Scratch Build – In Progress CryoCore - Final shots, July 5

I'm sure everyone's dreamed up dozens of mod's they'd like to do, or see, but few ever really pan out into an actual project. This was one of those things that started off with "i wonder if it's possible?" That leads to "i wonder if anyone would make this?" and finally, "i wonder if *I* can make it?"

The general idea started off with a couple things i've always wanted in an "ideal" computer. powerful, quiet, dust-free, good looking. the ideas in my head kind of flew around until it reached an extreme: 100% passive cooling, fanless, yet powerful enough to run crysis at 1920x1080 on high (or, dare i say it, VERY HIGH). not that i'd actually be playing crysis, it's more of a benchmark.

so, after about a year and a half's worth of doodling on napkins, and on the margins of my notes, and finally slapping together a sketchbook model (or seven) and countless numbers of revisions, i finally took the plunge. i did have to make a few compromises. first, it's not fanless. i can't eliminate the PSU fan, and i'm not daring enough to make a watercooled PSU. second... i'm not going to be running crysis at my current specs... yet. i left room for upgrades, though. This computer will be the culmination of two decades of pent-up insanity. i'm sure a lot of things i'll be doing haven't been done (sucessfully, at least) by anyone.

so, in the beginning: a pile of parts. i got these around december last year.

the black bars on the bottom are the passive watercooling pipes. the reason i ordered all this stuff so early was because the pipes, pretty much the most important part of this entire build, were discontinued. i managed to find one shop that still had them.

next up, shopping for acrylic. i wanted it to be roughly 1/4", which should be strong enough to hold up. i shopped around... and found that the prices suck. $125 for a 4'x9' sheet of acrylic. if i buy 100 sheets at a time. wew hew. the next best option is roughly $11 per square foot. and i'm in Vancouver. it's a major port city with lots of industries in the surrounding area. why is acrylic so expensive here? i just ended up buying 7 sheets of ACRyan acrylic from Performance-PCs.com. and here's (most) of the lines drawn on masking tape over the acrylic:

now that just sat there in the living room for a good month or so. lack of free time pretty much stalled everything.
i will pause here to mention that this article was of TREMENDOUS help in getting me started. thanks, Brett!
http://www.bit-tech.net/modding/2008/03/14/a_modders_guide_to_acrylic/1

now, how to cut it safely? preventing cracking is generally done by using a thinner blade with more + smaller teeth. that's easy to get. now to prevent the acrylic from being melted by the blade, that'll take some cooling. why not do it with a tilecutter? it automatically keeps the blade cool with a nice stream of water.

unfortunately, the blade made for ceramic isn't very good for acrylic. i don't want to replace the blade with an actual teethed blade, as i have no idea how much it could mess up the machine.


so, in the end, it had to be a table saw with manual, light spritzes of water to make sure the blade is cool. and yes, it does help. i got lazy with one of the panels and didn't spray, and at the end of the cut, the edge had a bunch of melted acrylic stuck to it. not a huge problem, since i just pulled it off after it cooled, but i'm sure it's something you want to avoid.


today's cutting done. there'll have to be a lot more in the future before i'm finished...

i'll leave it at that for now. i also polished the edges today, but that's worth a separate post.

i'm actually in the middle of exams right now, but luckily i have a 7-day separation between two exams and i decided to take today off. the next major update probably won't happen until may.

 

ok, i think i'll just post a quick mini-guide to flame polishing here. this was my first time doing it, and most of it was sort of just trial-and-error discovery.

so first, an introduction. most people avoid working with fire, and i think it's a skill everyone should be at least familiar with. it's like working with a knife. it's one of those fundamental tools that finally allowed mankind to outcompete those damned gorillas. stealing fire from the gods has really let us go places, like caves or siberia. unfortunately, ever since some people stole the lightbulb and electric heating element from the gods, fire has been frowned upon. with acrylic, though, it's still one of the best tools around.

so. polishing with acrylic. people will argue over sanding vs. flaming to polish the edges of acrylic. after having experienced it, i will, without a doubt, be a flamer. many people will argue that fire is dangerous. i agree. but i bet those people that argue that probably haven't worked with fire outside of a kitchen. my earliest memory of playing with a blowtorch was back in grade 4, when the whole class was making hot air balloons and trying to get them to fly. mine totally won, since my propane-fueled air was so much hotter than the other kids', who were using dinky little matches or lighters or whatnot to get theirs to fly. on the second flight, i set it on fire by accident, but it still went higher than everyone else's before it was reduced to cinders. good times. good learning experience.

so, back to present day. i will admit i suck at sanding. i have NEVER managed to get something shiny by sanding it. ever. even starting with 200 grit and working my way up to 4000 (should i have gone higher?), all i've ever managed was a very smooth (to the touch), hazy texture (on both aluminum and acrylic). screw that, i said. i'm going to try flaming. now, sanding WILL get you an incredibly smooth edge. if, for whatever reason, you need the edge to be absolutely flat and smooth, sand it. however, if your goal is merely to have a glassy appearance, then flame it. you don't even need a smooth edge to look "glassy." glass looks glassy even when it's heavily textured. flaming will give you that appearance really quickly, with little effort. i'm talking 100x faster and easier.

i first tested it on my scrap pieces left over from the cutting.


after a few tries, i quickly discovered how to do it properly. so here's basically what i learned:

First off, you should know that acrylic IS flammable. and by flammable, i mean "it will react with oxygen." it will not, however IGNITE. so don't worry about your entire piece of acrylic going up in flames. now, to avoid burning acrylic, follow these rules:
1. do not drive while flame polishing acrylic.
2. do not talk on the phone while flame polishing acrylic.
3. do not fall asleep while flame polishing acrylic.
if you can manage the above, the chances of actually burning or ruining your acrylic are pretty low.

next, find a nice stable surface, and clamp down your acrylic over the edge. like so:


at this point, i will say that you SHOULD do a bit of sanding before you flame. there's 2 reasons for this.

first, sanding will make it smoother. flaming will get things shiny, but will, for the most part, keep some marks visible. i just decided to leave the marks from the sawblade there. i just wanted it shiny, not perfectly smooth. it adds character, anyway. =P

the second, and probably even more important (to me), is that the dust makes a very handy marker. i will explain this below.

now for the actual polishing.
after you sand, clean off the dust from the sides you're NOT polishing. however, on the edges that you will be polishing, LEAVE THE DUST THERE. having your panel clamped down, you can fire up your blowtorch. keep in mind the hottest (and most consistent) part of the flame is at the tip of the inner glowing part. most people don't know this. it should be a healthy blue (nearly invisible) flame, and there will be a blue part in the center, and possibly clear/yellow/orange wavy part surrounding it. you want the tip of that inner part just touching the edge.

now here's the reason you should leave that sanding-dust on the edge. it's actually kind of hard to tell when it's been polished- you have to move the flame away and let it cool for half a second before it really appears to be shiny. the easiest way to know? if it still has some white powder, you need to heat it longer. just move your flame along the edge slowly enough so you see all that white dust melting and disappearing as your flame passes by. in the many acrylic guides i've seen, nobody ever tells you the exact speed (or anything more precise than "quickly") to move, so i shall: 1 cm/s. that's roughly how fast i ended up doing it after some practice. results may vary but i'm sure that's a good speed to aim for.

now, for a few things to watch out for. please keep these words in your head at all times: DON'T PANIC. thanks, HHGTTG!

first, as i mentioned above, acrylic does burn. on many occasions, just behind my blowtorch flame, there was a small orange flame and the smell of burnt plastic trailing it. don't worry. it's normal. there are no visible markings left from it (i.e. no black marks) and it puts itself out within a second (or at least, always within 1-2 cm of my blowtorch flame).

second, warping. your acrylic WILL warp slightly. again, don't panic. this is because the edge you're flaming is hotter than the rest of the panel, heat makes things expand, etc. etc. just finish your flaming, give it a good minute or two to cool down. if it's still warped after the whole thing is at room temperature, then you've got problems. i had about 10 pieces to go through, and all of them came out fine after cooling down.

i DO NOT recommend flaming for very thin pieces or very sharp angles, however. i flamed some 30 degree angles in my scrap test pieces, and a very slight bend stayed with it. my guess is that the very thin parts got heated all the way through, and being so thin and small, it couldn't hold its own shape as it was cooling down.

third, bubbles. this won't happen unless you keep the flame on for much longer than you should. as you can see the last picture of my OP, there were 2 pieces that had angles over 180 degrees. this pretty much forced me to keep the flame in the general area longer (probably 2-3 times longer) and a few tiny bubbles appeared when it cooled. for the most part, you probably won't encounter them.

fuel: i used a standard propane blowtorch. i don't know how hot butane gets, but i know there are several different types of fuels with different temperatures. i would recommend propane just because it worked for me. remember, hotter doesn't mean faster- it just means hotter. all you want is to slightly melt the surface, and you want to be able to do it at a nice consistent pace.

 

At first I thought I had a convert to the church of tile saws. I think the reason you had trouble is because you didn't have anything under the plexi. If there's no tray under it, the blade will take the path of least resistance and just roll over it. The more worn the blade, the more you have to force it to cut.

Any pics of your design? or is it a surprise?

 

there was a rolling platform for it, but it's made for tiles, not 48x48cm acrylic panels. i would have had to make multiple cuts to get it all the way across. the table saw worked perfectly on the first try, so i just went with it. i'll keep the design a surprise, for now. =)

if you count the number and shape of the panels, though, you can already tell it'll be a hexagon. no plain rectangles for me!

 

May 2nd - random craps.

first off, as part of my mod, i wanted something that looked like heavy-duty, sci-fi clamp thingys. to be honest, i searched all over google and couldn't find a picture of what i meant, nor any actual clamps/latches that fit the image in my head. just picture the kind of locks you'd have on a 20-ton steel door at a fallout shelter, or something that would be bolting down an escape pod from a spaceship or something.

in any case, i found something that fit pretty well: cabinet hinges.

now, i needed them at a particular angle. lucky for me, a 5mm piece of acrylic wedged into one of the joints holds it at (roughly) the angle i needed, so i just took some of the spare acrylic i had and glued it in there. i didn't like the rusty brown colour, so i grabbed some leftover black paint and covered it up.

i probably won't be putting those to use for a long time, until almost everything else is finished.

next up: bending acrylic!
i just used a heat gun. nothing special. i clamped down the top of it with wood, and just let it bend itself by gravity. it didn't bend right at the joint, though, but it still worked:


of course, it didn't turn out perfectly. oh well. good thing i only did 4 out of 6 pieces, i'll just fudge around with the last 2 to make everything fit. a lot of this project is kind of "go-with-the-flow" for me. i'll be surprised if it does turn out as good as i imagined. of course, i could try unbending and re-bending, but i doubt that'll make things anything but worse.


and here's a quick preview for you guys to see the general layout:


next up: the glue.
being both the cheapass and experimenter that i am, i decided to try my hand at homemade acrylic cement.
first, the sacrificial piece of acrylic.

this was the one i used to make wedges for my clampy-hingy things, as well as my experimental flame-polishing and bending test piece. you've served me well. into the jar you go:

huh. i didn't realize i'd end up with that much.


chili radish strips in soy sauce. with acrylic and acetone. yum yum. i'll let that stew for a few days and try to stick a few test pieces together. i hope this stuff works!

 

good start, I stay tuned

 

wow. my summer is apparently much busier than i thought it would be. this is going to be slow progress.

in any case, i did manage to get a bit done. first off, i put to use my homemade acrylic glue. i discovered a few things about it. first off, i can change its consistency. normally it's quite thick, and will barely even enter a syringe. it's a bit thicker than rubber cement. i just poured on some extra acetone, left it overnight, and the top layer becomes quite thin and watery, but still has some acrylic fibers dissolved into it, so it works quite well. if i leave it a bit longer, it gradually becomes thicker and thicker as the acetone dissolves and mixes with the rest of it. whoopee!

here's my first gluing. lots of surface area to cover.

unfortunately, it turned out a bit hazier than my test piece:

oh well. a side effect of working with all this glue is that my hands ended up being covered in a thin, healthy layer of acrylic. wew hew.


next up, i had to make my cooling rod supports. in the process of being lazy and not wanting to use more acrylic than i needed, i ended up making each piece about ~1cm thinner than i had originally planned. it still worked out fine, though. first, the glamourshot: i love flame polishing.


next up, the gluing. i finally found a use for my old textbooks! on that note, if you see anything you want, i'd be glad to sell it to you. shipping costs might be rather high if it's going overseas, though. being rather lazy, i didn't really bother to bend the acrylic to fit. i just used the glue in its thicker form to fill in small gaps.


and finally, the little dividers. for these, i made the glue much thinner and watery. it worked quite well.


in case anyone's been wondering what the hell all this work up to now has been for, here's a little preview:

i kinda just stuck them there to see if they fit.

and finally, a break from acrylic work. here it is guys! the warranty voiding! i'm sure this is a familiar sight to all of you around here.


i wasn't actually going to do this, but a brainwave just hit me about half an hour ago. i was planning to make a round cut in my acrylic, plus drill 4 holes to fit a 120mm fan filter on. on top of that, i would need another way to secure the power supply on to my case. then it hit me: i could just open up the PSU, screw the filter directly onto the PSU, and just make straight cuts on the acrylic. the filter itself could end up holding the PSU in place (laterally, at least. i'm still going to need one piece to stop it from sliding backwards), all is easy, all is well. yay! laziness truly is the driver of innovation.


and that's all for now. i'm hoping to have the front and rear panels cut by next weekend.

on that note, any suggestions on how to cut a 18cm radius arc in acrylic? jigsaw? dremel?

 

laser cutter?
also have you tried buffing acrylic with plastic/metal polish? it works really well

 

mini-update: power switch.

i just had a bit of spare time today, so i decided to try my hand at making my power switch. i haven't soldered since grade 8 tech class, but it went pretty well. the pin headers i thought i could get by with crimping, but no dice. they fell off with a light tug. one of them just flew off when i put them down. thus, in about 5 minutes, i had to master my soldering skills and drip just enough soldering on there so it would hold the wires in place but without getting in the way of inserting them into the plastic bits. all in all, it went quite well.

 

wow. almost an entire month and i forgot to update. well, it's been pretty busy, but i've managed to accomplish a bit.

first, the arcs:

this was achieved through the magic of.... a table saw! yes, the same table saw i've been using for all the other straight cuts i've done so far. it's not perfectly round, as it was essentially guided by hand, but the saw did manage to get a round cut out of it. wohoo! good thing i didn't need smaller arcs, i have no idea how it would have been possible then.

next up, the motherboard tray. the rear panel of it had room for 2 40mm fans... not like i was going to use them. that little panel was just getting in my way, and possibly blocking the dome i'd be putting over it. off it goes:

and if you're wondering about all the thumbscrews instead of standoffs there, i was rather bored for a week while the weather was crappy. i like to do the cutting and drilling and stuff outside, so it's easier to clean up.

 

....and a little bit more. i actually did manage to get a bit done over the course of this month, but i kinda forgot to take pictures. most importantly, i got the sides and the base glued up, so it's actually a hexagon now, instead of a bunch of random panels lying around. no picture of that to show you. whoopsies. let's skip ahead, to when i'm gluing the drive bay:

ah, those useful textbooks lending a helping hand yet again.

and here i'm just spraying a bunch of stuff. one of those things is the pump. given its location in the case, it was going to be rather visible and ugly, so i made it match the front panel.

i also sprayed the hose clamps to make them less ugly, but i ended up not using them at all.

 

Looks nice, more updates!

 

aliright, i got my motherboard and CPU in, as well as a few of the waterblocks. I'm just waiting for the video card- June 25th, Radeon 4870. =)

after reading an article about diamond thermal grease, i wanted to try to make some for myself. the original article's gone (probably dugg to death) but in short- he didn't even use that much diamond (i think the final mix was under 50%, probably around 33%) and he still got a ridiculous performance improvement. results may vary, of course. i bought my own diamond dust (it's actually pretty cheap) and some silicone grease.

i ended up not using it. i had already bought some coollaboratory liquid metal compound- another ridiculously high-performance (and risky!) compound to use. i was actually thinking of mixing the diamond powder with the liquid metal instead of silicone... but it didn't mix well, so i scrapped that.


i mixed up the diamond and silicone anyway, trying to keep the proportions of diamond as high as possible. i think i ended up with somewhere around 50-66%. the stuff turns out kinda greenish, and quite thick (probably twice as thick as regular grease, maybe a bit like shin-etsu). i only have about 1/4 tsp of it- but technically, that should be more than enough for an entire system if applied properly.

oral use only, guys.

so why didn't i use it? well, i was playing around with the liquid metal stuff. it's really quite fun to spread. you can squeeze a tiny drop of it out of the syringe, and it just sticks to itself (kinda like melted solder). and then it refuses to spread. i pushed it around a bit, and then i squished it. as soon as i did that, it just stuck and spread to the surface like crazy. it automatically makes this super-thin layer across the surface. after scraping it with a razor blade, it looks like the shiny side of aluminum foil. cool stuff. i was having too much fun playing with it to take any pictures. sorry, guys.

i also had an accident- being liquid metal, this stuff is highly conductive. and i spilled a drop on the motherboard. this is how i discovered the "spreading" technique- when i tried to wipe it up, it just coated the board. it looked like i had chromed a section of my motherboard- it was cool, but i freaked. lucky for me, a paper towel with alcohol just sucks it right up, and i managed to clean it all off. after that, i got pretty good at spreading it... so i thought, why not?

my original plan was to just use the liquid metal on the CPU- since it's covered by the cap anyways, it would be easier and safer than the other chips. but i got used to it, so i used it on the northbridge as well, and i'm probably going to use it on the GPU too.

for the RAM waterblock, since there was just so much to apply, and RAM doesn't get too hot anyway, i just used Arctic cooling MX-2.

here's my motherboard, in its half-finished glory:

seriously, i love enzotech blocks. the MIPS block looks really nice as well.

and finally, a bit of tubing. i had tygon 1/2" tubing- but damn, that stuff kinks so easily! not to mention other little niggles, like its sticky texture (can you say dust magnet?) and slightly yellowish tint. chemically, i'm sure it's very high quality stuff- it's generally the first choice at the laboratories i work at. tygon's everywhere. but i want my comp to look nice, and i don't want to use anti-kink coils if possible. i think they're ugly. hose clamps are also quite ugly, no matter what type. spring clamps aren't as bad as the screw-in type, but both aren't going to win any beauty contests. especially not next to my nice and shiny enzotech blocks. there's also the fact that they take up a little bit of space, and the northbridge was running really close to where my RAID card was going...

for some reason that i can't remember, i had a bunch of swiftech 7/16" tubing lying around. it was supposedly lower quality, but i liked the colour better, and it was more kink resistant than the tygons. also, the fact that it was slightly less than 1/2" meant that i could trade in the hose clamps for a little bit of swearing, and end up with a much nicer looking loop.

 

The bulk of my work is done, now. i'm just waiting for the video card to hit the stores, then i can finish everything off. then again, there's actually quite a bit to finish off- cable management, the top panel, filling and leak testing, installing stuff and stress/heat testing... oh well, one step at a time.

so, i more or less finished work on the dome.

it's kind of hard to see in the pic, but there's 24 screwholes drilled in there. lots of scratch marks from me making sure all the bits of metal were all gone. at the top, there's a flap of metal i bent over, for all the cables exiting my case. i had originally planned many separate holes, but then i realized that the power cord/plug would also have to go through. and there would be lots of cables: keyboard, mouse, LAN, monitor, audio, power, IR receiver, gamepad... possibly a few more that i'm forgetting. then, of course, came the problem of upgrading and troubleshooting... anyway, it would have been a total mess. i decided to just open up a flap at the bottom, and let all the cables spill out from there. much easier.

then i mulled for a bit, wondering about how i was going to make a foot-long cut in a stainless steel bowl. i had just about decided that it was impossible (with the equipment i had on hand), when i realized it was completely unnecessary. i just had to fold it over- and no sharp edge to deal with!. i ended up doing it using two clamps to bend it most of the way, and a hammer to finish it off. it actually turned out pretty well (on the outside).

however, there was still the problem of sharp edges to deal with. cables would be running out of this hole. the folded edge took care of one, but the other two had to be dulled. i sat in the garage filing away for an hour or so, and once again i realized that it was both futile and unnecessary. back to the home-made acrylic cement i went. i dug deep with the syringe, grabbed the really thick and jelly-like stuff from the bottom, and just slathered it all over the edge. after it dried, it would completely cover the sharp edge, and i could sand down any excess. woot!


with all that done, painting was next. nice smooth white to match the front bezel:


and finally, attaching the big black "clampy" things i made oh so long ago and nearly lost by forgetting that i had dumped them in some random corner of the living room underneath a huge pile of other miscellaneous unused computer equipment. that was merely a sentence fragment for those of you who like to keep track.


next up, the last few little details i needed. first, was the hard drive "pods."
somehow, i had to fit six hard drives inside this case. "crowded" doesn't quite describe it all. i managed to squish them in as two sets of 3. anyways, to hold them in, i didn't really have enough space to properly screw them in anywhere. i decided to just make a little pod for them to sit in, lined with black foam.

and there they are, inside the case:

as you can see, the one on the left makes for a VERY tight fit. i'm sure the one on the right will also be a very close shave once the video card's installed.

lastly, i just had to fix up the front bezel a bit. to make room for the cabling, i had to stick the drives out a bit. this meant i had to add a little border beside the drives, to make sure they didn't slide backwards, and so that it's not ugly.

two little pieces of acrylic did the job quite well.

and that's it for now!

 

well, it's been a while. i'm nearly done! most of the past 2 weeks has been just waiting for parts to come in. first off, i got my hard drives:


yippee!

now, cabling was pretty hard. with very limited space to work with, i ended up pulling the SATA cables out of their sockets every time i turned my head. thus, i decided to fix it like so:


since the "back panel" was actually inside the case, it also made for a really tight fit. luckily i had a 90-degree angle plug lying around for the one (microphone) jack that wouldn't fit.


after i did this, it was another week of waiting for the final parts...

 

Radeon HD 4850!

oh, and a blu-ray drive. i heard it ran pretty hot, but that was because it's on a single-slot cooler which ran slow and quiet. watercooling will fix that easily... and hopefully give lots of overclocking room.

i had loads of fun spreading the liquid metal around:


and in case anyone's wondering... a Danger Den Maze5 block fits the 4850 perfectly.


some of the RAM blocks had to be cut down, though...


...aaand the finished product. time to stick it inside.

i have to rant about these ramsinks for a bit.... they suck. they are high quality, pure copper, but the adhesive pads stink. i followed the directions carefully, and only about 2/3s of them actually work. the rest kept falling off at the slightest jiggle. never trust the packaged thermal tape.

with that, the internal part of the watercooling loop is complete!


and now, at the very end, i can finally install the external part of the loop... which is what this entire build is based upon.

 

there's half of it! the first half really scratched up my fingers and took me almost an hour and a half to do. i have also learned to avoid compression fittings. they sound convenient- nice and secure, no hose clamp needed, whatever. in reality, they are damn near impossible to screw in unless the hose is coming out perfectly parallel to the barb. the slightest angle/curve in the hose made a simple 5 second job into a 5 minute struggle. and with 24 compression fittings to go through for these 6 cooling rods... i was swearing like hell by the end of it.


fortunately, i learned a bit from the first 6, and the next 6 went by much faster. still a decent amount of swearing involved, but probably less than an hour of it.

time for filling! i used PC-ICE non-conductive fluid. i don't really trust its non-conductiveness, but since it's not water, most forms of life will have a hard time growing inside of it, plus it's a lubricant! and it's blue. yay. for some reason the camera captures the blueness much better than it is in real life. i think it's the first time my camera's made something look better than it actually is.


aaaaand my first leak. at least i was fortunate enough to have found it before i turned on the pump, and that it's relatively high up, so i didn't have to drain the whole thing to fix it. the culprit? the splitter block. i thought it wasn't tight enough, so i tightened the fittings on to it even more. still leaked. a little more tightening.... and i heard a crunchy noise. taking the whole thing apart, it seems like threading in the acrylic had turned to powder. good thing i happened to have a delrin splitter. it's doing much better.


now i'm just doing the full 24hr leak test, and getting all the bubbles out of the system.


there was one hiccup that really freaked me out. the 12 cooling rods (6 per side) are linked in parallel, with the water entering at the top of the first rod and exiting at the bottom of the last one. i did this so that the hottest water would be at the top, and the coolest at the bottom, theoretically making for better passive convection cooling. here's the snag: they all drained down the first 3 or so, with no water flow going to the last 3. i couldn't think of what to do, short of draining the entire thing and redoing it so that it entered at the bottom and exited from the top. it *should* flow equally to all fo them- if there was an equal amount of water pressure in all of them. unfortunately, the air content stopped that from happening.

then it hit me- if i poured coolant directly into the last rod, that would instantly increase the pressure as well as get rid of the air, forcing water to flow through all the rods. so i unscrewed the fitting on the final rod, stuck a funnel in there, and kind of hung the fitting over the funnel as well... and started pouring. after only a couple mL, the coolant was pouring out from the previously-empty fitting. i somehow managed to screw it back in without much leaking.

unfortunately, i was too excited at getting it to work and saving myself from a nightmare of redoing my entire loop that i forgot to take pictures.

so now, i'm just running it to get the last few bubbles out and to test for any slow leaks.

 

so, apparently that last little technique didn't get rid of all the air... at least, not all in one go. i had to repeat it 2-3 times, turning off the pump in between each, before i could get the last few bubbles out of the system. so, i took pictures this time.

here was the problem:

that's with the pump running. as you can see, water was flowing up to the second last rod in the pic, but not the final one.

and here's the solution:

i turned the last rod into a makeshift fillport...

and added extra coolant until it was running through to the last fitting at the top. i had to repeat it 2-3 times because, while the pump was running, it added a bit of extra pressure and allowed the air bubbles to stick around... somewhere... in any case, when it was off, the coolant would just drain back down a bit, and it took a few times before it was truly topped off.

now for some final shots:




night shot, UV cathodes turned on


and nightshot from above, UV on.


yay! and i made this final post on cryocore. there's just one issue, my HD4850 crashes under load. it's certainly not a temperature problem, since it's running at about 45 C max under load, a good 35 degrees less than the stock cooler. nor is it lack of power, since i have a corsair HX620 powering the whole rig, which is more than enough. i certainly hope i didn't damage it, but since many other people are having similar problems, i'm guessing it's a driver issue. just waiting on the next driver release (the 4800 series isn't even listed under AMD's official site yet!) to see if this gets fixed up.

and, on a final note, the cooling system works! i can't believe it! everything except the PSU is being passively watercooled via those 12 rods. my first stress test showed the CPU and GPU temps never going above 50... and the day i did it was one of the hottest days in june ever recorded in my area. i'm surprised how effective it is. the system is almost dead silent. if there's nobody else in the house and i lie down quietly, i can still hear a fait whir of the pump, but other than that, it's practically noiseless. mission accomplished!