Ground Loops and multiple power sources

[pmrid]

I'm going to see whether it is possible to eliminate ground loops in my mount/scope/CCD systems. But it's not an easy thing to do I gather. I was wondering what precautions or preventative measures others have taken.
In my EdgeHD rig, for example, I run the Titan/Gemini I on it's own 18VDC supply. Although the Gemini power board has a power-out connector, I don't use it because everything else on that rig expects 12 VDC and I expect the power out connector on the Gemini would be at the higher voltage - not stepped down to 12.
Then there is a big ATIK camera, and a filter wheel, plus 3 dew heaters. These all want 12VDC which I supply through a separate supply. There is also a JMI focuser which wants 12 volts and because I think the 12VDC supply is close to its limits, I also have a second 12 volts supply running this alone.
The cabling for all this equipment is a mixture of serial cables (Gemini, FW, focuser) and USB (CCD and Lodestar).

Peter

[rally]

Peter,

Are you having RS232 communications problems or data corruptions ?
Or what is the problem you are trying to resolve ?

Bit hard to offer advice if the problem isn't detailed.

I am guessing that its not actually ground loops, since most of your electronics probably isnt grounded.

A photo of your wiring might help

Here's a starting list of things to think about

Avoid unnecessarily long cable lengths, cut out extra power cords if they arent needed, avoid bundles with multiple loops.
A bundle of multilooped AC cable as I have seen on many rigs adjacent to a bundle multilooped data cable is asking for trouble as they may have been inductively coupled.
Try to get the shortest cable that will do the job in every case - no bundled loops.

Having a large number of independent fully floating power supplies might result in there being some differences in relative voltages between them.
That could upset detected voltage levels - RS232 was a plus 12v/minus 12v system which tended to be industrially robust whereas USB is just +5V - so there isnt as much room for floating voltage levels.

So you could try getting a single dual power supply that can supply all your needs at 18v and 12v.
Try to avoid running data and AC cables together in a single bundle to prevent cross talk/interference.
Crossing them is OK but running them along in parallel is not.

Only use good quality shielded cables and make sure the shielding is actually connected !
You could try to use RF suppressors on all your cables if that is the sort of problem you have.
Measure the relative differences between say the negatives of all your power supplies to see if there is anything to read.
Of course if they are all truly fully floating as I would expect with switchmodes, you could easily get spurious readings !

Are your power cables and the type of connections being used up to the job ? - nothing worse than having cables and connectors that due to their design, diameter, joint method, contact area or quality do not pass the required current and not only create a voltage drop for the device in use, but also cause the voltage to fluctuate on other supplies.

You could try measuring the voltage drop across the DC power cords from supply to device to see if there are any bad connections. In a perfect world it would read Zero, but in reality there will be some measurement, it just wants to be very low - say 0.1v or better.

If its more complicated then using an oscilloscope might help, but this is a specialist field.
Many microprocessor devices have been made and designed without sufficient RF suppression - usually achieved by shielding and capacitive suppression around the board.
If you have such a device, you might need to replace it.
You might also try deleting devices and or cabling and switching off power supplies until the problems stops - at least you might be able to identify the culprit

What sort of dew heater system do you use - some of these can cause problems.

Like most problems of this sort - its is most likely a complex combination of a whole bunch of insignificant possibly undetectable things - none of which on their own would be any cause for concern, but together or at certain times or in certain very specific circumstances they can cause havoc.

Good luck in your detective work.

Rally

[wasyoungonce]

Peter it's an issue that I have never really looked at.

Some users have issues, some don't, I don't. But the Gemini FAQs#3 is a good starting point. This has good advice and a link to another article on ground loops.

There have even been threads in IIS (http://www.iceinspace.com.au/forum/showthread.php?t=78973)about this. specifically with this (http://www.iceinspace.com.au/forum/showpost.php?p=774214&postcount=40)as a solution.

Usually, from my limited experience in this area, the ground loops issue is usually cause by one PSU being "not right" and using one PSU for all is a better "thing to do". Maybe use the above tests to find the culprit PSU.

edit:
I'll add...you shouldn't get a ground loop thru RS232 on Gemini if using a Gemini BT device (that takes pwr from Gemini).

[pmrid]

Thanks for the thoughtful responses Brendan and Rally. It is not a partiular problem I'm thinking about. It's a matter of caution. With the amount of lolly ties up in gear these days, it just seemed sensible to eliminate a possible source of malfunction or damage.
Peter.

[tlgerdes]

Like the others are saying, in your case I think you are chasing a ghost.

Ground loops occur when you have multiple ground points and differentials voltages between those points.

If you only have one power source ie one 240v circuit to your obs, then you would generally only have one ground point.

Once you have two or more power sources ie 240v + battery, or 2 x 240v circuits with different longs lengths of cable (ie greater than 50m difference), then you have different beast to deal with.