e-H2O pt.1

[Goatee]

I'm trying something a bit different this time, my inspiration came a while ago after reading about the development of a heart in a box.

I acquired an e-GPU and this made me think this would be pretty cool to do the same with an e-water loop to provide cooling on the go for a small water cooled PC that could be plugged into a different loop when at home.

It remained a concept floating round in my mind until i spotted some medium sized safe cases reduced in my local DIY shop.

=

I acquired a couple of boxes with the intention of using one as "the heart in a box".

I tried a number of different methods to glue a frame to the case, but being polypropylene nothing sticks, its like the opposite of excrement to a blanket.

I settled on drilling four small holes in the base of the case where the moulded feet are. I could then add o-rings to retain the waterproofing.

I then feed M6 screws from the outside to secure the large metal plate to the base of the case.

I was originally going to use a nut on the inside to secure the plate, but I received an early fathers day present of a nut riveter.

This allowed me to put a threaded insert into the metal plate to attach the screw to. I was amazed at how well this worked and how secure the plate is.

.

I offered up the radiator to the plate to mark where it will fit.

The fans sit just under the case lip.

Not pictured is the fact I have added 8 x 10mm standoffs to the radiator to give the radiator clearance from the retaining base plate.

I drilled some holes to match the standoffs to allow me to firmly attach the radiator to the base plate.

I added a couple more components to show the intended plan.

I'm awaiting a temp / flow monitor being delivered tomorrow I can then test the quick disconnects I am trialling to make sure its suitable (and don't restrict flow too much).

 

[Goatee]

Update time!

My friendly amazon guy arrived mid afternoon so once I had completed my fathers days activities (shout out to all the dads) and put the boy to bed this evening I retired to the garage to plumb some stuff up!

Here is the setup:

Its pretty cool being able to remove the base plate and work on the loop. Everything is so easy to access.

The base line setup is:

• 1 x 240mm Rad, 30mm thick
• EK-XTOP SPC-60 PWM - Plexi
• Aquacomputer Waterblock for Aquaero 5 (its not needed but gives a really nice return feature to firm the whole loop up)
• Thermaltake Pacific Tf1 Temperature and Flow Indicator
• Various fittings and tubing

I achieved a fairly stable 136 L/h which equates to approx 0.6 G/m

Next I added the QDC in (these ones):

Getting around 115-120 l/h which is a drop of 15 L/h or 0.06G/M. It was more difficult to bleed using my T-line but still looking OK.

So how do they work as a quick disconnect?

Pretty damn good!

I'm trying to source a couple of QDCs from Alphacool and Koolance to do a comparison. If anyone has any in a draw and fancies letting me borrow them for testing please drop me a PM. After a bit of testing versus watercooling specific fittings I will decide where to go next.

Not a bad test.

 

[Goatee]

A set of Alphacool Eiszapfen arrived. They are a lot smaller!

Here's the result of the Alphacool fittings

However the amount of fluid after a disconnect is more significant and the disconnecting process feels less "secure" which is quite hard to quantify.

I reset the loop again and am back to 136L / M with no QDC's, so comparing apples to apples looks like the Alphacool fittings are about 3 l/m or ~20% less restrictive.

The Flowfit fittings are quite a bit cheaper in the UK and seem to seal better, but are bigger and more restrictive. So I'm weighing up which solution I go for. I could potentially use a thicker rad with the alphacool fittings, so I might try that out then make a decision.

Thoughts?

 

[Valantar]

I might be missing something completely here, but is that radiator mounted almost directly onto the metal plate? If so, won't that block pretty much all airflow through it? Or is there some way to lift the rad up to give it more breathing room when in use?

 

[Goatee]

I might be missing something completely here, but is that radiator mounted almost directly onto the metal plate? If so, won't that block pretty much all airflow through it? Or is there some way to lift the rad up to give it more breathing room when in use?

I haven't shown it in a picture, but I am using some small standoff's (like those used for motherboards) screwed into the radiator. I then screw through the baseplate into these to secure the radiator.

It gives approximately 10mm of clearance between the baseplate and radiator to allow airflow and also means the two threaded rivets don't foul the radiator.

 

[Valantar]

I haven't shown it in a picture, but I am using some small standoff's (like those used for motherboards) screwed into the radiator. I then screw through the baseplate into these to secure the radiator.

It gives approximately 10mm of clearance between the baseplate and radiator to allow airflow and also means the two threaded rivets don't foul the radiator.

Sounds like it'll help somewhat, but airflow will still be very heavily restricted. Remember that air pressure drops by around half for every 90-degree turn it has to make, so the actual "suction" force into your radiator likely won't amount to much unless your fans have very high static pressure. Your fans look like they're set to pull rather than push air, which sounds best in this scenario, but you'll still likely be pulling hot air back down into the case around the sides of the rad - not all of it, but some. The air velocity coming off the fans isn't likely to be very high given how restricted the intake is, which sounds like it'll increase the amount of hot air being recirculated. Also, with an airflow path this restrictive, your fiddling with flow impedance seems like a very odd thing to be tweaking - I doubt it'll make any difference whatsoever given the limited airflow available through the rad. Lifting the rad another 10mm is likely to make a far bigger difference than any 10-20L/H difference. Of course, all of this is pretty much pure speculation and I'd be very interested to see this setup tested "free standing" (i.e. with the rad sitting on its side or something) vs. mounted to the plate vs. mounted to the plate and inside the case. I'd be surprised if there aren't significant losses in thermal transfer/cooling capability between the three.

 

[Goatee]

Its a good point, but as I only have one open side in the "case" the air has to take some turns to get back out

I am planning to fit some baffling to direct the airflow to avoid recirculation but I haven't quite figured out which way to direct it yet. I'm thinking to direct it out the "front" of the case furthest away from the lid, I may well increase the radiator standoff heights or put a thicker rad in instead. (I have about 100mm total height including the baseplate, so ~95mm height, so there are some options).

In terms of the fans, I'm using some San Ace fans which are broadly similar in terms of performance to Noctua NF-A12x25. They shift a lot of air with a good degree of ooompf (I'm aware that isn't a SI) but on the down side aren't the amazing cookies and cream colour.

	Size	Spped	Noise	Airflow	Airflow	Static pressure	Static pressure	Power use
PRODUCT	mm	RPM	dB/A	m³/h	CFM	mm H2O	PA	mA
NF-A12x25 PWM	120x120x25	2000	22.6	102.1	60.1	2.34	22.9	140
San Ace	120x120x25	1850	24	99.6	58.6	2.31	22.7	130

I had the fans plugged in while testing some of the QDC's outside the case. Air flow seemed "good" (SI?) but I like your suggestion and I could trial the three orientations you suggest with a waterblocked GPU and CPU to give some indication of air flow impact.

My testing with QDC's was solely to see if the cheap ones would be sufficient, which is seems they might be. But as I plan to have four pairs of them (2 on the heart in a box, 2 on case) I might just go with the watercooling specific ones rather than my budget ones, as they are smaller.

Really appreciate the feedback.

 

[Valantar]

[snip]

That looks like an excellent fan - what kind of bearing does it have?

Baffling for optimal airflow paths sounds like a good idea, definitely. I have another suggestion, though it might not be possible in practice, and even if it is might complicate your build quite a bit: hinges. More specifically, mounting the radiator on hinges along the side closest to the pump/auqaero, so that when in use you can lift the opposite side up somewhat for better airflow. There's no doubt this will interfere with some other hardware in some way, but some lift ought to be possible (30 degrees? 45?), and should give a far better airflow improvement than taller standoffs. It might be doubly beneficial as this might give some opposite directionality to the air intake and exhaust, alleviating recirculation somewhat. I am of course assuming that this unit will be stationary when in use, otherwise this would be ... problematic. Preventing it from rattling around when moved would of course also be required if only one side of the rad was securely mounted, for which I would suggest the following: instead of the standoffs on the opposite side from the hinges, get some neodymium magnets - the kind with screw holes in the middle - and screw them into the base plate. Then screw a large nut or steel disc onto the matching screw hole on the rad. Stick some foam padding in the lid of the case (I'd recommend this anyhow, to protect the tubing and loose QDCs), and the combination of these two should keep the radiator in place, but loose enough that you could lift it up when in use. All you'd need then is something to prop up the rad to maintain its angle, and you should have a pretty decent little cooling box!

I realise I might have made this a lot more complicated for you now - sorry! I just found this to be a very cool little project

 

[Goatee]

That looks like an excellent fan - what kind of bearing does it have?

Model number is 9S1212M4011 - Double ball bearing. I really like them (I had seven in my loop previously).

I realise I might have made this a lot more complicated for you now - sorry! I just found this to be a very cool little project

You have now got me thinking..... watch this space

 

[Valantar]

Model number is 9S1212M4011 - Double ball bearing. I really like them (I had seven in my loop previously).

Sounds like a slightly better Gentle Typhoon - slightly quieter (3-4dBA) and slightly more airflow and static pressure, but the same potential for "grindy" bearing noise at low speeds. Doesn't bother me with my GTs at least

You have now got me thinking..... watch this space

Will do!

 

[bubbl3]

San Ace datasheet, third from top is the fan used here: https://www.sanyodenki.com/archive/document/product/cooling/catalog_E_pdf/San_Ace_120S25_E.pdf