Night Fury - SR-2 [dual LGA1366]

[E-curbi]

Very cool motherboard bro, that SR-2. My dream rig a few years back.

I see so much amazing gear I want to build with, yet in the end cannot justify spending the dollars if only for the sheer pleasure of the build, yet when complete doesn't really help with getting my work out any faster, lol - the enthusiast vs the pragmatist. lol.

I have some of that amazing ddr3 from Corsair - back in the day like Scotty, I heard they fetch high prices the longer you hold onto them, so I'm planning on selling my Dominator GT 2133Mhz kit for $1000 in about ten years.

The EVGA SR-3 DARK will actually mount up to my chassis without any mods, whereas the Asus Dominus and Gigabyte LGA3647 - Xeon W-3175X boards are way too large and extended.

Actually spun the idea of getting the SR-3 Dark around in my head, worthy of about 3hours of pure entertainment planning the logistics, before arriving back at reality.

Good luck with your build plans.

 

[bogmali]

I am working on mine as well, took it out from an older DangerDen case and mounted it on a Phobya WaCoolIT bench. Just waiting for a GT-730 card and I can start crunching with this baby again. Excuse the ghetto mounting of some parts

 

[phill]

Hopefully this isn't de-railing the thread too much, but @bogmali how do you find the temps with the motherboard with them water blocked?

 

[Dinnercore]

Hopefully this isn't de-railing the thread too much, but @bogmali how do you find the temps with the motherboard with them water blocked?

I´d like to know myself, that would be good point of reference for me.

 

[bogmali]

Hopefully this isn't de-railing the thread too much, but @bogmali how do you find the temps with the motherboard with them water blocked?

I´d like to know myself, that would be good point of reference for me.

Back when I was tuning this setup for F@H competition runs, I used Lavalys' Everest (which is now known as AIDA64) to measure the NB at least. I have not run AIDA in a while so idk if it still has that feature

 

[Darmok N Jalad]

Back when I was tuning this setup for F@H competition runs, I used Lavalys' Everest (which is now known as AIDA64) to measure the NB at least. I have not run AIDA in a while so idk if it still has that feature

HWiNFO64 will measure more stuff than you may ever care to know.

 

[phill]

Back when I was tuning this setup for F@H competition runs, I used Lavalys' Everest (which is now known as AIDA64) to measure the NB at least. I have not run AIDA in a while so idk if it still has that feature

They are a little on the juciy side for power consumption that is a fact.. I was just curious since I've only had mine on air cooling since the day I've bought it but I've not thought I'd need to uprate any of the cooling. That said, all of my crunching PCs need a damn good clean anyways

It would be interesting to find out whenever you next use it just out of interest I love these setups, EVGA did know what they were doing when they made these boards

 

[Dinnercore]

I´m itching to complete this build, the wait is killing me and it´s only been two days since I last worked on it...

To keep myself busy I decided to delid my CPUs despite what I said initially. I got the time, might as well take the extra step.



If you want to delid these Xeons keep in mind that they don´t fit in some delidding tools due to their slightly larger substrate compared to the Core-series chips. I ended up using a big vice.



I used a No. 10 medical scalpel from Braun to get the indium solder off the die. I´d say a No. 22 will be fine too, but they are a bit big.



Marked them A and B to keep track of which socket they were in. I have never dealt with dual CPUs before and thought they might want to stay in the socket they used to work in

All that is left to do now is polish the surface a bit. Let´s just hope there will be no unexpected delays with my mainboard block or I might get bored and start to sand the die too.

 

[TheMadDutchDude]

Weren’t those chips soldered??

 

[R-T-B]

Weren’t those chips soldered??

He did mention taking a medical scapel to Indium solder... so yes.

 

[Dinnercore]

The waterblock for the board is gonna take a while it seems. These old parts are made in small batches which means they just now started the production of the batch my order is in. But I think it will be well worth the wait.

In the meantime I only did some minor touches, I installed some things in the case and did a quick testfire of the fans + fan-controller setup. I was a bit worried that those big 180mm fans will draw too much power for a single cable so I split them up in 2x4.




For the drive bay area I had some trouble with the front door of the case not being able to close when I install the reservoir so I put it at an offset further back into the case, same for the fan-controller. This way I can still close the door at the cost of not being able to close off the gap between the two with one of the drive bay covers. So I got an inline temp. sensor with display and will put that somewhere between the two.



In addition I also got a controller for RGB since the board does not support it and will re-use RGB-lighting strips from a previous PC of mine to create that glowing light for the top and front.

I can´t stand the wait, but it is what it is. Atleast GPU prices are falling with each passing week

 

[Dinnercore]

Finally the waterblock is done. Holy XYZ that took a while, but I think it was worth to wait for it.



And, this is sitting on an open GPU box Yes I now have all the parts and can start putting this thing together. It´s 4:30am, I could not sleep at all with everything sitting around so I´m getting to work now.

 

[HammerON]

Sub'd for an awesome build!!!

 

[R-T-B]

I am working on mine as well, took it out from an older DangerDen case and mounted it on a Phobya WaCoolIT bench. Just waiting for a GT-730 card and I can start crunching with this baby again. Excuse the ghetto mounting of some parts

Is it bad that I am tempted to try to figure out where in my city you live, just to gawk at this?

 

[bogmali]

Is it bad that I am tempted to try to figure out where in my city you live, just to gawk at this?

Closer than you think.......its not functional atm and I'm still working on it

 

[Dinnercore]

I see many people in here, I hope no one expects a tidy professional and ultra-clean build from me because what you will see soon is my typical chaos.

Spend ~ 12-13 hours working on it now and there were many complications. As I had finally mounted the mainboard in the case and I looked at the 16 fittings total in there, all kinda close to each other I felt for the first time like its growing a bit over my head. I only had 2x 90° fittings, all the other ones are straight... I think I will re-work this loop soon, for now tho I want to get this whole thing to power up atleast.

Anyway I completed the loop and currently I´m in the process of filling and leak-testing. And I encountered a problem with my loop. The dual D5s can´t fill it.



There is still air trapped in them, as the reservoir is right in the middle of the whole loop instead of on the bottom. Can´t really do anything about that since it´s mounted in a drive bay. When I turn them on they just recycle the water in the reservoir and build a bit of pressure at the far end towards my external radiator but not enough to get the water all the way through.
I don´t want to run the pumps with the air inside of them and even if I do they currently can´t push through the whole loop as I got air trapped everywhere that seems to be causing lots of back-pressure.

I tried each pump individually, and they both build positive pressure in the right direction, so the orientation is fine. So far everything is leak-free and I measured all the vertical tube length where it has to push water 'up' to be ~ 1.4m. I have less than 5m total loop length and 'only' 2 CPU blocks, a single GPU block, a 240mm rad, 280mm rad and the external radiator. The total tube length includes the routing to the external radiator.
So in total I thought the dual D5s can handle this... Seems like they can´t? EKWB has dual D5s in serial rated at 7m total head pressure, so I´m now a bit nervous if they can handle my loop.

Maybe it will sort out when I get rid of the air, but how I´ll do that I don´t know at this point.

 

[R-T-B]

Closer than you think.......its not functional atm and I'm still working on it

Heh. Now w1zzard knowing my screwy sleep and wake cycle suddenly makes sense... /tinfoilhat

 

[phill]

I see many people in here, I hope no one expects a tidy professional and ultra-clean build from me because what you will see soon is my typical chaos.

Spend ~ 12-13 hours working on it now and there were many complications. As I had finally mounted the mainboard in the case and I looked at the 16 fittings total in there, all kinda close to each other I felt for the first time like its growing a bit over my head. I only had 2x 90° fittings, all the other ones are straight... I think I will re-work this loop soon, for now tho I want to get this whole thing to power up atleast.

Anyway I completed the loop and currently I´m in the process of filling and leak-testing. And I encountered a problem with my loop. The dual D5s can´t fill it.



There is still air trapped in them, as the reservoir is right in the middle of the whole loop instead of on the bottom. Can´t really do anything about that since it´s mounted in a drive bay. When I turn them on they just recycle the water in the reservoir and build a bit of pressure at the far end towards my external radiator but not enough to get the water all the way through.
I don´t want to run the pumps with the air inside of them and even if I do they currently can´t push through the whole loop as I got air trapped everywhere that seems to be causing lots of back-pressure.

I tried each pump individually, and they both build positive pressure in the right direction, so the orientation is fine. So far everything is leak-free and I measured all the vertical tube length where it has to push water 'up' to be ~ 1.4m. I have less than 5m total loop length and 'only' 2 CPU blocks, a single GPU block, a 240mm rad, 280mm rad and the external radiator. The total tube length includes the routing to the external radiator.
So in total I thought the dual D5s can handle this... Seems like they can´t? EKWB has dual D5s in serial rated at 7m total head pressure, so I´m now a bit nervous if they can handle my loop.

Maybe it will sort out when I get rid of the air, but how I´ll do that I don´t know at this point.

When they quote the measurements, isn't that down to how high the liquid can be pushed?? I've really never taken any notice of it and even with single pumps in loops, I can't say I've ever noticed any issues otherwise when it comes to having the bubbles.. When the system is on for a few hours, all the bubbles works its way out of the system and it's then fine
My loop on my 5960X seems to make all sorts of noise at the best of times, so to be honest I never really take much notice

 

[Dinnercore]

When they quote the measurements, isn't that down to how high the liquid can be pushed?? I've really never taken any notice of it and even with single pumps in loops, I can't say I've ever noticed any issues otherwise when it comes to having the bubbles.. When the system is on for a few hours, all the bubbles works its way out of the system and it's then fine
My loop on my 5960X seems to make all sorts of noise at the best of times, so to be honest I never really take much notice

Yes but they assume a closed system with no air in it. Water does not compress or expand under pressure, air does. And I don´t have little bubbles, I got air pockets trapped between the lower radiator, the pump and the upper radiator. Which still would be no problem for a dual D5 to push out IF the impellers would be the lowest point in the loop and fully under water.

However I could not mount a regular pump + res combo in the bottom of the case as there is no space for it there. So now I have an air pocket from the lower part of the loop (240mm radiator and mainboard block) that just doesn´t want to go away.
Currently the impellers are only half covered with water and so they make some splashes but only throw air around in a circle, because the air wants to stay on top and the loop back-pressure is pushing that air right back at the pumps.

What I´ll try is to keep the loop closed and take the bay out, flip it upside down and put it a bit lower. Once everything is filled it should be fine. Now that I´m typing this I wonder why I didn´t do that 3 hours ago... Guess sleep is not as optional as I want it to be.

 

[phill]

Is there a way you can rock the rads/res to try and push the air out?

Would it be an option to run the water loop without the components (i.e. just the water loop with no hardware in, just in case of a leak or pressure build up and tubing coming off etc.??) would that make any difference or help at all?

It's kinda hard to help with cooling as being there to see is sooo much easier

 

[Dinnercore]

Is there a way you can rock the rads/res to try and push the air out?

Would it be an option to run the water loop without the components (i.e. just the water loop with no hardware in, just in case of a leak or pressure build up and tubing coming off etc.??) would that make any difference or help at all?

It's kinda hard to help with cooling as being there to see is sooo much easier

I´ve tried rocking back and forth, putting it upside down, putting the external rad upside down, shaking it gently. Nope.

Running just the loop would = a total rebuild. With the mainboard block involved and VRM cooler etc. this is the last thing I´ll try if all else fails. There is definitly no leak, all tubes are connected and I´ve already pushed the water through a full cycle by opening a port on the external rad and sucking the water in with a very long tube. So, no leaks, water makes it through the whole loop and all tubes are connected like they should.
However there is still air trapped in the middle, which sits between a partly filled radiator and the pump/res combo. So It has water on both ends but air trapped exactly where the pumps sit...

I feel like I´m making progress tho, by manually pumping through using just my lungs I can get a little bit of air out each time and refill the reservoir. By now I have 1,4l water in there and it still has room for more. I hope I can work the big air pocket out like this, because I tried to remove the pump/res-bay and noticed I can´t. There is a lip that only allows it to be pushed out, which is not possible because of the tube lengths.
Well it´s a thing you gotta learn the hard way. Next time I won´t place the pumps like that.

 

[phill]

I´ve tried rocking back and forth, putting it upside down, putting the external rad upside down, shaking it gently. Nope.

Running just the loop would = a total rebuild. With the mainboard block involved and VRM cooler etc. this is the last thing I´ll try if all else fails. There is definitly no leak, all tubes are connected and I´ve already pushed the water through a full cycle by opening a port on the external rad and sucking the water in with a very long tube. So, no leaks, water makes it through the whole loop and all tubes are connected like they should.
However there is still air trapped in the middle, which sits between a partly filled radiator and the pump/res combo. So It has water on both ends but air trapped exactly where the pumps sit...

I feel like I´m making progress tho, by manually pumping through using just my lungs I can get a little bit of air out each time and refill the reservoir. By now I have 1,4l water in there and it still has room for more. I hope I can work the big air pocket out like this, because I tried to remove the pump/res-bay and noticed I can´t. There is a lip that only allows it to be pushed out, which is not possible because of the tube lengths.
Well it´s a thing you gotta learn the hard way. Next time I won´t place the pumps like that.

I don't suppose there's a bleed hole or something on the rad that you can possible use to help with the trapped air? I know of some rads having them but otherwise it might just be a case of running the loop perhaps and letting the air work it's way out..
With those style of pumps/res combo's I've never used. The link in my post above where I show the updated build are typically the ones I use. I try to separate CPU and GPU loops as well as when both are dumping a lot of heat into one, I try to control it so it's not getting too hot and then fans need to spin faster etc..
Are there any way of running the loop with the caps off of the tops of the res to get any air out that way when the loop runs? I'm unsure, just talking hopefully not rubbish but something that might help Trapped air can be a nightmare in water loops

 

[dont whant to set it"']

Beasty build this one is for sure with respects to both altitude and speed.
let:typo

 

[Dinnercore]

Thank you for the suggestions and keeping my spirits up @phill !

I figured it out, it was just the airlock in the pumps themselfs that hadn´t yet cleared out. I was to big of a sissy to just let them run half-dry for a while. I tilted the whole tower a couple of times sidesways and turned the pumps on and off many times, eventually they got started.
And once these dual D5s start eating the water, they go absolutly crazy. I did not expect to see whirlwinds inside of my tubes - through nearly the whole length of them:

That is only speed 4/5. I can now rest in peace and catch some sleep while this thing does it´s final bleed and leak test.

 

[phill]

You can also try slowing it down if you'd like to see if that made any difference to the trapped air These pumps can move a serious amount of water and what we have in these systems are pretty much nothing for them in comparison Sometimes a bit of doing the wrong thing helps it out, just glad its sorted for you now

I'm real glad you have it all sorted out Get the CPUs and GPU loaded, run it for a few hours, you'll see all the air eventually work out of the system as the heat introduced will get rid of the air bubbles I'd really love to update some of my water loops, but....
Anyways, so pleased it's now all up and running Awesome work

Sorry, I don't think I put in enough smilies