Watter chillers

Community Laser Talk
21

Another area you may explore is a Peltier chiller. I built a DIY inline water chiller for a CW-300 based on devices found on Amazon for my 40W OMG. When running the laser for extended periods (say 6 hrs) it manages to keep the max temperature under 25C. There are no concerns regarding refrigerants and the devices are really pretty inexpensive. Further, if needed more than one unit could be installed with little effort. A couple of examples on Amazon: 12V 240W Thermoelectric Cooler Peltier Cooling System, 6-Chip Refrigeration Cooling Cold Plate Cooler, Water Chiller TEC or Peltier Thermoelectric Peltier Cooler Semiconductor Refrigeration with Easy Installation Quick Cool Down Water Chiller for Small Space Plate Cooling (4 Chip 6 Chip Optional)(6 core)

I’ve been pretty happy with the performance. One imporant note: My setup is in my basement where the ambient temp is a very stable 18 - 19C.
I was going to say there are no moving parts until I remembered the 12V PC Fans…

22

Those seem intended to cool the air, with water carrying off the waste heat. Running the elements “backwards” will reduce their already low efficiency, because the air heat sink isn’t designed to dissipate the total operating power.

If the room temperature is below 20 °C, then an ordinary (and much larger) water-to-air radiator will provide equivalent cooling without the 420 W (!) dissipated by the Peltier elements. If you already have one of those, how much did the laser water temperature rise without the Peltier cooler inline?

Not having any refrigerants is a definite plus, but it’s not clear Peltier elements add much cooling capacity for their tremendous power requirement.

23

Peltier devices heat or cool depending on polarity; from Wikipedia on Thermoelctric Coolers use the Peltier Effect. I haven’t looked in some time, but there used to be picnic coolers using Peltier devices plugged into 12V auto accessory plugs. Basically they transfer heat from one side to the other, thus can be either a heater or a cooler. The one I built, and that are referenced on Amazon chill the water running through the heat exchangers and the fans cool off the ribbed aluminum heat sinks. Really, very handy devices when heating cooling requirements are within their capablilty.
Photo is of my DIY setup.
IMG_6336

24

That’s serious-looking!

It’s been a while since I did anything with Peltier modules, which was a piddly thing holding MOSFETs at constant temperatures during tests:

The 10% efficiency figure seems about right, although surely things have improved since then. If the Amazon modules run at 20%, then you’d get 80 W of cooling from a 400 W input, although a larger temperature difference reduces the efficiency.

Laser tubes seem to run at 10 to 20% efficiency. At 15%, a 40 W laser at full throttle dumps 250 W into the cooling water, so a 20% Peltier would burn over a kilowatt to extract the heat.

You have a great test setup! Measure the inlet and outlet water temperatures, multiply the difference by the flow rate, and you can figure the cooling power. Divide by the Peltier supply’s input power and you’ve got the efficiency, which would be a really interesting number! :hot_face:

25

Your project can certainly provide some sort of baseline.
Testing my system sounds like an interesting experiment. I can certainly get a pair of thermocouples added to the system to look at inlet/outlet temps. Given how I set it up and knowing over time there is a gradual increase in cooant temp, it’s already clear that the system only mitigates the increasing temperature rather than halt it. It might also be interesting to see if the thermal paste I carefully squeezed very firmly between the Peltier plate and the aluminum thermal blocks actually helps, or maybe even hinders. Only thing stopping me is the 24hr day, most of which is already committed to other time-consuming projects/hobbies. :slight_smile:

26

I applaud anyone willing to build their own chiller unit - but be warned, careful thermal control of the water is needed. The large the range of your water temperature control ± your set point, the wider a power spread you will see with your laser, as colder water temps lead to higher power output.

27

I already have a crappy dual thermocouple meter reporting from the inlet and outlet fittings. If I can find the flowmeter in my pile o’ stuff, perhaps I can calibrate it and measure the results of the icemaker that should arrive in a few days.

So many projects, so little time! :grin:

28

++10
For me, it’s a constant battle with my ADHD, staying on task and actually completing a project!
One of my major time challenges is with the ceramics I make. So, of course I wandered down a rabbit-hole this morning exploring the idea of building my own 3D ceramic printer! :rofl: Geez…

29

These aren’t that hard to refill. If it does not have a port, you can get a bullet tap to add refrigerant. You have to go by weight and low side pressure, there won’t be a high side port. Now if this was a chiller that used R134a, that is simple and cheap to get from the auto parts store.

The CW5200, however, uses R410a, which can’t be sold OTC to unlicensed people but there are plenty of people on Facebook Marketplace who will sell you some. Difficulty is it’s pretty expensive by lb, and most commonly in 25lb bottles. You don’t need that much, so find a guy with a proportionately priced 90% empty bottle. And borrow a gauge set for R410a

If the pressure never went to 0, you can probably get by without pumping it down. Just add more refrigerant on top of what’s there.

If you didn’t see any oil leaking from somewhere, you don’t need to recharge the oil, it’s all still there. Which is good, that’s a different purchase and total guesswork on how much you might need to add, with no feedback if you’re adding too much or too little. These take several oz of oil so a leak capable of draining even 1/3rd of that would be a very obvious puddle of oil. No puddle, all the oil is still there, just add R410a

30

With regards to other chiller ideas- you really need to know the load wattage and calculate for BTU/hr. It is a pretty high number. An aquarium chiller is probably nowhere near enough.

It has to have the BTU/hr. Otherwise the temp will outrun the chiller on long full-power runs and typically just keeps rising every minute. The tube performance will fall off with higher temps.

“Can I run it through a radiator first on the return path to bring it to ambient, then add a small chiller after that?” Well, the prob is, if you have a 25C room and need to run the tube at 20C, the only way this works is the tube has to have >25C water coming out of the tube so the radiator can work. We already said we wanted the tube at 20C but this does not allow us to do that.

“Could I run the water really slow so the chiller gets it 20C in but gets >25C on the tube outlet?” That would defeat the point of chilling. The bore is now too warm on the exit side, AND has a temp gradient down its length (bad).

I’m not sure how many watts of heat a DC-excited glass 100W tube makes. You could measure the wall current for the power supply in actual constant full-power test, deduct 100W for the actual tube output, and the rest has to be lost heat- but, SOME will be in the power supply and is not a cooling load. Maybe 15%? 25%?

If you knew tube voltage and current, you could calc it, but you would have to measure the actual (not datasheet) tube voltage directly which requires a special meter and it’s kinda dangerous.

Anyhow, spitballing, maybe a 100W tube is 10% efficient, takes in 1KW. 900W needs to be removed. 1 kW = 3412.142 BTU/hr, so you need 3070 BTU/hr. A 1/10hp aquarium chiller is 1000 BTU/hr, it will lose- and, again, once it is in a cooling deficit, the way refrigeration works, it doesn’t transfer heat much faster from a hot “cold” side, it will keep rising. It could go to like 40C.

The coolant should not be within 10 deg of the wet bulb temperature in the room. This guarantees condensation will not occur, which can damage the optics.

Peltiers would be tough. The performance drops off rapidly with delta-T from the two faces and running anywhere near their rating:

perf_vs_current
perf_vs_current625×625 65.6 KB

Note that 20C water and 25C ambient is not the final delta-T. The water block has to a bit cooler than the water for heat to transfer. The drop on the hot side, to air, is much higher. If this transferred EXCEPTIONALLY well, that might be a 15C-20C delta-T total (water 5C below ambient, plus the two combined thermal resistances of the sides). If you ran the Peltier at the rated power, your COP is only 0.5 (need 1800W of electrical power to the Peltier and have to reject 2700W). Well, that’s gone into the weeds, this is SO MUCH heat and… well, no practical size of heatsink is going to give a delta-T of 15C at 2700W.

You’d need an array of Peltiers. If we shoot for a COP of 1.75, with REALLY good sinks, they’d have to be run at 20% of Imax. You’d need 8.57x “60W” 40mm/40mm Peltiers run at 20% of max current.

At this point 1400W is being dumped out of those sinks into the air, basically the output of a hair dryer. But these numbers ONLY work if the delta-T is really, really low. Like the output air can only be a few deg C over the room. That would require a massive CFM of airflow across the sinks to keep the exhaust flow minimally above ambient.

And that’s unrealistic, actually. This is going to have a higher delta-T, like 25C. Shoot for a COP of 0.75, 20x “60W” 40mmx40mm, 2100W of heat through the heatsinks. They’d have to be really good sinks.

31

There is a prob adapting dehumidifiers etc- they do not have a bypass valve like a chiller does. This means that they need several minutes to bleed off pressure from the high side or else they will stall upon a short-cycle restart. Even meeting that nominal waiting period, frequent restarts will likely shorten the compressor’s life. So, you want to calc the thermal mass of water needed to restrict water temp change to whatever deg C you want to live with. The water temp will rise while the compressor is cycling, and this affects tube performance- likely you would see “banding” on rasters where the tube output decreases as the water warms over a few minutes, then the compressor starts and the next raster lines will start burning deeper. In a horizontal band.

32

Hi, We have a ThunderLaser Mars90W. Built in China, purchased through local distributor here in NZ who also does the yearly service on our machine. Distributor knows his stuff and has even visited the factory in China so we get good support. Still on the original C02 tube. We have had for 7 years and no problems with chiller unit which came with Laser Machine. I know ThunderLaser have agents in the US. Maybe worth looking them up, just a thought.