Wednesday, 30 March 2011

A few photos of the setup:

20 litre oil drum, grabbed from a skip behind a car servicing centre. Hose and fittings bought new, can be a bit pricey, but Western Australia's a bit ridiculous price wise since they got rid of their manufacturing base.

Luckily the windows at my mum's place are pretty badly fitted, so there was a gap big enough to get the hoses through.
Just above is the car port, so the bank is shaded throughout the day. This seems to be important.

The very ghetto air pump housing. Taken from a old vacuum cleaner, stuck in some tupperware. Worked fairly well minus the fact that whatever plastic that is just does not want to adhere to any kind of glue, at all. Hence the tape and string, and it still leaked a bit.
Close enough is good enough.

RESULTS! (of cooling system prototype)

Actually got all this last week, but most of it was being gathered on the day I was flying out of Australia, so was gathered while running round packing, and this is the first chance I've had to get it down properly.

I did two basic experiments.


Filled the 20 litre drum (30 cm tall, deepest point 1.1 m, shallowest 80 cm, in dry sand over sandstone) with hot water. This was in the form of 3 x 7 litres bought to boiling on the stove and poured into the tank through the connector hose. Should've measured the water temp at start and end of filling (took a minute or two to get it through the funnel) but I'm going to assume it was about 90 C. The ground temp was around 24.5 C.


First 7 L volume at 9.40 am, second at 9.57, third at 10.12. Measured tank temp at 10.15 and was 77 C. This is a total of just on 5.75 mega joules.

Full data:

time water temp (C)

10:15 77
10:30 74
10:50 70
11:05 67
11:20 65
11:30 64
11:45 63
12:00 60
12:15 59
12:35 57
12:45 56
13:00 56
13:40 53
14:00 52
14:30 50
15:00 48
15:35 46
16:35 43
17:00 43
17:45 41
18:45 39
20:00 38
21:00 36
23:20 34

00:15 33
08:20 29
11:25 28
13:15 28
14:00 28
16:00 27
17:15 27
18:30 27
23:35 26

08:45 25


And 2:

More relevant to the actual cooling system, I plugged the whole thing together and ran it for about 8 hours. The setup was: air being sucked from the top of the house through about 9 meters of 18mm garden hose into the air pump housing, which was a vacuum cleaner pump and motor inside a tupperware container with the attached electronics.

This meant the air was passed over the motor itself, which added a lot of heat to it. Dumb for a proper installation, but good for fully testing the system.

Air flow was about 2 litres / second, in that it took 20 seconds to fill a 40 litre bag.

The air then passed through 3 m hose, bubbled through about 30 cm vertical of water, and exited into the house through another 3 meters hose.

Air In Air Out Air House
11:15 50 24 30
11:30 52 24.5 30
12:05 54 24.5 31
12:40 53.5 24 32
13:25 55.5 24 29
14:40 50 24.5 28
15:20 49 24.5 27
16:20 46 25 26
17:05 48 24.5 26
19:10 45.5 24.5 25

'Air in' being the temp coming out of the hose from the pump, ie just before it enters the tank. Pretty high due to the motor, the general air temp was about 34-36.

Air out is the air coming directly out of the tank.

Air house is the what came out into the house. At about 13:00 I wrapped the whole last length of hose in insulant and moved it out of the sun. I was surprised by how much heat the air picked up through only a couple meters of rubber hose, but there you go.

So I guess the two most striking things are that even with a fairly basic interface between the air and water (vague attempt at a diffuser, not really) the air dropped all it's heat, even at 55 C.

And that even with only 20 litres of water, which was only ever meant as a quick test, the water temp didn't really come up over the ground temp. True that 2 litres of air per second isn't much (I could've turned the motor up a lot more, but didn't want to 'splode it) but at an average of 50 C, the air was twice the temp above 24 that it would probably usually be running, so is therefor the same energy as 37 C at 4 litres per second, which is quick enough to replace 50 cubic metes of air in about 3.5 hours.

Which aint bad, frankly. It aint bad at all...