Peltier Cooling Questions!

[yaboc]

i'm going to have 226w and 80w pelts in the loop. my current WC system is 3/8, should i upgrade to 1/2? will it be the same if i just use 3/8 to 1/2 converters on tubing that goes to and from 3/8 components. will the performance be the same as using 1/2 fittings? im also goin to use heater core that will fit 2 x 120 on each side, and im planning to run em @ 7v. do u think the temps will be OK? HC is 11 x 9 1/2 x 2 inches.

now about the condensation proof. can i use liquid electrical tape to seal the blocks around and back of the PCB + fittings?

first time (only the GPU) i used ACIDIC silicone nad it was really hard to remove. i used Goof Off, and my question here is if it could damage the PCB? there are still some small "stains" of silicone visible on the graphi card PCB.

LMK what u think. any input is greatly appreciated. thanks

yabOC

[Farabomb]

:beerm8s: Welcome yaboc :beerm8s:

Glad to have you posting. :)

You're going to have a hell of a lot of heat with both your GPU and CPU TECed. I would go with a 1/2" system with a good pump. Your heatercore sounds like it's up to the task. With 4 120s at 7v you should be good.

depending on your TEC block you can go with precut neoprene or just neoprene sheets. GE silicone or phenoseal will take care of the gluing tasks and liberal amounts of dielectric grease will take care of the socket. I used 3M electrical coating on the back of my motherboard for insulation.

[yaboc]

thanks for nfo Farabomb, but the silicone is pretty hard to remove, so i was thinking if i could use somehow the electrical tape to glue/seal the blocks around. i know it's ok to use it on the back, but can i do the same and seal the blocks around with it (space where the neoprene connects with the mobo )? i'm going to be using MCW50-T and MCW462-UHT which already come with precut neoprene sheets.

on the other hand im using Quiet One 1200 Pump and these are some of the specs:
max head: 6.1 ft
gph: 296
Watts: 25.5

also when i had my reg wc running for a little while the pump was making gruzzlelin noise (like the sand was goin throu ). any ideas on how to fix that?

o yeah for now im not thinking about doing any fan shrouds for the rad, maybe later. is it ok to run mentioned earlier HC withouth the shrouds?

im gonna be using meanwell s-350-12 @ 29A to power the 226TEC and 80W off PSU

thanks

Ill post more question when get the heatercore (prolly next week) and start assembling everything.

btw this forum ROXORZ :flame: :duel:

[Player0]

GE Silicone II isnt that hard to remove, you just have to pick at it some. You can also by solevent for it, it will take it right off for you. I would avoid electrical tape. It will become brittle and come loose and shrink. You really want to use neoprene gaskests, di-electric grease and silicone. They are the best materials for the job.

Not familiar with your pump, so can't comment on the noise. Either it's normal, or theres some air in the system, or the pump is struggling and not getting enough water. Also, pumps will make weird noises if they are not sitting right side up.

You dont need the shrouds for the radiator, they are mostly there just for ease of mounting the fans.

[yaboc]

hey i tried some permatec 66b rtv silicon, but when i realized it was acidic i tried to remove it and it was really pain in the @ss to get rid of it. there's still some on the pcb that i can not remove. i will give a try to liquid electrical tape first (wouldnt hurt? right) and if that doesnt work im gonna have to go for geII. also do u think tiz ok to use Goof Off on PCB? cuz i did that. thanks

yabOC

[Player0]

rtv silicon wont be too bad. Ive used it before. Its usually just messier. Goof Off is also very strong, but shouldnt cause any damage, although I use Goo Gone, which is much less potent ;)

[soccrstar]

is this absoluttely ncesary?

http://www.vr-zone.com.sg/?i=224&s=2

[Player0]

OMG no that guys insane!

[mdzcpa]

yaboc:

Check out this LiquidNinjas article here:
TEC Upgrade (http://liquidninjas.com/reviews.php?op=showcontent&id=20&title=Swiftech%20Quiet%20Power%20Thermo%20Electric %20Cooling%20Upgrade)

I built a system very close to what you are doing in that article. it will address many of your questions including reliable but practical condensation prevention, TEC powering, etc.

[mdzcpa]

Originally posted by soccrstar
is this absoluttely ncesary?

http://www.vr-zone.com.sg/?i=224&s=2


Complete overkill for sure!!

[yaboc]

thanks for input guys

thanks for link to article mdzcpa

im gonna read it and post sum more questions

yabOC

[yaboc]

hey mdzcpa, so u actually used geII . i was thinking if i could sub this with liquid electrical tape? what do u think? were u able to remove the geII completely off the mobo? nice article btw. i'm still waiting for mah relay switch. is there any other way to connect dedicated psu to atx psu so they will turn on at the same time ?

[mdzcpa]

I only used the GEII around the based of the ZIF socket. Any silicon would be fine. I think liquid E-tape is to thin to do the job. It functions well as a "coating' versus A "sealing' agent. You could get away with it, but you'd need to apply a few layers onto the socket/mobo joint. Silicon would be much easier.

The GEII pops right off cleanly in big strips when you want to remove it.

Relay is the only way. The Meanwell itself will draw it's power directly from it's own AC cord, and power up via a relay. You can make one or buy one. But, whatever you do, NEVER pass on using a relay. If the system and water pump ever shuts down unexpectedly, for whatever reason, while your pelt coninues to remain ON, we're talking major meltdown. Not good. Powering your TEC PSU from a powerstrip is a bad idea.

That said, if the relay is the only thinig missing, you could power up the TEC PSU from a powerstrip temporarily just to get you up and running. Of course this is exactly how I melted my system....waiting for my relay:D So be careful if you do this.

[yaboc]

thanks for info again mdzcpa

just got back from HomeDepot and got myself 1/2 ID tubing, two fittings for my pump (1/2) and GEII :) i dont know but the other (acidic) silicone i used first time on the TEC GPU was really hard to remove. I hope GEII us easier to work with. Ehh in all that rush i forgot to pick up two clamps heh. So now im just only waiting for the heatercore, and im til i get the realy as u said im just gonna power up the meanwell from the powerstrip til i get the realy (of course im not gonna leave the comp on when im not gonna be aroound :)).

o and yeah i have thicker wires for the TEC's but they're copper, and the ones that r goin out of tec are silver and it's like a bunch of little wires twisted together. is it okay to connect em with copper wires? i dont have soldering iron, any ideas on how to connect em neatly? thanks. more questions to come. :)

yabOC

[mdzcpa]

Feel free to connect them together.

Although I strongly urge you to pick up just a cheapo soldering iron and solder, you can get away without it.

Just give yourself enough bare leads at the end of the wires so that you can twist them together real good. Then fold the twisted pair down and seal the connect with shrink tubing. Don't forget to put the piece of shrink tube on before you twist the wires together;) I use a blow dryer on "hot" to shrink the tubing real tight.

Don't use crimp connectors. Bad idea...I've seen too many come loose inside of a case. You won't want to mess around with that happening with the Meanwell powering stuff.

I'd pick up a cheap iron, solder, and shrink tube fro your local radio shack or electronic store.

[yaboc]

i might get some cheapo slodering iron, but prolly not any time soon. do u think i could use the luqid thermal paste and paint the wires around? o yeah and anyone know a place where i can get 1/2ID quick connects for cheap? (i no dis stuff aint cheap)

yabOC

[mdzcpa]

The liquid elec tape will work...but apply it thick in several coats.

What quick disconnects are you referring to? Are you talking about the hose inserts?

[yaboc]

here's what im talking about, but i want em for 1/2 ID tubing and for about $10 each...;)

yabOC

[dino-new]

I'm running a 226w and a 172w pelt. I needed two thermochill HE120.3 rads to effectively remove the heat out of my system. That's with normal 120mm fans.

Devin

[mdzcpa]

should there have been a link there yaboc?

[yaboc]

nice system dino-new, im only going to use 11 x 9 x 2 rad with 4 120s @ 7v. i'll check the temps once i get it all installed and if the temps are goin to be too high im gonna have to think of some better way to remove the heat from the system

mdzcpa here (http://xoxide.com/qudi.html) is the link. sorry about the other post.

i just came back from radio shack and got myself some n3wb soldering iron kit for $8 + few resitiors to do vdd, vdimm and gpu voltaged mods on abit board. also i will get the wires soldered to the pelt. hopefull some pics and definately :) more questions coming next week

[mdzcpa]

Sounds like it's coming together nicely:)

I'd skip the quick disconnects if going 1/2 ID and run stems.

Looking forward to the pics:)

[dino-new]

That sounds like a pretty big rad. Should do the trick! Good Luck.

[yaboc]

hey guys, i have quiet one 1200 pump. any input on quality / noise? i was running it for about a wekk a sometimes it was getting pretty loud, but i think it's because i had (maybe) some air bubbles in the loop. anywya im thinking about upgrading it, which pump do u think is the best choice? im looking for good and quiet pump with 1/2 (doesnt have to be 1/2 cuz i can convert it). which one is better eheim 1250 or 1048? LMK what u think thanks

[DeRailur]

Can this be moved to the extreme cooling section? I think this would help get that area moving :)

[Synthohol]

;)

[yaboc]

hey guys i got the heatercore today, already cut the fittings and got the tubing (1/2) fit nicely on it. not gonna do alot today as i have school in the morning. anyway LMK which is the best pump for wcing cuz im thinking about getting new one.

btw i checked the voltage on both spots (followed motherboardfaq.com directions) and the readings on the first one are 1.40 and the second is 1.68. they said the dioff should be around .1v and here it's .28v. is it normal? so the true volatage is the averagew of both right?

ehh i need to get new clamps. i was hopin the old ones 3/8 will werk

yabOC

[DeRailur]

SEAL STRING IT UP!!!! :drinky:

WOW, that has got to be the ugliest thing I have ever seen.

[Farabomb]

Pumps? I prefer Eheims but they are costly. You can go with swifty's option if money's tight. You can use 2 zip ties with the ends oppisite to each other as clamps.

[yaboc]

which eheim model? im considering 1250 or 1048. which one is better in terms of performance and noise?

[Farabomb]

The bigger you get the louder you get. I have the 1250 and the 1060 (now 1260 iirc). With good mounting the 1250 is damn near slient but the 1060 is pretty loud. It's not like a Delta 80mm SHE but it's noticible but not too annoying.

[mdzcpa]

I like the 1250:) My fav pump.

But, I've also had great recent success with the newest swiftech mcp600. It's aweful quiet for a 700lph 10.5ft head pump.

[yaboc]

ok i wasw testing the setup today (pics tomorrow) and everything worked fine until everything fell from the table (!^%@^%):drinky: :drinky: :drinky: :drinky: :drinky: . now the mcw462 cold side isn't getting cold anymore? any ideas? i just hope i didnt feck up the peltier. i also didnt solder the wires comin from the psu to the pelt, but it worked beofre. need help immediately. oh and btw im changin this stupid pump.:2guns: :2guns: :2guns:

yabOC

[dino-new]

I'm running the 1262 Eheim. It may be a little louder but dang, 900 gph. I can succesfully run my cpu and gpu pelts parallel. That way each get's freshly cooled water. If I did it all over I might get the 1264 eheim though. I think it's like 1100 gph!

They're kinda big though, If you want it all to go in the case, you'ld probably be best off to get the swiftech pump. It's real small and real good for it's size.

Devin

[Player0]

Its not the GPH that counts, its the pump's head rating which is most important (and most overlooked). The MCP600 works great because it has a really high head per GPH. Thats why SwifTech choose it, they have some good heads over there.

"I can succesfully run my cpu and gpu pelts parallel. That way each get's freshly cooled water"

Thats actually a very incorrect way to think about watercooling. Beleive me, its such a common misconception, that you even have experienced guys beleiving it. Water circulates (ideally) very quickly through the system. Water that enters the CPU block is not 10 degrees hotter when it comes out. Water that leaves the radiator is not 10 degrees cooler. In fact, with high flow rates and pressure, the temperature at the inlet and outlet of any device on the watercooling loop will be almost immeasurable. Fractions of a degree tops.

This has been discussed and proven elsewhere, and to death, it's simply part of the 'laws of watercooling now'. There's a great site that lists them, but damned if I can remember it.

It's for this reason that parallel doesn't work as well as some people think it should. Dino, the reason you mgiht see better temperatures doing things this way is not because of getting 'fresh-water', but probably because you have less flow restriction, improving head and flow through that leg of the system where you have twice as much inner diameter working. In fact, upgrading your tubing to 5/8" would probably give you even better results. There's know doubt that such a big pump is really struggling with 1/2" tubing. You're really just killing the pressure by running like that, and you'd be assured much better performance with 5/8" OR, beleive it or not, using a smaller pump.

But apart from flow area, there are no performance benefits to using parallel like this over series. This is because again, the water in a cooling system is balanced temperature all around, especially when you have a lot of flow. The water in the system is going to rise to some average temperature based on input wattage and output wattage. It will reqch equilibrium and will not change, and the system will act like a single entity, like theres no flow or water at all.

There will not be temperature differences at any point in the system, so you will never get the benefit of having 'fresh water'. I know, its kind of a weird concept, but it really is proven fact at this point. Its just something that most people have to learn as they go. I know I thoguth the same as you. Just look at my gallery, I have tons of projects using parallel, and after years of testing, series DOES work best.

[yaboc]

i dont know y but the pelt is getting the power but not getting cold. help!

[Player0]

You might have cracked the peltier. Check your power connections too. If the peltier cracks, it will no longer function at 100%. In fact, only half or a fraction of the peltier will still power.

[yaboc]

i dont know, i tried to resolder the wires with no luck. i placed order for new peltier and getting eheim 1046. any input the noise of this pump? thanks

[dino-new]

player0,

Actually I've noticed that the water coming out of the CPU 226w pelt is about 3-5c hotter than the water going in. I noticed a considerable drop in the temp on the gpu when I went parallel instead of feeding it warmed water from the cpu pelt. It droped about 5-8c.

[Player0]

How did you measure the temperature drop accross the CPU block? I've had plenty of pelt waterblocks and have never noticed a temperature difference.

Im not sure why you noticed such a performance difference moving to parallel. Again, it could be due to increased flow rate, but there could be other reasons too, such as less pressure on the mounts or something. Anyway, I can't explain your numbers. Do you have pictures of the series set up?

[dino-new]

Yea I'm not sure if I have pictures of the series loop, but it was pretty straight forward. Just from the rad to the cpu then to the gpu with as little hose as possible (no kinks). I measured the rise in temp by checking the difference between the parallel and series loops. I switched back and forth a couple of times and yep, there was a better temp on the gpu when in parallel. The cpu stayed the same pretty much though, no temp difference there. I beleive my success in going parallel is due to the higher flow rate of the pump.

[yaboc]

ok guys i have to choose between eheim 1046 and Hydor L30. Since i'm waiting for the pelt + gaskets it'll give me some time to get pump.

yabOC

[mdzcpa]

Hmmm...the water that passes through the TEC block isn't in there nearly long enough to rise by 5c. Unless, of course, your flow rates were very poor before you went parallel.

As Tom said, water temps anywhere in the system should be almost identical....if you have proper flow rates that is.

[Player0]

The size of the pump has nothing to do with how well parallel works. One leg will get some flow, the other leg will get the rest. To do parallel PROPERLY, you would need to reduce the tube size through the leg. So have 5/8" tubing until the Y split and then go to 1/2" or something. And the tube lengths would have to be equal as well.

Mike, whats the link to that site with all the watercooling 'laws', I cant find it.

[dino-new]

Mine seems to work great in parallel!:D Anyways, I tried the ceramique and my temps got worse. I switched back to AS5 and my temps are waaaaay lower :cool3:

[mdzcpa]

I know what you are referring to but I can't find it either:confused:

[yaboc]

ok i can get hydor l30 for $35 or eheim 1046 for the same price. i never owned em, but i hear some good stuff about hydors. LMK what u think

[illmatik]

Where have you found the hydor for that price?

[mdzcpa]

I'd suggest the Eheim 1250 or Swiftech MCP600 over either of those units. The 1048 is a tad weak. The l30 has nice flow rate at 1200lph, but the head is only 75", compared to the MCP600's 700lph but with a head of 126"...that's a lot more pressure which is helpful for the restictive nature of waterblocks and rads.

My 2 cents.

[yaboc]

i already bought Hydor L30. but ill consider the models u mentioned when ill be upgrading . now waiting for new pelt + pump. eh im not gonna do leak testing cuz i already did it and everything was fine. btw i couldnt get the res compeltely filled because it was really hard to "catch" air bubble with the opening. the air bubble was just moving around the sqare thingy that under fill n bleed hole.

[yaboc]

ok guys today i went thru horror.

i recieved everything i needed to set up tec system. i put the gaskets, pelts on the block and then followed mdzcpa's FAQ.

i used GE to cover outer and inner sides of the socket. put the bottom piece of the gaskets on the socket, filled evrything with the dielectric grease (pin holes, around the socket, in the socket). still everythign was set up on the table.

at the same time i was preparing my R9500 for MCW50. when i was screwing on the block i hear a "crack" and then noticed R9500 bent like C :). i was like OMFG and removed it really fast and loosen the screws.

so i attached the MCW462 on the CPU, then removed it to make sure it had good contact with the core. then i put it back on together with R9500 and started the pump + meanwell.

First the system shut itself off after few seconds. I started it again and it was fine, but nothing on the screen. Just black screen.

I thought it must be the Radeon. I unplugged it , removed the tubes from it, attached the stock HSF and went away to test it at my friend's house and it was fine, thou the back of the card was gettin really hot (probably because I didnt apply any thermal paste - i was in a hurry). So the card was fine (we ran 03 Mark on it). So i thought maybe it's CPU. Tested it at friend's house again and it was working.

So it turns out to be a motherboard, because the ram was fine. I tried to clear CMOS multiple times and run cpu with stock HSF and nothing on the screen (DEAD BIOS?). Anyway I'm going to RMA the board but first I 'd like to get some input from you guys if you think the dielectric grease in the sockets , pin holes around it + GEII aroundoutside and inside of the socket could cause that . I really don't know why it didnt work and i'm even more afraid of going through the same process with the new mobo. Maybe I put too much ? Not enought? Didn't change something in BIOS? I followed article so i tried all of it looked pretty much the same. Let me know what you think. :drinky: :drinky: :drinky: :drinky: :confused: :drinky: :drinky: :mad:

[mdzcpa]

What mobo and CPU are you using again?

A couple of things to try:

The dielectric grease should not be the problem at all. That's what the stuff is designed for. Contact should not be prevented. However, sometimes when the socket slug is full of grease, the CPU wants to "rise up" again due to pressure in the socket. Make sure you hold the CPU down into the socket as you lock the lever.

Did you damage anything on the mobo when mounting the MCW462 posts? Did you use the washers on the back? DId you make sure the posts had clearance around the 4 holes on the front? (these are AMD mounting issues). Check for short problems.

Did you pretest the TEC system at all before mounting? Do you know for sure that the pump is flowing well and that the cold side of the TEC blocks is getting cold? They should get frosty pretty fast.

Can you connect a PC speaker to the mobo so that we can listen for BIOS beep codes?

[dino-new]

OK first off, what's GE? If it's a konformal coating glue then you just glued you socket in place and then you wont be able to insert the cpu properly after removing it.

So I'm curious, what's the GE you're talking about?

Devin

[yaboc]

mdzcpa:
AMD 2500+
NF7-S

I made sure the CPU was in good contact with the socket. The only thing i was wodering was how much of die electric grease should I put in the socket whole. Fill it up to level up with the socket pin area? (i thought it would be too much so i just filled it enough to cover all the little chips etc). i don't think i damaged the mobo while installing the block and yes i tested it before installing and the cold plate was getting cold. I also made sure and took off the block two time to check if it has good contact with the core. I used washers and everything. I'm using Hydor L30 and the flow was good as i didn't have the res filled completely so i saw the movement and it was pretty fast. i tested graph + cpu + ram at diff comp and they were working. I already RMAed the mobo. i can't think of anythign that could cause this problem. LMK

dino-new:
GE Silicone II i just used shortcut.

[yaboc]

buMp + help.
i recieved the mobo yesterday. should i try to set it up again or order bios savior first? ehh
im gonna try the mobo with stock HSF.

any help is appreciated.send me email or use AIM @ yabko

[illmatik]

Only put dielectric grease in the pinholes!! if you put it in the socket center you will short the board again (thats what you did b4 if thats where you put it). Only use neoprene in the center quare of the socket with a thin bead of silicone to seal it/ hold it down

[Synthohol]

Originally posted by illmatik
Only put dielectric grease in the pinholes!! if you put it in the socket center you will short the board again (thats what you did b4 if thats where you put it). Only use neoprene in the center quare of the socket with a thin bead of silicone to seal it/ hold it down hey bud, the purpose of dielectric grease is to not conduct electricity, i dont see how it can short anything out ;)

btw, congrats on your 200th post! :)

[Player0]

Originally posted by illmatik
Only put dielectric grease in the pinholes!! if you put it in the socket center you will short the board again (thats what you did b4 if thats where you put it). Only use neoprene in the center quare of the socket with a thin bead of silicone to seal it/ hold it down

Uh, thats not true. Ive done it all the time.

[dino-new]

I'm writing this reply on a computer with dielectric grease in the center of the socket. If it's truely "dielectric" then it won't conduct electricity ;)

[soccrstar]

what happens if u dont put dielectrical grease in the pin hole?

[dino-new]

You run the risk of having a short between the pins on your cpu since condensation could develop on them from underneath. That is assuming your seal is 100% effective, which most likely isn't.

[yaboc]

i sealed the outer + inner socket edge. put dielectric grease into the pinholes + inside center of the socket,but forgot about the neoprene patch covering it. could that cause the sHaaBanng? the new mobo works with stock HSF, im going to set up the peltier ONCE again tomorrow. will keep u updated.

[mdzcpa]

You really do not need the neoprene in the center slog of the socket. Fill it up with dielectric grease, and you'll be just fine.

From what you've said, you've done everything right.

However, there is one thing that you did not clarify. Did you put the mounting posts onto the NF7 very carefully? There are numerous components on the board very close to the mounting holes and some folks were having problems mounting the posts without damaging the board.

Here are some NF7 pics I took with my board:
http://www.liquidninjas.com/mdzcpa/reviewphotos/NF7/Picture%20019.jpg

http://www.liquidninjas.com/mdzcpa/reviewphotos/NF7/Picture%20020.jpg


And here is a thread on the subject:
HERE (http://liquidninjas.com/bbs/showthread.php?s=&threadid=4952&highlight=nf7+holes)

[yaboc]

hey mdzcpa, i was don't think i damaged the mobo during installation of any components, cuz i was paying attention to every step and yes i noticed the standoffs were really close to the little chips on the board. i mad sure they're werent touching any of those chips. i'm gonna looks for smaller washers to put them under the stand off just to be on safe side. thanks

[mdzcpa]

What I did was add additional felt washers between the post and top of the mobo to give the post more clearence.

You can see how I did this best in the 1st pic.
http://www.liquidninjas.com/mdzcpa/reviewphotos/NF7/Picture%20071.jpg

http://www.liquidninjas.com/mdzcpa/reviewphotos/NF7/Picture%20075.jpg

http://www.liquidninjas.com/mdzcpa/reviewphotos/NF7/Picture%20076.jpg

hope this helps.

[yaboc]

thanks for pics mdzcpa. i'm going to the store to get some washer. i also need sum metal nuts cuz the plastic ones from swiftech just don't work (had hard time removin them). anyways i also don't have two scres from the mounting kit of the block and im just using some other ones from z-chip block mounting kit

http://http://www.dangerden.com/images/z-chip%20block/zmb1_large.gif

so im using two scres from swiftech on opposite corners and another set of these on the remaining corners.

do u think this could affect anythin?

also about the dielectric grease and the middle of the socket. do i have to fill it up with dielectric grease up to the level of the pin holes area? or just cover it a little bit? cuz the temp sensor thingy is stickin up so i dont know if i should cover it to or something, or just bend it a little . LMK

yabOC

[yaboc]

ok everything's installed, but now when i turn on the pc it shuts itself down after few seconds. any ideas? the block is mounted correctly, and i power up the pelt one mis or so before powering on the pc so it gets cool. also this time i used little piece of neoprene to cover the center of the socket after applying dielectric grease. any ideas. i thought this could be because there was no fan connector connected to the mobo so i did connect stock HSF to the fan connector but still the machine awas booting up for few sexs anbd then followed by shutdown. any ideas?

[soccrstar]

that happens to me when i vmodd

um could b u put some AS (artcic silver) paste on the mobo?

also improper mountin' causin' thermal protection to shutdown pc cuz of too hot?

i sometimes forgot to turn on the pump b4 the pc (not good at all)

a64s dont like heat either

try reseatin' the ram/video card

[mdzcpa]

For the socket, skip the neoprene and fill it almost totally up with dielectric grease.

As far as the mounting is concerned, you MUST use screws of equal length always always always. And, you must use the same springs at all for corners. if you do not do this, uneven pressure will give you bad temps, and potentially crack the core!

Even pressure is critical!

If the mobo won't power up, reset the cmos and try again. I'd remove the tec block and check to make sure the core is okay and that the thermal paste pattern is totally even.

[yaboc]

mdzcpa, i will take off the pelt block and try to boot with stock HSF.

ok i will remove the neoprene patch

ehh i only have 3 screws from the original mounting , and i have longer ones (4+) that fit but they're too long and i dont have a lot of springs to for all of em). ehh

off to look for some material

[yaboc]

ok the pc was shuttin itself down cuz there was no contact between core and block (ehh i thought i mounted it good),

but now i made sure it's sitting good.

HOWEVER the BAD NEWS is my sfotmoddable 9500NP is BAD. It takes numerous time of rebooting ( + a lot of luck) to see not cheakboarded POST screen and get to BIOS to see the temps.

I guess it's BAD if it checkboards at POST?

but with two pelts 226 + 80 in the loop and two 120 at full blast it was around 37C (CPU) anything wrong with that temp. not enought pressure on the block? ( i really didnt want to overtight it). LMK what u think.

[mdzcpa]

When mounting the block, it is important to screw in the bolts until they "bottom out" inside the mounting posts. So there should be no risk of overtightening them. The pressure is applied by the springs. This is why original bolts and springs is important.

Temps seem a bit high...I assume 37c is load. What is the idle temp?

[yaboc]

mdzcpa my graph card is broken , and i'm gettin cheakboarded screen at post and it's really hard to see so there was no point of installin OS on my new raid. but when it booted few times properly it was giving me 37c in the BIOS right after boot (seems kinda high) . but only two fans (push) were workin (pull (2x) were disconnected)
still seems knda high, i didn't expect this temp. there was some review on the net where mcw462-uht + mcw50t were tested in the same loop. anyonme got the link to it?

about the screws, i'm missing only 1 from the original packaging. (using 3 orig and 1 too long but fits the bolt. ( i guess it's better to use 4 longer ones then), ill post a pic as soon as i get my camera back.

yabOC

[dino-new]

If you're using the swiftech 5002 block then you really need the original screws, I would call them up and have them send out some more asap. The screws are that long to set a specific tension with the springs. If the screws are too long then it wont have enough force against the cpu.

The checkerboard is not a good thing at all. Sounds like there's a problem there.

I'm running a 3.0c cpu and a swiftech 5002 waterblock and my load temps are about 10c with a water temp of about 30c.

[yaboc]

im gonna email swiftech if they could send me the screw and if not im just gonna cut the longer one to the same lenght as the rest. anyway i have 2500+ at 1.8Ghz (default 2500+). the system doesnt have hard drives yet and still the temp is about 37 which is very high. im waiting for new card.

do u think the temp could be read wrong by the thermal probe in the socket ? i bent it a little bit and covered everything with the dielectric grease.

i also think the screws are maxed out and hit the bottom of the standoff. so any ideas y the temp is so high? might be the rad + fans. im still workin on this setup.

the loops is in series : pump > mcw462uht > mcw50t > rad > res

yabOC

[soccrstar]

Originally posted by yaboc
but with two pelts 226 + 80 in the loop and two 120 at full blast it was around 37C (CPU) anything wrong with that temp. not enought pressure on the block? ( i really didnt want to overtight it). LMK what u think.

what kind of RAD do u have?

i have a BIX n BIP in series and it gets wicked hot and doesnt do its job at all

[yaboc]

soccrstar it believe it '79 chevy HEATERCORE 9 1/2 x 11 x 2
i bent few pins when i threw it on the ground when the whole setup fell down from the desk :) , but it's an easy fix.
i ordered bios saviour for my mobo for about $30 shipped and gettin delay timer kit. now i need to figure out how to mount the rad. still waiting for my new graphic card. then i need to remove the stock fan from meanwell cuz when it turned on it sounded like jet place, and mount two 80 or 72 at 5v


what r ur temps?

[mdzcpa]

When I was still running my AMD set up, I had CPU idle of about 8c and load about 28c. That was with a MCW462-UT, MCW50+T powered by a meanwell at 13.5v. My rad was the rather unimpressive 676 swifty rad. I ran my two 120mm at high rpms though, so it wasn't really quiet.


I think your having a pressure problem from a bad mounting. Get that fresh kit from swiftech and give it another go.

Your temps are too high so soon after boot.

Perhaps your flow rate is real bad. Did you test the rad to make sure it flowed well. If flow is very poor, it could explain the high temps so quick.

[yaboc]

i called swiftech and they will send me 4 screws ( i asked only for one ) at no charge :) (gotta love swiftech). as for the heatercore i never tested it for good flow, it was brand new so i thought it will perform well. my tubing is pretty long as I didn't do final cucts to fit perfectly in the case, so maybe that's the reason, however the water was moving pretty fast when there were few bubbles in the res so i could "see" the flow. also on the side how do i remove the stains off the tubing? what kind of loop do u guys sugest , series or paraller?

btw how do i test rad to see how well the water flows through it?