Server room & humidity...

[slick101]

I had a service call regarding "high" humidity in a server room. I was sent to check out the RTU which is a brand new York 410A. Now the T-stat was set for 20*c and room temp was 20*c. I had a 15*f TD across my coil and there was a ton of condensate coming from my trap (which leads me to believe that not only is it cooling, but dehumidifying aswell.)

All in all, the RTU is working great. But apparently their instrumentation tech complained of high humidity. For all I know he could be overweight and ran to work. But I've suggested the use of a hydrometer to read the actual RH in the space. I'm not familiar with hydrometers, as I've never used one before..could some1 please explain the use of one? how long do I have to take a reading for etc etc.

Also, if the RH is too high and the T-stat is satisfied, then could I add a duct-heater into the system that could activate when the stat satisfies on temp, but not RH?

Thanks!! - SlicK

[teddy bear]

A dehumidifier is a better choice than a heating element. A dehumidifier removes moisture and adds heat with a 66% savings in energy. Quest is line of dehumidifiers made for this application. thermastor.com Regards TB

[powerup]

Great sense of humor slick, haven't laughed that hard in a while. Try using a sling psychrometer or check out one of the digital ones they have. Check out all the duct work. also room for any negative pressure. Don't assume great amounts of condensate means all ok.

[jayguy]

... and there was a ton of condensate coming from my trap (which leads me to believe that not only is it cooling, but dehumidifying aswell.)

be very careful on this job!!! you can get yourself into a whole bunch of hot water with server rooms if you are not familiar with the application (this exact application...not what ASHRAE or any other governing body generally says). you need to find out what the RH actually is, what the RH is supposed to be, and all of the other loads. generally speaking (you see, i used the word "generally"...here is where your spider man sense should be tingling...maybe you take my advice...maybe you don't ) server rooms do not have a high latent load (or any at all except for the IT tech or some small infiltration) so you should not have to have hardly any condensate at all. i would investigate the requirements and then "where is all of the humidity coming from?" i understand that you want to "fix" the problem but perhaps it is not wrong where you think it is. perhaps the room is not sealed well enough...perhaps the unit is oversized (this is where your thinking is going and you may be right).

good luck.

[slick101]

Ahh icic.. good call I'll discuss it with my foreman and bring these topics up

[superfittertech]

be very careful on this job!!! you can get yourself into a whole bunch of hot water with server rooms if you are not familiar with the application (this exact application...not what ASHRAE or any other governing body generally says). you need to find out what the RH actually is, what the RH is supposed to be, and all of the other loads. generally speaking (you see, i used the word "generally"...here is where your spider man sense should be tingling...maybe you take my advice...maybe you don't ) server rooms do not have a high latent load (or any at all except for the IT tech or some small infiltration) so you should not have to have hardly any condensate at all. i would investigate the requirements and then "where is all of the humidity coming from?" i understand that you want to "fix" the problem but perhaps it is not wrong where you think it is. perhaps the room is not sealed well enough...perhaps the unit is oversized (this is where your thinking is going and you may be right).

good luck.

what he said plus check if unit has an economizer or osa damper bringing in humid outside air

[bob_scheel]

Is the AC package designed for a computer room or just a standard commercial package? What RH% are they looking for? Usually its the big printers that cause them trouble when the RH is high. (the paper jams more)

Not sure if the computers these days care about RH%. With the death of big computers in these rooms in favor of small PC sized servers or remote data processing it is not uncommon to find the existing AC package way oversized. We recently removed a Liebert unit from a computer room that had two. Now works fine on one unit and rarely runs the second compressor in that one.

Check the walls for holes. The telephone / data cable guys love to make holes in the drywall big enough for the whole reel of wire.

BTW the Lieberts dehumidify by forcing the heat on via a humidistat. The AC runs to control the temp thus in the process lowering the humidity. Not very efficient but it works and is VERY common.

[JRINJAX]

For driving down humidity, the Delta across the evap coil needs to be closer to 19 instead of 15 [lower airflow].

You also need to make sure the room is well sealed above the ceiling due to the law of partial pressures creating a pressure imbalance across the walls to the room when you reduce the atmospheric mixtue by removing the water vapor from the air [inside the room]. The humidity outside the room is actually a pressure pushing towards the room.

Check for hydrostatic pressure too if it is a slab-on-the-ground foundation.

You might need a server room unit with "Active" humidity control.

[Airmechanical]

We recently removed a Liebert unit from a computer room that had two. Now works fine on one unit and rarely runs the second compressor in that one.

sounds like that was a bad decision

there goes your redundancy

i would of left that system there in case of the primary unit breaking down!



.

[tpa-fl]

What size server room are we talking about? And how much equipment's in there? Are we talking about the back of a file room that someone's thrown two computers in or racks upon racks of equipment? Are the floors and walls finished or raw concrete?

For many server rooms, I'm cranking the blowers about as fast as they can go because dehumidification is usually one of the last things I need in there. I even find myself upsizing the evap coils to provide more efficient cooling and reducing dehumidification capacity. Computers & electronics are a 100% sensible load, no latent with them whatsoever.

If you're getting "a ton of condensate", it's not coming from the computers. Look for infiltration & bad seals. Computers also don't need fresh air, clamp down the fresh air dampers. 20c/68F's not all that cold. With regards to humidity you need to measure it with a calibrated instrument/meter. Personally, server rooms are the one place I LOVE to have "high" (~50-65%) humidity. As long as it's not condensing, it keeps the static electricity down. Weird random problems* happen in server rooms and offices when there's static electricity around.

Also...does that system have a low ambient kit installed? If you're dealing with a real computer room, that air con's going to be running in the middle of December.


*True story. It was a newly-renovated office building. New walls, carpet, data/com and electrics. Existing RTUs. For some reason the brand new phone system kept dropping phone calls. The installers couldn't figure it out. They replaced every single piece of the system, from the phones all the way to the switch, even new trunk lines from the street. Still kept dropping calls. They were also having issues with computers randomly hanging.

I got called in to fix both after the installing contractor couldn't resolve it.

The first thing I noticed was that I got a nasty zap of static electricity within the first 5 minutes of being in their office. Why was this such a big clue? This is Florida, and there was a thunderstorm pissing down outside. There shouldn't be static electricity. Checked the deltaT on the RTUs servicing the office, all were within norms (18-23F drop). After getting zapped a few more times I ran over to the nearest drugstore and picked up a humidifier. Filled the tanks, cranked it on high and told them to keep refilling it...and give me a call in 2-3 days to see how things are. 4 days pass, no call. So I call them. Everything's working fine, no phone system or computer glitches.

BUT, that doesn't address the actual problem. What ended up happening was the RTUs servicing the common hallway were dead. Hallway had big picture windows that received plenty of sun in the afternoon. The heat load from the hallway was being picked up by the tenant's ACs, thus high sensible load, long runtimes = very low humidity. The phone installers changed out $30K+ of equipment, not to mention their time, and never figured out what happened. Sometimes humidity is a good thing.

[bob_scheel]

[QUOTE=Airmechanical;1984818]sounds like that was a bad decision

there goes your redundancy

i would of left that system there in case of the primary unit breaking down!

Liebert's have redundancy built in. They actually had 4 systems now 2.

Customer changed a main frame/mini computer room into a node for a remote system. Just a few PC servers there now and a couple of big printers. They wanted the space for storage. I don't argue just make suggestions <grin>

[slick101]

Good stuff. All of these posts help in some way. Long story short - there are a bunch of potential issues that need to be checked off before jumping to conclusions regarding this problem.

The server room is only about 300square feet, but has quite a few racks of electronic equipment running. The next time i'm in the area, I'll be poking around and will soon get back to you all.

Thanks - SlicK

[tpa-fl]

Good stuff. All of these posts help in some way. Long story short - there are a bunch of potential issues that need to be checked off before jumping to conclusions regarding this problem.

The server room is only about 300square feet, but has quite a few racks of electronic equipment running. The next time i'm in the area, I'll be poking around and will soon get back to you all.

Thanks - SlicK

Whip out your handy dandy ammeter and see how much power the equipment's pulling. Look at the size of the racks and how much room there is for future extra equipment. If it's a server room for a single company this isn't as much of an issue. If it's a data center type setup, expect that place to get filled with equipment until the racks are full. Overall, the trend with computers and datacom equipment is lower wattage, so don't go overboard with oversizing for future expansion. I like seeing inverter-based mini-splits in server rooms for this exact reason -- it runs perfectly with the current load and will ramp up if/when future equipment is added. Also reduces the temperature swing associated with HVAC cycling.

[JRINJAX]

Posted in wrong post

[Airmechanical]

bob_scheel;1985123Liebert's have redundancy built in.

what happens when the blower motor goes bad?



.

[bob_scheel]

[QUOTE=Airmechanical;1985903]what happens when the blower motor goes bad?

Bldg maint is there 24/7 and they are capable of changing it. They have a spare ready to go. Worst case they open the door to the adjacent office and prop it. There's not much heat load left in this "computer room" now that all of the big machines are gone. About 4 PC sized servers and a couple of big line printers. The management decided that it was costing too much to maintain/repair the older Liebert so out it came.

[shifferbrains]

Generally there aren’t many people that actually work in a server room at any given time so very little or no outside air makeup is required. The first thing to check is the actual RH in the room and at what level they need to keep it. I like to use a sling psychrometer.... It’s easy to use and doesn’t cost as much as the really good digital RH meters and it’s dead accurate if you use it right. Check for outside air dampers if there are any. Check for any openings where air from outside the room might be coming in. Also check to see if the AC unit has a humidifier that could be malfunctioning. Also ask them if they leave the door open.... Server room doors should be closed at all times. Most server room’s operate at several degrees below normal room temperature and it is very hard to maintain temperature and RH if the door is open.

In order to control humidity in a room with your cooling coil you also need to reheat the air after leaving the cooling coil or else you will drive the room temperature down during de-humidification. This will also tend to drive the RH up. Which means while you are running the cooling coil to remove moisture at the same time without reheat you are dropping the room temperature which is trying to raise the RH.

Most IT types have funny ideas about server room temperatures. They tend to want to over cool the room. I recently had a problem with high RH in a server room. When I arrived there to check it out I noticed the room felt colder than it normally did. I took the temperature in the room with my digital thermometer. It was 65 deg F. I checked out the AC unit and everything worked fine but it just wasn’t keeping up with the humidity load. I asked the Server Room Manager why he had the room temp set so low. He was new there and was standing in for the regular guy that was on vacation. He said at his last job they kept the server room’s as cold as possible. I told him I would like to raise the temp setting up 2 degrees to see if we can get this RH under control. He agreed. The next day I checked back and the room temp was 67 and the RH was at set point. Personally I don’t think they need those server rooms any colder than 70 degrees... Lol but try telling that to some IT types.

[sfarrell12]

I know this may be a simple way of looking at it, but I know from expierience I have seen porr designsd of server rooms in buildings with no vapor barriers, and single pane windows in the server room. Before you think it is the equipment make sure you have little to no infiltration of outside air. Just my .02.

[The Doctor]

I hope your excursion with the York goes better than some other applications with RTUs in a server room. For example, the Predator has had issues operating late fall/early winter with time delays related to outdoor air temp and minimum off cycles which were precipitated by low pressure (no fan cycling). That much said, it was the only time that the unit got behind, which in that case was temp, not humidity. But it can still "generate customer concern"
It was the exception and not the rule, but no fun while on call in the dark on a ladder with strong winds, new site. yada, yada, yada.

Is this RTU set up for low ambient? Is the outdoor air minimum at 0% ?
Don't ya just love it when people park a comfort cooling unit on a server room?

And isn't the latent load in a server room generally 10%/ sensible 90% ?

As for producing condensation, I guess I've not measured the difference in the volume of condensate produced here vs. that from comfort cooling, but if I were you, I would expect to continue to see condensate being produced any time you're cooling. Not sure why it's being said about "no condensate".

[workingman]

sounds like that was a bad decision

there goes your redundancy

i would of left that system there in case of the primary unit breaking down!



.

Doesn't that train of thought go against all IT Logic?? I havent met an IT person yet who can figure out that it only makes sense to have redundancy if it costs you big $$$$ if the servers go down.

I have an account that has 6 Lieberts running 100% cooling 24/7. If one goes down, the servers start going down in about 10 minutes. If the whole room goes down, it is costing them $2,000,000 (yes that is million) an hour.

They are putting in another Liebert only cause they are adding MORE servers. When I mentioned adding units for redundancy, they just looked at me like I was stupid......