Build log: Corsair 600T

[Sprucemoose]

Quite eager to move forward, so I hooked the pump up to the Aquaero connected to my old computer. Ready for some leak testing:





Remember, I had changed the layout for the reservoir a little bit by adding a 90 degree rotary fitting...bad move!

The bloody thing started leaking. First, I though the leak might come from the connection between the res and the fitting, but it soon became evident that it was from the fitting itself. I even tested this specific fitting doing a test-run outside the case when I still had the XSPC reservoir, and it didn't leak at that time.

I'm quite sure that the leak comes from this part of the fitting:





Now, the leak didn't have any real consequences, but I had to redo the reservoir bit. I rotated the reservoir and added my last straight fitting. I also moved the intake to the bottom of the two ports instead of placing it on the top. I had noticed that the upper intake produces a lot of bubbles when the water-level is not at the very top. Displacing the intake down seemed to eliminate the problem.





I've run the new setup for about an hour, and it seems water tight for the time being.

 

[Fragment85]

Sucks about that fitting, but things are looking great!

 

[Conumdrum]

I tried my res like your is setup. Doesn't bleed well at all. I changed it to inlet at the top with a barb in the res and a 2-3" piece of tubing of the inside barb.

Martin discovered it, here is his review:
http://martinsliquidlab.org/2011/02/25/swiftech-mcp-35x-reservoir/

 

[Sprucemoose]

I tried my res like your is setup. Doesn't bleed well at all. I changed it to inlet at the top with a barb in the res and a 2-3" piece of tubing of the inside barb.

Martin discovered it, here is his review:
http://martinsliquidlab.org/2011/02/25/swiftech-mcp-35x-reservoir/

Yes, I've been to that review prior to ordering the reservoir. I figured I try it with side inlet for the time being. If I am to swap to the top, I need another 90 degree rotary fitting, but I'm not really in the mood for another one of those leaking ones

I wonder if the sponge restricts flow much, because I only measure 1.2GPM at max pump speed. When the loop is bled I might try getting that sponge/grating out of there.

 

[m0r7if3r]

Yes, I've been to that review prior to ordering the reservoir. I figured I try it with side inlet for the time being. If I am to swap to the top, I need another 90 degree rotary fitting, but I'm not really in the mood for another one of those leaking ones

I wonder if the sponge restricts flow much, because I only measure 1.2GPM at max pump speed. When the loop is bled I might try getting that sponge/grating out of there.

Really shouldn't affect it at all, I would leave the sponge alone...it's there to stop vortexes, not bleed the loop.

 

[Conumdrum]

Flow meter...pfffttt. Those kinds restrict flow..........

Sponge is fine.

 

[Sprucemoose]

Ok, now then...things have turned to the worse. Bled the system, and everything looked good. Hooked up the cables and bumped into a hardware problem right away.

I have screen corruption during boot and in BIOS. Two green garbled bars running vertically, one in the middle, and one one down the right edge. I´m using an old single link DVI-D cable which have worked beautifully on my old system (don´t have a VGA or other DVI cable lying around, I´m afraid). Looks like the plug in the card is a dual-link DVI-I port.

What I have tried so far:
Tested with the displayport/DVI adapter, but still same problem.
Switched to second set of PCI-E power cables, no luck.
Took out the card, reinserted in slot, no luck

Switched to the iGPU graphics in BIOS and used same cable connected to the motherboard: All ok.

Switched back to graphics card and booted into a Windows which I had preinstalled on an SSD for my laptop. The whole screen had garbled graphics (this time tiny horizontal lines), but visible enough to navigate. No graphic drivers installed, though.

Then, when I was about to do a clean install of Windows to do graphics driver installation, the PSU went up and died

I was worried that it might be the motherboard, but even when using the short-circut trick on the 24 pin plug to check if the pump would start, it was all dead.

So what do you guys think about my graphics problem? My PSU is obviously dead now, but was it the cause of the problem in the first place? Is the graphics card defect? Is it as simple as a bad DVI cable? Any help and input you can give me will be greatfully received.

EDIT: PSU seems not as dead as I thought. Suddenly there was some power, but it was short-lived. Now it seems pretty much dead again.

 

[m0r7if3r]

I would try a new PSU, first and foremost. Bum one from a friend, pull one from your storage bin, RMA the one you have, SOMETHING, but try a new PSU. It could be that your psu was outputting awful ripple from some manufacturing defect (though I believe corsair does a stress test on all units before they ship ) and preventing it from functioning correctly. It could be that this killed your graphics, don't worry, just RMA anything that's dead (test as much as you can) and see what comes out of that...then work from there.

 

[Sprucemoose]

Just so you know what kind of graphic corruption I experienced in BIOS, here is a picture. It acctually looks to me that the two halves of the screen experience the same corruption (i.e. vertical bar on the righthand side of both halves). Does this mean anything in particular?

 

[Conumdrum]

Corruption like that is usually a sign the card is bad. The 8 pin mobo plug wasn't put in the PCI 8 pin connector?

I got mine backwards once. Mobo would not boot.

 

[m0r7if3r]

Corruption like that is usually a sign the card is bad. The 8 pin mobo plug wasn't put in the PCI 8 pin connector?

I got mine backwards once. Mobo would not boot.

lol how'd you even manage to get that to fit? I make that error every time I put my computer together and am always saved by the fact that they're shaped differently...

 

[Conumdrum]

Long story. 100% modular on the PSU, so I had a PCI plugged in on the Mobo connector on the PSU side. The PSU side is clearly labled as is my connectors, but they don't have the special pins on the PSU side. Silly mistake.

 

[Sprucemoose]

A small update to my issue:

Bad PSU:
I have no idea what is up with the PSU, some times it works, some times it does not. I can´t see any real pattern to the problem, and I haven´t had it shut down by itself after it starts. It seems to appear when disconnecting power and/or flipping the power switch.

Concerning cabling and plugs:
The PSU has 2x 12-pin connectors for either E-PCI (8-pin) or CPU12V (4+4 pin, only 8 pins are really wired on the 12-pin connector). It also has 1x 8-pin for CPU12V (4+4 pin) exclusively. The 12-pin connectors on the PSU should be interchangeable.

I had connected both the CPU12V and E-PCI to the 12-pin connectors on the PSU, but changing the CPU to the 8-pin did not do anything to rectify the problem (nor should it).

Testing with a different PSU:
I took out the 620W Corsair PSU from my old computer, it had 2x6-pin connectors for E-PCI, but the GFX-card came with adapters. Now, 620W is less than what the card is supposed to need (Powercolor states 750W), but it should not be needing that much just to get into BIOS, right? Anyway, the same problem persisted.

GFX drivers:
I managed to install Windows, I then tried to install the latest Catalyst drivers from the Powercolor website (even though I think it´s a hardware problem, not a driver issue). When restarting, I got past the post, but the screen just went black. Had to do a system restore in safe mode to get out of it.

Conclusion so far:
Even if the PSU is acting weird, the replacement PSU didn´t solve the problem. I´m starting to believe that it is not a power-issue.

The motherboard/RAM/CPU seems to be working properly as I can install/boot to Windows.

So that leaves me with 2 probable causes 1) a ruined GFX card or 2) bad/wrong monitor cable. I´m off to buy a dual link DVI-cable, and I´ll try to borrow an analog VGA cable from somewhere. If that fails, I´ll borrow a fanless GFX card from a friend just to see if that will resolve the problem.

 

[Gritash]

Your issues

Wow Spruce sorry to see the run of bad luck. I have an AX850 myself but I have never experienced any issues. Even when i was bleeding system w/out connecting mobo, was really uncomfortable shorting pins.

To compare psu's you would have to use same output, ie 850w vs 850w. by the sounds your rig is thirsty and w/ pumps and extra fans you cant really go lower than 850w.

Reseat the gpu and look maybe try placing case on the side, strain will be reduced on PCI_E slot. Also maybe clean the graphics cards pins w/ acetone(?). The 90 Degree fittings are tricky, i had to use a pliers to tighten it fully, but i had same nightmare scenario where it leaked @ 4am it wasn't nice..

Also try another DVI cable, sometimes if cables have been in place for long periods the cables can become used to working within the angles that they are currently positioned. If you are using an adaptor then try another one.

BUT you still cannot rule out bad Gcard so maybe try for RMA?

 

[Sprucemoose]

Ok, I've tried a brand new dual link DVI-D and a VGA cable producing the same result. Going to remove the 6990 and put in an older ATI card later tonight. If that works ok, I've isolated the problem to the gfx card. The PSU needs an RMA too, I don't know if the two problems are related, though.

Testing with a 620W PSU as I did should definately be enough as idle power consumption is pretty low.

I'll probably run the computer using the iGPU while RMAs goes through. Will post again as soon as I've tested another graphics card.

 

[m0r7if3r]

You've done a good job of isolating the problem here...only thing I can think of to try that you haven't is using the same gfx card in a different slot (when you have the baby-card, make sure you test all the slots to remove the mobo as any potential cause of anything)

 

[Sprucemoose]

You've done a good job of isolating the problem here...only thing I can think of to try that you haven't is using the same gfx card in a different slot (when you have the baby-card, make sure you test all the slots to remove the mobo as any potential cause of anything)

Thanks for that input. I acctually thought that you had to use the 1st PCI-E slot when running only a single graphics cards. I only thought that second (and third slot) where for Crossfire/SLI.

But if the baby-card as you call it runs correctly in the first slot, wouldn´t that be proof of that the 6990 is broken? Do you think I should try the 6990 in different slots too?

 

[m0r7if3r]

Yea, if it works, that means there's a problem with the 6990, but it's good to check everything anyways IMO

 

[Sprucemoose]

The new build tragedy is coming to a resolution. Switching to a different graphics card in the same slot (Slot 1) solves the graphical corruption.

Regarding testing different slots:
The manual of the motherboard says that slot 2 is only x8 speed, which means that if the graphic card works here, there could still be problems with the 6990 as it then will be able to transfer at lower speeds, right? In that case, testing different slots wouldn´t result in a more accurate decision. On the other hand, if it didn´t work, I would know nothing more than I already know now. I´m inclined to leave it as it is, blaming the graphics card (...and PSU)

One more thing: Remember that I bought the card for less as it had been returned by a customer? The story was that the metal bracket was a bit bent, so that the card wouldn´t fit in the case. Could this all be a cover for some guy overclocking it past its threshold, then returning it stating a different reason. The overclock BIOS switch is still in the default position, but the sticker that was supposed to cover it is gone. Any way to test and potentially reset the BIOS somehow?

 

[m0r7if3r]

techpowerup has a TON of bioses, you could try those