The system is still in design and conceptualization stage, but I'm getting very close to buying the parts. Just need a couple more things figured out, and we're ready to go. The Step-By-Step page will contain pics.
Well, the refrigerent oils I looked at held their thin viscosity until -20 celsius (when they begin to firm up), but they didn't disclose the freezing point. I'm looking into using a synlube (polyol ester lubricant, click here) instead, but the problem is this stuff is expensive ($40/gallon with contractor's discount, which I don't have) and they don't list its viscosity. The freezing point on this stuff is listed at N/A (huhu).
I've been working immersion for a while now too. I have the refrigeration unit already, I picked it up a week ago, and the computer components, but I have been stuck on what coolant to use.
What I have gathered is that most oils will degrade components, seals, and component caseings within a year or so. Which is fine if this is just an experiment or you don't mind replacing your system every now and then. They are pretty thick at the temps we are looking at anyway. I just got an email today telling me to forget alcohol too, which came as a dissapointment since that was what I thought I would probably end up using. Eugene has a pretty impressive resume in the area of chemistry and cryogenics. I looked at some of his research and it deals with pretty serious alcohol chemistry, so I am gonna trust him on this. I found him randomly here. Knowing now that he started a cryogenics company and that he is in the ranks of the OC'ers, I will assume he has "expert witness" status and has considered all of this before I showed up, and regrettingly accept his advice. Thanks Eugene.
Moving on, DOT5 brake fluid is out too I think, it is pretty crude stuff. So, after getting rid of all the oils that I have found so far(mineral, transformer, machine....), alcohol, and brake fluid, that leaves on my list fluorinert. Maybe Syltherm, but I would need to contact Dow's techies about it before I pay that much or put my board in it. It is after all, just a silicone oil. I called 3M yesterday, but with it being good Friday, there were hardly any techs to talk to, so I only spoke to the operator. I ended up leaving a message for the most appropriate tech guy after getting a few too many voice mails. Hopefully he will call early next week. The operator put me in touch with my local rep for sales info too, but she was gone also. The op would not/could not give me pricing, and kept dodging the question.
Spring break is almost over, and so is my time. Eugene recommended trying to contact local recycling centers or the electronics industry for used Fluorinert. I will try to look into that next week, but the semester is going to start in full throttle. I just dismantled and "prepared" an old psu and container to use as a guinea pig in an immersion test. I am ordering a pelt to use as both the load and cooler, so when it gets here, I will need to have some fluid ready. I've got my fingers crossed, but I am gonna go get some drinks and forget about this for a while. It is, after all, the last night of break for me I need a camera too
As a P.S. There are several threads that deal with this, let's try to keep them together and just use one. I don't have time to keep searching for this stuff. Also, this lifted my spirits a little. What are we doing? lol
I've read that thread; not much was of use (rather, not much I didn't know already).
I have a budget of about $200 (remaining) so Flourinert is out.
Also, I'm going to be using a low-viscosity oil for this purpose. And changing the mobo out after a year is pretty much standard anyway, but if I were really concerned (which I am) I'd just seal around the all the capacitors, and other sensitive areas. The only things that need cooling on a motherboard (or any other card) are the heat producers, and capacitors aren't on that list.
I'll be posting Step 2.1 in a bit; gotta write it up. As for keeping this to one post, I'm fine with it, except for the fact that the other posts I've seen on this topic are about a year old... Can't have that.
Thanks for the interest guys, I'm glad to see there are still people out there interested in extreme cooling
This is an excerpt from the email from Eugene. I think if you could keep water out of the alcohol, by hermetically sealing it or using dessicants(which is/was my plan), it would be alright. But I am not sure how the components or PCB will respond to it physically. I don't know for sure what he means by aggressive.
What are you thinking about using to coat the board? Bodosi dissolved plexiglass and coated his psu with that, but it still did not work. Here is one correspondance with Bodosi where it is discussed, and another about booze, I mean alcohol. I was planning on using conformal coating because I already have some, but the coolant would have to agree with it and not dissolve it. Bodosi still gave the thumbs down, but I would like to try it. He went to an indirect water solution to cool his psu now, no more immersion for him it seems
Nice site, nice domain name. Lesson learned, never drive with alcohol in you or your car. Put it in your computer
Disclaimer: The emails and excerpts were chosen and edited for clarity and to compact the info.
Hypothetically, sure. It wouldn't even need to be distilled if the board is that well sealed. But I don't think I would take that risk. And if you use water, you will not be able to cool it below 1C without an antifreeze or similar additive anyway. I would like to reach -10C ambient, and maybe put some TEC's in there for the CPU's to get them colder. I don't know though, cause that will draw a huge amount of power. I think dipping it in a tub of conformal would work pretty well, if you used some dielectric grease appropriately, but it would be really expensive to get that much coating. I need a sponsor. Or a lottery ticket.