Jump to content
inTank Media Baskets

Iceprobe chiller in stock Nano Cube (no drilling)


artarmon42

Recommended Posts

As a spin-off from my MH thread, this one will journal my attempt to put an Coolworks Iceprobe into my Nano Cube (12G DX ). I suspect that over time, both threads will have cross-overs (as you can't talk about MH without talking about cooling the tank etc), so it might get confusing for new readers 8)

 

So let's begin...

 

I bought the Coolworks Iceprobe and Controller.

Here it is with everything out of the box (I put a Rio 90 in front as a size reference).

C1-IceProbe.JPG

 

If you've seen one of these before, you'll know that the top heatsink/fan is too big to fit into the back chamber of a Nano Cube 12. The back area has a clearance of 2.5 inches, and the heatsink (the largest component) is 3.75 inches.

 

So, off comes the top...

C2-RemoveTop.JPG

 

It's actually pretty well made, and was heart-breaking to have to rip it apart. They've done a nice job drilling and threading wires through the heatsink for that nice clean look.

 

To fully detatch the heatsink, you'll need to loosen 2 screws from the bottom. It needs an allen-key type instrument to remove, but brute-foce an a star-head bit will also work :-*

C3-RemoveBottom.JPG

 

Here's the top fully removed from the bottom.

C4-Dismantled.JPG

 

There are 2 important (to this DIY mod) parts of the Iceprobe.

The first and foremost is the probe itself. I think it is made of aluminium (for conductivity) and coated with some type of plastic (which is important as you're sticking this into the water).

The second part, and arguably the most expensive part, is the peltier. Unfortunately, at least on mine, the peltier is stuck to the (ueselessly large) heatsink.

So in short, the only thing you can salvage from a dismantled Iceprobe is the probe and the wiring.

Link to comment
  • Replies 57
  • Created
  • Last Reply

So the first order of business is to get a lower profile heatsink/fan. By my calculation, you need one that is at most 2.5 inches (length or width... you have to get the sucker into the back chamber which only has a clearance of 2.5 inches) and less than 1.25 inches high (otherwise the splashguard will get in the way and the hood won't close). Standard CPU-sized heatsinks won't work (too wide, too high). The only ones that seem to fit this dimension are either:

1) GPU coolers (devices to cool graphics card processors)

2) liquid-cooled waterblocks (you either know what it is because you're an overclocker, or you'll have to figure it out yourself... it's not relevant to this mod).

I ended up buying the GPU cooler that pushed the most air around.

C5-NewTop.JPG

 

Here is a picture of the old and new heatsinks for size comparison purposes.

C6-SizeOfNewTop.JPG

 

Because the original Peltier was stuck to the old heatsink, I had a buy a new one. I suspect what I bought is the same one Coolwork uses (40x40mm 51W rated).

C7-NewPeltier.JPG

 

The hardest part of this mod is this one hole I drilled (I'm joking... it was pretty easy to drill, but that's the hardest bit). The probe screws (which went into the original heatsink were spaced 2 inches apart. The holes of the GPU cooler is spaced 2.5 inches apart. So I had a drill one hole.

C8-DrilledTop.JPG

 

So it was now time to connect up all the wires.

When I started dismantling the wires in the old heatsink/fan, I found this device. It looks like a temperature shut off sensor (I guess as a safety method to kill the peltier if it went crazy).

C9-SafetySwitch.JPG

 

Because I like living dangerously (seriously, there was nothing I could connect it to and I didn't know how it was calibrated) I didn't wire it to my new setup.

Link to comment

It has also got a cool LED to show you when the unit it on.

C10-Wires.JPG

 

Using thermal paste (included with the GPU cooler), I sandwiched the peltier between the probe and the heatsink/fan.

C11-ConnectedTop.JPG

C12-ConnectedBottom.JPG

 

Note to others who might try this... read the Peltier instructions to figure out which one is the hot side and which one is the cool side. Don't just throw out the instructions like I did the first time :rolleyes:

 

With everything (silicon washer, nut, etc), I call him NanoIceprobe (hope I didn't violate any copyrights :angel: ).

C13-NanoIceProbe.JPG

C14-NIPBottom.JPG

 

After the drilling, the next most difficult bit was connecting all the wires back together (new peltier/fan to the original cable). I suck at soldering, so here's the ugly job I did...

C15-Solder.JPG

Link to comment

Following the Iceprobe instructions, I calibrate the unit.

C16-Calibrate.JPG

 

Once calibrated, I dumped in hot water...

C18-Cooling1.JPG

It may not be clear in the picture, but the vital stats are: time is 7:31pm, room temp is 76, water temp is 89.8.

 

12 minutes later...

C19-Cooling2.JPG

Vital stats are: time is 7:43pm, room temp is 76, water temp is 87.6.

 

Basically a drop of 2.2 degrees in 12 minutes!

 

19 minutes later...

C20-Cooling3.JPG

Vital stats are: time is 8:02pm, room temp is 76, water temp is 85.6.

 

A drop of 2 degrees in 19 minutes. Not as great, but remember this is in a container without water movement. I found that if I moved the problem closer to the probe, the temperature was much cooler (measurement was 84.2).

 

Through the 30 minutes of this test, I was hand-checking the heat dissipation. The fan was definitely blowing warm air (not as hot as my MH exhaust, but noticeably warmer than room temperature). The copper fins of the heatsink were also warm (not hot, definitely can touch and hold without pain).

 

C21-HeatSink.JPG

This picture shows the temperature of the heatsink (probe against the copper base). The temperature seemed to be stable at 91. Replacing the Iceprobe's old-style heatsink/fan with current overclocking technologies (copper, fins & fan design, etc) seems definitely possible.

 

Interestingly, while the probe itself was cool to touch (not cold) the water around the probe felt "chilled" (like water from the fridge).

 

As an after-thought, I did a little experiment to compare the chiller against "normal" heat dissipation.

C22-Ambient1.JPG

C23-Ambient2.JPG

It dropped 0.7 degrees in 10 minutes.

Obviously I didn't do a double-blind scientifically-validated test, but annecdotal and circumstantial evidence shows that the NanoIceprobe works.

Link to comment

I'm going to run some more tests tomorrow (plus seal up all the exposed wires), and then hope to put it into my tank on Sunday.

 

Comments and feedback welcome!

Link to comment

Vent the fans through the hood.

 

I dont think the probe will be as effective, since it counts on conduction (and any water that chances by it) in order to cool down. Still, a great mod tho..

 

I wish you were in LA, id hire you..

 

Chris

Link to comment

Yeah I saw your other thread (great minds think alike I guess :) ).

 

I was thinking of falling back to the waterblock plan if the heatsink/fan doesn't work, but I was trying to minimize cost (I had already bought the Iceprobe).

 

I think all up, this mod cost me around $190 ($150 for the Iceprobe + Controller, $15 for the heatsink/fan, $25 for the new peltier).

 

Waterblock method would have probably cost around $350 ($150 for the Iceprobe + Controller, $25 for the new peltier, $150 for the liquid-cooling kit (waterblock, radiator/fan, waterpump, coolant), $25 for the PC powersupply to drive the liquid-cooling equipment).

 

If I was a business (something you might want to think about for nanocustoms), a liquid-cooling kit that uses a titanium "pipe" (instead of the copper waterblock) would be a perfect cooling solution. For me, it's not worth the money/effort to custom make one titanium pipe when there are so many off-the-shelf parts I can play with.

Link to comment

Like I mentioned in the other thread. You may have a problem with the salt creep and water evaporation on the copper surfaces.

 

If you could make an acrylic 'case' for the heatsink aspect and only allowing air from outside the tank into it, you may avoid a problem.

 

I'm just not sure how well it will disperse heat when it's rusted and caked in salt.

 

I like the modd and I wish I had the cash to throw one together right now and test it along side you!

 

I'm definitely going to try and find some time to do your MH mod as you did an amazing job. Now, only if I could have YOU do it so I don't break anything... :)

Link to comment

I wonder how much heat could be passed or cooled through the glass in the tank. It's possible that you could use some method of cooling against the back pane and keep all the equipment outside the tank. You could also still stay low profile.

 

Maybe 10 PC fans? :)

Link to comment

Yup the copper is a concern to me. Maybe I'm optimistic but I don't think it's likely to rust and flake into the tank. It is not directly touching the water, and hopefully the fan is generating enough airflow to counter condensation.

 

Salt caking it up is another matter altogether :o

But that's one advantage of this journal. I can track and report on it over time.

 

I think I mentioned it on the MH thread, but the big problem is that the DX hood doesn't have any vents for the air/water (only the lighting). This keeps the evaporation low, but hinders natural cooling (airflow over the surface area) and consequently the ability for the NanoIceprobe to vent properly.

 

The easy fix is to prop the front flap up slightly so that air can flow more freely. If the NanoIceprobe was only going to be used in Summer, that might be acceptable. But as I'm using the NanoIceprobe to counter the MH heat, opening up that flap every day is going to void the original objective (keep the clean lines and one-piece look of the Nano Cube).

 

Even if the tests are successful tomorrow, I'm still looking for ideas to provide better venting. Right now, my gut feel is that I need to dremel a small slit in the back of the hood ("out of sight") to provide adequate airflow...

 

 

Also thanks for the comments about the MH mod. It is pretty simple, and hopefully I've provided enough instructions/photos. Let me know if anything isn't clear when you get around to it ;)

Link to comment

Well I don't have a DX hood, I have a 12g OG hood and there isn't as much venting.

 

I'd have cut some more vents.

 

How long is the photo period?

Link to comment

Actually I now believe that a non-DX hood is the best for both the MH and NanoIceprobe mods. The only down side is that you won't be able to have a "reliable" fuge mod because of the lack of splashguard in the back.

 

3-4 scroll fans would do the trick: 1 fan exhausting out each side vent, and one in the back blowing air through the hole in the back of the splashguard. Then you use the other hood back vent to create circulation for the NanoIceprobe.

 

Look into it... I'm pretty sure that would work.

 

 

 

 

My target is to have a continuous 8-hour photo period with the MH (I currently run 2 photo period in the morning and night because the water temp gets too high for my liking after 4 hours). Thus the necessity of the chiller...

Link to comment

Oops, I forgot to address my "reliable" fuge mod comment.

 

The 7W PC that I ripped out of the Red Seas clip, has reliably been velcro'd to my fan ever since I installed it. Of course on my DX, it is behind a splashguard.

 

Without a splashguard, the moisture/creep might mess up they adhesion (not the velcro, but the adhesive that holds each velcro piece to the fan and PC). You could "hack" the mod and zip-tie it or something. But that would be rather ugly IMHO ;)

Link to comment

You need to test this with bringing the water BELOW ambient temp to see if this will really work. You're also working with a much smaller water volume which will give you much faster results. I can pour 90 degree water in a small tub with 76 degree ambient temp and it will be "chilled" on it's own. I'm not trying to bring you down or be mean, just would like to see you be sucessful in this.

 

It comes down to, can your chiller pull more W of heat out then the lights/pumps/ambient air can add.

Link to comment

Yup, that's what I'm going to test this afternoon/tonight.

 

Any and all comments welcome. This is a hobby for me, and people poking holes in my ideas so that together we can come up with better ones is what I'm hoping for (otherwise I wouldn't bother putting it in journal).

 

As long as it doesn't get personal, I'm all good for "constructive feedback" :)

Link to comment

The peltier you're using will cool better then the one that chris from nanocustoms is using(at lower wattage), but he also has a more appropriate cooling solution for his, which will make it more effective. The thing with TECs is if you can't cool the "hot" side enough, the "cold" side will feel the effects of this. The sink/fan combo you have picked is made for much less then 51w.

 

After you test this, look into a low profile rack mount server heatsink. Something that will give you more then 5cfm.

 

nexfan03_1821_1508393

 

Overall Dimension: 70 x 70 x 23mm

FAN Dimension: 70 x 70 x 10 mm

Current: 0.3A

Fan Speed: 4500 rpm

Air Flow: 30.79 CFM

 

 

My cube comes monday so I have no idea how much space is actualy back there to work with.

 

BTW the sink pictured is ment to work with a CPU which puts out up to 89w. While you're testing, use thermal tape so you don't have to drill the sink. When you're done use some thermal epoxy from artic silver.

Link to comment

I'm not really worried about a fuge light honestly.

 

I'm using the ZEOvit system to controll my nutrients so I don't need a fuge.

 

I'll look into getting my hands on the parts and trying it. I'll continue watching this thread about the chiller for sure!

Link to comment

So I put the NanoIceprobe into the back of the tank today. To summarize: it did not work as planned, and was unable to compensate for the MH heat.

 

But I'll go over what I did, some mistakes (that I corrected) and then look to the power of the community to help me figure out how to overcome the problem(s).

Link to comment

So, what you're saying is to have 2 scroll fans blowing out the sides (exhaust) and one sucking in from the outside and blowing that air into the MH area?

 

It sucks that your initial test failed; but I'm confident you'll figure it out, you seem like a smart guy afterall.

 

Anyways, keep us posted!

Link to comment

To detail the installation process...

 

As I was going to put the NanoIceprobe in the back compartment, I needed a "quick" way to get the cool water to the main tank (before it dissipated). I use a Rio 90 with the funky connectors...

C24-ReturnPump.JPG

 

The picture shows how I connected it together. Once joined, I rotated the pump down (so it sucks water from the bottom).

 

To make room for the probe, I re-arranged my back chambers. Heater is now in the leftmost chamber. The suction-holder for the Ebo (seen in the rightmost chamber) was removed after I took the picture.

C25-BackReorg.JPG

 

Turns out that the neoprene and silicon washer is too big to fit into the back chamber. They're easily trimmed...

C26-CutToFit.JPG

 

Here it is, installed...

C27-Installed1.JPG

 

You'll notice how it's a nice tight fit above the chamber walls.

C28-Installed2.JPG

 

And I velcro'd the Iceprobe Controller to the back...

C29-Controller.JPG

Link to comment

I won't post all the data I collected (don't think anyone cares about the details anyway :ermm: ). Obviously the important point here is that the Iceprobe is no match for the power of a 70W MH :-*

 

But here are some things I learnt from the failure...

1) I made the mistake of keeping the Controller's probe (that tells the Iceprobe when the shut off) next to the Iceprobe at the start. Obviously stupid, because when the water around there cools (as it would) the Iceprobe would shut off. Putting the Controller's probe in the main chamber is much smarter :rolleyes:

2) Without the NanoIceprobe, the MH was heating the system up ~1 degree every hour. With it, the temperature increase was reduced to ~0.6 degrees every hour. Still too much (4+ degrees in an 8 hour photo period). My target is :blush:), the heatsink/fan that I used isn't powerful enough. I was measuring the air temperature around the heatsink/fan and noticed it was starting to get up to 100F. The combination of the peltier's heat plus the MH heat plus the lack of venting, surely put a dent in the fan's ability to distribute heat.

 

I terminated the experiement at the 2 hour mark, when it was obvious that the Peltier could not cool the water faster than the MH was heating it up, and the overall target cooling was not going to be met.

Link to comment

Chronicles,

The 70mm fan is too big to fit in the back chamber. 2 inches (50mm, the size of the Icebreq) it the biggest that'll barely fit.

 

From my tests though, I don't think any fan will be enough. The lack of venting (to/from the room temperature air) is the killer I think. MH puts out crazy heat!

Link to comment

Sean,

(regarding the scroll fans for the OG splashguard)

Yup, I think so, based on my memories of how the old splashguard was designed.

Link to comment

If you mounted an exhaust fan above the iceprobe, I think you could remove a lot of the heat, do you already have this setup or are you pushing air into the unit?

Link to comment

Archived

This topic is now archived and is closed to further replies.

  • Recommended Discussions


×
×
  • Create New...