-
You heard it here first: Black Friday 2024 has arrived!
- Status
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
- Apr 14, 2002
- 20,603
- 6
- 52,915
Clay may just flat out not work under vacuum pressure it may eventually pull through so here is what I would do.
The hole that is cut in the cooler I would remove at least a full inch of the Styrofoam filling around the cut area creating a 1" cavity perimeter inside the 2 inner and outer cooler walls.
I would cut a hole in a working surface like a wooden table or piece of wood large enough to accept the wire bundle going through it. and also large enough the entire cooler can sit on it on it's side.
Clean all the wires going through the cooler opening so there is absolutely no residue from the wire production process left behind on the wires so adhesive will completely adhere to them.
Run the wires through the cooler exactly where they are supposed to be, use something like a slick plastic plate cut to fit around the wires on the outside of the cooler, use only the flat part of the plate so a good seal can be achieved, use something like Gorilla tape to fasten the plate to the outside of the cooler, and use enough of the same Gorilla tape to seal around the wires and the plate.
Turn the cooler on it's side over the hole that was bored in the wood surface, then simply mix up a large batch of clear marine grade 24hr setup epoxy pour it in around the wire bundle and completely fill the void, it will flow into the area of Styrofoam you removed and when completely hardened will also strengthen the cooler in the process.
The plastic plate taped outside will keep the epoxy from leaking out, if it doesn't around the wires, you could use a thin layer of 5 minute epoxy just enough to seal any leaks and then use the 24 hour epoxy after, but if you are particular enough you can completely seal it all with the tape.
Fill the void completely until it is fully filled to the inside cooler wall, then leave it overnight to cure, it will be an absolutely permanent solution to the wire pass through.
As an explanation, Does that make sense to you?
The hole that is cut in the cooler I would remove at least a full inch of the Styrofoam filling around the cut area creating a 1" cavity perimeter inside the 2 inner and outer cooler walls.
I would cut a hole in a working surface like a wooden table or piece of wood large enough to accept the wire bundle going through it. and also large enough the entire cooler can sit on it on it's side.
Clean all the wires going through the cooler opening so there is absolutely no residue from the wire production process left behind on the wires so adhesive will completely adhere to them.
Run the wires through the cooler exactly where they are supposed to be, use something like a slick plastic plate cut to fit around the wires on the outside of the cooler, use only the flat part of the plate so a good seal can be achieved, use something like Gorilla tape to fasten the plate to the outside of the cooler, and use enough of the same Gorilla tape to seal around the wires and the plate.
Turn the cooler on it's side over the hole that was bored in the wood surface, then simply mix up a large batch of clear marine grade 24hr setup epoxy pour it in around the wire bundle and completely fill the void, it will flow into the area of Styrofoam you removed and when completely hardened will also strengthen the cooler in the process.
The plastic plate taped outside will keep the epoxy from leaking out, if it doesn't around the wires, you could use a thin layer of 5 minute epoxy just enough to seal any leaks and then use the 24 hour epoxy after, but if you are particular enough you can completely seal it all with the tape.
Fill the void completely until it is fully filled to the inside cooler wall, then leave it overnight to cure, it will be an absolutely permanent solution to the wire pass through.
As an explanation, Does that make sense to you?
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
yes, that is not a bad idea, and your right, that would work.
epoxy does not leave any room for error correction, but I don't want to have to keep redoing the seal on the ice chest either.
I guess this would fall into the same category as the military connectors, if I have a connector that I just can not get sealed, than I will use a set of the air tight military connectors to solve the issue.
same goes for this I will try the clay and silicon, but I fully understand that it may be too weak for this application, if so, I will most likely go for the epoxy idea, as I know it would work.
epoxy does not leave any room for error correction, but I don't want to have to keep redoing the seal on the ice chest either.
I guess this would fall into the same category as the military connectors, if I have a connector that I just can not get sealed, than I will use a set of the air tight military connectors to solve the issue.
same goes for this I will try the clay and silicon, but I fully understand that it may be too weak for this application, if so, I will most likely go for the epoxy idea, as I know it would work.
- Apr 14, 2002
- 20,603
- 6
- 52,915
You're right epoxy won't leave room for error correction but all your bundle should not need changing, if you had to swap out a peltier or fan you'd do that with the already available feed wire, and the epoxy would seal that opening and be a strong vacuum seal, reinforcing where you weakened the cooler by having to cut into it.
You've already had a cooler fail under vacuum, remember?
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
yes, this is true, and I do need the rig to be strong, and that hole needs to not give out on me. the wire going through it will be handled all the time, as things are plugged into the system, and unplugged. so if a seal was to fail on this build, this would be the one to do it. so your suggesting I use the real runny liquid epoxy that comes in two parts, that mix as it comes out of the duel tube container. and I thought that drilling the hole through the ice chest was scary, this tops that totally. so all I need to do is about 16 more wire checks, and epoxy the hole :lol:
- Apr 14, 2002
- 20,603
- 6
- 52,915
The runny epoxy is still thicker than you may be thinking is not like water, but it will fill the voids over it's cure time that's why I recommended the 24 hour cure type as it will have time to get into all the cracks and crevices.
One caution to share:
Test the epoxy on the Styrofoam you get out from around the hole of the cooler, to make sure it has no side effects on the Styrofoam, I know it won't get a reaction from the plastic cooler walls, but it might from the Styrofoam?
One caution to share:
Test the epoxy on the Styrofoam you get out from around the hole of the cooler, to make sure it has no side effects on the Styrofoam, I know it won't get a reaction from the plastic cooler walls, but it might from the Styrofoam?
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
cant put this stuff on any kind of foam, at all.
it's like putting gas in a foam container, it destroys the foam.
the good thing here, is that the area I drilled through is only about 3/16" thick, and it is all plastic, the foam is about a 1/2" away from the hole, and covered by the ice chest plastic. I drilled the hole in the same place that the drain hole once was for the ice chest.
it's like putting gas in a foam container, it destroys the foam.
the good thing here, is that the area I drilled through is only about 3/16" thick, and it is all plastic, the foam is about a 1/2" away from the hole, and covered by the ice chest plastic. I drilled the hole in the same place that the drain hole once was for the ice chest.
computers_are_freaky
Distinguished
- Aug 11, 2015
- 601
- 0
- 19,110
What about linig the inside of the chest, on the edge of the foam, with something else, like a sealent or something, then using the epoxy. That way you get the same effect with no epoxy-foam contact.
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
ok so reality speaking here, the TEC's that your using in this rig are 12726's these are the same TEC's I am using?
I ask because it is stated that the amp draw on these TEC's in your rig is 26 amps.
so, do you think that 3 of the same TEC's that you are using, if keep a little cooler, would be enough to handle my system?
if not, I do have 3 12730's that I could add to the system, for a total of 6 TEC's, and I know that would do it, but I would like your opinion, because you have been using them for some time now.
I am starting to understand these TEC's a little better, as well as how every manufacture seems to have a different rating system for there TEC but the reality is that I have TEC's now from three different manufactures and it's the amp draw that the TEC is capable of handling that determines how strong it is.
a 12726 is a 12726 no matter what the manufacture wants to call it, there may be a little variance of about 30W from TEC to TEC, from the different companies, but they can still only do so much with there product to make it a little better than the other guys.
I ask because it is stated that the amp draw on these TEC's in your rig is 26 amps.
so, do you think that 3 of the same TEC's that you are using, if keep a little cooler, would be enough to handle my system?
if not, I do have 3 12730's that I could add to the system, for a total of 6 TEC's, and I know that would do it, but I would like your opinion, because you have been using them for some time now.
I am starting to understand these TEC's a little better, as well as how every manufacture seems to have a different rating system for there TEC but the reality is that I have TEC's now from three different manufactures and it's the amp draw that the TEC is capable of handling that determines how strong it is.
a 12726 is a 12726 no matter what the manufacture wants to call it, there may be a little variance of about 30W from TEC to TEC, from the different companies, but they can still only do so much with there product to make it a little better than the other guys.
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
http://www.datasheetlib.com/datasheet/1210848/tec1-12715_hb-electronic-components.html
just for the info the manufactures of TEC's do the testing of there products inside a vacuum chamber, for lots of reasons.
same goes for a lot of the computer products, there tested while inside of a vacuum chamber, for lots of reasons.
just for the info the manufactures of TEC's do the testing of there products inside a vacuum chamber, for lots of reasons.
same goes for a lot of the computer products, there tested while inside of a vacuum chamber, for lots of reasons.
- Apr 14, 2002
- 20,603
- 6
- 52,915
26 amps at 16.6v DC, mine are only getting 12v DC so they are pulling 20 amps @ 200w, mine do not go sub zero because I don't let them go that low, but even with the limitations I've got on mine, it has been down to 4C, with all the TECs powered and if I had not caught it when I did and shut down the 3rd TEC it would have continued dropping.
If you supply them their spec'd voltage of 16.6v DC, they will take you sub zero very fast, I guarantee you!
Mine are getting 4.6v DC less than they are spec'd for, if I supplied the full 16.6v my air coolers would be in trouble, forcing me to run 2 fans in push/pull wide open all the time increasing the noise level.
That is IF, my air coolers could handle that much wattage, it is very difficult finding the actual TDP rating of the original Thermalright Ultimate 120 Extreme cooler IE. the Thermalright TRUE.
The supplied 12v allows my setup to work as it is configured for, but the peltiers are capable of more, with more voltage applied to them, that additional 4.6v will shock you with what they can do!
I hope that answers your concerns?
- Apr 14, 2002
- 20,603
- 6
- 52,915
I do however know the TDP for the newly designed Thermalright True Spirit 140, which is 360w TDP, which is going to be the coolers that replace the Thermalright 120 Trues I am running now.
I am going to replace them in hopes the new heat pipe design with it's obvious better cooling layout of the heat pipes through the cooling fin field, will allow me to use them with the cooling fans they come with, instead of the 110cfm cooling fans I am presently running, which should give me the same cooling capability as I presently have, with lower speed fans, meaning lower total noise.
My noise level presently is far from intolerable, but if I can lower it, I will!
If I am going to modify my setup at all, I will get three of the Thermalright True Spirit 140s and do it all at once.
Plus I do have additional improvements planned, as I have a new insulated reservoir design on the drawing board, but it will require replacing the Azza 1000 case presently housing 2 of the TEC assemblies, as that case is not wide enough to house the new reservoir.
So it will literally be a major overhaul of the entire cooling setup!
The power supply powering the peltiers is going to be retired, because it is approaching 10 years old, I have already bought it's replacement, I will also transfer out the 2 active D5s with the brand new shelved backups and let the used D5s be my backups.
I'll also be replacing all of the tubing runs in the chilled water and the radiator graphics cooling, which I also already have on hand, I got tired of buying 10' here 20' there, and usually some of it was crimped to the point it couldn't be used, and bought an entire 100' roll of tubing, so that expense is already covered.
I am going to replace them in hopes the new heat pipe design with it's obvious better cooling layout of the heat pipes through the cooling fin field, will allow me to use them with the cooling fans they come with, instead of the 110cfm cooling fans I am presently running, which should give me the same cooling capability as I presently have, with lower speed fans, meaning lower total noise.
My noise level presently is far from intolerable, but if I can lower it, I will!
If I am going to modify my setup at all, I will get three of the Thermalright True Spirit 140s and do it all at once.
Plus I do have additional improvements planned, as I have a new insulated reservoir design on the drawing board, but it will require replacing the Azza 1000 case presently housing 2 of the TEC assemblies, as that case is not wide enough to house the new reservoir.
So it will literally be a major overhaul of the entire cooling setup!
The power supply powering the peltiers is going to be retired, because it is approaching 10 years old, I have already bought it's replacement, I will also transfer out the 2 active D5s with the brand new shelved backups and let the used D5s be my backups.
I'll also be replacing all of the tubing runs in the chilled water and the radiator graphics cooling, which I also already have on hand, I got tired of buying 10' here 20' there, and usually some of it was crimped to the point it couldn't be used, and bought an entire 100' roll of tubing, so that expense is already covered.
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
4Ryan6
The power supply powering the peltiers is going to be retired, because it is approaching 10 years old, I have already bought it's replacement, I will also transfer out the 2 active D5s with the brand new shelved backups and let the used D5s be my backups
I like this idea a lot, I think that I will adopt this to my rig as well, after 3 years of operation I will put new pumps on the cold side of the loop, but I will most likely put the old pumps from the cold side on the hot side and shelf the old hot side pumps for backups.
thanks for another great idea.
The power supply powering the peltiers is going to be retired, because it is approaching 10 years old, I have already bought it's replacement, I will also transfer out the 2 active D5s with the brand new shelved backups and let the used D5s be my backups
I like this idea a lot, I think that I will adopt this to my rig as well, after 3 years of operation I will put new pumps on the cold side of the loop, but I will most likely put the old pumps from the cold side on the hot side and shelf the old hot side pumps for backups.
thanks for another great idea.
- Apr 14, 2002
- 20,603
- 6
- 52,915
@ toolmaker_03:
The shelved pumps should still have plenty of life left as long as the ceramic bearing and bearing cup are in good condition, so don't just assume the pump is OK, check inside the bell housing of the motor winding body to be sure the back side of the impeller has not touched the inside of the housing anywhere, you can plainly see scarring if it has.
The magnetic field produced by the main motor body holds the impeller dead center in magnetic suspension, it is the spindles length with the ceramic bearing bonded to the end of the spindle and the impellers bearing cup, that sets the operational gap between the bell housing and backside of the impeller.
The D5 pump IMO is a pure design of genius and engineering, the way it works is literally flawless as long as the bearing surfaces are good.
Failures are usually preceded by increased pump noise, if on your inspection you see any scarring inside that motor bell housing the bearing surfaces are wearing out and the pumps life is getting very close to the end.
You almost have a perfect case study situation toolmaker_03, to observe pump bearing longevity under extreme hot and cold running scenarios, but to be a fair test, the study would need to be started out with brand new D5 pumps.
The shelved pumps should still have plenty of life left as long as the ceramic bearing and bearing cup are in good condition, so don't just assume the pump is OK, check inside the bell housing of the motor winding body to be sure the back side of the impeller has not touched the inside of the housing anywhere, you can plainly see scarring if it has.
The magnetic field produced by the main motor body holds the impeller dead center in magnetic suspension, it is the spindles length with the ceramic bearing bonded to the end of the spindle and the impellers bearing cup, that sets the operational gap between the bell housing and backside of the impeller.
The D5 pump IMO is a pure design of genius and engineering, the way it works is literally flawless as long as the bearing surfaces are good.
Failures are usually preceded by increased pump noise, if on your inspection you see any scarring inside that motor bell housing the bearing surfaces are wearing out and the pumps life is getting very close to the end.
You almost have a perfect case study situation toolmaker_03, to observe pump bearing longevity under extreme hot and cold running scenarios, but to be a fair test, the study would need to be started out with brand new D5 pumps.
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
well, I really don't know, there are reasons to think that either would cause excessive wear on the pump.
for the cold side there is the hall effect of the cold water messing with the magnetics, this should make the magnetics a little stronger than normal.
for the hot side there is the added heat from the hot water that could cause more stress on the pump motor.
the reason I liked the idea of changing the cold side pumps out every three years, is that I don't want to do a tear down of the cold side just to replace a pump.
so yea, this would be a good time to check the pumps, to see just what they can really handle, but to be honest if the pumps last 3 years on the cold side, and 3 years on the hot side, I would be really happy with that.
for the cold side there is the hall effect of the cold water messing with the magnetics, this should make the magnetics a little stronger than normal.
for the hot side there is the added heat from the hot water that could cause more stress on the pump motor.
the reason I liked the idea of changing the cold side pumps out every three years, is that I don't want to do a tear down of the cold side just to replace a pump.
so yea, this would be a good time to check the pumps, to see just what they can really handle, but to be honest if the pumps last 3 years on the cold side, and 3 years on the hot side, I would be really happy with that.
- Apr 14, 2002
- 20,603
- 6
- 52,915
@ toolmaker_03
My D5 that failed after 4 years was the ambient pump on the graphics loop the below ambient pump got it's start running chilled water and it has been doing so for over 3 years, but I have no idea how much longer that pump will last and as you changing it out will be more complicated, because when it gets changed so will all the tubing on the cold side.
I do think the cold side is better on the pump simply because it aids the pump staying cooler, but you will be running below zero and using anti-freeze and the glycol does have lubricating properties so I guess we will see.
My hot side concerns are because of the peltiers, as you will be running very hot on the hot side, much hotter than any CPU or combination of CPUs and GPUS could ever heat that water that the pump was designed to handle, so I think the higher temperatures will be a faster killer.
I'm curious, you showed all your work removing that copper step and machining it to level on the base of the water block, but did you modify the flow inside of the water block?
The reason I ask is CPU water blocks are designed for CPUs, but you are running CPU water blocks on peltiers that are almost as large as the base plate of the water block, which means the flow has to be adjusted to be maximized across the entire base plate for cold or hot pickup.
When I said in the beginning there was a lot you had to learn the hard way, I truly hope that's not one of those things because I thoroughly covered that in this thread.
There were more reasons to modify the flow of the water block than freezing the block which you did not have to be concerned with because of the glycol, the second reason was faster cold pickup to get the cold the peltier was producing away from the peltier as fast as possible and to the insulated reservoir.
The insulated reservoir is where you want the cold to build up, not the water block, that's why I was concerned regarding your packing all that clay around the water blocks, you may have approached that from one point of view condensation and ice buildup, but actually created the perfect environment for holding in the cold, which is great for the reservoir but not the water blocks.
Because if the cold takes too long to get away from the peltier because it builds up in the water block it will cause the peltier to get too cold and stall out and since you are after really cold temperatures it was even more imperative to modify the flow of the water block.
So did you?
My D5 that failed after 4 years was the ambient pump on the graphics loop the below ambient pump got it's start running chilled water and it has been doing so for over 3 years, but I have no idea how much longer that pump will last and as you changing it out will be more complicated, because when it gets changed so will all the tubing on the cold side.
I do think the cold side is better on the pump simply because it aids the pump staying cooler, but you will be running below zero and using anti-freeze and the glycol does have lubricating properties so I guess we will see.
My hot side concerns are because of the peltiers, as you will be running very hot on the hot side, much hotter than any CPU or combination of CPUs and GPUS could ever heat that water that the pump was designed to handle, so I think the higher temperatures will be a faster killer.
I'm curious, you showed all your work removing that copper step and machining it to level on the base of the water block, but did you modify the flow inside of the water block?
The reason I ask is CPU water blocks are designed for CPUs, but you are running CPU water blocks on peltiers that are almost as large as the base plate of the water block, which means the flow has to be adjusted to be maximized across the entire base plate for cold or hot pickup.
When I said in the beginning there was a lot you had to learn the hard way, I truly hope that's not one of those things because I thoroughly covered that in this thread.
There were more reasons to modify the flow of the water block than freezing the block which you did not have to be concerned with because of the glycol, the second reason was faster cold pickup to get the cold the peltier was producing away from the peltier as fast as possible and to the insulated reservoir.
The insulated reservoir is where you want the cold to build up, not the water block, that's why I was concerned regarding your packing all that clay around the water blocks, you may have approached that from one point of view condensation and ice buildup, but actually created the perfect environment for holding in the cold, which is great for the reservoir but not the water blocks.
Because if the cold takes too long to get away from the peltier because it builds up in the water block it will cause the peltier to get too cold and stall out and since you are after really cold temperatures it was even more imperative to modify the flow of the water block.
So did you?
- Apr 14, 2002
- 20,603
- 6
- 52,915
@ toolmaker_03,
Actually in your setup the same applies to the hot side you want to pickup the heat and get it away from that peltier as fast as possible, remember you are not cooling a CPU with those CPU water blocks, you are cooling a peltier and if the hot side gets too hot because the water block cannot remove the heat fast enough, it will stall out the peltier as well.
Have you even at least wondered why my setup actually works?
It works because I found the balance of the hot/cold sweet spot of the peltiers, which means my peltiers are always producing the cold I need and do not stall out.
The hot side doesn't get too hot affecting the cold side, and the cold side doesn't get too cold affecting the hot side, because when either of those scenarios happens the peltier will stall, simple as that.
I would ask you what you think, but that is really irrelevant at this point, because either you've created a cold production marvel, or a stall out monster, that's why I was somewhere in all this talk, suggesting getting the cold and hot loops running first, and then see what may be needed going forward, not just assuming and building the entire package all clay packed and then pushing the power button.
Packing all that clay around an untested setup is IMO getting way ahead of yourself, but if you modified the water block flow then all may be OK, but even my first water block flow modifications were not enough I had to go back and modify even further, which is the operational testing phase of the build, which you seem to be skipping altogether?
Actually in your setup the same applies to the hot side you want to pickup the heat and get it away from that peltier as fast as possible, remember you are not cooling a CPU with those CPU water blocks, you are cooling a peltier and if the hot side gets too hot because the water block cannot remove the heat fast enough, it will stall out the peltier as well.
Have you even at least wondered why my setup actually works?
It works because I found the balance of the hot/cold sweet spot of the peltiers, which means my peltiers are always producing the cold I need and do not stall out.
The hot side doesn't get too hot affecting the cold side, and the cold side doesn't get too cold affecting the hot side, because when either of those scenarios happens the peltier will stall, simple as that.
I would ask you what you think, but that is really irrelevant at this point, because either you've created a cold production marvel, or a stall out monster, that's why I was somewhere in all this talk, suggesting getting the cold and hot loops running first, and then see what may be needed going forward, not just assuming and building the entire package all clay packed and then pushing the power button.
Packing all that clay around an untested setup is IMO getting way ahead of yourself, but if you modified the water block flow then all may be OK, but even my first water block flow modifications were not enough I had to go back and modify even further, which is the operational testing phase of the build, which you seem to be skipping altogether?
toolmaker_03
Honorable
- Mar 26, 2012
- 2,650
- 0
- 12,960
ok, as stated I will have the clay at the start of the TEC's cold side to begin with, and I will slowly add little bit, by little bit, of clay, until I hit that point where the radiated heat from the hot side, just warms the surface of the clay to prevent any freezing. I do not want to add too much clay to cause the clay to heat up internally and warm the cold side of the TEC.
you found that balance point, by warming the hot side of the TEC just enough, that the cold side did not cause the TEC to stall.
I am going to insulate the cold side just enough, that it does not cause the TEC to stall.
its not that I have not tested it, but I do not know exactly where that point is with the clay, I found with the foam on my test build, so I know that it exists. I have to be very critical about how much clay I add to the white area too little and the cold side will freeze the hot side stalling the TEC, too much and the hot side will heat the cold side and stall the TEC. so yea, I do get it, I am only taking a little different approach to the issue.
you found that balance point, by warming the hot side of the TEC just enough, that the cold side did not cause the TEC to stall.
I am going to insulate the cold side just enough, that it does not cause the TEC to stall.
its not that I have not tested it, but I do not know exactly where that point is with the clay, I found with the foam on my test build, so I know that it exists. I have to be very critical about how much clay I add to the white area too little and the cold side will freeze the hot side stalling the TEC, too much and the hot side will heat the cold side and stall the TEC. so yea, I do get it, I am only taking a little different approach to the issue.
- Status
TRENDING THREADS
-
-
-
-
Question I need some help in deciding which PSU to buy
- Started by impuls5000
- Replies: 12
-
Facebook
Twitter
Instagram