PDA

View Full Version : sizing window unit for home office



gobo
17-07-2009, 06:27 AM
i'm a software/hardware engineer who did some a/c work out of high school long ago. this one, while very basic, is driving me nuts. i have a 200 sq ft room used as office and electronics workshop. using the 30 btu/sq ft guide, i've estimated the load to be ~8900 btu. (unshaded southwest corner, small windows, minimal insulation, east texas.) the additional 2900 btu is desktop computer, server, etc. for seven years i ran a 12,000 btu ge unit that while oversized i had no issues with it. late last season the fan bearings gave out and i decided to replace the unit. at that time the stock on store shelves was pretty thin, but i found a deep discounted samsung 12k. the first thing i noticed was the short cycling this unit did. it was the end of the season, so i ignored it.

here we are half way through summer and this thing is still ferociously short cycling. of course i only see this short cycling when i'm at home at night. during the heat of the day i'm not real sure how it behaves, other than my server stays running.

so i'm down to two options. (1) run the unit until it beats itself up short cycling and be certain to not make the same mistake again. (2) grab a smaller unit while there are still some available.

anyone willing to share their thoughts?

if the best plan is to replace it, then what is the guide for sizing error? a 10k btu unit is 10% oversized and a 12k is 25% oversized. i suppose it's safe to say 25% is beyond acceptable. but i certainly don't want to go too small. when the workbench is fired up, there can be as much as 1200 watts dissipated.

btw, the sensible loads (electronic) are measured power. the current draw of the samsung 12k unit is within 5% of rating plate.

thanks.

The Viking
17-07-2009, 04:24 PM
Can you adjust the airflow / air direction to stop the short cycling?

Or, if you can live with some temperature fluctuation, locate the temperature sensor and either give it a higher mass (by submerging part of it in water or strapping a small metal bar to it) or insulate it (by wrapping it in tissue paper for example)

gobo
17-07-2009, 07:30 PM
air flow is one of the issues. the unit is mounted in an alcove surrounded by book shelves. not the best place for an a/c, but its the best window to block. (yeah i know, i asked for it...)

actually, i was thinking about moving the sensor. i've noticed that most of these electronic controlled units use an RTD right in front of the evaporator coil. i could extend that out into the room by a few feet and get it out of the blast. if the wire is too long, the calibration will be off, but i don't use the display to determine room temp anyway. if the values of the rtd are published in the service manual, and i can find a manual, i could compensate for the extra wire.

this project is really bugging me. i've put a tc probe right next to units rtd and when it short cycles i'm reading ~77 deg. i can't figure out if i'm stuck in a bad location (the old ge was a side discharge unit which was a benefit for its location), i've forgotten how to add, or all of this is a side effect of the software trying to make the unit more efficient.

thanks,
g

The Viking
17-07-2009, 07:34 PM
As long as you use good cable and proper connections (read: solder) there will not be any real life issues with extending the sensor cable.