Problem with "phantom" AC voltage on DC part

Forum 2005-2010 (read only) Hardware Troubleshooting
1

Dear Arduino forum friends,

Though I've been playing with Arduino from some time now, this is my first post in the forum. I usually find all the necessary information without writing here just reading you guys (you've done every possible combination! :-D), but this time I'm facing a problem that I'm not able to easily find its solution. Let me summarize it and let's see if anyone can help me.

I've developed for a friend's company a couple of devices to remotely control relays through ZigBee. The local device has an Arduino Duemilanove, a XBee shield with a XBee Pro Series 2 working as ZigBee coordinator, a Ethernet shield and a custom board with an AC/DC PSU (220VAC to 5VDC) and a custom circuit with two relays. The power supply unit, the relays and the circuit are build on a prototyping single sided strip board (something similar to this http://img-europe.electrocomponents.com/largeimages/R1004334-01.jpg). The remote device is the same without the Ethernet and has the XBee working as ZigBee Router. All is packed in a plastic box with an external 5dB antenna connected to the ZigBee PRO with an UF.L connector. Everything is controlled by a remote PHP application though web and works flawlessly 24/7! :smiley:

You can see in the following pictures what the boxes looks like

Local box:

Remote box:

Well? everything seemed perfect, but now comes the problem. My friend told me that when installing the remote box, having it branched, at one moment he touched the metallic part of the antenna (ground) and a metallic post and had an electric shock (not big, but significant). He measured with a voltimeter what he had there, and between the antenna's ground and an external “good” ground he had 92 VAC!! there should be only DC voltage on that part! :o Similar thing happened on the local box (~50 VAC)

I went to see my friend to measure things on the remote box and see that between ANY point in the DC part (arduino, relays circuit, ZigBee,?) and an external ground there was 92 VAC. The DC voltage was perfect, but there's this huge AC ripple (¿?) that we don't know where it comes.

We tried without the XBee connected and still had the same problem (then this has nothing to do with RF)

The power supply is from TRACO POWER (http://www.tracopower.com/fileadmin/medien/dokumente/pdf/datasheets/tmlm.pdf) and it says in the datasheet the the isolation is 3,000 VAC and the ripple less than 100mV.

The circuit designed is this one. The drawing doesn't include the antenna (connected to the XBee PRO) and the Ethernet shield (only present on the local box)

I don't have the boxes with me now (my friend is using them even with this ripple as they work perfectly) though I'm buying the same stuff to build another box and try to reproduce the error again and find a solution.

Any ideas where this problem might come from? :-? need extra ground? bad design? need better prototyping board? the PSU is maybe a SMPS and needs output filtering? anything else?

Any comment is welcomed!

2

I've had the same 'problem' with two different switch mode power supplies (SMPS), both of them were fine. I could feel a tingle in my finger as well. I was running a cheap PC based measurement device and some other thing powered by said power supply. The PC side is grounded properly, the SMPS is floating! It turned out that the "common ground" connection was not good (bad connector).

Here's some more info on it:

http://www.ce-mag.com/archive/06/ARG/grasso.htm

It seems the 'bridge capacitor' mentioned in the 2nd link can explain this.

I'm no expert in power supplies myself and was quite surprised about this fact. There's always the option to properly ground the (-) side of the power supply (wire to a cooper water pipe or whatever).

If the device is sitting on a shelf somewhere and usually nobody touches it, it wouldn't worry about it anymore.

3

Thanks for the quick reply. My doubt of connecting the (-) output of the PSU with the external ground is the potential isolation isues... the DC part is no more isolated of the AC part (common ground). In the links you've provided is suggested to put an isolation condensator between these two parts. I'll take a look to it...

Other comments/ideas are welcomed! :slight_smile:

4

The way I understand it, class Y capacitors are already in there. They are in the pF range, therefore the max current (at 50Hz!) is limited and not lethal to humans. You only feel the tingling.

+--------------------------o-------------+
| | |
| | |
| ------ |
| ------ + ---------+
o | | AC/DC |-------- (+)
220V ~ +----------+ | |
o | | | SMPS |-------- (-) "GND-2"

"GND-1" ------ +----------+
+--------------------------o-------------+

In my case I measured (and sensed with my finger) about 120V ~ between "GND-2" and what I call "real ground" (PC case). Therefore I think it can be assumed that "GND-1" and "GND-2" are connected. The capacitors also work as a 1:1 capacitive potential divider. Maybe there is also another "bridge-capacitor" in between.

5

This is common and nothing to worry about. The regulations allow for about 2mA leakage current like this. You will only measure this with a high impedance voltage measuring device like a DVM, a neon screwdriver or you fingers.

My doubt of connecting the (-) output of the PSU with the external ground is the potential isolation isues... the DC part is no more isolated of the AC part (common ground)

Electrically this make no sense, there is no need to think that grounding the DC side in any way couples the AC side to your DC circuits. Ground is ground full stop you don't get any voltage across it.

6

This is common and nothing to worry about. The regulations allow for about 2mA leakage current like this. You will only measure this with a high impedance voltage measuring device like a DVM, a neon screwdriver or you fingers.

Really? 2mA is the threshold of muscle spasm (well with DC it is).

I say 'ground your power-supplies unless they are double-insulated'. I thought that was the requirement for domestic appliances.

7

Well, the power supply specifications says it shouldn't have more than 0,25mA of leak current.

I'll ground everything on the DC part as soon as I build another board (and reproduce the same problem) and let you know if this AC voltage goes away. The power supply doesn't have any ground pin...

8