Turning on and off the air conditioner consumes more energy?

[Psinter]

The air conditioner is set up to put the temperature at 72F. Which I hate because that's cold, but whatever. The air conditioner turns on and works for about 6 minutes when it automatically turns off because the temperature reached 72F. It stays off for 10 minutes and then it turns on again. Repeating that cycle of turning on and off for about 12 hours a day.

Because the cycles are not exact it means that it turns on a little bit less than 36 times in a 12 hour period, but certainly more than 32 times in that period according to my observations.

Does turning on the air conditioner for 12 hours and have it work for 6 minutes to then have it turn off for 10 minutes before it turns on again for 6 more minutes consumes more energy than leaving it turned on for a longer timespan?

 

[SunThief]

You'll probably get a more technical answer from someone, but here are a few thoughts:

• When you note that the air conditioner repeatedly "turns off", I'm assuming you mean that it's just the compressor that goes off, while the fan still stays on. Aghh, this was wrong, I was thinking of my car. (Forgive me, it's morning here.)

• A running compressor uses more energy than just the fan portion.

• There is a temporary start-up surge when many devices/components turn on. But although the associated energy draw can be very high, surge times tend to be very short.

• I would think you would need an extremely high number of repeated turn-on cycles for the associated surge to amount to a significant energy impact.

As an aside, I'm pretty surprised that your system would cycle so frequently! I'm guessing that might suggest deficiencies in the room's insulation, etc. (??)

 

[CWatters]

I don't know about AC but for heating systems it depends if the heater can modulate it's output or if the burner can only be on or off.

In the UK gas/lpg fired heaters can modulate/vary their output but oil burning heaters can't. This means the only way an oil fired domestic heater can control it's output is to cycle the burner on and off with a varying mark/space ratio.

I suspect AC systems are also non-modulating. So some form of cycling is normal.

If during the design phase you get the sizing wrong that can cause short cycling which can be inefficient for an oil burner.

Need an AC expert to comment if your AC is doing that.

 

[CWatters]

Ps: Sometimes you can reduce short cycling by changing the hysterisis on the thermostat at the expense of accurate temperature control.

 

[russ_watters]

Does turning on the air conditioner for 12 hours and have it work for 6 minutes to then have it turn off for 10 minutes before it turns on again for 6 more minutes consumes more energy than leaving it turned on for a longer timespan?

No. There is almost no "warm up" time for an air conditioner (a few seconds).

As @CWatters suggests, a decent programmable thermostat will have a setting for maximum number of cycles per hour and time between cycling to reduce equipment wear. But 4 cycles per hour is not unreasonably short.

 

[Psinter]

I still can't get my head around it. Let me see if I can understand with some math.

The electric bill comes in kWh. Assuming everytime the AC's compressor turns on, 1kW is consumed for about 5 seconds. Then it drops down across the next 10 second timespan to about 500W and it keeps working like that for 6 minutes. After 10 minutes it turns on again, consuming another 1kW for 5 seconds and drops again to continuous 500W for 6 minutes. The cycle repeats for 12 hours. For my arithmetic I will ignore those 10 seconds of power drop.

Because the 1kW was consumed for 5 seconds only, it won't come in my electric bill as 1kWh, right? My doubts are sort of located on that initial surge.

Like if 5 seconds are $\frac{1}{720}$ hours and kWh is kilowatts multiplied by the time in hours, then the electric bill accounting for those 5 seconds of 1kW consumption should be $1kW * \frac{1}{720}h = \frac{1}{720}kWh$.

So if that initial surge were to happen, say 33 times in a 12 hour period, the electric bill accounting for those initial surges in the 12 hour period would be around $\frac{1}{720}kWh * 33 = \frac{11}{240} kWh$. Right? Whereas the working time which might be 6 minutes for 33 times in the spoken period, would be 0.5kW (1/10 hours) * 33 = 33/20 kWh. Which is 1.65 kWh. At the end it would be more because the first time it is turned on it runs for more than 6 minutes, but let's ignore that. I want the math of the constant cycles.

Was my math here right, or did I still manage to screw up such a simple calculation?

 

[mfb]

Right.
5 s of 1 kW are 5s*1kW=5000 J, while 6 minutes at 500 W are 360s*500W = 180,000 J. The initial surge is a small effect for the power consumption of one cycle.

11/240 kWh will cost something like 1 cent each day or less.

 

[russ_watters]

This math does not, of course, account for the extra energy spent over-cooling the space...or if the controls are able to increase the average, the loss of comfort of the temperature being more variable.

 

[Psinter]

This math does not, of course, account for the extra energy spent over-cooling the space...or if the controls are able to increase the average, the loss of comfort of the temperature being more variable.

But we don't care about that.

Right.
5 s of 1 kW are 5s*1kW=5000 J, while 6 minutes at 500 W are 360s*500W = 180,000 J. The initial surge is a small effect for the power consumption of one cycle.

11/240 kWh will cost something like 1 cent each day or less.

Thanks for confirming.