Something that can cool a near extreme overclock at temps below 60c

[Somerandomtechyboi]

I just wanna overclock the balls off my poor 45nm wolfdale c2d and pentiums xD

 

issue atm is cooling at 1.7v because just idling with 1.7v in the bios has my e8400 running at 85c idle on both my whimpy 92mm tower and my ghetto "waterloop" with a cheap all metal chinese block

 

I have validated that it isnt the solder going bad or something cause with the poorly placed waterblock the ihs does get burning hot, thermal paste is gd900 which performs like arctic mx4 but dirt cheap

 

Is there any kind of cooling solution thatll keep this under 60c that isnt just making a chiller out of an ac?

 

My first thought is using a car radiator and a chinese or potentially thermaltake w2 block in a loop, i have done a test run with the ghetto loop and with the cheapo chinese metal block it performs abit worse than my 92mm chinese tower

 

The config of said ghetto loop is a repurposed plastic container as a res + aquarium pump in the res at the very bottom of the loop and the waterloop on the cpu as the highest point of the loop, no rad yet since its just a test run and the water did get very warm after ~5 mins of prime95 at 4.45ghz 1.42v, is there anything wrong with the config of the ghetto loop other than the fact its missing a rad?

 

My other cooling solution would be a used aircooler, found this decently beefy looking scythe cooler but idk the name for ~11$

Spoiler

 

What would perform better? A ghetto loop v2 with a chinese copper block (does have microfins)/thermaltake w2 block + a massive car radiator or an old scythe cooler?

 

I have doubts on whether or not that scythe cooler can maintain any decent temp and near extreme overclock conditions (1.6v+) but i am pitting it against a cheapo and prob horribly designed 92mm chinese tower

[Bombastinator]

The bottleneck here seems to be the block.  You have to get the heat out of the chip and into the water before you worry about how to get the heat out of the water.  It seems real possible to me that increasing the radiator might not do much.  The thing a chiller can do is reduce the water temperature to just above dew point which creates a greater gradient so a block will be more efficient for what it is, but even that is limited. Air cooling is limited by the envelope of the air cooler.  If you’re ghettoing you might be able to leverage the superior block of the air cooler by watercooling the air cooler.

Edited February 21, 2022 by Bombastinator

[Somerandomtechyboi]

7 hours ago, Bombastinator said:

you might be able to leverage the superior block of the air cooler by watercooling the air cooler.

I wasnt talking turning an aircooler into a watercooler but it actually is feasable, except it seems like my 92mm has garbage heatpipes so rip to that plan, but maybe the scythe cooler would work, i do have some acrylic that i ended up never using but not much and i have seen someone turn an aircooler into a watercooler before, i do have someone selling a used cheapo chinese block with copper and microfins for about the same price that i can get a cheap aircooler with beefy asf heatpipes like raijintek themis evo

 

Is this good enough or would a raijintek themis evo or even that scythe cooler turned into a waterblock be the same or better performing? And if i cant mount it with the regular mount then ill just resort to zipties, if i do upgrade to a w2 block ill most likely repurpose the chinese block for nb tec cooling so i can push fsb abit higher on subzero nb, atm on this ds4p i can manage 556fsb on my e8400 stable so not bad but could be better

Spoiler

 

7 hours ago, Bombastinator said:

The bottleneck here seems to be the block.  You have to get the heat out of the chip and into the water before you worry about how to get the heat out of the water.  It seems real possible to me that increasing the radiator might not do much.  

Yea i figured it was prob the block, im just gonna get a massive rad just so its futureproof so incase i go x58 or newer then i can still cool them well enough for near extreme oc aswell, plus the bigger the rads the cooler it runs after a prolonged period of time which will be helpful for dailying near extreme oc since i only start considering 1.7v and up unsafe for daily on 45nm core, ofc not all the time cause id rather not obliterate my parents power bill but theres the option i guess

¯\_ (ツ) _/¯ 

 

Prefferably i would want to run as close as possible to ambient but i dont see <50c as being a possibility unless i go the w2 block route, and the w2 block is sufficiently expensive for me so maybe an aircooler into a waterblock might be more feasable

[Demonic Donut]

I'm not familiar with the physical aspects of the chip, but I would delid and do direct die if possible. Preferably with liquid metal. Reducing the distance between die and cooler will definitely help. The thinnest, most conductive interface is best.

 

If you're looking for best 24/7, a water loop will beat an air cooler. Cost to performance, a good air-cooler will win though.

 

Are you trying to stay as cheap as possible? Willing to splurge on anything? Mounting pressure is your friend too, a good mount will help more than the best pastes or differences between mid to high end water blocks etc.

[Somerandomtechyboi]

41 minutes ago, Demonic Donut said:

If you're looking for best 24/7, a water loop will beat an air cooler. Cost to performance, a good air-cooler will win though.

 

Are you trying to stay as cheap as possible? Willing to splurge on anything? Mounting pressure is your friend too, a good mount will help more than the best pastes or differences between mid to high end water blocks etc.

Im going cheap as possible here and willing to go extremely jank solutions so in this case water has better value than air cause i can find cheap asf water blocks and car rads in my area

 

No i do not intend to splurge on garbage i dont need but basically just go bare minimum and i really mean bare minimum aka making garbage tube adapters with bottlecaps kind of money saving

 

I also found this interesting video

Apparently shoving water through the heatpipes of an aircooler results in really good performance, not to mention i can keep my acrylic since i use that for writing on my bed while studying and it would be a massive inconvenience to lose that

 

I do still intend to use my cheapo 92mm chinese tower for potential tec cooling on the nb for higher fsb capability when ocing cpus with low multi but most of the time i expect to deal with cpus that have multi >9 anyways so im gonna contemplate on whether or not i should f around with the poor 92mm tower, but then again if im gonna have a massive car rad then i dont think cooling a 72w tec + nb heat would be much of a problem, the nb on these giga are very efficient voltage wise since they barely need any volt to get decent fsb, i run 556fsb with 1.4v nb volt while my sad p5q can run only 490fsb at 1.56v, i think its something to do with nb timings and they are apparently bad on these giga board which makes benchmarks abit slower, i might f around with mchbar but i dont think ill bother unless cpu performance is affected

[ApolloX75]

Honestly I used an ebay H100, modified the bracket to fit LGA775 and then oc'd my QX9650 to 5GHz and pumped 1.8V into the damn thing and never went near 80C.

 

THIS was my stable just shy of 4.6GHz run. I wasn't able to stabilize the chip at 5GHz unfortunately, too much transient voltage between cores even my 790i FTW couldn't tame it.

 

But yeah, H100 from ebay or something in that class and some rockin' 120mm fans. More than enough to cool a C2D like an E8400. I have an E8600 I'll be throwing on the oc bench soon when I get some more components to build something that cools even better than the H100 solution. 

 

But for now my focus is my daily and getting it back up and running after my pump failure.

[Somerandomtechyboi]

8 minutes ago, ApolloX75 said:

Honestly I used an ebay H100, modified the bracket to fit LGA775 and then oc'd my QX9650 to 5GHz and pumped 1.8V into the damn thing and never went near 80C.

If even an h100 which is a mediocre aio for all i know can do that then maybe i can do better with a janky custom loop, after all the car rad im using is gonna be far bigger than the rad on an h100

 

I guess this gives me hope that maybe i can run near ambient at near extreme overclocks (<50c) with a jank custom water loop, what kind of temps did you get when running that 9650 to 5ghz?

[ApolloX75]

10 minutes ago, Somerandomtechyboi said:

If even an h100 which is a mediocre aio for all i know can do that then maybe i can do better with a janky custom loop, after all the car rad im using is gonna be far bigger than the rad on an h100

 

I guess this gives me hope that maybe i can run near ambient at near extreme overclocks (<50c) with a jank custom water loop, what kind of temps did you get when running that 9650 to 5ghz?

Mid 70's when benching, but it was never stable enough to complete any tests. I was able to run 3D Mark and Prime95 with the QX at almost 4.6 though, and it maybe hit high 60s to low 70s at peak.

 

I doubt you'd get decent results on an automotive rad, maybe if it was a trans cooler you might, I just can't see getting good flow rates through any decently sized car rad unless you've got a pretty powerful pump. Mind you, you did say you were using an aquarium pump, so that might do it. I've never experimented with aquarium pumps I've always gone the small 12V rotary vein or ceramic impeller pumps that I can drive off a header.

 

I ended up building a little loop using a cheap $15 generic copper block and $20 copper 240 rad off Amazon and got even better results with a little pump/res combo. Mind you I've got over a decades worth of fittings, tubing and various other water cooling components that have been collecting in my parts bins so I didn't have to front any money for that kind of stuff. 

 

My H100 is pushing at least six years old now, it's not working so well anymore hence why I decided to just build a little custom loop for testing and OC'ing on the bench. 

[ApolloX75]

Oh, yeah forgot to mention. The reason I decided to buy a cheap used H100 in the first place way back in... I think it was 2017 or so was because I couldn't get my QX or even my little Xeon off the ground using an old TX3 or a big old OCZ cooler. Then I had built a little 1/4" tubing loop for my old Athlon XP system and saw ridiculously low temps from that and it set the lightbulb off in my brain.

 

So yeah cheapo air coolers, or even older ones that may have been "top notch" back when these C2Ds were new or newish? They're garbage compared to even a cheap modern Hyper 212. I was able to hit 3.8 or 3.9 with the Xeon using a Hyper 212 modded to 775, and then I switched to the H100 and drilled the brackets out so it'd line up with 775 and bingo. Overclock heaven.

 

The whole "ThEy DOn'T maKe 'Em LiKe tHey UseD tO" shtick doesn't always apply, especially when you're talking about coolers.

 

 

[Somerandomtechyboi]

21 minutes ago, ApolloX75 said:

I doubt you'd get decent results on an automotive rad, maybe if it was a trans cooler you might, I just can't see getting good flow rates through any decently sized car rad unless you've got a pretty powerful pump. Mind you, you did say you were using an aquarium pump, so that might do it. I've never experimented with aquarium pumps I've always gone the small 12V rotary vein or ceramic impeller pumps that I can drive off a header.

Well its one of those 2$ aquarium pumps that claim 1000 l/h flow rate and with the outlet not restricted by a bottlecap adapter it does seem to flow decently well though i may need more flow, if its a water resistance issue could running pumps in series solve that? Cause theyll both work together and combine their strength, it runs on 240v a/c so unfortunately no overvolting shenanigans

 

My plan for the car rad is to run 2 of these pumps in serial and either use a cheapo chinese copper block with microfins or my 92mm tower with sawed off heatpipes to flow water through them (im leaning towards the 92mm tower mod)

 

Ive also found that my e8400 seems to top out ~4.7ghz cause after that it needs crap tons of voltage, its stable enough for benching at 1.504v load and 104 pcie since thats where the performance comes from. Ill go an upgrade to a xeon x5260 (xeon binned e8600) and see if i can do higher cause from the hwbots ive seen it is an impressive cpu and i can get one for 5$ which is cheaper than those stupidly overpriced e8600

 

1 minute ago, ApolloX75 said:

The whole "ThEy DOn'T maKe 'Em LiKe tHey UseD tO" shtick doesn't always apply, especially when you're talking about coolers.

Interesting, i did find a cheap scythe cooler w 5 heatpipes from the 775 era, how would that stack up against a 212 evo?

Spoiler

Fin density seems pretty meh but its a beefy cooler with 5 heatpipes, and then theres the god awful push pins

 

 

Btw how do i get my pcie to clock higher? Cause holy crap pcie freq brings a shtload of performance, at 4.7ghz just moving pcie freq from 103 to 104 results in a 7 point gain, and apparently this thing performs similarly to a 3770 single score wise which is kinda funny, maybe at 5.2 or something i can get near a 4770k xD

 

Ive already tried overvolting the crap out of the clock drive from 700mv to 1000mv and still pcie 105 is not stable, any way to get pcie 105 or over to be stable? Cause it seems like pcie freq capability increases the more you raise fsb, and this ds4p caps at 556fsb

[ApolloX75]

5 hours ago, Somerandomtechyboi said:

Well its one of those 2$ aquarium pumps that claim 1000 l/h flow rate and with the outlet not restricted by a bottlecap adapter it does seem to flow decently well though i may need more flow, if its a water resistance issue could running pumps in series solve that? Cause theyll both work together and combine their strength, it runs on 240v a/c so unfortunately no overvolting shenanigans

 

My plan for the car rad is to run 2 of these pumps in serial and either use a cheapo chinese copper block with microfins or my 92mm tower with sawed off heatpipes to flow water through them (im leaning towards the 92mm tower mod)

 

Ive also found that my e8400 seems to top out ~4.7ghz cause after that it needs crap tons of voltage, its stable enough for benching at 1.504v load and 104 pcie since thats where the performance comes from. Ill go an upgrade to a xeon x5260 (xeon binned e8600) and see if i can do higher cause from the hwbots ive seen it is an impressive cpu and i can get one for 5$ which is cheaper than those stupidly overpriced e8600

 

Interesting, i did find a cheap scythe cooler w 5 heatpipes from the 775 era, how would that stack up against a 212 evo?

  Reveal hidden contents

Fin density seems pretty meh but its a beefy cooler with 5 heatpipes, and then theres the god awful push pins

 

 

Btw how do i get my pcie to clock higher? Cause holy crap pcie freq brings a shtload of performance, at 4.7ghz just moving pcie freq from 103 to 104 results in a 7 point gain, and apparently this thing performs similarly to a 3770 single score wise which is kinda funny, maybe at 5.2 or something i can get near a 4770k xD

 

Ive already tried overvolting the crap out of the clock drive from 700mv to 1000mv and still pcie 105 is not stable, any way to get pcie 105 or over to be stable? Cause it seems like pcie freq capability increases the more you raise fsb, and this ds4p caps at 556fsb

IIRC PCIe clock frequency messes with a lot of devices. So it may not be the frequency itself that is the issue, but one of the components that is running on it. That being said with boards this old there isn't a whole lot on that bus, as they still use a PCI bus for slower hardware like the Southbridge.

 

I think the Northbridge and the PCIe slots themselves are the only major components, so it would be your NB or your GPU that isn't liking the frequency changes.

 

As for your pumps, flow rate is important yes but so is head pressure. You could have 3000 l/h flow rate but if your head pressure is .2 m, you're not going to be able to work against any restrictions worth a damn. So check that as well, running them in series will increase the flow and pressure but if it's only little by little you'll need a bunch.

 

That scythe cooler might work alright, but I find the older the heat pipe design the less effective it is.

 

I won't get into RAM and FSB since from the sounds of it you're running older DDR2 boards anyways and those are a lot less picky than my DDR3 boards so that shouldnt be causing you any grief. I will say that every board has a frequency hole where it will top out or get stuck, and every board is different.

[Somerandomtechyboi]

6 minutes ago, ApolloX75 said:

IIRC PCIe clock frequency messes with a lot of devices. So it may not be the frequency itself that is the issue, but one of the components that is running on it. That being said with boards this old there isn't a whole lot on that bus, as they still use a PCI bus for slower hardware like the Southbridge.

Well why does this ds4p max at 105 but my giga g31m s2c can push pcie to the moon (195 and maybe higher with pll crystal modification), and for all i know most p45 are capable of running the pcie and pci unsynced so its not pci destabilizing to sht

 

The gpu does 195pcie on the giga g31m s2c

 

6 minutes ago, ApolloX75 said:

As for your pumps, flow rate is important yes but so is head pressure. You could have 3000 l/h flow rate but if your head pressure is .2 m, you're not going to be able to work against any restrictions worth a damn. So check that as well, running them in series will increase the flow and pressure but if it's only little by little you'll need a bunch.

Im guessing that increasing head pressure means making the tubes smaller?

 

6 minutes ago, ApolloX75 said:

I won't get into RAM and FSB since from the sounds of it you're running older DDR2 boards anyways and those are a lot less picky than my DDR3 boards so that shouldnt be causing you any grief. I will say that every board has a frequency hole where it will top out or get stuck, and every board is different.

Im on a giga ep45 ds4p and caps at 556fsb on my e8400, might go higher with a better bin cpu

 

My rams are 1x3 extreem dark 1066, supposedly these rams come with nanya d die or something like that though timings suck at high freq, i can run 1480mhz 7-9-9-10 trfc 132 at 2.2v stable on my p5q though giga board seems to suck ass at overclocking rams but still does 1400+, literally just stole the timings off hwbot and adjusted them aka used as a reference

[Bombastinator]

14 hours ago, Demonic Donut said:

I'm not familiar with the physical aspects of the chip, but I would delid and do direct die if possible. Preferably with liquid metal. Reducing the distance between die and cooler will definitely help. The thinnest, most conductive interface is best.

 

If you're looking for best 24/7, a water loop will beat an air cooler. Cost to performance, a good air-cooler will win though.

 

Are you trying to stay as cheap as possible? Willing to splurge on anything? Mounting pressure is your friend too, a good mount will help more than the best pastes or differences between mid to high end water blocks etc.

Most pre I core stuff was soldered which can make delidding problematic.  THS compound went crap with haswell and then improved again slowly.  I’m not sure where this chip is on that timeline.

[ApolloX75]

1 hour ago, Bombastinator said:

Most pre I core stuff was soldered which can make delidding problematic.  THS compound went crap with haswell and then improved again slowly.  I’m not sure where this chip is on that timeline.

They're soldered as far as I am aware all the way up to at least Ivy Bridge. I don't remember if Sandy is soldered or not though.

[Bombastinator]

I don’t either.  Haswell for sure wasn’t. It’s famous for its exceptionally bad TIM

[Bombastinator]

10 hours ago, Somerandomtechyboi said:

Well why does this ds4p max at 105 but my giga g31m s2c can push pcie to the moon (195 and maybe higher with pll crystal modification), and for all i know most p45 are capable of running the pcie and pci unsynced so its not pci destabilizing to sht

 

The gpu does 195pcie on the giga g31m s2c

 

Im guessing that increasing head pressure means making the tubes smaller?

 

Im on a giga ep45 ds4p and caps at 556fsb on my e8400, might go higher with a better bin cpu

 

My rams are 1x3 extreem dark 1066, supposedly these rams come with nanya d die or something like that though timings suck at high freq, i can run 1480mhz 7-9-9-10 trfc 132 at 2.2v stable on my p5q though giga board seems to suck ass at overclocking rams but still does 1400+, literally just stole the timings off hwbot and adjusted them aka used as a reference

I think in this case head pressure is a bit like friction or resistance, but in fluid, so smaller tubes and sharper bends will increase it, and if it gets too high the pump won’t be able to move the fluid worth a damn any more.  Head pressure is going to be a real hard term to google I think, because that same word happens in several more popular things. 

Edited February 22, 2022 by Bombastinator

[Somerandomtechyboi]

Well i think im gonna have to upgrade pump to accomodate a massive car rad but i have found some cheap used ones that seem to have decent flow and head pressure max

 

5 hours ago, Bombastinator said:

Most pre I core stuff was soldered which can make delidding problematic.  THS compound went crap with haswell and then improved again slowly.  I’m not sure where this chip is on that timeline.

E8000 is from ~2008 i belive so def soldered

 

For all i know delidding is fine as long as theres a sudden shearing pressure on the solder like using one of those chinese delid tools, heat will also work but shearing the solder would prob be more convenient so i dont need to put my cpu on a stove or something, atm i dont think its needed as the solder is fine at conducting heat

[Bombastinator]

18 minutes ago, Somerandomtechyboi said:

Well i think im gonna have to upgrade pump to accomodate a massive car rad but i have found some cheap used ones that seem to have decent flow and head pressure max

 

E8000 is from ~2008 i belive so def soldered

 

For all i know delidding is fine as long as theres a sudden shearing pressure on the solder like using one of those chinese delid tools, heat will also work but shearing the solder would prob be more convenient so i dont need to put my cpu on a stove or something, atm i dont think its needed as the solder is fine at conducting heat

It’s possible, it’s just problematic. Soldered chips have pretty decent heat transfer to the ihs. The next step the heat takes is from the ihs into the cold plate, then from the cold plate to the water, which is where an efficient block becomes important.  Liquid metal is indium and gallium. The gallium has several issues as it reacts with some other metals.  It is important to not have those metals in either the cold plate or the ihs, as it can cause big problems. Then you have to worry about galvanic action between the other parts. Again, doable, but care must be taken. What metal the radiator is made of can matter for that one.

[thrasher_565]

21 hours ago, Somerandomtechyboi said:

If even an h100 which is a mediocre aio for all i know can do that then maybe i can do better with a janky custom loop, after all the car rad im using is gonna be far bigger than the rad on an h100

 

I guess this gives me hope that maybe i can run near ambient at near extreme overclocks (<50c) with a jank custom water loop, what kind of temps did you get when running that 9650 to 5ghz?

the point of buying and building a hi end water cooling loop is nose. most people dont want 70dbs loop....

[Bombastinator]

25 minutes ago, thrasher_565 said:

the point of buying and building a hi end water cooling loop is nose. most people dont want 70dbs loop....

So you’re saying everyone wants a smaller nose then.  This went poorly for Michael Jackson.

 

anyway there is also the SFF issue where you can’t get enough cooling surface area into one area but can get it into a different one. Watercooling also gets more efficient as the amount of heat increases.  Eventually it overwhelms air cooling.  There’s just too much heat to move with conduction. heat pipes have moved that goalpost to the point that  Normally the last is practically never an issue with home based systems and most non home based ones.  Even cars weren’t bound by this one until engines got well over a hundred horsepower.  Even then it was to a degree the SFF issue but writ larger.  There was a tank destroyer in world war 2 that had twin air cooled engines.  It was a big machine though. 

[Somerandomtechyboi]

59 minutes ago, thrasher_565 said:

the point of buying and building a hi end water cooling loop is nose. most people dont want 70dbs loop....

Its not exactly gonna be noisy but then again im the type of guy to buy cheap server fans just because they are cheap and decent quality static pressure + airflow fans, and when overvolted they become f ing screamers. Ofc im not gonna do that, porb just gonna connect the fans to a boost converter to adjust speed since i plan on using 24v server fans, shouldnt be that noisy, atleast better than my aircooler at full tilt but boy oh boy if i ever try running all of the fans that ill need (prob ~12 or so) at full tilt then poor connectors are gonna start melting unless i run them in parralel in 4 fan groups, even then id still need to mod a beefy wire internally to the psu just to make sure sht doesnt start melting xD though at this point i might aswell hook it up to a secondary psu cause power consumption or i just get really low power ones and just leave at 12v just so power draw isnt an issue and full tilt shouldnt be needed anyways

 

3 hours ago, Bombastinator said:

Then you have to worry about galvanic action between the other parts. Again, doable, but care must be taken. What metal the radiator is made of can matter for that one.

Literally all car rads are made of aluminum but idc since i think ill just use my 92mm chinese tower as a test block, after i get a proper block ill run antifreeze, cars have been mixing metals for awhile so antifreeze should mostly prevent galvanic corrosion

[thrasher_565]

1 hour ago, Bombastinator said:

So you’re saying everyone wants a smaller nose then.  This went poorly for Michael Jackson.

 

anyway there is also the SFF issue where you can’t get enough cooling surface area into one area but can get it into a different one. Watercooling also gets more efficient as the amount of heat increases.  Eventually it overwhelms air cooling.  There’s just too much heat to move with conduction. heat pipes have moved that goalpost to the point that  Normally the last is practically never an issue with home based systems and most non home based ones.  Even cars weren’t bound by this one until engines got well over a hundred horsepower.  Even then it was to a degree the SFF issue but writ larger.  There was a tank destroyer in world war 2 that had twin air cooled engines.  It was a big machine though. 

haha ya typo... meh ill keep it.

 

well thats were thinking out side the box helps. i mean why stop at a water cooling cpu block why not add fins and heat pipes to that so it can also use air... i mean lm and and dlided seems to be the best but also cost the most. then there's not just using ambient air why not use ac air and so on...

 

the point was that noise to alot of people mater. and wc has increased in cost by alot.  even thow we dont use 120mm thicc rads anymore...  alot of wc companies started with parts they used for refrigeration and got in to pc wc on the side.

 

still dont understand why we barley use 2x 360 rads in loop's these days when back then it was a minimum of 2x 360 rads for 580 and a i7-920...

 

a new keyboard is in the mail hope it will help with the typos...maybe...

Edited February 23, 2022 by thrasher_565

[Somerandomtechyboi]

1 minute ago, thrasher_565 said:

still dont understand why we barley use 2x 360 rads in loop's these days when back then it was a minimum of 2x 360 rads for 580 and a i7-920...

Most ppl just do it for looks rather than performance cause cpu oc is dead nowadays and gpu oc is very easy so loop doesnt really need to be that good, those old i7 were nuclear reactors, same goes for the gpu, not to mention back then you really needed to overclock to gain the most out of your system, ofc volt wasnt a concern cause with enough cooling you can theoretically daily upto 1.7v cause 45nm and 32nm being literal bricks cause very voltage tolerant and hard to kill

 

I personally despise hardline cause annoying to set up and bend the tubes, extremely generic and you see it everywhere with those sad excuses of wannabe gaming pcs with pathetic specs for their cost (reffering to those generic youtube pc builds), very annoying to upgrade, good luck doing basic maintenence like taking off the block without dissasebling the entire waterloop, etc.

 

If ppl go for soft tubing then they are on an extreme budget, they prefer to not sacrifice everything for aesthetics, or be like me and do both + overclocking

 

7 minutes ago, thrasher_565 said:

well thats were thinking out side the box helps. i mean why stop at a water cooling cpu block why not add fins and heat pipes to that so it can also use air... i mean lm and and dlided seems to be the best but also cost the most. then there's not just using ambient air why not use ac air and so on...

Having a good waterblock will always beat out a diy solution like shoving water through aircooler heatpipes, but diy solutions like said water cooled aircooler can perform pretty well, ive made up my mind and ill go shove water through my chinese cooler heatpipes once i got all the watercooling equipment like the rad and upgraded pump cause no rad = rip cooling, cheapo 2$ pump = rip cooling unless you go motorcycle rad

 

Im also gonna use large diameter tubes just so waterflow from the pump to the rad is not ruined but ofc more regular size 3/8 tubes for shoving water into the heatsink, i may still have a use for the 2$ pump maybe as an assisting pump right before the aircooler block thingy just to make sure waterflow is good cause car rads are absolutely massive after all

[thrasher_565]

4 minutes ago, Somerandomtechyboi said:

Most ppl just do it for looks rather than performance cause cpu oc is dead nowadays and gpu oc is very easy so loop doesnt really need to be that good, those old i7 were nuclear reactors, same goes for the gpu, not to mention back then you really needed to overclock to gain the most out of your system, ofc volt wasnt a concern cause with enough cooling you can theoretically daily upto 1.7v cause 45nm and 32nm being literal bricks cause very voltage tolerant and hard to kill

 

I personally despise hardline cause annoying to set up and bend the tubes, extremely generic and you see it everywhere with those sad excuses of wannabe gaming pcs with pathetic specs for their cost (reffering to those generic youtube pc builds), very annoying to upgrade, good luck doing basic maintenence like taking off the block without dissasebling the entire waterloop, etc.

 

If ppl go for soft tubing then they are on an extreme budget, they prefer to not sacrifice everything for aesthetics, or be like me and do both + overclocking

 

Having a good waterblock will always beat out a diy solution like shoving water through aircooler heatpipes, but diy solutions like said water cooled aircooler can perform pretty well, ive made up my mind and ill go shove water through my chinese cooler heatpipes once i got all the watercooling equipment like the rad and upgraded pump cause no rad = rip cooling, cheapo 2$ pump = rip cooling unless you go motorcycle rad

 

Im also gonna use large diameter tubes just so waterflow from the pump to the rad is not ruined but ofc more regular size 3/8 tubes for shoving water into the heatsink, i may still have a use for the 2$ pump maybe as an assisting pump right before the aircooler block thingy just to make sure waterflow is good cause car rads are absolutely massive after all

hmm ill show you something. i remembered this years ago and could not find anything about it ahs it just come up as hardline tube bending...so i when to yahoo... and when to an old oc site and found it.

 

 guide by voigts on http://www.xtremesystems.org

http://www.xtremesystems.org/forums/showthread.php?239246-How-to-Bend-Form-Tubing-for-Your-Setups

How to Bend/Form Tubing for Your Setups

    I'm currently wrapping up some changes on my setup, and I thought I would share a trick which proved very useful that I recently picked up here on XS. I can't find the thread now, but the idea is one of the handiest I've come across in years of watercooling. It was only mentioned in a thread, but there isn't an entire thread about it. I used this several times in redoing the lower part of my loop, and it works great. Not only can it save you from having to use elbows and such, but it can help you avoid having situations where the tubing is pressing against barbs and wanting to kink due to tight bends.

    If you want to form tubing for tight bends or any bends, you can shape it the way you want. You simply make sure it somehow it is in a steady position in that shape, and then immerse it in boiling water and then in ice water several times and it will form how you like. I found a spring for $3 at a local ACE hardware that fits perfectly inside of the 7/16” id tubing. A spring is necessary in order to make sure the tubing doesn't kink while forming. In the mention of this idea I originally saw, the guy used a rope, but a rope would leave marks on the tubing whereas a spring doesn't.

pic 1

I am using XSPC 7/16" id 5/8" od tubing in my setup. For the loop in the pic above, I put the tubing into a plastic Tupperware cup, immersed it in boiling water for a minute or two, then ice water, and repeated the process a few times. Obviously if you use a cup or the like, it needs to be able to handle boiling water. It would probably be easier to use a jar like a pickle jar since the tubing did have a tendency to want to work its way out of the cup due to the taper.

pic 2
pic 3
pic 4
pic 5

Once heated and cooled like this a few times, the tubing holds it shape.

I have a very tight spot where one radiator feeds the other radiator. The space measured from the outside (not the inside) between the barbs is only 2 1/4” (about 57mm). The inside space between the barbs is only about 1 1/4”. Tubing would normally kink in a bend this tight. Since the piece of tubing was very short, and the bend had to be so tight, I simply used a pair of tongs to hold the tubing while forming. Using tongs did leave a small indentation on the tubing, but it is no big deal.

pic 6
pic 7

The tubing formed so that just sitting the outside was only 3” apart. I had no kinking problems putting it at 2 1/4” apart.

pic 8
pic 9

I also used this on my pump to GPU connection so that the tubing wouldn't be pushing against the pump.

Maybe this will be a help to others as well.

 

 

anyway i think doing a diy is fine but i think having different id is going to bottle neck but who nos un till you try it.

 

 

[Somerandomtechyboi]

14 minutes ago, thrasher_565 said:

hmm ill show you something. i remembered this years ago and could not find anything about it ahs it just come up as hardline tube bending...so i when to yahoo... and when to an old oc site and found it.

 

 guide by voigts on http://www.xtremesystems.org

http://www.xtremesystems.org/forums/showthread.php?239246-How-to-Bend-Form-Tubing-for-Your-Setups

How to Bend/Form Tubing for Your Setups

    I'm currently wrapping up some changes on my setup, and I thought I would share a trick which proved very useful that I recently picked up here on XS. I can't find the thread now, but the idea is one of the handiest I've come across in years of watercooling. It was only mentioned in a thread, but there isn't an entire thread about it. I used this several times in redoing the lower part of my loop, and it works great. Not only can it save you from having to use elbows and such, but it can help you avoid having situations where the tubing is pressing against barbs and wanting to kink due to tight bends.

    If you want to form tubing for tight bends or any bends, you can shape it the way you want. You simply make sure it somehow it is in a steady position in that shape, and then immerse it in boiling water and then in ice water several times and it will form how you like. I found a spring for $3 at a local ACE hardware that fits perfectly inside of the 7/16” id tubing. A spring is necessary in order to make sure the tubing doesn't kink while forming. In the mention of this idea I originally saw, the guy used a rope, but a rope would leave marks on the tubing whereas a spring doesn't.

pic 1

I am using XSPC 7/16" id 5/8" od tubing in my setup. For the loop in the pic above, I put the tubing into a plastic Tupperware cup, immersed it in boiling water for a minute or two, then ice water, and repeated the process a few times. Obviously if you use a cup or the like, it needs to be able to handle boiling water. It would probably be easier to use a jar like a pickle jar since the tubing did have a tendency to want to work its way out of the cup due to the taper.

pic 2
pic 3
pic 4
pic 5

Once heated and cooled like this a few times, the tubing holds it shape.

I have a very tight spot where one radiator feeds the other radiator. The space measured from the outside (not the inside) between the barbs is only 2 1/4” (about 57mm). The inside space between the barbs is only about 1 1/4”. Tubing would normally kink in a bend this tight. Since the piece of tubing was very short, and the bend had to be so tight, I simply used a pair of tongs to hold the tubing while forming. Using tongs did leave a small indentation on the tubing, but it is no big deal.

pic 6
pic 7

The tubing formed so that just sitting the outside was only 3” apart. I had no kinking problems putting it at 2 1/4” apart.

pic 8
pic 9

I also used this on my pump to GPU connection so that the tubing wouldn't be pushing against the pump.

Maybe this will be a help to others as well.

 

 

anyway i think doing a diy is fine but i think having different id is going to bottle neck but who nos un till you try it.

 

 

With soft tubing its literally just hook it up wherever without f ing around with boiling water

 

Though hard tubing does have the advantage of being able to be formed unlike soft tubing so it can increase waterflow depending on how you form it, pretty much just like playing around with car exhaust headers and seeing which flows best