TEC/Peltier CPU Chilled Water Cooling

[toolmaker_03]

mine is the STC 1000 so it only has a 20A circuit this is a great find.

 

[4Ryan6]

mine is the STC 1000 so it only has a 20A circuit the is a great find.

I've been running it the better part of the day, and it works flawlessly!

The 30a relay barely gets warm under load.

 

[4Ryan6]

The DWH7016M remembered the settings I programmed into it yesterday even after the machine and thermal controller was completely powered down for the night, it obviously has a built in memory, and that's great!

Back to set it and forget it, and that's just the way I like it!

Booted this morning at 20.1c coolant temperature and it was at 10c in less than 20 minutes, office ambient this morning is 25c.

 

[rubix_1011]

I ran an STC-1000 as a PID for a 2000w eHERMS coil to maintain brewhouse mashing temps...didn't like having 1600-1700w being pulled through it all the time, so I switched to a REX PID with a Solid State Relay. I recommend something like this if the STC-1000 decides to give up the ghost (just makes sure if you use the SSR that the PID is SSR-enabled, some are not).

But I use the STC-1000 dual stage controllers for all of our kegerators and fermentation chambers...they are simple to use and work great...especially for the cost.

Edit: took a look at this later...noticed you linked the 30A version...most of the standard >$20 STC-1000's are only like 6-7A, so likely the biggest difference there.

Carry on.

 

[4Ryan6]

Edit: took a look at this later...noticed you linked the 30A version...most of the standard >$20 STC-1000's are only like 6-7A, so likely the biggest difference there.

Carry on.

Until I found the 30a most everything I had run across only went to 15a and that would never support a 20a peltier, I did consider buying one of those and swapping out the relay to a 30a.

That's when I stumbled across the one I linked to and bought it immediately.

FYI: FrozenCPU has never reimbursed me for my purchase they failed to deliver, even after they said they would, so I am just writing them off as being a bunch of liars and crooks!

 

[RyanML86]

Is there any reason to put a tec unit on a aluminum waterblock to chill the water in a pc water cooling system, instead of trying to make the place to put the peltier between the cpu/gpu and waterblock? Thats the idea i am having and add a heatsink and fan to the hot side to cool it down. Would you tell what you think of the idea

 

[computers_are_freaky]

Some people do, but this way is much safer, becuase if the cold side starts to cause condesantion on it, in one case it will just drip onto the waterblock, in the other it will fry all your hardware. Also, Ryans way let's you get things a lot colder, because you can put on more peltiers and have a cold storage reservoir, whereas the peltier directly on the CPU will cause serious problems and hassle on start up.
Basically, it works, but Ryans is the bigger (and better as far as temps and reliability goes) of doing things. I'm not saying you can't, because people have, just that you need to seriously look into it before just throwing a peltier between your CPU and cooler.

 

[4Ryan6]

Is there any reason to put a tec unit on a aluminum waterblock to chill the water in a pc water cooling system, instead of trying to make the place to put the peltier between the cpu/gpu and waterblock? Thats the idea i am having and add a heatsink and fan to the hot side to cool it down. Would you tell what you think of the idea

Regarding the Aluminum water block to my knowledge they are all channel designed to cool the hot side of the peltier, and you end by " add a heatsink and fan to the hot side to cool it down.", so what exactly is the Aluminum water block for?

Regarding direct mount, OCN it would be best to get information from those that actually do or have done that.

When you start gathering information at OCN, go past the Oohs and Aahs and read all the comments and posts in the threads like, "My chip died unfortunately", to me that's very interesting information because my chip is not expendable!

 

[4Ryan6]

Some people do, but this way is much safer, becuase if the cold side starts to cause condesantion on it, in one case it will just drip onto the waterblock, in the other it will fry all your hardware. Also, Ryans way let's you get things a lot colder, because you can put on more peltiers and have a cold storage reservoir, whereas the peltier directly on the CPU will cause serious problems and hassle on start up.
Basically, it works, but Ryans is the bigger (and better as far as temps and reliability goes) of doing things. I'm not saying you can't, because people have, just that you need to seriously look into it before just throwing a peltier between your CPU and cooler.

They do take precautions and insulate the motherboard socket, all around the socket, and in a tower setup the back side of the graphics card as well, because dripping water from the ice melting off of the peltier will drop down on the backside of the GPU, so it has to be protected.

To a certain extent I guess I'm kinda lazy, as I like the set it and forget it type of computer operation, of course lazy did not get my setup built, so maybe it's daily operational lazy, is the better choice of words.

 

[4Ryan6]

I have my own theories regarding direct die mounting but what has surprised me is their end result overclocks are not earth shattering, I was expecting record breaking overclocks and have seen some less than mine, I'm sorry but I have to scratch my head on that one?

When I was experimenting with ice water cooling, I decided one day to load up the cooler with ice and drop the coolant temperature very low and then boot the machine, to my surprise it took close to 5 consecutive boot attempts for the CPU to get hot enough to support the overclock/coolant temperature without crashing.

The CPU was 100% overclock stable but the CPU had never been cold booted with that low a water temperature and it took multiple boots to get it going, so after that I allowed the water temperature to rise overnight, booted with no problems and then added the ice to the cooler.

So it is a curiosity to me wondering how they control a direct mount, do they power the peltier first then boot?

I am only curious as I have zero intention of going that route, but the CPU produces heat as soon as it is powered up, the peltier does not produce cold instantly, the peltier takes a short time to produce it's cold so does the CPU getting hot cause the peltier to stall in that situation?

If the peltier is powered up independently first and is already cold before booting the machine, how many boot attempts does it take to get the machine running?

So RyanML86 you'll probably find here, that we have more questions ourselves than answers?

 

[toolmaker_03]

wow, ok some of this I can answer, they both go active at the start up of the system, but my setup had a water block cooling the hot side of the TEC.
as far as stalling, I never had the CPU stall the setup, but mine only ran for 6 months before the motherboard fried from condensation.

that was my fault, I should have covered the whole board with insulation.

 

[4Ryan6]

wow, ok some of this I can answer, they both go active at the start up of the system, but my setup had a water block cooling the hot side of the TEC.
as far as stalling, I never had the CPU stall the setup, but mine only ran for 6 months before the motherboard fried from condensation.

that was my fault, I should have covered the whole board with insulation.

What are you using to insulate the motherboard with this new sub zero build of yours?

 

[toolmaker_03]

it is my hope that the vacuum chamber will stay dry enough that I will not have any condensation occurring inside of it.

so, nothing :lol:

 

[schmuckley]

it is my hope that the vacuum chamber will stay dry enough that I will not have any condensation occurring inside of it.

so, nothing :lol:

IMO;You should probably insulate.

 

[4Ryan6]

it is my hope that the vacuum chamber will stay dry enough that I will not have any condensation occurring inside of it.

so, nothing :lol:

IMO;You should probably insulate.

I totally agree!

You've already lost one motherboard, I would not take the chance, I know you are in hopes the vacuum chamber will eliminate that possibility, but what if it does not.

 

[computers_are_freaky]

Couldn't you just use some spray on non conductive silicon? I've heard of other people using this method.

 

[toolmaker_03]

well that is a good point.
if the vacuum chamber does not eliminate the condensation, than there is no point in having it at all.
but I am confident that it will prevent condensation from occurring, the only concern is springing a leak in the vacuum chamber, allowing normal air to enter the chamber, this would be worst case for this build.
as the air would instantly freeze on every thing it touched, but that is why I have the camera inside the case, and there is a good possibility that even if this happened, I could turn every thing off, open the chest, dry every thing inside chest, close it and reseal it, vacuum it down, and every thing would still run fine.

 

[4Ryan6]

well that is a good point.
if the vacuum chamber does not eliminate the condensation, than there is no point in having it at all.
but I am confident that it will prevent condensation from occurring, the only concern is springing a leak in the vacuum chamber, allowing normal air to enter the chamber, this would be worst case for this build.
as the air would instantly freeze on every thing it touched, but that is why I have the camera inside the case, and there is a good possibility that even if this happened, I could turn every thing off, open the chest, dry every thing inside chest, close it and reseal it, vacuum it down, and every thing would still run fine.

How long will it take to pull a vacuum on the chest?

I ask out of curiosity of running daily, how much added startup time does this require?

Will the chest actually be able to hold the vacuum, or will there be slow leaks requiring the pump to constantly run?

 

[toolmaker_03]

that is a good question.
to start the ice chest will hold a vacuum for up to 3 days before the gage starts to move.(with no leaks)
I only run the vacuum pump for 3 min a day, just long enough to keep a vacuum state on the ice chest constant.

the part that takes time, is the first time I go to vacuum the ice chest for this system, I will vacuum the ice chest for 3 min a day, for 10 days, for the first time that I vacuum the ice chest only, I do this to dry the inside of the ice chest.
there is a humidity gage in the ice chest and I will show how the humidity in the ice chest goes down a little each day, after the chest is dry, as long as I keep a vacuum on the chest, It should stay that way.
I hope this answers the questions.
if not, I will try again.

 

[4Ryan6]

that is a good question.
to start the ice chest will hold a vacuum for up to 3 days before the gage starts to move.(with no leaks)
I only run the vacuum pump for 3 min a day, just long enough to keep a vacuum state on the ice chest constant.

the part that takes time, is the first time I go to vacuum the ice chest for this system, I will vacuum the ice chest for 3 min a day, for 10 days, for the first time that I vacuum the ice chest only, I do this to dry the inside of the ice chest.
there is a humidity gage in the ice chest and I will show how the humidity in the ice chest goes down a little each day, after the chest is dry, as long as I keep a vacuum on the chest, It should stay that way.
I hope this answers the questions.
if not, I will try again.

You can shorten that tremendously by putting a box of baking soda inside and it should absorb the excess humidity.

You can buy a box of baking soda that is designed to be put in a refrigerator with a peel off seal so the contents stay in the box.

 

[toolmaker_03]

I was going to use desi packs, but yes that is the idea, with the desi packs it only takes 6 to 7 days to dry the chamber completely. that is also with me only running the pump for 3 min a day, If I increased the amount of time that the vacuum pump is running, I could shorten the dry time, but I don't know how long is too long to leave the pump going, I know 3 min does no damage, but I do not know what longer periods of time under a dynamic vacuum could do. I know I am not explaining this correctly, but there could be issues that occur with longer periods of time under a dynamic vacuum, meaning the vacuum pump being active for longer periods of time than 3 min.
if this where a industrial model vacuum chamber, it would have heated walls, and I could dry the chamber in a couple of hours, with the vacuum pump on.

 

[4Ryan6]

Too much vacuum may crush your cooler like a beer can, I mean those coolers are not designed for what you are using it for, and you cannot change the outside air pressure, so you had better stay on the safe side.

 

[toolmaker_03]

:lol: yea, that is what happened to the first ice chest, but I think it had more to do with the 12 holes I had drilled into it for the wire, for the test build I kind of drilled holes as needed, and where the wires where, so as a result the chest became rather weak. so as I when to vacuum it down, the side of the chest cracked open and collapsed. scared me real bad at the time, because it was all quilt, and then a loud snap, and I knew that the chest had seen its last days of being vacuumed. so to correct this issue, I am only drilling one large hole this time, and I am drilling it in the same place that the drain hole for the ice chest already is.

 

[4Ryan6]

:lol: I'll bet that was a shock!

Probably pretty loud too!

 

[computers_are_freaky]

Just an idea here, in the event that the cooler can't handle the outside presure, even with the fewer holes.
Why not use 2 chests? One inside the other. Have your current cooler as is, but then have a second one at 0.5 bar all around it. You could use your micro-controller to alternate some valves or something and make sure you keep the presure stable in both cases. That way, the outside presure is much less. Would be much bigger and more points of failure, but it might be a solution, in the event that a cooler can't handle the presure.