Quad Head Dual Rotary Cooler Project

[Gray Mole]

Thanks guys

The displays are very accurate thankfully. I have had one of these be 5c off so it's nice to get a pair matched. I mean, these are more about reference to boot up the pc and there's 4 K plugs out the back for real diagnostic measuring but having them be this close is great.

One is around 1-1.5c warmer than it should be. The other is more like .5 colder than it should be.

Since I'm not that picky about these that's fine with me. Cold and warm about the same variation now.

I'll have to keep that issue with power in mind though. The cascade that has one of these reads off but I also ran the wire by the pipe for a few inches. If in future I make sure I get distance from the pipework maybe I won't get interference and they'll be more accurate. At least they don't blank out like the 5v displays were. That was really annoying

Can't lie, dealing with captube problems can be a real nightmare.

For the brazed or soldered stuff it's easier to just scrap the pipe and cap and replace it, whether you 'tag' the new captube in or run it all the way to the filter. Can lead to a lot of waste when you're playing with tweaks and stuff but I think it's worth it for the advancements you can find. I use and scrap a lot of captube. Thankfully it's not that expensive.

It's hard to deal with evaps when you have a bad or blocked cap. Mostly it's easier to remove the evap from the system, drill and new hole and clean up blasting out any bits that drilling leaves as well as filing/sanding the hole so no little bits can break off and go into the system. Run the new cap and braze as normal, purging of course. More or less as much work as brazing a new head.

I had another prepped evap with a captube tail and I just tagged it in. Over the years I've gotten good at joining captube, so it's not as big a deal but joining can be tricky. Easy to overbraze/heat and end up with blocked cap.

Was pretty scary on the other evap too, only had 1.5" of cap tail to work with, if I'd blocked it I may not have had the length to redo it again and would have had to drill and run new.

Gray

[MaRtIe]

depending what resistance the sensors give out, you could possibly calibrate them yourself with a pot or resistor. i thought about it with this crappy one that come with the SS im working on now.

[Gray Mole]

Yeah I was thinking about that, especially before I swapped the probes. One was reading almost 10c higher than the other

Just a quick peek at what the sealer finish looks like. i need a few more coats before I'm happy with it, but it'll come out like this...



And yeah, I may still put a pot on one of the displays to equalise them. For a degree or 2 I'm not that concerned though. 5 or more, definitely.

Gray

[mandolin hot]

Thanks guys

The displays are very accurate thankfully. I have had one of these be 5c off so it's nice to get a pair matched. I mean, these are more about reference to boot up the pc and there's 4 K plugs out the back for real diagnostic measuring but having them be this close is great.

One is around 1-1.5c warmer than it should be. The other is more like .5 colder than it should be.

Since I'm not that picky about these that's fine with me. Cold and warm about the same variation now.

I'll have to keep that issue with power in mind though. The cascade that has one of these reads off but I also ran the wire by the pipe for a few inches. If in future I make sure I get distance from the pipework maybe I won't get interference and they'll be more accurate. At least they don't blank out like the 5v displays were. That was really annoying

Can't lie, dealing with captube problems can be a real nightmare.

For the brazed or soldered stuff it's easier to just scrap the pipe and cap and replace it, whether you 'tag' the new captube in or run it all the way to the filter. Can lead to a lot of waste when you're playing with tweaks and stuff but I think it's worth it for the advancements you can find. I use and scrap a lot of captube. Thankfully it's not that expensive.

It's hard to deal with evaps when you have a bad or blocked cap. Mostly it's easier to remove the evap from the system, drill and new hole and clean up blasting out any bits that drilling leaves as well as filing/sanding the hole so no little bits can break off and go into the system. Run the new cap and braze as normal, purging of course. More or less as much work as brazing a new head.

I had another prepped evap with a captube tail and I just tagged it in. Over the years I've gotten good at joining captube, so it's not as big a deal but joining can be tricky. Easy to overbraze/heat and end up with blocked cap.

Was pretty scary on the other evap too, only had 1.5" of cap tail to work with, if I'd blocked it I may not have had the length to redo it again and would have had to drill and run new.

Gray

Thanks Gray, changing a captube really makes me nervous, especially with the evaps that have a significant length of captube inside, like the stepper ones. Sorry to be off topic again, but is there any guide or good pics showing a "tagged" cap tube extension? If I were to give it a try, I would use a short segment of 1/16" copper tube, with just a big enough I.D. so that the cap tube can fit inside and the rest can be filled with brazing material.

[Gray Mole]

http://www.xtremesystems.org/forums/...Vapo-mod-Guide

If you scroll down you'll see how I do it in that thread.

I put in the captube into either 1/4" or 3/16" pipe and squeeze it into what looks like a 'tag'. crimp the ends but don't pinch the captubes. put the in so they're just barely not touching in the middle.

Then when you braze, only braze the ends, make sure that no or very little braze flows into your tag. That ensures it won't get blocked by braze.

Holds the captube in place when you pinch it inside a tag, and if it's crimped shut nicely you can flood it out with nitrogen, then return the pressure to zero and braze without mucking up the insides.

[sdumper]

Show off

Nice work

[V2-V3]

Impressive work there

Cant wait to see the mounting and insulation

[Gray Mole]

Your wish is my...Well I don't have a lot of choice now do I?

So, some mounting/plate/insulation stuff...


















The plates I've got more or less prepped. Studs installed. The heads just barely fit in the studs with the insulation now so they really guide themselves. Slide 'em in and the the plate and nuts.

The wingnuts are only there to make life easy for my side of things. Space issues mean between cards there's no room for that, so flat nuts and washers, and foam tape on the plates above. Should be as close to condensation proofing as can be done. 2 layers of tape outside, tape 'shimming' on the plate where it's machined low, then 10mm insulation sheeting. All glued together and airtight. Head mount contacts the inner layer, then the outer head insulation contacts the foam sheet. All very low profile. then the outer foam protects the studs since they're installed in the cold transfer plate.

On top will have another layer of foam to seal the top of the mounting plates, they could get cold, so they get insulated. I'll send it out assembled. Makes his job a bit tricky but it's too much assembly required otherwise.

His side of things for the cards he'll have to do. I don't have them here so can't measure and cut for everything on the card. I'll send loads of everything needed but shouldn't be too bad. Means putting the case up higher to make it easier to insulate for the cards.

Once these are on the cards, it should be no real issue. Put the cooler in place, slide cards in. More that enough flexibility to get the cards in.

I honestly think I'm overkilling the insulation a bit, but only on the top. The side tape is ideal but no more than I think necessary, the backing will be 'normal' but the size of foam on the surrounding ring, the sheeting, it's all bigger than needed but it's also not getting in the way.

From the 'cold' pic you can see why I'm not that concerned about balance. I wanted good balance but a couple degrees will disappear. Even a 50w load imbalance will be compensated easily but the plates when they cool down. Beyond that, they'll help slow the temp swings. If it's benching then it's amazing. 24/7 will still be good but heat transfer is compromised by a transfer plate, as always.

So yeah, promised some pics and finally doing it. The work takes a while but the pics do eat into the worktime. Sorry that the pics aren't in order, would make more sense that way but there's work to do and the daylight's burning

Edit: That foam ring is just scrap and it's not cut to size so it bulges out I'll be cutting to size shortly, when the system's warmer I'll be doing the final polish and install, so I'll do my stud cutting, foam sizing, etc. then.

Gray

[Gray Mole]

Oh yeah, I guess a few pics of the cooler IN the case would be good too









Hope those are ok, I know my pics always suck.

Just finishing up painting. Sand, paint, sand, paint, and hope it's pretty when you're done.

May finish today, may not but I'll sure be close

[n00b 0f l337]

Very nice. I've been working with waterblocks like that for a while, works very well, just need to get the mill running to be able to make copper tops to braze on. But I like your swapability!

[Gray Mole]

I toyed with that idea, just a big common evap for the whole card, but then if they change anything at all it's instantly obselete.

I am hoping this waterblock is ok when they do the 7990 setup and perhaps even future dual chip cards but you never know. I would be inclined to drill and mod it even if it's not. Just depends on other component locations.

It's working out better and worse than I wanted (as always) but for the most part it's as good as my target. Around -35 unloaded per head, around -10c at 300w per head.

The use of the cold plate is a mixed bag but for giving complete balance, it's ideal. The heads weren't any more than a degree or 2 apart for the most part but they'll be even better with it.

Projects like this aren't common and that's something I guess. They're a lot of work and often (at least with me) I don't charge anything like you should. But it's not just another cooler. A chiller would be similar but harder to deal with, and not as pretty and I suppose not as 'efficiently cold' since direct will always give a bit more.

Looking forward to putting the plate on for real now to test for condensation. I didn't even put paste on for that 'ice shot'. Nice to see the surfaces are so close that it's freezing up that well

Gray