...
The exit has the same principle as the rear of a rocket engine - for a moment let's forget the fans and think only in passive air circulation. Since the hot air expands, it will hit the walls.
But those walls angle will make it go out, instead of returning.
Since the air exits, it will create a very low pressure in the middle, pushing the air from the back.
Oh, now I get it. I should do more testing
. Btw, I ordered some temperature sensors (two), only need something to heat in the middle section. Should be fun 
I'm thinking about using a
but I can't find the equivalent of a boxed fan here in germany, only round ones (I want a boxed one!)
What sensors have you ordered?
I had a lot of trouble that way. They were both expensive and hard to use.
I've been advised in MSP430 forum against the thermo-resistors type, and those such as DHT11 are too complex for my knowledge (and expensive for experiments).
As for heating, as I PM'd you, I'm thinking in heated water, for now.
But I've thought some months ago in this: http://www.iceinspace.com.au/63-292-0-0-1-0.html
BTW, I've found out yesterday's night that a 7ohm @ 10W resistor and a 12V@1A power source don't mate well, although they really produce heat. Too much of it!!
Read somewere once, many years ago, that if we powered a couple of pins on an old pentium 100 processor, and shorted some other, it would heat up stable, but can't find that anymore. Just this: http://www.instructables.com/id/USB-Heater-or-How-to-Upgrade-Your-Coffee-Cup/ which as nothing to do with...
