Corroded Computer Equipment (Twice)

[m1dql]

Hello,

I have a problem that hopefully someone with a little more expertise in chemistry / corrosion will be able to help me with.

I'll outline the situation:

We have a remote site that has a single server installed, the room it is in is not air conditioned, there is a UPS and monitor with it. The room is not used for much else, some bits of old equipment mainly.

In November last year, after the server had been in place for about 2 1/2 years, we started having problems. In the end (after a number of interventions) our maintenance supplier decided to replace the server on a temporary basis. On return to their workshops it was found to be severely rusty inside. We decided to purchase the loan unit. I was there when it was installed and there was no obvious signs of rust.

In the last few weeks problems have re-started. We have had the room checked for damp and this evening I moved the server into another part of the building. The new server now has significant rust. On top of this the surfaces feel greasy, if it was skin you would say it was clammy. I have a recently healed cut and this greasy coating has irritated the area.

Before I install another new server I would really like to know what is causing this. I've worked in IT for 15 years and admit to having no idea. Any help greatfully received.

Thanks

Steve

 

[blacksmith37]

Get the corrosion products analyzed; I expect the equipment would generate enough heat to prevent moisture condensation on itself.
Potential problem (30+ yr ago) very large new IBM mainframe; a small fire on another floor burned vinyl insulation releasing HCl. Although no smoke was seen on the floor with the mainframe, gold contacts continued to loose continuity. Despite serious cleaning, the mainframe had to be replaced a year later. Moral- if you find halogens , you have a problem.

 

[unclesyd]

To get rust as you described in the OP you have to have water or some chemical vapor present. I would suspect moisture from the normal ambient air temperature swings. Green and slimy usually indicated mold, mildew, or bacterial contamination. Depending on the species of same they could and or produce acids that will accelerate the formation of rust.
On way to attack this is to either air condition or dehumidify the room air. Depending on the room size small air conditioner could be used as a dehumidifier.
If possible another way to mitigate the rusting process is us a vapor phase inhibitor in the server rack. Z would contact someone like Cortec and discuss you problem with them. If you do call make sure you mention the green slime.

http://www.cortecvci.com/index2.php

Are there any things in or on the old equipment that could cause corrosion problems?

What is the magnitude of the temperature swings and relative humidity?

 

[racookpe1978]

Is it in an industrial location or warehouse?

If in an "office" building, is anything being kept nearby; or even next to a bathroom/sewer connection/janitorial closet that is releasing a chemical residue? Is the air stagnant, or is a vent fan drawing humidity or contamination in from outside the clean space?

 

[KENAT]

Any kind of pests likely to be present in the room? I'm guessing rat pee or similar may not be good for it.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[cloa]

What part of a computer server has steel? Aren't they made of stainless steel where you need a structural component?

 

[MikeHalloran]

Stainless steel is not common in computer construction.
Carbon steel is better for shielding and much cheaper.
The preferred screws are not stainless, either; they're carbon steel so you can pick them up with a magnet after you drop them between the daughterboard connectors, and not have to disassemble or invert the computer.

Punch a hole through the wall and install a $100 cheapo air conditioner.
Be sure to get one with a mechanical temperature control.
The zoomy solid state controls on the cheapest models don't remember they were 'on' after a power outage.


Mike Halloran
Pembroke Pines, FL, USA

 

[EdStainless]

I have installed computers on manufacturing floors where moisture and dust were killers. In the end I bought a bar sized refrigerator and put the computer inside. I used a hole saw to cut access holes for cables and then foamed them in afterward.
Don't set the frige cold, some parts don't like being too cool.

You can have the slime analyzed, but in the end it still is corrosive.
Syd on on the mark, there are lots of potential causes. Sealing the machine away is the easiest solution.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[tomwalz]

My first thought was HCl or similar. We had sealed acids stored in an outside room. Sealed HCl leaked vapors causing the kind of rust you describe. It takes very litte.

Muriatic for cleaning?

Were the floors once cleaned with muriatic?



Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[EdStainless]

burned equipment. a blown capacitor, cleaning solutions
lots of sources of corrosive fumes.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[bimr]

You do not note the climate or industrial exposure.

If you have corrosion occurring, the first approach would be to investigate the use of a dehumidifier. If you keep the humidity below 40%, then there is not enough moisture in the air to promote rusting.

You can install the equipment inside of a control enclosure supplied with conditioned air or nitrogen, or dehumidify the entire room.

http://www.bry-air.com/applications/corrosion_prevention/default.html

 

[Guest102023]

Concur with unclesyd, there must be something in the air.