Help with problem at the South Pole!

Status
Not open for further replies.
G

good john

New member
I am the electrical foreman at the Amundsen/Scott South Pole Station and am trying to solve a problem with one of the scientific experiments here. It involves two computers that will randomly freeze up with no rhyme or reason. Both computers are on UPS power which is supplied via an isolation XFMR. THD is 3.7% topps and I have recorded voltage surges of up to 180v RMS. I added a bonding jumper from the rack back to the main area grounding point which did not help. Grounding is a big problem here seeing as we are on over 9,000 feet of ice. Power for the building is three phase,480V, and a #500MCM single conductor is used to bond this building with the powerplant about a quarter mile away. I have been told that people have touched the equipment rack and received a static shock which had the same effect on the computers but most of the time no one is around when the problem occurs. Some of the solutions I have been given range from too low humidity (which is next to zero here), harmonic distortion from the UPS systems,to the obvious grounding and static build up issues. I'm not trying to fix the problem, only understand what is going on and give the engineers that are here in the summer some useful information. Thanks!!!
 
Few Inputs...

Do both computers freeze at the same time?

What is the voltage rating of the computers and what is the voltage you are getting in the socket to which the computers are connected?

cheers.
 
You work at the South Pole and you wonder why the computers are freezing? It's cold!

I couldn't resist.

Why are you so sure that the computers freezing has anything to do with the power feeding them? I would think it is more likely that there is just a problem with these computers. If it was a problem with the distribution system, wouldn't ALL the computers behave the same way, and not just these two?
 
So are we invited to visit in the summer?

If you can put a meter on the feed to the computers it might give some additional information to the Engineer about what's going on.

Is voltage showing up between Neutral and Ground?

Does the voltage surge show up the same time of day, which happens to be when some test in another department is done every day?

Is the only ground for the building that 500MCM back to the power station? I think the resistance on this ground conductor is .0265 ohms @ 32 degrees F for a 1/4 mile. Maybe that has something to do with it.

I am thinking it might be some sort of a grounding issue. I don't know what appropriate means there is to ground a building sitting on a glacier. It might be similar a building built on sand. Grounding rods don't do much. Using the rebar in the building's footing works as a Ufer ground.
 
I would make sure my BONDING from the receptacles for the computers, to the panel is solid. I'm sure you've tried other computers on these stations?

what is your voltage reading between the nuetral and ground? Thinking 1.4V or there abouts is acceptable. If it's 28 to 30v, then they need to tied together back at the service.
 
I have a suspicion that this may be a software/IT problem and not related to the power supply.

However it is worth trying the following though.

Is supply voltage is correct for the UPS, for example if you have a UK type 230 volt supply, and a USA 208 volt UPS (OR VICE VERSA) then the voltage might be just within tolerance most of the time, but out of tolerance often enough that the UPS runs frequently on the battery which has worn out.

Presuming that you have a suitable test meter, and the knowledge to use it safely, try measuring the voltage at the UPS input.
Since you mention an isolating transformer, is it possible that this has the wrong tapping selected?

Is the UPS in good working order? some types have an automatic bypass that connectes the load direct to the incoming supply, in the event of a fault within the UPS.
This also happens if the UPS is overloaded.
Any voltage fluctuations or drop outs would then feed through to the computers.
If you have one, try a different or spare UPS

Do you have any alternative source of power of the correct voltage and load capacity? such as an emergency generator, a portable generator, an invertor worked from a vehicle battery etc. If so try powering the computers from this alternative source to see if any improvement results.

What is the computer used for? is it connected via any long copper cables to distant sensors, data loggers etc. Any such could introduce undesired voltages via the screens of data cables.
Whilst one would hope that the rules as regards grounding/earthing have been followed, so far as possible, I understand that grounding is quite a problem at the South Pole, since ice is a poor conductor.
"ground voltage" could be very different from one end of a site to the other, and if long data cables introduce "ground" voltage from a distance, then that affect the computer.
If this is supected to be the problem, it might be worth trying short range radio or fibre optic for data transfer.
 
My experience with the same situation with computers and the addition of a true UPS corrected all the voltage problems. I put voltage recording equipment on the input and output of the UPS. I remember seeing voltages dips at 95volts and spikes of hundreds of volts on the input. The output stayed rock solid at 120v. Buy a cheaper UPS and you're really getting nothing more than a battery backup which does nothing for filtering transient voltages.
 
I don't buy the "bad ground" theory.

I am betting either there is a problem with the UPS dropping offline for a cycle or two, or something wrong with the computers. Since it is both computers, I would be looking at the UPS.

Some UPS units have a setting where they come back online automatically after a fault. I might try disabling setting that so if it faults, it shuts down and you know what happened.

Grounding will not solve your static problem, but bonding might.
 
Get a Mac. :D

I would also lean more toward a software issue. I'm not the slightest conviced computers care anything about the grounding.

And I think computer power supplies do an excellent job of filtering the incoming power. Combined with a UPS, I think something obvious would have to be wrong with the power before it caused a computer problem.

Who is pointing toward power problems? Someone associated with the software these are running?

Steve
 
good john said:
...Both computers are on UPS power which is supplied via an isolation XFMR. THD is 3.7% topps and I have recorded voltage surges of up to 180v RMS.
Click to expand...

How isolated is the isolation transformer?
Bonding is your friend. If you have a truely isolated power system, we likely have found your problem.
 
Low humidity is a big problem in those environments and I agree that bonding will eliminate most problems. One point that I think is a good one was the instrument control lines that likely run out of this building and into the telescope. I am only guessing at which project these computers are for but lets assume they are for the new telescope. The wind down there can cause a hell of a static charge on the instruments and bring that back to the computers or related instruments inside. I am the formeman at a similar station in Greenland with similar problems and that was one problem that messed me up for a little while last year. The problems were intermittent and hard to trace because of that. Turned out to be a charge that would build up during big wind events and arc onto a pipe when it got big enough.
Sounds like they may have a good crew down there if you are asking questions like that. Interested in working at the other pole?
 
Status
Not open for further replies.
Top