1
\$\begingroup\$

Application

  • 20 gallon aquarium cooler
  • Ambient temp: ~22C
  • Tank target temp: ~17C
  • Tank houses an axolotl and needs to be kept at cooler temperatures then currently in the basement.

Current Peltier Set Up:

(Top to Bottom):

  1. 80x80x38mm fan running at 5700 RPMs and 76CFM
  2. 80x80x20mm copper fin heatsink (0.5mm fin thickness and 40 fins with a 3.5mm bottom thickness)
  3. 2-TEC1-12706 hot side towards heatsink, cold side down towards water block (Imax: 6.4A, Umax: 15.4V, Qmax: (dT=0) 63W, dTmax=68C)
  4. 40x80x12mm water block centered under the heatsink (surrounded on the sides with 20mm styrofoam and 10mm styrofoam at the back)
  5. ~26mm thick styrofoam
  6. Wood base
  • All power is supplied by an AC/DC converter (12V 20A 240W)
  • Power to the system is managed by a W1209 Temperature Control Module (Relay)
  • Water flow is achieved by a 4L/min water pump (slowest I can find)

Issue and Solution Thoughts

The current cooler will only reach 18C and that is with it having a long continuous run time. My hypothesis is that the water does not have enough time in the water block for good thermal exchange. That seems to me that we either:

  1. Need to increase the dT (between the water and block) [larger Peltier module]
  2. Keep the water in contact with the block for a longer period of time. (Slower circulation pump or longer block.)

Proposed Solution (new Peltier set up):

(Top to Bottom):

  1. (qty 1) 80x80x38mm fan running at 5700 RPMs and 76CFM, (qty 2) 80x80x25mm Fan running at 4500 RPMs and 59CFM
  2. (qty 1) 80x80x20mm copper fin heatsink (0.5mm fin thickness and 40 fins) with a 3.5mm bottom thickness, (qty 2) 80x80x27 aluminum fin heatsink (0.8mm fin thickness and 26 fins with a 6mm bottom thickness)
  3. 5-TEC1-12715 hot side towards heatsink, cold side down to block (Imax: 15.6A, Umax: 15.4V, Qmax: (dT=0) 150W, dTmax=68C)
  4. 40x200x12mm water block centered under the heatsinks (surrounded on the sides and back with 20mm styrofoam)
  5. ~26mm thick styrofoam
  6. Wood base

I can provide a CAD sketch or pictures of the current cooler if that helps. I have tried to use the Peltier cooling formulas but my experience is minimal with those calculations so I came here hoping for some advice/direction.

Will the proposed solution (higher capacity coolers, longer waterblock and heatsinks) solve my problem or is there another variable I am not accounting for?

UPDATE
I will order a 10k thermistor (https://www.adafruit.com/product/4890#description) this weekend and install between the plates when I strip it apart. I am seeing about a 1.5-2C drop from inbound water temp entering the block and the output to the tank. (Yes, I do expect some thermal loss for the ~1m travel, it will be insulated once I have a steady set up).

Forgive my weakness in thermodynamics but why do you want the water going through the block quicker rather than slowly? I realize that Al has about a 5x lower specific heat than water but will it really transfer the heat and hit an equilibrium that quickly? I have the option to go from 4L/min to a 5L/min, 8L/min, or 12L/min pump, I don't know if I can diffuse anything over 8L/min currently. (pun intended).

\$\endgroup\$
5
  • \$\begingroup\$ P.S. I am using Arctic MX-4 Thermal paste at the hot and cold junctions of the Peltier Cooler. \$\endgroup\$
    – Alfiero
    Commented Jun 15, 2022 at 17:24
  • \$\begingroup\$ How hot is the hot side of the peltier getting? You can only achieve so much dT between the hot and cold sides and the efficiency gets terrible. If you have more than 20 C dT across the peltier, I would focus on cooling the hot side better. \$\endgroup\$ Commented Jun 15, 2022 at 18:42
  • \$\begingroup\$ If you had two independent (as far as possible) cooling systems like the current one, then they would be able to keep the temperature at the desired value and if one system failed then it wouldn't get higher than 18°C. \$\endgroup\$ Commented Jun 15, 2022 at 19:15
  • 1
    \$\begingroup\$ 20 Gallon is a lot. Cooling that much water needs quite a bit of power. The peltier might not be up to it even when properly heatsinked. A compressor cooler is loads more efficient and not necessarily loud (think fridge) \$\endgroup\$
    – tobalt
    Commented Jun 15, 2022 at 20:09
  • \$\begingroup\$ Measure the temperature of the cold block. If it's significantly below its water outlet temp, then your idea is correct, water is flowing through it too fast. If it's like 17 or 18C, not so much. (You can easily restrict the water flow rate with a tap, and an overflow valve if the pump doesn't like working into high pressure) \$\endgroup\$
    – user16324
    Commented Jun 16, 2022 at 13:56

1 Answer 1

1
\$\begingroup\$

The circulation speed vs. heat removal rate has an asymmetric knee.

  • With circulation too slow, the tank will heat faster than the slowly flowing very cold water can absorb. Go slow enough, and eventually the water on the cold side of the Peltier will even freeze up!

  • With circulation much faster than necessary, the pumping losses will be heating up the water excessively. The pumping losses are always there - after all, the entire power that goes into the pump ends up as heat in the water!

    Making the pumping efficient is desirable: use large tubing and minimize flow resistance by using larger heat exchanger blocks on the water side of the Peltier.

The slowest pumping speed you chose may be too low already: try with a faster pump and see if things improve. Measure!

If the Peltier is able to remove the heat generally faster than the water is absorbing it from the environment, then you'll still initially have outlet temperature higher than the target of 17C - much higher. But as the water slowly cools down, and the inlet temperature goes below 22C, the outlet temperature will be coming down as well - at a bit lower rate the inlet temperature. That's because heat removal rate will drop as the water gets cooler: that's due to \$\Delta T\$ going down

So, making the water stay longer in contact with the Peltier will not necessarily improve anything. Minimum pumping rate is desirable to cut pumping losses, but it almost always reduces the heat removal rate. So pump as fast as needed to get sufficient heat removal, but no faster. You should measure heat removal rate for various pumping rates, and choose a pumping rate that provides highest removal rate per measurements. And then decide whether to add another Peltier module to speed down the cooling and as a backup.

\$\endgroup\$
3
  • \$\begingroup\$ @tobalt It is a good amount of water to cool but only needed for a few months during the summer. The cooler had dropped the temp from 22C to 18C overnight and can keep it there as long as it runs about an hour with 30-60min breaks \$\endgroup\$
    – Alfiero
    Commented Jun 17, 2022 at 1:23
  • \$\begingroup\$ This makes a lot of sense, thank you. I need to pull more data, when I was first prototyping this set up, I had input temp ~25C and post cooler temp around 18C. I do not have an easy way to measure the temp on the cold and warm side of the Peltier module as I dont have a flat temperature sensor to place between the junction of Heatsink/peltier and peltier/water block. Any other idea/method? I will gather input and output temps tomorrow and post them here. I do have a few other small pumps lying around with diff l/min to test as well. @ Kuba \$\endgroup\$
    – Alfiero
    Commented Jun 17, 2022 at 1:33
  • \$\begingroup\$ One more question regarding your 2nd to last paragraph. I understand the part of the rate of cooling will decrease as your delta T goes down. The part I am a little confused on is if the peltier unit is removing heat from the water block quicker than the water in the water block is absorbing it from the environment (are we talking heat gain from the pump/tank/tubing or are we talking about peltier cooler cooling the block quicker than the water can offload the heat?). Sorry if that seems like a dumb question but I'm trying to follow the enthalpy path and getting hungup. TY \$\endgroup\$
    – Alfiero
    Commented Jun 17, 2022 at 1:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.