Server Room

[rude]

Hi guys

I have a problem i guess with a server room. The customer said the room got a bit warm a few days ago when temps reached 36 degrees C.

The model number of the unit is Fujitsu AOT 18AZCL 4 Way Cassette.

Now the 1st issue i saw was suction pressure reading 200kpa on R22. Bit low i thought. 2nd issue was the condenser coil was quite badly block with dust. I did blow out the coil with nitrogen but i think ill have to remove it and wash it properly with water. So after i had done this i checked my air on and air off temps.

Indoor
Air On - 21c
Air off - 12.5c

Outdoor
Air On - 28.5c
Air Off - 36.1

I didnt want to go adding refrigerant because i didnt spot any leaks and the unit has previously run fine.

So any solutions or ideas if the system is running ok? or is there anything i can do to improve it.

My 2 thoughts would be to clean condenser and weigh in a new charge. Or could thisthe unit be to small for this particular room.

Thanx

[nike123]

For evaluation of proper function you need to measure indoor unit air in and out WB and DB temperatures at high fan speed and/or amount of condensate in some time period when compressor is working.
Then post your findings here.

[rude]

I knew you were going to say that. I didnt have a wetbulb tester with me.

[Thermatech]

If you read through archive threads about low humidity in server / computer rooms you will understand a bit better about the problems of using standard comfort cooling split systems in this type of application.

But for the slightly poor cooling performance you report suggest make sure the indoor coil is really clean & the fan is set for maximum speed to ensure maximum air volume through the coil.

For short of gas I tend to look at the compressor discharge pressure & temperature.
For a normally charged R22 split system you could expect arround 15 to 20 bar discharge pressure & arround 70 to 80 deg C discharge pipe temperature at the outside ambient temperatures you have.This will result in a discharge superheat of arround 25 to 35.
So if the discharge pressure is lower but the discharge temp is higher then the the system may be SOG & it would be worth recovering the charge to check weight.
For instance if the DSH is 50 then system is SOG but if DSH is less than 10 system is overcharged.

On some small split systems there is no discharge service connection point & the liquid service valve is after the expansion device so no good for measuring high side pressure.So then you are limited to looking at the discharge temperature.

My ball park rule of thum is, if the discharge temp is over 100deg C the system is SOG for sure & if its arround 90 deg C then suspect SOG but also based on other data like suction pressure & SH.
I am sure that some engineers here might disagree with my method but it works for me.

If you recover the full charge & find it is correct then you have to start looking for a partial blockage as possible cause.

[Sridhar1312]

air of indoor unit at 12.5 C should be in order to cool server room.However you mentioned as cassette which normally does not have air quantity enough to take care high sensible heat prevalent in the server room.

[rude]

Thanks guys yeah i think its working as best it can in not an ideal situation.

[chemi-cool]

Cassette is not made for server room cooling.

There is a special way of cooling the servers cabins via air tunnels under the floor .

There are many companies that make these servers air conditioners.

Look here for example: www.emiconac.it

[Tigermoth]

Hi Rude. Looking at your test data I think that you may have a evaporator icing problem. The difference between the "air on" temperature and the evaporating temperature on these split units is approximately 20C. So your test data indicates about +1C evaporating if the unit is behaving normally. Very close to creating ice build up on the evporator. If the thermostat has been set below 22C, the unit will most likely ice up, especially at night, or under cool outside conditions.

Checking superheat is a little more involved. First you need to determine the metering device type, fixed restriction type (cap tube or orifice) or varible restriction (TXV). A casette would probably be an orifice. These types of units don't have a relatively constant superheat (as a TXV would) because the flow rate across the orifice depends on the pressure differential across it. So, on a hot day the superheat is low (high pressure differential) and on a cool day the superheat is high (low pressure differential).

You really need to use a charging chart to get it right. The technique depends on the manufacture's requirement but I have used a generic chart that gets me very close when I don't have the manufacture's information.

Using the generic chart I just take the suction line temperature at the inlet to the condensing unit, along with my suction pressure. I calculate the "actual" operating superheat from these. Then I measure the ambient dry bulb temperature onto the condenser and the "on coil" dry bulb at the evaporator inlet filter. From these two temperature the chart will tell me what superheat I should have at the inlet of the condensing unit. If my "actual" is higher I add refrigerant, if lower I take some out, until I'm within 2C of the chart value. It works a treat even in cool 15C ambient installations.

I chart that I use it similar to the one found on Trane split units but I have converted it into an SI version for my location.

One other item that may be useful on determining problems with air conditioning condensers, is that they are normally designed for a 17C to 20C difference between the "air on" temperature and the condensing temperture at full load. So anything higher than this means you have problem. I don't base my decision on the condensing pressure reading, as it is low in low ambient temperature conditions. Just use the condenser TD it tells a nice story.

On the units without a high side pressure port (don't use the one on the liquid line if the liquid line has to be insulated or if it sweats it's not really the HP) I have fairly good luck with taking the temperature of a U bend on the condenser. Choose one that is in the centre of the bunch that are similar in temperature. The first section will be getting rid of the discharge gas and will be hot, while the end will usually be cooler because of the subcooling, between the two areas the refrigerant will be condensing and at the saturated temperature.

[DEVIL]

The problem is simple

1rst it's true a fujitsu cassette is not for server room, for a cheep alternative i would recommend a daikin FTK that works like a charm in cooling

but u have it installed so let's see the problem:
if the unit would have low air volume transfer (because of the dust on the evaporator ) u would have a lower output of temperature lower then 6 deg C, and that would make the unit ice up, but u should have around 9-10 deg C depending on unit age and so on.
The unit might also ice up because of low refrigerant and the expansion in the first part of the evaporator should be aggressive and lower then 0 deg C ,and that might be the case seeing that u have only 2 bar pressure so judging by this the probability is high that the unit is low on refrigerent


BTW Thermatech


For a normally charged R22 split system you could expect arround 15 to 20 bar discharge pressure & arround 70 to 80 deg C discharge pipe temperature at the outside ambient temperatures you have.

i have no clue what temperatures are that 70-80 deg C, no r22 system reaches that temperatures

[rude]

I did check to see if the indoor coil was icing up and i couldnt see no ice forming after an hour or so. It did look quite clean as well. I also thought if it were undercharged the leaving evap temp coming out of the 4 outlets would vary. But they were the same.

They do keep the unit set at 18c which isnt the best i spose but apparently after speaking to the other techs the units worked fine like this previously. I think ill just monitor it on my next maintenance, and maby even recover and weigh a fresh charge in.

If that fails just by looking at the unit it does look underisized compared to other units i have seen in server rooms.

[Makeit go Right]

You might also ask them if some extra servers have been added to the room in the last few weeks.

[DEVIL]

Adding or not extra load the system should work as it's designed to work.

The load just modifies the idle time of the unit or in the extreme case it can't keep the temperature.

Probably it's undercharged and not designed for the task at hand, probably not icing up because of the large amount of heat in the room, but discharge temperature should be lower