An airconditioned greenhouse with a COP of 50.

Woke up to a house filled with smoke today. Luckily the house wasn't on fire. Unluckily, many other people's houses were on fire. Victoria has been gripped by fires again. It seems that bad fire years in SE.au go on a roughly 20 year cycle (presumably connected to things like the Southern Hemisphere Oscilation (el nino). We had big fires in 1939, 1967, 1983, so we're overdue by that rather unscientific estimate (we had some bad fires in Canberra in 2003 too).

On a happier and totally unrelated note, I've been pestering dnh to get me a car radiator for some time. The idea was to use a radiator to move heat to/from the greenhouse thermal store from/to the air.

So dnh, his father in sin, Andrew, and I went to a wrecking yard to find my a good quality radiator. After some searching we found one in a Toyota or Mazda (can't remember which now) which was still full of coolant. So we unscrewed it and bought it. The radiator had two fans attached, one with 4 fins (arranged in two skew pairs - odd) and one with 7. I presume the second fan is designed to also draw air through some extra heat exchanger, perhaps the airconditioner condenser.

The next challenge was to work out how to interface it to the existing plumbing. We stopped at Bunnings and scoured the shop to find a suitable fitting. It turned out that a 1 inch PE riser tube fitted inside the rubber hoses well enought that we felt we would be able to tighten the hose clamps enough to seal.

I assembled the unit at home and set it up on a trickle as an open circuit to wash out the remaining coolant (and gave the new fruit trees a good starting soak). whilst this was happening I set the radiator up at the door to my workshop and wired up the fan. Sure enough I was rewarded with a blast of cool air.

At this stage I only had my IR thermometer handy, which indicated that the radiator was cold (which I could tell by touching it :). The IR thermometer is not useful for measuring the air temperature, and in hot weather humans feel moving air as cold (due to sweat evaporating).

I decided I needed to set up a closed loop in the greenhouse and measure the temperature. A bit of fiddling and scrounging for suitable pipe fittings later and I had my prototype:


Using an alcohol thermometer I measured the air temperature to be 30C and the air coming out of the radiator to be 23C, so I am getting cooling. So I decided it was time to set up something more convenient and out of the way. I ended up hanging the radiator from the rafters and splicing into the creek water supply. As a nice freebie, I seem to have a fairly good HAF resulting from this arrangement saving me another fan.

Having the water coming out a separate line, and at a relatively low flow rate allowed me to measure the cooling power indirectly. At a rate of 500ml / 4s the water was heated from 22C to 26C which works out as 4C (1kcal/kg C) (kg/l) / 8s or about 2kW cooling (or about half a tonne for you old farts) when cooling 30C air to 23C giving a heat exchanger rate of about 280W/C. That's a lot of cooling power. And I'm getting that cooling power from about 60W electrical. I think this idea may be practical for house cooling.

Tomorrow is going to be stinking hot too, so I'll hopefully get a better idea of how effective this approach is in practice. I am only using one of the fans, presumably using both fans would double the cooling capacity. I probably would need to increase the water flow rate.

Coincidently dad just bought a stack of DS18S20 digital thermometers. I should set up a few to monitor things over a long period. I'm still not sure of the best algorithm for cooling the greenhouse. In hot weather I want to cool the air, in cold weather I want to heat the air (or use the underbed water lines). It is tempting to just leave the fan running all the time and let the water settle to the average temperature. On the other hand the fan is noisy and uses 85W, corresponding to $85 of electricity a year. It would be more efficient to heat the cutting benches directly in cold weather, and if the temperature difference between the tank and the air is too low then running the fan will do nothing.

Eventually I'm going to want to link all the heating and cooling logic together inside the house and GH (for example, to prevent the whole house fan from running when the airconditioner is running). This is probably a good spot to start investigating such a system. I'll probably use 1-wire bus on the grounds that it is well entrenched, robust, cheap and provides most of the input/output needs.




[æ]