Build Log - Corsair Obsidian 900D

[TeaLeaf]

Ended up ordering the PWM SP Quiet Edition versions of the fans so the variable fan speed hooks up better with the fan controller.   I'm trying a Corsair Link as it fits nicely inside the case and is software controlled - and hence I won't need to break up the nice clean look of the front of the case.  



I'm not putting a bluray/dvd/cd drive into this build either.    

If I have to go with a controller other than the Corsair Link then I'd probably wait for supplies of the Lamptron FC5 v3 to become available.

[TeaLeaf]

Started on the build this afternoon as most bits have now arrived.

-confirmed that a 360 rad does not fit in the front of the 900D under the 5.25" bays (unless you drill some new holes), so:
-removed 12 HHD bays from the lower compartment and switched 8 of them (inc the 4 hot-swap HDD bays) to the upper compartment under the 5.25" bays
-replaced the stock 140 and 3x120 fans with the AF140QE and the 3xSP120QE fans.    Red accent rings selected & fitted!
-played around with the radiator positioning, looks like I will almost certainly have 480 in the top and 360 in the bottom compartments
-fitted gaskets to the outside edge of the rads to seal the gap between fan and the radiator body to improve airflow

Currently trying to decide which way around to put fans and the PSU to get the best visual and the best airflow.   The red ring faces out when you you have the fans in an INTAKE orientation whereas when they are used to EXPEL they have the red rings facing the inside of the case.  So I'm probably going to do a bit of both with intakes being: 3 x SP120QE in the front, 3 x SP120QE on the bottom 360 rad (push config); and Expelling fans should be visible through the case side window: 1 x AF140QE by mobo and 4 x SP120QE in push configuration on the top 480 rad.

The PSU fan on the AX1200i sucks air in then expels it through the rear of the PSU case.    So I can either have the fan sucking clean air in from the dust-filtered side panel (in which case it appears upside down), or have the fan facing the interior of the case (where it now appears the right way up from a logo & name perspective) where it will end up sucking in air just expelled from the 360 rad.    I know it should be the former, but the latter looks so much nicer through the window! :blush:

I'll try to bash some piccies up tomorrow.

[Chaosphere]

I couldnt fit my rad in the front of my case with any HDD trays at all there, so I cut the HDD trays out (not removable) and mounted the HDDs to the side panel, just in front of where the mobo tray ends, using rubber grommits to stop contact between the drives and the case. Looks gorgeous, I think, and meant I could cram my rad in

But, I put all my stuff in the smallest case I could find to do it with

[TeaLeaf]

Delidding is scary, ok? :crazy:

I have a cheap clamp (metal mechanism but faced by me with pieces of mdf) on my woodworking bench.   Clearly this is not strong enough as I thought it would have been as the mdf faceplates have started to rip themselves out of the screws that secure them when I bash the CPU with the hammer.   Think I need to go and use a proper metal vice that will not suffer the same movement problems, I reckon most of my hammer blow force went into ripping the screws out :sad:

It's darned annoying when you psych yourself up to do something crazy (like delidding your new cpu) and then it doesn't work!

[lionheart]

You sir have far bigger danglies than I.

[sulky_uk]

You sir have far bigger danglies than I.

probably the best reply ever on dMw

[BrotherTobious]

Fair play TL can't wait to see the rig built

Sent from my Nexus 4 using Tapatalk 4

[TeaLeaf]

I borrowed a metal vice this morning from a neighbour and got back to delidding my Haswell CPU.

Tools:
Hammer
Block of wood
Metal-faced vice (not metal mechanism with wood-faced)
and my new Intel i4770K cpu:
[ATTACH=CONFIG]2328[/ATTACH]
I clamped the CPU into the vice upside down (with it's heatsink 'wings' aligned parallel to the two vice faces).    
The vice clamps the CPU at the very bottom of the heatsink (ie as close to the PCB as possible) so that the PCB extends over the vice face and is flush with the top of the vice face.
The mounted cpu should look like this (image courtesy of overclock.net as I forgot to take a picture whilst it was in the vice!):


I put a square block of wood completely flush against the side of the PCB and on top of the vice.
I placed a cloth bag under and around the vice so that if the PCB/CPU did come off it would be caught by the bag and not fly off and damage itself.
I then used a standard claw hammer to hit the wooden block horizontally (from right to left in the above picture).

It took 3 raps with the hammer and the CPU popped off and was caught neatly by the cloth bag, undamaged.
The trick to this is definitely to aim to deliver a sharp shearing force to break the 'glue' on the IHS.  
The result was this nice dlidded cpu:
[ATTACH=CONFIG]2329[/ATTACH]

Check out the amount of TIM and black 'glue'.  It's not brilliantly apparent from the pictures, but the IHS did not have fantastic contact with the CPU and the volume of TIM was way too much to maximise cooling potential.   Both the CPU/PCB and the IHS will now get a thorough cleaning with some Akasa TIM cleaner which I have left over from previous builds.

I have some Coollaboratory Liquid Pro Liquid Metal TIM arriving tomorrow (I hope) which I will put directly on the CPU die.   As it is a conductive TIM, I will also mask off the area on the left of the cpu with the exposed contacts and give it a good coating of nail varnish to insulate the contacts from accidental TIM contact and a short that would fry my CPU.    

Once the delidded cpu is installed in the motherboard then I'll be simply placing the IHS back onto the Liquid Pro-topped cpu die and the Haswell locking mechanism will hold the IHS in place and flush on top of the cpu die for good thermal contact.     I have some Noctua NT-H1 TIM which I will be using on the outside of the IHS to give the waterblock a good thermal contact patch.

The black stuff around the edge of the IHS did appear to 'lift' the IHS off the cpu die a little, further degrading the performance of the TIM, so I am sure that having delidded and cleaned the cpu & IHS and reapplying some Liquid Pro & NT-H1 I will get significantly cooler core temps from this process.   From reading around the typical results from this kind of delidding and application of these TIMs is about 28-30 degrees improvement in the cpu core temps, which can only be good.

So, I'm very relieved at having got through the delidding process unscathed.   As I have never delidded a cpu before it was a worry, but having completed the delid I can honestly say that it was remarkably simple and easy, something anyone could do.

Difficulty Factor: 1/10
Sphincter Factor: 10/10

[sulky_uk]

well done mate, was on tenterhooks reading waiting for the "and i broke the flipping thing" remark but well done

[TeaLeaf]

That black stuff around the edge is just plain nasty.   It doesn't disolve in TIM cleaner, so I ended up having to scrape it off the surface with a razor.   It's sort of like a black mastic, but much harder, more like rubber.  No wonder it wrecks the IHS spacing.

[Tutonic]

Well done sir - you're a braver man than I.

[kregoron]

The black stuff is plain crazy, tried a great assortment of chemicals when removing mine, (even tried a few acid ones, which didn't help) so a razor blade was the only way. I wonder what its made of.

Tho i cut my IHS with cutting wire (its a little like sandpaper, just in a thin wire form)
Help tremendously with Haswell heats issues, as im it suspecting it to be one of the culprits causing the general low OC capabilities of the Haswell.

[Ranualf]

That black stuff is very reminiscent of something we use in construction, it was a 2 part epoxy resin sealant, and it stuck to everything like poo to a bed blanket!
Basically it looked like a sausage, which you loaded into a "gun"...a fat assed mastic gun basically http://www.sealantsandtoolsdirect.co.uk/silicones_and_sealants/marine__auto_and_specialist_sealants/sikaflex_298_sika_marine_deck_bedding_adhesive_600ml_box_of_20_P26185.html?gclid=CJSg84Tzi7kCFUXKtAodsTAAdg

Just remembering how nasty it was to get off skin makes me shiver, as i was using it in the middle of summer, dressed in full overalls with rubber gauntlets on, and i still managed to get some on myself!
Alas my work mate didnt bother to use any ppe, and had some on his hands, his face(he smoked) and arms... nothing took it off ... NOTHING, we tried the chem remover stuff- nope, diesel..nope, petrol.. nope.. so he rolled a cig, popped it into his mouth and pulled out his lighter... i grabbed his hand and asked, "your`e not going to set yourself on fire now are you?  considering that you got petrol over you hands and arms..."

he dropped the cig

about 6-8 weeks later he still had some of that black goo on his arms, despite a daily scrub with a floor scrubbing brush (basically the skin Underneath died, and it pealed off... eventually)

Good times

Also, brave man TL, brave man, challenge accepted and beaten, i`m just glad the "operation went as planned".. so many what if factors..hope its worth the effort with the new rig!

[TeaLeaf]

Graphics cards arrived this morning!  EVGA GTX770 2GB.   I would have loved to get 4GB versions but there was simply too many problems locating blocks that would fit the non-reference PCBs, so I had to stick with a 2GB reference design in the end for simplicity's sake.
[ATTACH=CONFIG]2330[/ATTACH]
The waterblocks arrived too (both the front waterblock & the rear plate).
[ATTACH=CONFIG]2333[/ATTACH]
So it was off with the ACX cooler and a clean up of the EVGA TIM.    Interestingly the thermal pads they used under the ACX cooler seemed to only hit half the RAM cards, there is clear indentation on the smaller chips to the left of the picture, but almost no marks whatsoever on the vram chips.
[ATTACH=CONFIG]2331[/ATTACH]
The waterblocks went on real smooth, I trimmed the vram thermal pads slightly to ensure a trouble free fit and put some Noctua NH-1 on the gpu itself.   It reallyu needs two hands to pull the whole lot together as a single screw holds both blocks plus the card itself together, but I managed to balance it on the box to keep everything aligned so that the screws went in ok and the thermal pads did not move before tightening.
[ATTACH=CONFIG]2332[/ATTACH]

Tomorrow is (hopefully) dry-fit day where I start running pipe between components and seeing how the loop looks.

[TeaLeaf]

Been trying to build the loop today and failing miserably.    One of the XSPC G1/4" (1/2" ID, 3/4 o/D) compression fittings fits fine into the mobo's mosfet cooler block, but at the other end of the plastic on top of the block stops the compression fitting screwing all the way down - and I don't want to take a dremmel to it and end up putting a hole in what used to be a watertight block!    

Now trying to source some thinner fittings.    They're tricky to find in stock without compromising the looks (I'd gone for black nickel XSPC), so I'm looking for black or similar as a replacement.    Have ended up ordering some Koolance QD3 black 'no spill' quick disconnects' but there are not enough in the UK at the moment to complete the build so I have to wait for them to arrive with the UK distributor from Germany on Tuesday next week, so I'll not be completing the loop until next week now :sad: