TEC build log start

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

toolmaker_03

Honorable
Mar 26, 2012
2,650
0
12,960
http://imgur.com/cd15M1a
ok so I have finished all my testing and decided on a general build for this unit.
this build will be sealed inside of a chest so I will not have physical access to the computer case, I would like some input on the general idea that I am presenting, I use antifreeze so freezing the coolant is not a issue, and seeing as I am increasing the cooling capacity of my water cooling system, I have also chosen new clocks for my hardware, the CPU clock will run at 5Ghz, and the GPU's will be running at 930Mhz, I have already tested theses clocks and found them to be stable for my hardware.
 
ok so I have decided to save a little, and build a custom frame and mount system for the radiators, I decided to do this because, I realized that I needed to build a custom frame and mounting system for the TEC assembly anyway.
I will be building the frames and mounts for the TEC's and radiators out of aluminum angle of varying sizes from 3/8" to 1"
the bottom plate that all the different mounts will attach to for this build is going to be a piece of 1" ply wood cut to 28" wide by 72" long.
as far as modding goes, that will be all of it for this build.
so I need to finish building the reservoirs, thanks to ryan for showing how his was made.
I need to build frames for the radiators and TEC's.
last thing is to make simple angle mounts for the PS's and fan controllers.
wow, that's a lot of stuff to keep me busy for some time, I will have pic's as I complete each piece.
 
I forgot to mention that one large reservoir is much easier to insulate and though I could have insulated it more than I did, but for my needs I wanted it to be able in about 8 hours to return to ambient, Why?, so I would not have any cold boot problems.

My concern for you with the temperatures you are planning to run is cold boot problems, have you investigated into that subject?
 
I was looking at OCing with LN2 and I think they said cold boot bug was like -100 (I'm guessing °C, but it wasn't specified), but it fluctuates depending on other temps, model and so forth. That might just be for that cpu or only with notrogen, but still, I doubt you'll get THAT cold.
 
For my experience if you remember back when?

I was using the Coleman Cooler with 10 gallons of distilled water and ice in it, when I had added ice and pre-chilled the coolant down before booting, it took numerous attempts to boot before the CPU got hot enough to hold it's own, thermal shock to the CPU is something you will have to overcome.

It works completely problem free booting at ambient and then gradually lowering the coolant temperature the CPU is much more cooperative that way, and will run with the coolant temperature down as low as you desire to run it.

Keep in mind we are talking about a CPU designed to run in a range of ambient temperatures, ambient temperatures taking into consideration that human being can actually survive in those same temperatures, so whether a PC could actually boot totally outside in the weather at the north pole?

You may need to research LN2 users to find out how they boot with extreme cold, they may boot without the LN2 and add it once the CPU is operational, or use extreme voltage to boot, and that is not something I am going to do.
 
When I cold boot my machine each morning the coolant has risen to ambient, the electricity starts flowing to the peltiers upon booting and they chill and start producing cold, it usually takes about 30 minutes of natural ambient air (without my office AC running), and my coolant temperature has dropped to 10c.

I can drop it faster if I need to by running the office AC, or switching on the 3rd peltier TEC assembly (that was added later and not covered in my thread), which can drop to 10c in half the time.

My 5ghz overclock can idle at ambient but has to be at least 8c below ambient to support stress load and I normally run at 15c below a 25c ambient.

I hope that information helps you.
 


You need to understand that LN2 cooling is extreme cooling not for everyday use, those guys go through some CPUs, and usually have about 5 on hand to run their overclocks, they run voltages that would scare you to death.

Unless you had plenty of money and did not care how many CPUs you sacrificed to reach your overclocking glory goal.

I don't know of any that have applied LN2 to everyday use?

Liquid Nitrogen boils off quickly so it could get rather expensive attempting to use it daily.

The highest record holders overclock in teams, one is on the computer running it while the others are keeping the LN2 supplied.

At least that what I've seen from the competition videos I've watched.

I am actually referring to a cold boot, which has happened to me numerous times no where near LN2 temperatures, I just mentioned LN2 because of it's temperatures and TM_03 wanting to run sub-zero.

If he goes to cold boot his CPU hitting it with sub-zero coolant temperatures he will have booting failures, 2, 3, 4, or even more attempts to boot up before the CPU can handle those temperatures, and each CPU has different booting tendencies.

But in my experience both Sandy Bridge and Ivy Bridge had the same problems.

 
Yeah, you're right. Was just something I remember seeing/hearing. I hope no one uses LN2 daily, I mean shees, that would mean having to set up you cooling again every few hours. Loads of hassle :lol:

How would you go about calculating the heat loss of your res over time? That way, you could calculate within a reasonable margin of error how much to insulate your res'. If you figure it out or already know, please tell me so I can plan my own res in more detail.
 


That was over 3 years ago when that reservoir was made and insulated, and there was no calculating involved in insulating it, just acquiring what would actually work good to insulate it.

The thread covers the actual build of the reservoir but not the insulating of it, insulating is totally dependent on the moisture resistant insulation material that's available to you where you live?

The insulating material was like a rubber foam and actually added about a 3/8th of an inch or 10mm full insulation layer to the outside of the tube, and that works great.

 
No, I don't mean in your case, I was wondering if there was a formula that uses the thermal properties of a material to calculate how much heat would "leak" per mm thick layer, that way you could find the optimal thickness of insulation. Of course you can just insulate it, that will work, as proven by you and arthurH, but my question was for optimal insulation and whether or not a formula existed to calculate it. Yeah, there would be a lot of variables, but it was just a thought.

I ask mainly because I "plan" to also make a custom res, however I want a window in the front of mine but also insulation, so I was planning a similar concept to double glassing windows, but with plexi and was curious to see how much air I would have to leave in between the sheets for good insulation.
 
What do you need a window on the insulated reservoir for?

To know the coolant level?

Once all your installation leaks are resolved the coolant level won't suddenly drop and if you desire to check it take the top off and look.

The only calculations I did were volume to know what size and length of PVC pipe would hold the coolant volume I needed it to hold.

I am a full fledged member of the KISS club, Keep It Simple Stupid! :)

If you have to have a see the coolant use a sight glass, it is the exact same concept of a very large style percolator coffee pot, the sight glass tube shows the coffee level in the pot, the same concept will work with a reservoir and not compromise the cold too much.

It is a useless addition IMO, and totally not necessary, but it is your project and your prerogative.
 
cold boot, CPU does not start at 0C, I did not have that issue as of yet, because I always started the TEC at the same time I started the system.
with my test rig I always shut it down after I was done mainly because it was covered in ice, and I did not want to create any more, as it would turn to water as it melted, but after a night, 8 hours of sitting there off, the water was always back to ambient temps by the next day.
with the final rig, I was planning on getting the insolation thick enough that condensation did not occur, so that I could run it 24/7 like I do now.
but until I have that at the right thickness, I will have to shut it down each day, and I will have to wait for my coolant to warm up before I can start it again.(with insolation that could take as long as 5 days)

thanks.
 
thoughts?
well I was really considering what you have stated and my water pumps are on a separate PS so I always left my pumps running after shut down for the hot side to cool down I would think this also served to warm the cold side.
Question lets say I am running at 0C on the water temp for the cold side at idle and I plan to shut down, if I turn the TEC's off could I heat the cold side water to 20C with the hardware, before I actually shut the system down?
I know that I can, I am really asking if you think that it will happen to fast, or if it will take a few min for the hardware to warm the water?
 


I would think that would work OK because your water blocks can only transfer heat or cold to the coolant at a certain rate, so it would not be like shocking the hardware, but without power to the peltiers the heat will transfer through the peltier.

Expansion and contraction are going to happen, but the long term effects of that, are an unknown variable as to if it will have negative hardware effects?

Unfortunately, You'll just have to find that one out the hard way.

 
Now this I have studied (a few days ago in fact)
Q = M *c * delta T
Basically:
Q = energy input
M = mass of water
C = thermal constant or specific heat or something, water is 1 calorie over gramms*°C
Delta T is the delta temperature of the water

So, multiply the mass of water (was it said before that 2 litres was ideal?) you have in gramms by 20°C to reach room temperature by 1 cal over grams*°C(1 cal is the same thing as 4.186 joules) and it will give you 167440Joules

Now, watts = j a second, so now divide 167440J by the TDP of your hardware at idle and voila, you have the amount of time it will take in seconds (the unit of measurement, it sadly took a lot longet to learn :lol:)
 
If you run the loops without the peltiers on, the peltier function will invert, now producing power. I don't know what in influx of power would do to your PSUs. You could maybe sort this by attaching a small chargable battery that you flip a switch and connect when you turn off the PSUs, that way you are charging the battery. Don't know if this will work, my knowledge of electronic circuits is limited, it's just an idea
 
the units do have a power surge protection on them, so I would think that it would prevent the return power from the TEC from damaging the PS, they are designed for travel trailers so there is a battery in the circuit on those, so it should be no issue for me to add one on this circuit at all, I will look into getting some small 12V batteries for this build and give it a try.
thanks, I never considered this might be a issue, but I understand how it could be.
 
back to system running temps, its is not like I plan to run this system at below 0C temps, the goal is to be able to clock my hardware further while still keeping the same load temps I have now, my hope is that at load all my hardware will be running between 25C and 50C. so at idol I may have to shut down one or even two of the TEC's to keep the hardware running at or above 0C on temps, I will pay close attention to the hardware temps at idol to begin with, and find out what that right setting for the TEC's cold side is with the hardware at idol. I can adjust the water temps on the cold side by slowing down the fans and pumps on the hot side, and letting it heat up, I can adjust my water temps on the hot side from 30C to 70C with slowing the fans down or even turning them off, and turning the speed controllers down on the pumps.
I hope that will be enough variance to properly control the TEC's cold side.
 
loop design the cold side is all fittings and soft copper tubing so that it can handle temps below 0C.
the hot side is all fittings and 3/8 ID 5/8 OD norprene tubing, its a smaller version of car radiator hose, it can handle temps of 100C +.

I started collecting this stuff about 3 years ago and using it in my builds, I used to be a toolmaker for industrial machinery like CNC's and bonders, now I do testing of electronics and minor repairs. so I have gotten to see a lot of different types of industrial tools used for manufacturing, well almost anything.
so I would guess that this is why I decided to use the vacuum chamber approach, to the issue of condensation.
it does need to have a chamber dry, done to it before the first hardware start up, but after that it should be a self contained dry atmosphere for then on, as long as I maintain a static vacuum on the chamber, meaning that I turn the pump on once a day for about 3 min.
 
Hey Ryan, to answer your question, I will use a real world example:
Why do people mod cases?
Why paint hardware?
I am devoted to getting the best bang for the buck for whatever I do (I do not choose what I do based on cash, but whatever I do do I try to get the best bang for the buck), but I'm not above spending a few extra bucks to make things look much better. My "current" rig (will be assembled tomorrow but will have fans and ssd arriving in january) is housed in a brandless budget case I got for free from a scrapheap, howerever I am willing to invest a few days and a few bucks to make it look awesome, therefore I would do it for any future pc as well. My plan is to get a friend of mine with a CNC machine to cut me the pieces and then I'll glue them and insulate it. I have priced it out at about 20€ for everything, but it will be so much more awesome with a window and an awesome paint job (I have friends in art school to help me with that too).

In short, awesomeness is important to me, maybe more so than time. I will undoubtedly end up eating my words, but at the end of it it'll be worth it. Of course, my initial rig won't have this res, I will add it later. If you want, I can post the 3D model (will have to be mid-January though, going on holiday day after tomorrow).
 
tools used for this project.
I will give a tool list for each section of the mod work being done, as there are a lot of tools that are used for creating this build.
the tools for mods being done to the PC case
drill, drill bits of varying sizes, air compressor, dico cutter, cutting wheels, C-clamps, small metal files, and stop blocks.
the tools for building the reservoirs
drill, drill bits of varying sizes, hack saw, thread tap, sand paper, and small metal files.
well, I will have pics when I am done.
 
ok, so with this build I will be removing the cover on the 3930K CPU, to attach the water block right on top of the die, I have been looking at the liquid metal pro, for a TIM between the water block and the die, and was wondering if anyone has used this stuff before? I have seen it done on you tube, but still I use AS5 and have never tried this type of new stuff.
 
ok so I have a question about some of the info that I am presenting when doing the mods to this build should I be more descriptive about what I am doing or simply generalize the info like I have been.
let me give a example of what I mean, while making the reservoir I needed to drill holes in the plastic for the copper hardware, then tap that hole with a tap, so that the copper fittings could be screwed into the plastic. should I be explaining how I used a pair of calipers, on the copper fittings to decide what size drill bit I needed, so that I could tap it with a tap for threads?
or is this type of information inferred and not needed?
 
Status
Not open for further replies.