- Joined
- Jun 22, 2002
- Location
- Asheville, NC
Drench it in Marine goop?
lol Thats what I would do.
lol Thats what I would do.
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Is a peltier water chiller a dead idea?
- Thread starter raskren
- Start date
the marine goop will seal it up or you could buy a polcarbonate container (it will not crack like that acrylic one does). They are more expensive.
fwiw....I saw some reviews of that pelt chiller early this year...one guy found that it was adequate to keep his xp 1600 running stable at xp1800 speeds. In his case a noisey fan and hs gave better performance. Sorry but there are very few 'undiscovered' cheap solutions to water chilling.
wj
fwiw....I saw some reviews of that pelt chiller early this year...one guy found that it was adequate to keep his xp 1600 running stable at xp1800 speeds. In his case a noisey fan and hs gave better performance. Sorry but there are very few 'undiscovered' cheap solutions to water chilling.
wj
Any way to get a large washer, and a nut on the inside of the Res for the barb? That should give it a larger area for silicon, and spread the load a bit when the barb is tweaked sideways, and whatnot. Maybe a large washer on the outside also, sandwiching the plastic with lots of caulk
camel said:
When I built my reservoir I have to sandwich 2 large washers inside and outside on a thin plastic bottle to help spreading pressure (when pushing/pulling tube to/from the barb), and get better sealing. I use a Step-Drill (about $25 from Homedepot) to drill the hole (a picture of the Step-Drill is shown on the "Water-cooling Reservoir Theory and Construction article on the front page http://www.overclockers.com/tips1089/index05.asp). It worked great for drilling a big hole for a 1/2" NTP barb on a thin plastic.
- Joined
- Jun 14, 2003
- Location
- North Carolina
People.....Why not gain access to a water fountain/"bubbler"? (this thought has crossed my twisted mind on occasion). Just kiddin'
Seriously though, why not put the TEC on a seperate water block either before or after the water flow to the CPU?
This is something I've been picking my brain over for the last two months, and it seems logical enough. It keeps the water cooler before getting to the CPU, (since the radiator/heater core does "all" the ambient temp cooling), it will help to remove the heat generated by the pump, and if you got one that would run on say, 50-70 watts, you could easily integrate it into the power supply for the 'puter. (i would think a 400 watt PSU would be sufficient). Actually, the only things i need to finish this prototype test is the waterblocks and pump (standard watercooling kit is all thats needed).
Seriously though, why not put the TEC on a seperate water block either before or after the water flow to the CPU?
This is something I've been picking my brain over for the last two months, and it seems logical enough. It keeps the water cooler before getting to the CPU, (since the radiator/heater core does "all" the ambient temp cooling), it will help to remove the heat generated by the pump, and if you got one that would run on say, 50-70 watts, you could easily integrate it into the power supply for the 'puter. (i would think a 400 watt PSU would be sufficient). Actually, the only things i need to finish this prototype test is the waterblocks and pump (standard watercooling kit is all thats needed).
- Joined
- May 27, 2002
- Location
- california
An idea i have been throwing around is putting a swiftech MCX462+T with a high output 92mm and a socket A water block togeather. The heatsink and waterblock will be bolted togeather using the socket A mounting holes. The TEC on that is a 226 watt.
- Joined
- Jun 14, 2003
- Location
- North Carolina
Here it is, people.........
Basically, i only sat for an hour and came up with this idea.
Yes the pic is crude, but it gives a general description of what i'm thinking.
Basically, i only sat for an hour and came up with this idea.
Yes the pic is crude, but it gives a general description of what i'm thinking.
- Joined
- Jun 14, 2003
- Location
- North Carolina
OOPS!... How do i get the pic to show?
- Joined
- Jun 14, 2003
- Location
- North Carolina
Sorry, people. The pic is BMP format, and since i can't upload it normally, e-mail me for a pic. it gives the idea some light, and will possibly spark some new interest in this form of cooling, since it will only take a few hours to build from scratch.
- Joined
- Jun 14, 2003
- Location
- North Carolina
Well, first tests are completed, with really good results. what i did was fill up the system with HOT water (120ºF), fired up the pump, then turned on the PSU for the 72 watt TEC. At first, the temps wouldn't drop fast enough, so i took the TEC off the 5 volt side and tied it to the 12 Volt side.
What happened next was almost unbelievable.....Temps dropped at a rate of 8ºc per hour, based on no load, total capacity of 1 gallon. I'm not sure what that would calculate to in C/W, but its gotta mean something. I'm just waiting on the waterblock right now, but everything looks to be golden. Flow rate is a meager 360 gal/hr with only a 1 ft. head so, this may pose a flow rate problem.
Updates will follow in a new thread.....wish me luck..
What happened next was almost unbelievable.....Temps dropped at a rate of 8ºc per hour, based on no load, total capacity of 1 gallon. I'm not sure what that would calculate to in C/W, but its gotta mean something. I'm just waiting on the waterblock right now, but everything looks to be golden. Flow rate is a meager 360 gal/hr with only a 1 ft. head so, this may pose a flow rate problem.
Updates will follow in a new thread.....wish me luck..
- Joined
- Jun 30, 2003
- Location
- Columbus, Ohio
*BUMP*
- Joined
- Jun 14, 2003
- Location
- North Carolina
Well, I know i'ts been a couple days since i last posted but things have been busy. upon further tests, including the radiator, i was able to cool the block by three degrees on each pass until 36º below ambient was reached (estimating that a 220 gph pump, pumping with 2 ft of head, will circulate approx. 1 gallon every 36 seconds) under no load, on 226 watts of peltier. Now, how this will compute into actual cooling efficiency will have to see........
- Joined
- Jun 14, 2003
- Location
- North Carolina
Sheesh......took me long enough to finish it, but its done.
First off, let me state that my knowledge of thermal dynamics is minimal at best, so i might screw up a word or two.....im not in any way perfect......
Ill start with the parts layout, which consists of one Vector model bc-28 portable cooler/warmer, one little giant model pe-100 water pump, one scavenged heater core out of a 91' cavalier, a DTEK white water W/B (borrowed), hoses and assorted fittings to connect it all together, and one used 350 watt Antec.
Dropped the water pump inside the cooler, cut holes in the top to accomidate the wire and hoses, fitted the pelt with a scavenged 4 pin power connector and removed the original wiring (theres a fan inside that leaves much to be desired), and added a 80mm fan. ran the hoses to the radiator and waterblock (which was fun making a 2 to 1 splice pipe to come off the w/b), and ran the outlet from the pump directly into the waterblock. powered everything up and let it cook.
Ambient around here is about 27-30ºC and i wanted to keep the cpu around that at first, so i ran the cpu at stock and came up with these numbers:
CPU Temp 22º
System Temp 29º
CPU Speed 133Mhzx11=1466
I let it bake with Prime95 for three hours, and came back to this:
CPU Temp 29º
System temp 34º
Now.....is this safe to bet on? or am i barking up the wrong tree?
I'm asking because if the CPU is pulling about 45 watts and the temps climb like this in three hours, whats it gonna do when i crank it back to where it was (200x10.5=2100)?
Think i need more pelt power.....(thinking tim allen thoughts)
First off, let me state that my knowledge of thermal dynamics is minimal at best, so i might screw up a word or two.....im not in any way perfect......
Ill start with the parts layout, which consists of one Vector model bc-28 portable cooler/warmer, one little giant model pe-100 water pump, one scavenged heater core out of a 91' cavalier, a DTEK white water W/B (borrowed), hoses and assorted fittings to connect it all together, and one used 350 watt Antec.
Dropped the water pump inside the cooler, cut holes in the top to accomidate the wire and hoses, fitted the pelt with a scavenged 4 pin power connector and removed the original wiring (theres a fan inside that leaves much to be desired), and added a 80mm fan. ran the hoses to the radiator and waterblock (which was fun making a 2 to 1 splice pipe to come off the w/b), and ran the outlet from the pump directly into the waterblock. powered everything up and let it cook.
Ambient around here is about 27-30ºC and i wanted to keep the cpu around that at first, so i ran the cpu at stock and came up with these numbers:
CPU Temp 22º
System Temp 29º
CPU Speed 133Mhzx11=1466
I let it bake with Prime95 for three hours, and came back to this:
CPU Temp 29º
System temp 34º
Now.....is this safe to bet on? or am i barking up the wrong tree?
I'm asking because if the CPU is pulling about 45 watts and the temps climb like this in three hours, whats it gonna do when i crank it back to where it was (200x10.5=2100)?
Think i need more pelt power.....(thinking tim allen thoughts)
- Joined
- Jun 14, 2003
- Location
- North Carolina
to answer the question about solid state chillers......yes they do, but its only a 72 watt unit, and with the one i will need is a 150+ watt unit, since im also dissapating heat from the pump, which is in the water to begin with (i chatted with one of my "Thermal Dynamics" buddies, and he stated that in order to keep the temp stable, id need at least 1 and 1/3 in pelt wattage as i want to dissapate. since the pelt in the system is only 72 watts, i plan on swapping it with the 226 watter i have, but im not sure if the PSU can push it.
Second, My temps at start-up were about 28ºC then climb to about 40º on air. as i stated earlier, my ambient can range anywhere between 26-30ºC, and my cpu at idle is around 38.5º, stressed its about 43-44º. (maybe this was a good idea after all).
Second, My temps at start-up were about 28ºC then climb to about 40º on air. as i stated earlier, my ambient can range anywhere between 26-30ºC, and my cpu at idle is around 38.5º, stressed its about 43-44º. (maybe this was a good idea after all).
I need 225 watts of pelt chilling to keep my 2600 @ 2600mhz running around at 5~8 below ambient. The issue you will have is dumping the heat from the hot side of that 226 watter...you will need water (I'm pretty sure) cooling. I started out with 150 watts and found that on hot days I had temps continue to rise (on the cpu). with 225 watts I can control it (I also started this project last year with a 1600@1800 so I'm pretty sure you need somewhere around 175~200watts to keep it cool.
wj
wj
Similar threads
