L=220uH
R=100k
C=2.2uF

With the following values I was able to reduce the amplitude at 50Hz by a half:

Noise amplitude at 50Hz (a.u.)
Vibration isolation active - 0.046
Vibration isolation active and soldering iron on - 0.062
all power cables detached from vibration isolation unit - 0.039 (of course this measurement was just to check for repeatibility)

Is this the way to go? Do I make sth wrong? Do you have other recommendations?

Thanks in advance for you help and with best regards

 

The circuit drawing serves me as a reminder that the PE ground IS ONLY FOR SHOCK PREVENTION!!! what that means is that it does not need to be connected at all to anything beyond an isolating power supply. Really, it only needs to be tied to exterior portions of the supply that might become connected to the mains in the event of a failure of some kind. That also means that if the 12 volt supply can qualify as "double insulated", it does not even need to have a PE connection. That might possibly allow a noise reduction.

 

IMO the best method for a sensitive lab or medical clean power connection is isolation. I would using a medical quality isolation transformer to create a separately derived power system to break the galvanic connection on all instruments using the isolated power of the sensitive application and bond The Protective Earth (Ground) Conductor on the isolation transformer to the lab's safety ground if needed. The PE is usually not 'dirty' in properly working circuits, it's the Neutral, current carrying connections that cause the bulk of noise and hum in sensitive circuits IRT the PE connection. The PE connection is a reference "common' that should have current only in a fault condition.
https://www.mouser.com/new/hammond-manufacturing/hammond-line-isolation-transformers/
https://www.hammfg.com/electronics/transformers/line/178-240-89-98mg

 

IMO the best method for a sensitive lab or medical clean power connection is isolation. I would using a medical quality isolation transformer to create a separately derived power system to break the galvanic connection on all instruments using the isolated power of the sensitive application and bond The Protective Earth (Ground) Conductor on the isolation transformer to the lab's safety ground if needed. The PE is usually not 'dirty' in properly working circuits, it's the Neutral, current carrying connections that cause the bulk of noise and hum in sensitive circuits IRT the PE connection. The PE connection is a reference "common' that should have current only in a fault condition.
https://www.mouser.com/new/hammond-manufacturing/hammond-line-isolation-transformers/
https://www.hammfg.com/electronics/transformers/line/178-240-89-98mg

The exception to the PE being clean is when it shares an electrical conduit with noisy circuits such as light dimmers, or some electronic ballast devices. perhaps not very common, but certainly very real.

 

What is wattage of the AC/DC power supply? Can you make a photo of the converter? Is it posible to open it and make a photo? I suggest you to use an another AC/DC converter for experiment. Or an additional standard RFI filter with two-winding choke on the AC side

 

In addition to what "N" stated in post #7, after the isolation transformer the PE "ground SERVES NO PURPOSE AT ALL!! That is because the isolated portion of the system has no current path to deliver a shock to a person becoming connected to one of the isolated lines. Connecting the isolated circuit to any "ground" defeats the whole benefit of isolation.

 

@Pyrex

Here are some pictures of the power supply. It has a wattage of 60W. I also tried one with 120W which shows the same behavior.

Please note, that I soldered the connection from PE to secondary GND with the black wire here by myself. Normally this connection is given by the metal shielding as seen in the third picture.

@MisterBill2
Ok I get your point that you should not connect PE and secondary GND in any case. But somehow if I cut this connection, the 50Hz humming in my device is even worse. And a lot worse. Like by a factor of 10-100.

Thank you very much for your help!

 

OK, that means that some place in the system there is "hum frequency" current flowing where it should not be present. Probably in what should be a shield or "common" side of a signal carrying cable. This is where the sorcery gets involved, because if the cause were obvious it would have been fixed already.
One technique is to attach a wire to your known "ground" point, and then, with the hum going strong, grounding different places that in theory should already be grounded. At some point the hum should stop, and now you know the area that needs attention, because somehow that point that should be grounded is not. This process may easily be boring and frustrating and seem crazy to many folks, and certainly it can be a pain.

 

Try use an isolation transformer 230V/230V , 100VA at least. Disconnect GND from the PE and note the difference.
Try to use additional EMI filter on the DC side. You may found it in an old TV or computer. Like this-

Emi Filter for sale | eBay

 

Try use an isolation transformer 230V/230V , 100VA at least. Disconnect GND from the PE and note the difference.
Try to use additional EMI filter on the DC side. You may found it in an old TV or computer. Like this-

Emi Filter for sale | eBay

An isolation transformer for 175 watts cost $35 ten years ago!! Not a cheap item to use for the experiment. Searching for the missing "ground connection" is a lot cheaper.