DIY chiller

[hillegom]

I don't think slowing the water down will make the unit more efficient. It would be better if you could attach the peltier device to the glass. Because heat will travel through glass faster than plastic.
Because this device is on the outside, why not try a thin layer of that thermal paste they use on the cpu of a computer? That would make the heat transfer more efficient, Just don't get any into the water.

[pirate]

I was affraid to put it on the glass. I just see big cracks startng in my mind. It has a little bit of that grease stuff on it but not much maybe I'll try more. I was thinking that if the water was moving slower then it would be in contact wih the cold plastic for longer witch would turn into colder water. Maybe for my winter project I'll make a canister type on, so I guess I'm gonna have to start looking for all the parts now. Any other ideas out there?

[DiverDude]

Look for a water cooler -the kind that takes 5 gal jugs. What you have right now simply does not have the thermal capacity to deal with the volume of water you have to cool.

You could make a down and dirty chiller by running a few loops of tubing through the water-holding portion of the water cooler. However, you need to be very careful because you need to ensure that the chiller doesn't over cool the tank either. Now you're into measuring temperature and feedback loops.

Frankly, I'd suggest you implement some simple short-term fixes and shop for a used chiller. Point a powerhead at the surface to get is disturbed. Raise your lights if possible and/or look at shortening your photoperiod a bit.