Very interesting project! :D
Thanks for the link.
Printable View
Very interesting project! :D
Thanks for the link.
Here's a pic of the cut away, haven't installed the copper tube to direct flow to the die yet. This should make it easier to do.
Looks nice, i'd love to see a video of it once in action. I don't know if you could that but i would appreciate
I wouldn't be able to show the finished block as it will not be cut away like this one. I might be able to seal the mock up with some clear plastic and show the flow over the die, but it won't actually be connected to a mobo or anything....still might be cool to see. :D
Anyway, here it is with 1st positioning of the "stem" to direct fluid to die...may have to shorten it a hair and adjust angle a tad.
When you gonna try it, put it vertically to check how it is filled
I think I will try to put some clear plastic on it and pump some water thru it just to see where the flow hits and how different mounting positions effect.
Here is a shot of the inside...I won't braise it in until I am able to determine best length, angle...etc.
cool! you could make the flow visible by letting the block run with water, and then put some colour into the water. a video of that would look reaaaallly awesome. you would see where the colour goes first, and if theres a dead angle, or a place where the water almost doesnt move.
I have gone down this road before, but never got beyond the theoretical stage, since that's where this idea usually breaks down.
Some things you need to consider:
1. Your coolant will pick up ions from both your waterblock and your liquid/air heat exchanger.
2. It is not the traces you need to be concerned about corroding, those are easy to seal. It is the pins that connect the die to the organic package. Since the package is permeable, coolant will eventually seep under the die and corrode the pins.
3. Bowman recommended clear coating the entire CPU minus the surface of the core. The die is non-porous, and the coating will not adhere to it. Erosion is also a factor in stripping away the coating.
4. Erosion of the die itself is also a concern.
5. Waterblock: Ideally for heat transfer on a flat surface, you want multiple impingement jets striking the die at a perpendicular angle.
1. True, but this is a problem common to other conventional water cooling setups as well.Quote:
Originally posted by koslov
I have gone down this road before, but never got beyond the theoretical stage, since that's where this idea usually breaks down.
Some things you need to consider:
1. Your coolant will pick up ions from both your waterblock and your liquid/air heat exchanger.
2. It is not the traces you need to be concerned about corroding, those are easy to seal. It is the pins that connect the die to the organic package. Since the package is permeable, coolant will eventually seep under the die and corrode the pins.
3. Bowman recommended clear coating the entire CPU minus the surface of the core. The die is non-porous, and the coating will not adhere to it. Erosion is also a factor in stripping away the coating.
4. Erosion of the die itself is also a concern.
5. Waterblock: Ideally for heat transfer on a flat surface, you want multiple impingement jets striking the die at a perpendicular angle.
2. I will try to seal the organic material with several coats of acrylic clear coat paint up to the edges of the die. The under side of the CPU will be coated with dielectric grease as thick as possible.
3. I'm not coating the die with anything. I realize the coating will wear off in time. There may be other materials that will last longer.
4. Erosion of the die itself is inevitable, I'm afraid and see no fix for this. It should last for some time.
5. Some of the links in this thread have articles where they used multiple jets with no great improvements. I chose to keep it as simple as possible for now.
This whole thing started for the fun of doing it. I'm sure there will be mistakes made and the whole thing may prove to not be worth the effort as a viable everyday cooling solution. The things that fail, I will attempt to improve upon. I certainly don't expect to get it all right the 1st time if at all. One thing is for sure, I am enjoying the attempt and the discovery process. :D
Thanks for the very valid points that you made.
When do you think your going to have it ready to test TheDude ??
I'm anxious to see your results ! ;)
I'm cutting some plastic tonite, so hopefully will be able to see the waterflow very soon. As far as the finished block....It will be some time yet...hopefully before I have surgery on the 23rd. :D
Got it cut and mounted. It's going to be very hard for me to get the water flow on film, but I will try. You can see by this pic that I can't take closeups very well...I can see the spinach in your teeth from a mile away (27x real zoom) but no closeups. :D
Well....I just ran some distilled water thru it, but you can barely see it, just some movement on the die. It won't show up in a pic. It does seem to work well. Of course the test will be with the next block with a working chip in a pc so we can see some temps.
I am letting it run to test for leaks and see if anything gives out.
while running water through the block, put some colour in the res, so the first coloured water that enters the block will be visible, and you can see the flow. then of course the whole block is full of couloured water, but in the short time when the first coloured water enters the block you can get a good idear how the flow in that block is.
record this "first color flow" and play it in Slow motion this would be the best to see the flow ;)
Some colorant would help you to see the flow
koslov welcome to Xtreme :banana:
Dude, as you very well know Im not expert in this at all but im concerned about you filling up the entire waterblocvk with water. Personally, i think thats to big an area as you see the whole waterblock will be filled with water and the water escaping will be the cool stuff coming in as i presume your goingt lower the other pipe near the core, if you do this then itll be only taking aout that water and the other water will ehat up quickly but wont be taken out.
Hope you understand what I mean.
Craig
Not lowering the other pipe.;)
Ah good, just didnt want all the water to heat up, at least by not lowering other pipe al lwater can get out.
Do you think you'd get betetr temps if there was only a small area i.e both lowered and not a big area around for water to stay. I would think so as the water would be removed quicker.
Maybe something to try next time.
Craig
i also think that would work better... next time :D
Yep CCW idea is not bad, you could put the barb at a 90° angle
looking good my friend:D
ah you see im really a genius in disguise....obviously a very good disguise!Quote:
Originally posted by saaya
i also think that would work better... next time :D
Dude, I didnt read anywhere the flow rate of the oil over the core, sorry if its been asked/posted but what is it please?
Craig
Right now it's 1380 L/hr when in chiller it will be 800gph
There is a guy at another forum that I go to that has done this direct die stuff, maybe you can ask him some questions or whatever. His name is Zennzo (I think I spelled that right) and he is at pimprig.com
lol@ pimpin.org :rotf:
Im pretty sure he said pimprig.com
Craig
psychofunk,
Thanks, I need all the help I can get. :D
Came up with an idea for catching the coolant hitting the die on film. Going to have a hose connected to tank with dyed water, other end to block, then after it leaves block, it will dump into a seperate container, then be hand poured back into the tank.
That way it won't pool in the block and fill up and I can catch where it hits the die. Simple finger over the end of the hose to build pressure. I really want at least one pic of it, if not a short film at slow speed. Hope it works. :D
i know, but the first time i read it i thougt it says pimpin.org heheQuote:
Im pretty sure he said pimprig.com
updates?
Yeah...learning how to use this camera in slow motion video mode. Hope to have some pics soon....waiting on the new block to get here. Hope the oil arrives this week as I will be in hospital all next week. :DQuote:
Originally posted by CCW
updates?
It's Thursday morning. Looks like things are gonna be on hold for a while. I leave for Hospital Sunday and I have too many things to do to prepare for that...take animals to the kennel...pack....pay the bills up....all those little bothersome things you have to do when you will be laid up for a couple of weeks.
If I do manage to get anything done with this, I'll post it, but probably be a pretty dead thread for the next 2 weeks. By the time I'm back in the saddle, the special oil and new block should be here. Looking forward to having some fun with that! :D
Dont worry Dude, I will be here to *bump* this thread up! :D
Craig
i honestly wish you good luck dude! dont let em screw up your organs :)
Dude, instead of typing *bump* now and again in here heres a link i dug up, not sure if youve read it but it looked good:
http://www.spodesabode.com/content/article/directdie2
http://www.aoaforums.com/forum/showt...ght=direct+die
the latter has some good pics
Craig
EDIT - Dude, that AOAS thread is a long haul, over 16 pages long but is full of info and lots of greta pics of how it was set up.
Craig