I’ve been thinking about this problem for a long time.

First an sd card or xbee to log information.

Now this is where it gets fun. I conjecture you could use an outside temp and correlate runtime for your heating and cooling. Then you could build a predictive model and forecast your heating and cooling.

Then when you go greener you can actually prove you had an affect by looking at your predctive model.

I would add reed switches on windows to turn off the AC if you open a window.

woot! had a programmable thermostat for $7.99 the other day.

The brewtroller (highly customized version of sanguino varient of arduino) with the fermtroller firmware installed can control either six cooling zones or three cooling + three heating. Primarily intended for controlling fermentation temps for home brewers but could certainly be adapted.

http://www.brewtroller.com/wiki/doku.php

Okay, here’s what I want, multiple coordinated boards with awesomeness. I want to use my home server a lot more than I do now. So, let’s have the home server running an A/C control program [simple enough]. A controller in each room with temperature, humidity, and ‘is someone in the room?’ controls over Ethernet [yes, wired, they aren't moving and you don't want them moved accidentally therefor breaking the server logic]. Each one reporting back every interval to the home server. An additional one outside under rain protection, but the same information being reported minus the person presence piece and maybe add in barometric pressure while we’re at it so you can detect raining or just ‘muggy’ conditions. Now, add some servos to the ductwork in A/C at specific points to control the Baffles to put the A/C where you need it, not just the whole house [should reduce overall usage of the A/C unit]. Then you add another controller/servo that could open a vent to the outside as a ‘source’ of fresh cold/hot air if you needed it instead of having the A/C kick on, you could have Mother Nature help you. How many times have you wondered why current A/C systems don’t do two critical things: 1. if you’re cooling the house and it’s colder outside than you want it inside, but it’s hot inside, just use the outside air [using only the fans, not the A/C unit which is your biggest power sucker in the house] and 2. Just because I have it on Cooler mode, doesn’t mean when we have a freakishly cold night in the Summer, that I want it 40F in my house or have to manually change it to Heat mode just for tonight [and remember to change it back to Cooler mode in the morning or come home to a stupid hot house], why can’t the A/C control handle that? A setting like “set for 76-79F in the Cooling mode, but don’t let it drop below 50F, okay?”

The air conditioner already had a thermostat knob. Was it broken?

Words cannot express my feelings when I see hacks like this…..

Cool – you could have your Arduino thermostat post to twitter and facebook and logged on google’s cloud computing architecture with this bit of code

http://nimbits.blogspot.com/2010/07/data-in-connect-arduino-to-nimbits-data.html

using the Nimbits data logger.

Ooops, forgot about the LEDs. They’d need to be shut off most of the time, too.

@GreenLee
Because the nob on that one is a simple bi metal switch, and the AC unit is rather dumb.

As for the question about what I would add.

1) An outdoor temp/humidity sensor
2) A indoor humidity sensor
3) A way to open and close a window
4) A mositure sensor for the window, so it can be closed if it is raining
5) Code to figure out the skin load of the room, and predict how much and when cooling will be needed.
6) Code to check to see if simply opening the window would reach the temp/humidity set point.
7) A small LCD and some buttons to change the set points.
8) An alarm circuit that could notify me while I’m away if the AC fails to turn on, or turns off early.
9) A few temp sensors through out the room for better temp measurements.

My AC units in my apartment wouldn’t need any physical hacks, as they have IR remotes already, remote codes would need to be figured out though. Figuring out how to open and close my sliding windows with out damaging them could be interesting(fun).

Like GreenLee, I fail to see the novelty in inventing something that already exists. What is the value-add here?

a phone line/ internet connection so you can call/ping/whatever to turn on the AC when you’re planning on going there

@Grazz256

That’s very true. I was going with my personal experience with “whole house” fans. They had a timer, which matches your note of taking time, but we would open a few windows and within a few minutes the whole house had the air changed and could drop the house temp very quickly (~10 minutes total).

As far as the air quality, I agree that it would need to be accounted for as best we can. That’s why I noted the “muggy” test that could be expanded by more than just temperature and humidity, but I wouldn’t go as far as checking parts per million quality [at least not in version 1.0 :D]

I guess adding an ‘air quality’ retrieval module from the weather.com or other online source before opening the exterior vents would be a good test that wouldn’t require an additional controller, just a web service consumption.

I would make it also controllable via IRC, or have it read my emails so when I send myself an email saying “turn on” it would. I don’t always have access to the internet, but can email with my phone.

Thanks for documenting your great project. Nice DIY spirit. If your windows are sliders and have wooden or unsealed jams you could install a motor in the wall and have it pull the window open or push it closed. The easiest way in this case I can think of to drive it would be by using a gear rack. I would suggest a prototype though so you don’t accidentally drive your real window through the other side of the jam during your beta testing lol.

If you have a steel frame window, you could just leave the windows all the way open, and then build a window insert with an actuator to open and close a partition. I’m thinking: Making a frame from aluminum C-channel stock, with the open sides facing out on the top and bottom, and facing inward on the sides. You could then use a curtain system with a roll on the top(think of a solid version of the roll-up cages they use to secure store entrances at night). Implement a worm gear to counter the ability to force the curtain back up, and you should be secure…

As a former energy consultant, I find this hack a bit pointless. My current electric bill (mid summer is $80. We live in a 3 bedroom duplex with all electric power. Because the AC unit rarely gets used our biggest 3 electrical loads are the water heater (electric), the dryer (electric) and the stove. If the AC unit has a working thermostat, just use it. Most residential units here in CA don’t have an economizer, because the ducting for the unit isn’t in place for a split system. The cost savings would be minimal for a 3 tom system. We leave our AC/heat pump off most of the year to save energy. We can do this because of good thermal orientation and good window shading. Currently 91 outside / 82 inside. As far as the other controls, there are all available commercially at a price (timer control, themostat/ outside air economizer / motion sensing / zone control / window venting) Adding external controls to a window unit you can’t reach is productive, I just wouldn’t go over the top.

@Ben:

Are there commercial “window venting” units that can be used to retrofit an existing window, with no damage done to the existing wood framing/window framing? I see some really nice(and expensive) units that are meant for permanent installation, but don’t see any retrofit kits for renters, et al….

Feature suggestion (assumes addition of better display) – display of how much energy used (in $, kWH, lb carbon) – estimated from on time & wattage of the AC.

Display how much could save by turning the thermostat to slightly higher temperature.

When you adjust the temperature, give estimate of impact on cost, etc. (i.e. turning down the temperature by so much will cost you so much per day/turning temp up saves you so much per. day).

–
Another suggestion – after you have programmed the temperature profile, have it gradually raise the temperature (one day it is 72, a few days later 73, etc., gradually let you acclimate to warmer temperatures).

At the beginning of summer, 70 degrees feels really warm, a ways into summer it doesn’t feel so bad (and 80 or 90 feels hot). Take advantage of people’s adaptability to gradually shift towards less AC.

Connect an input to the home/business alarm system, when everyone has left the building, put the A/C into ‘hibernate’.