So, I finally ordered parts for my first WC build!

are you able to push your 3930k higher than 4.8?

I haven't tried yet.

I accidentally wiped out my overclock settings in BIOS when i had to reset it to stock to reflow my Indigo Extreme.

It's been so long since I overclocked my system that I've forgotten the Sandy-E overclocking tips and tricks. I'm going to have to read up on it again first. :p
 
are you able to push your 3930k higher than 4.8?

I accidentally wiped out my overclock settings in BIOS when i had to reset it to stock to reflow my Indigo Extreme.

Actually, I was wrong.

I still had an old 4.6Ghz profile in there I can build on. That one was tweaked for lowest possible temps at 4.6, so it had lower voltages than my 4.8

Just did a quick 20 minute prime 95 run at 4.8. Vcore maxed out at 1.464, and CPU temp maxed out at 70C. Prime 95 is an interesting stress test. Certain iterations are much higher load than others. My temps sit at ~58C most of the time at 4.8, but occasionally a particularly heavy iteration comes along and drives it up to 70...

I did try 5.0 and it instantly crashed on me when I started Prime95, which is odd, because I remember having it almost stable at one point on worse cooling (stable on cold days, not on hot ones) and I'm blasting th eAC today. Voltage may be slightly lower now than it was during those runs. I can't remember all the details though.

Next I'm going to try for 4.9 with the same voltages.

I'm trying to decide just how high I am willing to go on the voltages. Testimonials are all over the place. I've had this chip since it was new, and for much of it's live it has seen just over 1.45 at full load, and it is still running strong 5 years later. That being said, i really only see full load on all cores during stability testing, so it might actually not have spent THAT much time at those voltages.

Peoples advice on this matter seems all over the place online. Some thing even going over 1.5V would be safe with adequate cooling, others caution breaking 1.4V

Edit:

Nope. Instafreeze when starting Prime95 at 4.9 as well. I'll likely need more voltage. Will think about whether or not I want to do that.
 
Last edited:
Meanwhile, I just noticed an odd RAM problem.

I have 64GB of RAM installed. 8x 8GB sticks.

The BIOS sees and reports a stick of RAM as present in each slot, but when it totals the RAM up, it only shows 48GB available. This is definitely a new problem. I might have knocked into something when I had the parts out to install the water cooling loop. Going to have to reseat all the ram sticks and hope that fixes it.
 
Reseated all the RAM.

Broke off two RAM clips in the process

Man this motherboard is old and crumbling. They used to add lead to the plastic to make it stronger, but they can't do that anymore due to European RoHS regulations, so now plastic ages worse than it used to.

This never would have been a problem.in the past though, as a 5+ year old motherboard would have been hopelessly obsolete. Not anymore :p

Anyway. RAM was still missing when I started it back up. Surprisingly it came back when I returned to stock clocks, so it appears to be relates to my overclock.

Maybe the ram controller needs some.mote voltage?
 
Seems to have been an odd BIOS problem.

load up default settings, and then load up overclock again, and RAM is back...

I think I've decided to stay at 4.8Ghz for now.

I could definitely get 4.9 out of it, and might be able to coax it up to 5.0 with a little more voltage, but at this point, it's a geriatric motherboard and CPU, and I don't feel like pushing it unless I really need it. 4.8Ghz is plenty fast. 1.92 single threaded Cinebench R11.5 performance puts it in good company.

If I suspect I am CPU limited before it is time to upgrade (which judging by how easily plastic parts are breaking off, might be soon, but I hope it lasts until Zen launches so I can see what AMD brings to the table) I may revisit this, but for now 4.8 is fine.
 
I've got an Aquaero 5LT that I'm using for fan control; it's a pretty neat little device. I've got it set up with my radiator fans on the PWM channel and two temperature sensors; one ambient case and one coolant in the reservoir.

In the Aquasuite software I have a virtual sensor that is the difference between ambient and coolant temperatures and a custom fan curve that will ramp the speed up from 30% to 50% fairly quickly over a short change then gradually to 100% over the rest of the curve.

The initial quick ramp is where I have a difference of about 1 to 3 celsius in my setup and the rest happens after 3 celsius. I think full speed is hit with a 10 celsius difference (I've never seen this, I usually am maxing out at 6 celsius when running a GPU test + CPU test). It's a pretty neat tool and beats my old static numbers for the fans, haha.

This keeps it pretty much inaudible for me when just doing normal PC stuff and when gaming it gets a little windy but nothing loud, like running the fans at full blast.

Those seem nice. Cost a little bit more than $8 though :p

Does it output a PWM signal, or does it control fan speed based on voltage?

I might install a switch on the front of the case hooked up to the PWM lines so I can switch the fan control between the Chinese pwm controller based on loop temperature, the motherboard fan header based on actual CPU temp, and a manual fan input.
 
Seems to have been an odd BIOS problem.

load up default settings, and then load up overclock again, and RAM is back...

I think I've decided to stay at 4.8Ghz for now.

I could definitely get 4.9 out of it, and might be able to coax it up to 5.0 with a little more voltage, but at this point, it's a geriatric motherboard and CPU, and I don't feel like pushing it unless I really need it. 4.8Ghz is plenty fast. 1.92 single threaded Cinebench R11.5 performance puts it in good company.

If I suspect I am CPU limited before it is time to upgrade (which judging by how easily plastic parts are breaking off, might be soon, but I hope it lasts until Zen launches so I can see what AMD brings to the table) I may revisit this, but for now 4.8 is fine.

Had the same problem with my f2a88x-up4 a few times. I think it happened when increasing the bclk (which also increases memory clock), but also sometimes when increasing the core multiplier. Solution was to reduce clocks or increase voltage on the memory controller (northbridge?).
 
mas envious of your 3930k. mine is kosher at 4.5ghz 1.4v but above that ratio, no matter how much voltage i add, the system hard resets within seconds of starting a stress test. ibt peaks the cpu in the low 60'c max at its stable oc. i'm contemplating an upgrade to a 4930k.
 
Heh- I have never even tried to put my 3930 past 4.1, but Sabertooth starts getting cranky and dropping RAM. But, a hex 4G+ is overkill for my 7970 for the foreseeable future. :D
 
mas envious of your 3930k. mine is kosher at 4.5ghz 1.4v but above that ratio, no matter how much voltage i add, the system hard resets within seconds of starting a stress test. ibt peaks the cpu in the low 60'c max at its stable oc. i'm contemplating an upgrade to a 4930k.


Yeah, it's the original C0 stepping though. No VT-d, but better overclocking...

I wouldn't mind having VT-d
 
yep, my hex is a c2. what do you think about the 3960x. it seems on ebay they average $200 for the c2 stepping vs $300 for the 4930k.
 
yep, my hex is a c2. what do you think about the 3960x. it seems on ebay they average $200 for the c2 stepping vs $300 for the 4930k.

Havent really thought about it to be honest.

Ivy-E and Haswell-E don't overclock as high as Sandy-E could. Between Sandy-E and Ivy-E Intel appears to have changed their process and optimized more for power efficiency, and less for high clocks.

It is fairly common to find Sandy-E's that will hit 4.8, and not extremely rare to find ones that can be coaxed a little bit higher.

My understanding is that Ivy-E and Haswell-E rarely overclock above 4.5Ghz.

On the flip side they have slightly improved IPC.

So, as I see it, with Sandy-E I get ~4%-6% higher clocks (depending on how high it overclocks), buty Ivy-E only has about 2% IPC improvement, so I am getting better performance with Sandy-E than I would with Ivy-E.

Haswell-E and Broadwell-E parts have more IPC gains though, so once you look at those the upgrade is likely worth it. I'd skip Ivy-E though. The only improvement is likely to be Heat/Power use.
 
GPU Temps are amazing. Overclocked to 2063Mhz the Pascal TitanX never goes above 34C under load at a 74F (23.3C) ambient temp.

CPU temps I am less sure about. The CPU maxes out at 54C in Prime95 at stock, non-overclocked settings. I know these Sandy Bridge-E chips give off more heat than the latest chips though, so maybe this is normal? I'll have to google for others results.

Considering 74F is 23.3C, and my water temps are about 30C at load I guess that means I'm at a ΔT of 5.7C which isn't terrible. I guess I could improve on that by ramping up fan speeds higher, but I'm not sure I want to.

Just need to figure out if these CPU temps are right...



Oh,

I forgot to mention that I didn't use the paste EK included with the GPU block.

I went with some Thermal Grizzly Kryonaut instead.

Not sure how much of a difference this may have made.
 
Oh,

I forgot to mention that I didn't use the paste EK included with the GPU block.

I went with some Thermal Grizzly Kryonaut instead.

Not sure how much of a difference this may have made.

Ive got several tubes of the EK goo. I haven't tried it yet. It's a lot of work to get my monoblock off the mobo. A whole lot. I have plans on getting some more CLU(used it on my 280x crossfire setup), I wanna get some under my 6600k lid. Not sure how I feel about having the Gelid under it right now.
 
Ive got several tubes of the EK goo. I haven't tried it yet. It's a lot of work to get my monoblock off the mobo. A whole lot. I have plans on getting some more CLU(used it on my 280x crossfire setup), I wanna get some under my 6600k lid. Not sure how I feel about having the Gelid under it right now.

Yeah, I decided on the Thermal Grizzly Kryonaut because of the non-conductive pastes it seems to perform the best in every test I have come across.

I really wanted to make sure my GPU temps were as low as possible, as I am trying to make a single GPU work for 4K gaming, so I needed to really focus on GPU temps, or so I thought.

In retrospect, had I known I'd see 34C load temps overclocked, I would probably have been less concerned :p

CPU temps I was less concerned about. the 4.8Ghz I was hitting before the water loop was more than enough for me. I even sometimes turned it down to stock clocks just to get some peace and quiet, and it was still enough :p
 
Funny side note.

I was playing Fallout New Vegas last night. First game I've really played since getting the loop up and running.

I play it at 4K with vsync at 60hz and Ultra settings but no mods (yet).

I wasn't expecting it to load up my system at all (it is an older title after all) so I didn't even set OC settings, but it was really amusing to see that the GPU temp didn't budge AT ALL over idle temps even after playing the game for an hour :p
 
Funny side note.

I was playing Fallout New Vegas last night. First game I've really played since getting the loop up and running.

I play it at 4K with vsync at 60hz and Ultra settings but no mods (yet).

I wasn't expecting it to load up my system at all (it is an older title after all) so I didn't even set OC settings, but it was really amusing to see that the GPU temp didn't budge AT ALL over idle temps even after playing the game for an hour :p

Lol yeah. In FO4 1440 ultra, I have adaptive vsync set at 75. So vsync at 75, and if it drops below, turns vsync off. Played for a few hours yesterday @ 1404mhz on the 980ti, and it leveled off at 34°C.
 
Haswell-E and Broadwell-E parts have more IPC gains though, so once you look at those the upgrade is likely worth it. I'd skip Ivy-E though. The only improvement is likely to be Heat/Power use.

really appreciate the feedback. i might fish around for another 3930k or 3960x to play with since they're so affordable these days. great looking first water cooling build. my first looked like trash.
 
Seconded! My first loop, well let's just say there is no photographic evidence. :D
 
really appreciate the feedback. i might fish around for another 3930k or 3960x to play with since they're so affordable these days. great looking first water cooling build. my first looked like trash.

Seconded! My first loop, well let's just say there is no photographic evidence. :D

Ahh, but when you first did yours, were there as many good resources out there to learn from as there are today?

While having a custom water loop still puts us in an extreme minority of computer users, I get the impression it is a lot more accessible today than it once was.

Anyway, I do appreciate the compliments.

I think I had some practice in doing semi-custom things when I put together my 3x AIO cooler setup for my dual 980TI's. One H110i GTX in push pull up top, and two 980ti's with HG10 n980's attached to one H90 each and installed in the front.

The tubes on the H90's weren't long enough to reach all the way to the front of the case, so I bought four of the cheapest 140mm fans I could found, and cut out the centers to use the plastic surrounds along with extra long screws as spacers.
 
Ahh, but when you first did yours, were there as many good resources out there to learn from as there are today?

While having a custom water loop still puts us in an extreme minority of computer users, I get the impression it is a lot more accessible today than it once was.

Anyway, I do appreciate the compliments.

I think I had some practice in doing semi-custom things when I put together my 3x AIO cooler setup for my dual 980TI's. One H110i GTX in push pull up top, and two 980ti's with HG10 n980's attached to one H90 each and installed in the front.

The tubes on the H90's weren't long enough to reach all the way to the front of the case, so I bought four of the cheapest 140mm fans I could found, and cut out the centers to use the plastic surrounds along with extra long screws as spacers.

Doing the fan thingy, we call that plenum :)
 
Those seem nice. Cost a little bit more than $8 though :p

Does it output a PWM signal, or does it control fan speed based on voltage?

I might install a switch on the front of the case hooked up to the PWM lines so I can switch the fan control between the Chinese pwm controller based on loop temperature, the motherboard fan header based on actual CPU temp, and a manual fan input.

Haha, it did cost a little more than $8 dollars but I'm mostly OK with that :) It is a PWM signal on the fourth fan header for the 5 LT; the first 3 are DC control. In the newer 6 version all the fan headers can do PWM signal but they don't make one without a display and I didn't want to spend another $100 to just remove the display and use the board itself especially since I just needed one PWM port for the radiator fans. In the future I might have to go that route if I start going crazy with fans in push-pull and multiple pumps but I doubt that'll happen anytime soon! Next up is hard lining my system anyway, haha.
 
I've been working on making my own DIY manual PWM fan controller.

Right now fan speed is controlled automatically via water temp, but I want to install some custom switches on the front, allowing me to have the fans set in three settings, off, manual, and auto water temp.

The switching part is easy. I can just get some rocker switches and cut holes in a bay cover and install them. The manual fan speed control is trickier. It seems the only manual PWM fan speed control device on the market was the Zalman PWM mate, which had a design that made it difficult to integrate neatly into a case. And this is moot anyway, as it appears to have been discontinued and all but erased from the internet.

That's when I found this writeup over at the OCN on how to make your own DIY PWM signal generator, and decided to give it a shot.

I've never soldered circuits before, and I don't exactly have the steadiest of hands for fine detail work, so my soldering work is pretty rough, but it appears to be going well.

29403639391_05a4e6ab88.jpg
29194870320_8e1e1c1cf6.jpg



I'm having a momentary setback though, because of Amazon's stupid fuzzy logic. I searched for a 680pf capacitor, but because Amazon shows you similar things to what you search for, instead of theexact things you search for, I accidentally ordered 680 uF capacitors, which I failed to catch before I soldered one to my board. I'm waiting for the replacements now.


attachment.php




I really wish there were a way to shut off Amazon's fuzzy search. I NEVER want to see similar items. I ONLY want to see EXACTLY what I am searching for.
 
I really wish there were a way to shut off Amazon's fuzzy search. I NEVER want to see similar items. I ONLY want to see EXACTLY what I am searching for.
Yeah, all fuzzy searches should be thoroughly destroyed, imo. I don't mind being given suggestions based on what I searched, but showing me stuff I didn't search for and changing my search terms and categories because you think you know better is beyond irritating.

Anyway, I'm pretty sure there were instructions for making a pwm fan controller in these forums somewhere, too, though I could be wrong...
 
So, the PWM controller is done, and works great.e

It's butt ugly, but it works great.

Here it is, sealed up with some hot glue, just so nothing shifts around and causes a short:

IMG_20160908_152941.jpg


I ordered a 3 pack of extra drive bay covers from corsair so I could chop one up and install switches and the pot with a dial on it to control my fans.

IMG_20160908_214604.jpg



Middle switch: Fans On/Off
Right Switch: Fans Manual control or auto control based on water temp (if middle switch is on)
Dial: Controls fan speed via PWM

IMG_20160908_214619.jpg



Turns out th efront panel on Corsairs Obsidian series bay covers is textured sheet metal. It is glued somehow to the plastic part of the bay cover underneath. I pried the metal part off using a contractor knife. The goo from the glue was intense. Took lots of goo-off to clean both sides. Then I carefully drilled/dremeled holes to fit my two switches and my pot dial in the metal front, and went nuts with my dremel on the plastic to make everything fit in the back.

I was going to glue them back together again, but I think the screws that tighten the switches and pot to the panel do the job, so it's not needed.

The hole locations weren't exactly perfect. I measured them precisely, but kept slipping with the damned drill when I went to drill them. Should have used a punch to start the holes, but I couldn't find mine. I don't think I can tell that they are misaligned, but if it winds up bothering me, I have three more spare Corsair drive covers to re-do it on. :p

I wiped the damned thing down, but it appears to be statically charged, and clings to all the dist from my dremeling. Will have to do a more careful cleaning when I put it all together.

I had originally planned on using rocker switches instead of the bat-handle toggle switches, but every single DPDT rocker switch I found onlyhad about 1mm of cut margin on each side that the bezel covered. I decided it would be rather challenging to cut the metal in a straight line with 1mm bezel overlap, so instead I went with bat handle toggle switches. I'm not crazy about the look, but when covered by the black boot, they aren't terrible
 
Last edited:
Wow, looks great. I heart toggles.

Thanks!

And now it is installed!

I haven't broken out the DSLR yet, so here isa quick and dirty cell phone shot. Looks like I need to wipe down the front of it (the textured metal rubs off skin onto it very easily when touching)

It's not perfect, as I am not a seasoned case modder, but its good enough to be functional, without looking embarrassing. I like how the black switches and dial kind of blends with the Corsair Obsidian design, at least in as much as they don't look like they don't belong there.

IMG_20160911_202713-crop.jpg


Middle switch is a fan on-off switch. When flipped up, fans are off, when flipped down fans are on.

Right switch is Manual/Auto control switch. When down, fans are controlled by water loop temp. When up, fans are controlled by the dial on the left.

Dial in this picture is set at approx 40% resulting in ~750rpm.

I also ran the fan speed wire to the motherboard header, so I can see the RPM's in carious system monitor tools. I's kind of fun to twist the dial on the front of the case and see the fan speed readout change in OS.

Once I wipe it down and break out the SLR I'll post some higher quality shots.
 
While I was in there I also installed the backplate on my Titan, and added some rubber bushings to see if I could reduce pump noise.

Not sure the bushings did anything, but I made a discovery.

It seems like in only 3 weeks of use, my Ekoolant appears to be cloudy?

I know people usually say the plasticizer leeches from the tubing, but this isnt supposed to be a problem with the Primochill Advancer LRT I used.

You guys think this is cause for concern?


Pic from fresh install:

res_avatar.jpg



Pic after 3 weeks of use:


3wk_cloudy.jpg



Maybe this warrants its own thread?
 
Hmm.. I'd start a new thread and give the topic it's own space. I'm curious as well. I'd guess plasticizer leech too, but that's an awful short amount of time for it to be that pronounced. I also just switched my two year old EK tubing that was exhibiting some signs of deteriorating to Primoflex, based on the reviews. I love the flexibility and look, but never used the stuff before.
 
While I was in there I also installed the backplate on my Titan, and added some rubber bushings to see if I could reduce pump noise.

Not sure the bushings did anything, but I made a discovery.

It seems like in only 3 weeks of use, my Ekoolant appears to be cloudy?

I know people usually say the plasticizer leeches from the tubing, but this isnt supposed to be a problem with the Primochill Advancer LRT I used.

You guys think this is cause for concern?


Pic from fresh install:

View attachment 7772


Pic after 3 weeks of use:


View attachment 7773


Maybe this warrants its own thread?
Yeah, that's weird. I have primochill adv lrt black tubing in my build and I hadn't noticed any clouding, though I didn't add a reservoir until recently, so maybe there just wasn't enough water for me to notice. Also less than 3ft of tubing, and used liquid utopia + distilled water instead of ekoolant, if that matters.
 
Your loop was probably not as clean as you thought. Flush it and use fresh water and see how it goes? You might have to go thru a couple flushes. If your rads were new, they typically come with with tons of contaminants leftover from the production.
 
Hmm.. I'd start a new thread and give the topic it's own space. I'm curious as well. I'd guess plasticizer leech too, but that's an awful short amount of time for it to be that pronounced. I also just switched my two year old EK tubing that was exhibiting some signs of deteriorating to Primoflex, based on the reviews. I love the flexibility and look, but never used the stuff before.

Yeah, that's weird. I have primochill adv lrt black tubing in my build and I hadn't noticed any clouding, though I didn't add a reservoir until recently, so maybe there just wasn't enough water for me to notice. Also less than 3ft of tubing, and used liquid utopia + distilled water instead of ekoolant, if that matters.

Your loop was probably not as clean as you thought. Flush it and use fresh water and see how it goes? You might have to go thru a couple flushes. If your rads were new, they typically come with with tons of contaminants leftover from the production.

New thread started. I'll be re-quoting your posts there.
 
Last edited:
One more time, bit this time with Bitspower...

...because several of my XSPC bends started leaking after 2 years of use.

178399_IMG_20181206_225048.jpg


Hopefully Bitspower does better...
 
One more time, bit this time with Bitspower...

...because several of my XSPC bends started leaking after 2 years of use.

View attachment 125823

Hopefully Bitspower does better...
Only fitting I've had leak was a y fitting from a(lpha?)cool. I have a bunch of xspc rotary fittings, but they're the stupid fat ones so maybe that's why they haven't started leaking. Edit: Looking back, looks like you used the same ones (except you got the black nickel finish instead), so that can't be why. I always try to avoid putting lateral torque on my rotaries too, so maybe that's it?
 
Last edited:
Only fitting I've had leak was a y fitting from a(lpha?)cool. I have a bunch of xspc rotary fittings, but they're the stupid fat ones so maybe that's why they haven't started leaking. Edit: Looking back, looks like you used the same ones (except you got the black nickel finish instead), so that can't be why. I always try to avoid putting lateral torque on my rotaries too, so maybe that's it?

I don't think I have much in the way of forces or torque on my rotaries.

Theories I have as to why this happens include:

1.) Bad batch from XSPC back when I ordered them in 2016. (I had one out of box failure that leaked during my initial leak test, the rest were fine for 2 years until now when three started dripping leaking)

2.) Overheating: There was one gaming session where I accidentally disabled all fans, and played for a couple of hours with the fluid reaching inappropriately high temps. Maybe this somehow damaged the seals? Maybe this was even exacerbated by the fact that my office sometimes gets quite cold during the winter, resulting in thermal expansion/contraction?

3.) Constant full pressure from my D5 pump: I have the old style D5 with a 1-5 speed selector switch, which I have just left at 5 (because lowering the speed doesn't seem to impact sound level, so why not? I have noticed that most seem to run theirs at lower than full speed. Maybe running at full blast for 5 years eventually wore the seals?

That's all I can think of right now.
 
Decided to swap out the fluid while I am at it.

Rinsing it with some distilled water to get the old fluid out
 

Attachments

  • IMG_20181207_225513.jpg
    IMG_20181207_225513.jpg
    662.1 KB · Views: 0
I don't think I have much in the way of forces or torque on my rotaries.

Theories I have as to why this happens include:

1.) Bad batch from XSPC back when I ordered them in 2016. (I had one out of box failure that leaked during my initial leak test, the rest were fine for 2 years until now when three started dripping leaking)

2.) Overheating: There was one gaming session where I accidentally disabled all fans, and played for a couple of hours with the fluid reaching inappropriately high temps. Maybe this somehow damaged the seals? Maybe this was even exacerbated by the fact that my office sometimes gets quite cold during the winter, resulting in thermal expansion/contraction?

3.) Constant full pressure from my D5 pump: I have the old style D5 with a 1-5 speed selector switch, which I have just left at 5 (because lowering the speed doesn't seem to impact sound level, so why not? I have noticed that most seem to run theirs at lower than full speed. Maybe running at full blast for 5 years eventually wore the seals?

That's all I can think of right now.
I have a D5 Strong (only one speed), haven't had any issues. Overheating, I guess could be a possible cause, but I'm a bit skeptical about that too. My bet would be on bad from the factory, or otherwise normal wear from use.
 
im also running a pmp d5 strong but with the koolance 24v pump ocer. yeah it exists! i kid, it allows me to run either of my 12v d5s up to 24v which i have done for very long durations being that i forget i have turned the knob. it will drop the coolant level in my res when i crank it up to 24v but only briefly(negative pressure?). i never had a chance to run the d5 ocer before my rotaries leaked. the liquid temp has never really gone up either when i was monkeying with the pump speed. shit who knows, none of the above? ALL of the above? im baffled now that the bp rotary has failed. 2 of the 3 rotaries ive most recommended and still use (ek s rotary) have leaked. everything i thought i knew...i think Z has a baby spider monkey climbing all over his gear and THATS what keeps causing the leaks!


the 24v d5 ocer
https://koolance.com/pump-or-fan-speed-controller-ctr-spd1224m
 
Back
Top