[en7jos]

Morning All,

I recent ordered one of these cheap aquarium chillers from China to test on my 11 litre nano shrimp tank:
https://www.lazada.sg/products/aqua...oling-machine-i258367369-s401911804.html?mp=1

It cost me $66 SGD (roughly $49 USD, £38 GBP) delivered, which I think is a bargain.... if it works!

Seller's spec's (which I take with a large pin of salt for now) are:

• Voltage: AC 110-240V/50-60Hz
• Power: 72W
• Fit For: 8mm(Inner) Hose
• Suitable For Water Flow: 1-3L/min
  [*]Suitable Lift: Lower Than 3m
• Size: (L)X(W)X(H) 11x11x19cm / 4.33"x4.33"x7.48"(appr.)

This seller didn't give a max. tank size, but others selling same unit have said 30-35 litres (and I've read reviews left by people with 45l plus tanks which say it doesn't do anything).

I have three of these little shrimp tanks sat next to my desk here in Singapore. The local climate is hot, hot or hot! So daytime room temp's with air-con on low are 27-29C, and at night with air-con off can rise to 30-32C. Tank temps sit at about 28-29C which I brought down to 27C with a small fan on each with temp controller. The fans work great, but evaporative water loss is a pain with such small temp's requiring daily top-ups of about 0.3-0.5 litres. (Sorry, I can't work in Fahrenheit)

First tank has my yellow Goldenback cherries (breeding!), and the second tank has my new orange Sunkist cherries (more babies!). So these seem quite happy in the higher temp. But in the third tank I have crystal reds and I've had a slow die-off from my initial 10 down to 3. All other parameters seem good, so I could only think that temp was too high. I could have increased the fan cooling, but water loss would be too high - hence I wanted to try a small chiller.

First tests in a bucket of water looked promising and confirmed that it actually worked!

I put 4 litres of warm (32C) tap water in a bucket and set it going. Temp was down to 22C after about an hour, would probably have gone lower still, but it had proved it worked / chilled / didn't leak so I switched off. Flow rate was pretty fast using the included pump through short lengths of 9mm ID tube with no flow restrictions. The water flowing back into the bucket was about 0.2C cooler than that going into chiller.

I used my multimeter and thermocouple to check the temp reading on the chiller. It seemed to initially over-read by about 1.4C, so I adjusted the temp compensation by -1C on teh controller and then the readings pretty well matched my reference as the water cooler below 29C.

So time to connect it to the shrimp tank! The tanks have a small inbuilt filter partition at the back with a tiny submersible pump at the bottom. My initial plan was to use a separate pump to circulate the water through the chiller, but I didn't have the right connections, U-bends and tubing to hook this up and I was getting an air leak into the out-take from the tank causing bubbles in the tube and a noisy external inline pump. So I gave up on that. I had a spare small submersible, but that was too big by a few mm to fit in the filter compartment (and I didn't want anything in the main tank that could suck up shrimp!). So in the end I connected the chiller inline with the existing filter pump. Water circulation speed is noticeably lower, but still acceptable. Actually it's probably much better flow for the shrimp and plants which as the original flow was a bit too strong. So this setup works well, less noise, less piping into tank, quieter and less cluttered.

You can see the two clear tubes going to/from the chiller and into the black partition at the back of the tank in the photo below:

I have set the temp at 26C at the moment, but will reduce to 25C after a day or two (just want to be sure it works reliably first!). The chiller can handle this no problem in my small (13 litre total, ~11 litre water) tank.

The chiller works well and I am pleased - especially given the bargain price (which included pump, power supplies for chiller and pump, some hose and hose clips).

The biggest problem is that the temp controller has only 1 degree steps for the set temp and hysteresis loop, even though it reads the tank temp to 0.1 degree. So you can only set the temp to 24, 25, 26 degrees etc, but that's fine. The annoying bit is the hysteresis loop which I'll try to explain: it controls how the chiller switches on and off. If you set the temp to 26C, and the chiller turned on at 26.1+, then off at 26.0, it would constantly be flicking on and off. This is really bad for compressor chillers (fridges) that the temp controller is designed to work with, but fine for solid state chiller like this except the noise of constantly switching on/off/on/off would be annoying.

So the controller has a hysteresis loop which you can set at 1, 2, 3, 4 degrees etc. The smallest setting is 1 deg which means that the chiller turns on when the water temp reaches set temp + hysteris = 26 +1 - 27C. It then cools the water down to the set temp (26C) before turning off. Then the water needs to warm back up 1 deg to 27C before the chiller switches on again. And so the tank temp is constantly cycling between the set temp (26C) and set temp + hysteris (27C).

For my tank, it seems to take about 1 hour to cool the 11 litres of water down from 27 to 26C, and an hour or two for it to warm back up to 27C again. So this gives a 2-3 hour temp cycle over that one degree range. Is that good or bad? I don't know! But I would have preferred the hysteresis loop to be controllable in 0.5 or even 0.1 increments so I could reduce that temp fluctuation.

What do you think please? Is a continuous cycle between 26-27 degrees over the course of 2-3 hours ok? It's a very gradual change both up and down, so I think so (but I may look to swap out the controller to one with finer resolution).

Power consumption is pretty much exactly the stated 72W when running, almost zero when off. The fans are pretty noisy when the chiller is running - loud fan hum similar to a noisy desk fan. I think I can probably quieten them down a fair bit by modifying the case as it seems to be the slots on the front (rather than the fans themselves) that create the noise. They look to be standard PC cooling fans so this should be an easy mod and upgrade to make - watch this space....

Here's the full gallery of my photos of the chiller.

In summary...

Good things about this cooler:

• Price
• Build quality (seems good enough, but a few case screws seem to have stripped threads from factory so don't tighten 100%)
• Looks (neat and tidying design)
• Small size
• Ease of use (controller manual was included)
• Works very well to chill my nano tanks
• Cooling is slow and gentle, no rapid temp shocks
• Temp reading seems accurate (after setting -1C calibration in controller)
• Long cable to temp sensor so chiller could easily be positioned away from tank

Bad points:

• Noise - the fans are pretty noisy!
• Temp can only be set in whole degree increments
• Hysteresis loop can only be set in whole degree increments.
• Temp calibration only in 1 degree increments (but seems accurate enough after dialing in -1C calibration)
• Probably not going to have much effect in larger tanks (I guess anything much above 25-30 litres)

My plan now is to run it for a few days at 26C to see how it goes, then lower temp to 25C if all seems well. I'll open it up shortly to see what's inside and will almost certainly be modifying the fans and case to reduce noise. I'll also see if I can get a better temp controller that will give me better (to 0.1 rather than 1 degree) control over set temp and hysteresis loop.

 

[en7jos]

Ok, so I've timed the cooling and warm up cycles and it takes 1hr 15 min (75 min) for the chiller to cool the tank water down from 27 to 26C, and the same 1hr 15min for the tank to warm back up naturally to 27C (at which point the chiller kicks back in and the cycle starts all over). Temperature changes in both directions are slow and gradual over this time. I guess this will vary slightly throughout the day / night depending upon room temp; maybe teh warm up cycle will be a bit quicker when the room temp is a couple of degrees higher at night without the aircon on.

Question please for experienced shrimp keepers:
Is a one degree temperature fluctuation over the course of 75 mins ok for crystal reds?

Just wondering whether I need to swap the controller so I can keep the temp 26.0 - 26.5C (for example), rather than it varying 26-27C? Or is that 1 deg C variation already well within normal limits?

The remaining shrimp in this tank already seem much happier at their new lower temp - they seem much more active and out-and-about foraging, whereas they hid away much more previously. So I think this little chiller is a winner (at least once I've quietened the fan noise down a bit!).

 

[Oughtsix]

Great review! Thank you! Unfortunately I do not keep shrimp but I would expect to see a +-1 degree C over an hour period in the wild, wouldn't you??? Just my guess but it sounds good to me. What where the daily temperature fluctuations in the tank prior to the cooler? (I just reread your post, going from +-3 degrees to +- 1 degree seems like a huge improvement!)

Is the temperature controller integrated into the chiller or separate? If the temp controller is a separate unit you could probably replace it with one that has a finer adjustment for not too much $$. My shrimp ignorant gut feeling is that you do not need to though.

P.S. I use a DMM with a thermo-coupler for getting REAL water temperatures also. Aquarium thermometers seem worthless in comparison!

 

[en7jos]

Cheers @Oughtsix. I have a few mod's planned for the chiller over the coming week or two. First is to replace the noisy fans with some better quality ones to try to reduce the noise. Might also need to cut out a better vent window at the front as I'm sure the small slots aren't helping with cooling and are actually creating additional wind noise.

I'm thinking the 1 degree temp variation seems reasonable and well within what natural conditions might be, especially given that it is a very slow up and down cycle. But I still might see if I can replace the integrated controller at some point to get control down to 0.5C temp fluctuations.

Monday morning's task is to disassemble, upgrade the fans and see how airflow through teh cooler can be improved and quietened. Will post some more photos of the internals later in the week...

 

[monkeyruler90]

Very interesting read.
Whenever I had issues with my tank overheating I also found a good solution was to increase the tank volume. Going up to a 40 gallon instead of 10 gallon made it more stable for the daily shifts.
Do you use RO water and change water frequently? I wonder if it is helpful to build a sump like system that may be able to more ballast the temp swings without the chiller. It would also help with the evaporative losses since it would all take place in the sump

 

[en7jos]

Hi @monkeyruler90

Tank volume (or rather lack of water volume) is almost certainly part of the issue here - its a 13 litre tank containing about 11 litres of H2O! So on the small side to say the least. Bigger tanks would of course be nice in some ways, but then they wouldn't be my three little nano shrimp tanks next to my office desk, and a sump not practical for the same reason. On the other hand, water temps even in a bigger tank would still stabilise at about 28C as that's what the room temperature is. But I would need a proper compressor-type chiller to be able to cool a bigger tank down, so this mini chiller does works well with the nano tank.

I've now reduced the set temp down to 25C (77F) and the chiller seems to be able to cope with that, albeit that it now seems to run for more time than it is off (at 26C it was on/off for about 50/50 of the time).

New temp controller device hopefully arriving today which will give me ability to set temp in 0.1C increments, calibrate temp probe to 0.1C and set the hysteresis loop down to 0.3C minimum (current controller can only set all of these parameters to 1C). I also have quieter fans to install and will cut out the front grill of the case as I think it is the narrow slows that create noise and impair airflow. So I'll make one big hole with a grill to cover it instead (assuming my Dremmel makes light work of the sheet metal case).

I really want to quieten the fan noise down (like I said, it's right next to my office desk) and get tighter control over temperature stability, so fingers crossed these mod's work :nerd:

 

[ckbernie]

What an intensive and detailed write up about the solid state chiller. Nice work. Understood from some write up that solid state chiller efficiency will drop through time. May be you should monitor this and give us an update after few months. =) The Peltier is a consumable and I believe it's quality is propotional to it's price.

Sent from my SM-G975F using Tapatalk

 

[en7jos]

Ok, so finally got around to making some modifications to the chiller last week. Main objectives were to make it quieter and get tighter control over temperatures.

With the cover and digital temp controller unit removed, the guts of the chiller were revealed:

The black bits on either side are two 92mm fans blowing / sucking air through from back to front. The blue thing is what the aquarium water circulates through and this is mounted on top of the Peltier chip (the thin white line immediately below it). There's another aluminium heat sink below that and the copper tubes take the heat down into the big silver radiator. Removed from the case, you can see how substantial the radiator actually is!

First step was to cut out the grill in the front of the case as this seemed to be a big cause of air noise, and also flow restriction. Seems odd that they went with these small slots at the front rather than another big circular hole with wire grill over it, like at the back.

So out with the Dremmel and a cutting disc and I had a nice big hole, albeit square and not a nice circle like on the back. I neatened up the edges a fair bit more than shown in this photo.

Next step was to replace the stock (cheap Chinese) fans with two better quality (not so cheap Chinese!) ones. 92mm is awkward size so didn't have too much choice, but these looked like they would do the job. Plus they had nice white fan blades which I'm sure alone will reduce temp's by at least 2C!

Fan and radiator reassembled ready to go back into the case. The white foam is insulation to try and keep the ice in.

Someone also painted the case with Hammerite spray paint when I wasn't looking, so out with the hearing-aid beige colour scheme - woohoo!

New temperature controller also installed. Looks very similar to original one and exactly same size so a direct replacement, but importantly it has finer control of:
- calibration down to 0.1C (old one only to nearest whole degree)
- hysterisis loop down to 0.2C (old one only to nearest whole degree), so temp will now cycle 25.0-25.2 C, rather than 25-26 C.

New wire grill to cover the new big fan hole at the front which I think looks quite neat, even over a square hole.

So ready for testing!

Noise was definitely reduced with the new fans and after cutting out the vent hole on the front, but still definitely audible and too loud to have sat next to my work desk. So this still needs work. Next mod will be to add a resistance in the fan's power supply to drop the voltage down and hence reduce fan speed and noise. 100 ohm potentiometer is on order...

Unit looks much better now in black (well, dark grey) than it did in the yucky beige colour, so is far more discreet next to my tanks.

But the important thing is that I have been able to calibrate the temperature sensor (using iced water at 0C and against my Fluke meter) so I have some confidence that the displayed temp is correct to within about 0.2C (i.e. as accurate as I'm ever going to get!). I have reduced the set temp to 25C with a 0.3 degree hysterisis loop, so the chiller kicks in when the tank temp reaches 25.3C and turns off when it gets down to 25.0C. Previously it was cycling between 25-26C so hopefully the shrimp will appreciate the stability, and the chiller now operates in shorter bursts rather than an hour on and an hour off as it did before.

Now just need to reduce the fan speeds to hopefully get it much, much quieter (maybe even silent???). Watch this space! =]

For full gallery of pictures see here.

 

[Oughtsix]

Great post! I love it... down to the Hammerlite which is my favorite spray paint for metal objects!

Thank you for sharing!

Is there any room in the case for adding a bit of open cell foam to mute the noise a bit? It looks like it would be too tight. Since the noise is the sound of the air rushing through the case I kind of doubt how much open cell foam would reduce the noise.

What new temperature controller did you use?

 

[Leeatl]

Looks good . I would also like to know what controller you are using . I use the Inkbird ones , but they have the 1 degree between on and off .

 

[alifaziz24]

hi,

how the chiller performance now after a few months of use?
Planning to buy it for my soon to be CRS tank

 

[en7jos]

Hi all, apologies for the slow reply.

So, since my last update I made a few more tweaks to the chiller in the quest of silent operation. The better quality fans and cutting out the grills helped a lot with quietening airflow and improving ventilation, but the fans were still too noisy at full speed. So I reduced the speed by installing a resistor in the power line to bring the supplied voltage down from the 12v supply to something in the 6-10 volt range.

Finding the right voltage needed to be trail and error as voltage vs speed curves vary fan to fan. So I hooked up a 100 Ohm 10-turn potentiometer to be able to play around with changing the resistance and hence voltage supplied to the fan. (Note that reducing supply voltage is not the best way to reduce fan speed because the voltage 'burned off' by the resistor is turned to heat, so you need a suitably rated potentiometer to stop it overheating, and there is a risk of the fan stalling or not starting if the voltage is too low). But as a semi-temporary solution, it works well!

This is the pot I brought from Amazon (100 ohm, 10 turn, 2 Watts): https://www.amazon.sg/gp/product/B07MJDLSBQ/

I found that having the 100 ohm resistor set at 40% makes the fans run silently (can only be heard if you put your ear right next to them, nothing at all from my desk 1m away) whilst still giving more then enough ventilation. I also rewired them to run constantly at this level (rather than turning on-off with the peltier) to keep the chiller radiator temps low. Reducing the fan speeds hasn't had any notable impact on how the chiller cools the tank.

These are the fans I've been using as direct replacement for the original ones- Artic F9 series (90mm fans, well actually they're 92mm):

The F9 base model is the standard 2-wire version that I have currently running with the potentiometer reducing the supply voltage.

The F9 TC is a temp controlled version that has its own temp sensor and reduces the speed down on a pre-defined curve depending upon temperature. This removes the need for the pot, is more efficient and better for the fans as it uses PWN to control fan speed (not reduced voltage), so no risk of fan stalling or not starting. I haven't installed this yet, mainly because it has been working so well with the temporary potentiometer setup! The downside is that the speed-temp curve is fixed and designed for a PC case, so I have no idea whether it will give me enough airflow and silent operation. Need to try it and see, and I guess experiment with the position of the temperature probe.

The temperature controller I am using is an unbranded Chinese type called a "STC-1000". Seems quite widely available and most importantly it has the tight hysteresis control. Works very well, easy to use, direct replacement for the original controller and wasn't at all expensive. I think this is probably the most important upgrade of the chiller unit.

This lack of control of fan speed vs temp gave me an idea for the next part of this project (completely unnecessary and overkill admittedly, but a fun project for a future rainy day). I'm going to setup an Arduino to take care of the chiller temp control and fan speed control so that I can define exactly how everything turns on and off. That's what the third fan type is for - the F9 PWM. It has a 4th input wire that will take a PWM signal from he Arduino (from the PCs motherboard normally) to give full fan speed control. (The PST bit in the fan's name just means there is an extra connector for daisy-chaining the fans - not needed, but this version just happened to be cheaper than the one without, so I'll just cut that extra wiring off!).

This is waaaaay overkill and over-complication though. The 100 ohm potentiometer setup works just great - silent tank cooling.

The tank itself is looking somewhat green with algae on the glass and moss overgrowth after the inevitable Christmas neglect (water is super clean though of course, just not pruned or cleaned the glass, but the shrimp probably prefer it that way anyway!). I have the temp set at 25C which the neo's (few odd RCS, rillis and wild types I moved from other tanks) seem to be enjoying, but they're not doing any better than those in my non-chilled tanks which sit at 27C. The CRS are certainly surviving, but they are still not thriving and overrunning the tank with babies. I've seen a couple of berried females but the few juveniles I spotted disappeared. But maybe they're still there - like I said, the tank is currently an overgrown jungle!

I'm thinking that the CRS really need the temperature to be even lower though. Maybe I'll try dropping it to 24C and see how well the chiller copes?

Really interested to hear how others get on with these chillers please!

James

 

[Hellothere]

hello james

I just bought the chiller and it stopped working suddenly while i was doing a water change for my tank the fan stopped working ... after a few hours (6-7hours later) when i tested it again it was working again . But when i tried to switch it on again, the fan would not move again . Is the system a overheat or do i have to change any parts of it ? (Like motor?)

 

[Nihaopaul]

My chiller broke.. turns out it's the temperature probe on these things that go wrong.

Anyway, I removed the logic board and replaced it with an ESP32 and replaced the probe with a real one not a thermistor. So I have decimal places and an IOT device, since I don't need to keep my temperature between a range (I use it for plants) I just set it to cool to a temperature and then turn it off until it reaches a higher temperature through the microcontroller.

Pretty simple to hack..

Remove the logic board from the base board. And you have 3 pins needed, 1 turns on a relay (use a transistor and pull to ground with a resistor) and 2 pins will give you GND and VCC of 5v (it's a 7805 so you don't get much current).

Now mine doesn't look pretty but I do have an epaper display on it as well as hooked it up to esphome. Might trash the shell as it adds no value, wouldn't take much to add fanspeed with a MOSFET or pwm