Custom Air Conditioner Control Unit

by Chad Hauris, Updated 10/1/05

Special Feature Articles

Retro Electronics Home Page

Click on the pictures for larger views.

I formerly worked for KOCV, 91.3 FM in Odessa, Texas as Operations Coordinator and was responsible for maintaining the studio equipment. In Dec. 2004, we had been having a lot of air conditioner problems. The control room of the station also has a transmitter room which must be continually cooled to remove heat generated by the transmitter. Even in December, when it is 30 degrees outside, the room will rise to 90 degrees if the A/C fails. The A/C had quit and the room had been unbearably hot for a couple weeks.

An air conditioning contractor was called in and he said the problem was due to the compressor staying on constantly and overheating. One compressor already had burned out because of this. He said that what was needed was a device that would force the compressor to shut off for a rest break of a decent duration every so often. Even with a thermostat in place, the on/off cycles were too frequent and the compressor was overheating. If the thermostat was turned down low enough that the cooling was effective, the compressor stayed on constantly. He could not find such a device commercially available but, since I am experienced in electronics, asked me if I could design such a device.

So, I went to the ever-useful World Wide Web to find some circuits. I knew a 555 timer I.C. could probably do this job but I needed some schematics to help put it all together. I found a circuit at that appeared that it could do the job with some modifications. I used the extended duty cycle astable circuit which allowed you to have different time durations for the "on" and "off" portions of the cycle.

I used the calculator portion of the web-page to help determine the component values for the timing components of the circuit, R1, R2, and C. I decided to use a 3.5 meg ohm pot for each, with a 1k ohm resistor in series with each pot to provide a minimum value of resistance in the circuit. I also used a 1000 mfd. capacitor for C.

With these values I was able to get time values between a 1-second long pulse up past 10 minutes. It probably can do longer but I have not tested it for longer than a 10 minute on, 10 minute off pulse.

To hook up the I.C. circuit to a relay, I used an NPN transistor salvaged from the video output board of an old TV we got at Texas Recycles Day. The circuit for the relay-driver is at It calls for a 2N3904 transistor which is a small transistor...I used the more robust video output transistor just to give an extra measure of reliablility. The input to the transistor circuit connects to the output of the 555 I.C. circuit.

The relay then was connected to a terminal strip which also had the power connections. An indicator lamp was also connected to one of the relay contacts to show when the system is in the "On" mode. Another of the relay contacts was connected to the compressor circuit of the air conditioner system. The pulse unit is powered by a 12-volt d.c. "wall wart" power supply, and a fuse link is in the positive power lead. Here is a picture of the circuit being tested on the bench.

Testing the air conditioner controller.

The device was constructed at the Retro Lab and then I took it in to the station for actual testing. I set the potentiometers for a duty cycle of 10 minutes on, 10 minutes off. This results in very effective cooling action of the room. The temperature in the room is now held to between 58 and 67 degrees. The room temperature during the "off" period of the compressor never rises to an objectionable amount, due to the fan continuning to circulate air. The compressor is able to rest and cool off thoroughly during this "off" portion so that when it is pulsed "on" it creates ice-cold air like we had not had in the room for a long while.

After seven days of service, operation of the unit was good. The only problems encountered were during the initial test phase of the pulse unit, the lamp was not coming on.This was traced to flaky contacts on the relay. I was able to find a good 12-volt relay at the station...I wired it in, and operation has been perfect since.

Inside the controller unit.

Inside view.

Completed air conditioner controller.

Completed unit.

UPDATE: Feb. 3, 2005: The controller unit is still going strong, in continuous operation for 3 months now. The relay contacts are in series with the thermostat switch contacts, so that the thermostat can control the system when the relay is in the "on" mode. We have reset the pulse durations to 40 minutes "on" and 10 minutes "off" on the advice of our A/C contractor.

Update: August 2, 2005. Unit is still working fine. Have had no problems with air conditioning system.

 Control unit installed at KOCV

Device installed and working at KOCV.