M17x R2 7970m CrossfireX Ultimate Installation and Tweaking Guide

Check my post on T|I. Not booting isn't necessarily bad. AlienHack has a simular situation as you where his secondary card does not work in the primary slot. He has two different revision of Dell cards though so it makes me wonder what hardware changes if any did CLEVO make on the card. Can you take a photo of the rear of the card and post it up here please?

The link to the CLEVO vBIOS pack is in my sig. I think sangemaru tried but the 3D voltage was greyed out if I remember correctly.

...in other news, the AMD Catalyst 13.11 Beta9.4 is out now. Direct link for mobility is on the page.

Feature Highlights of The AMD Catalyst 13.11 Beta9.4 Driver for Windows
- Includes all Feature Highlights of The AMD Catalyst 13.11 Beta9.2 Driver
- May resolve intermittent black screens or display loss observed on some AMD Radeon™ R9 290X and AMD Radeon R9 290 graphics cards
- Improves AMD CrossFire™ scaling in the multi-player portion of Call of Duty®: Ghosts
- AMD Enduro Technology Profile updates: XCOM:​ Enemy Unknown, Need for Speed Rivals

 

Heres a photo of the back. I tried stock voltages with no luck. I just dont get why i had it running last night in CF but it would lock idle at 99% till i turned of UPLS now I cant even load. I must have done something

 

Its the same hardware revision so that's good. Try the uninstallation of the driver through safe mode like I suggested on T|I.

 

Uninstalled the drivers. I can load into Windows now without the drivers. Tried installing the new drivers and hit the same issue of black screenin right after the windows screen

 

Are you running Windows 8 by any chance?

 

nope running windows 7 64

 

Thats strange... I know a few other users experienced similar problem. sangemaru did as well but he was running Dell cards. Hopefully someone else can help.

Try disabling a few devices in the BIOS and see if it gets past the Windows logo. Start with the High Definition Audio. Also you may want to try AlienHacks vBIOS, PM him for a copy.

 

No luck turning off the devices

 

Well, this has power issue written all over it as I see it. One of your cards could be drawing a lot more current of an odd reason.
I had same behaviour and was able to flash the few times it allowed me to boot enough, but never worked alone and had problems starting my computer but it worked in safe mode or without Xfire as long as the crappy one was secondary. This secondary card was the one I had running alone as primary for almost a year.

Xfire did work in the beginning but suddenly stopped after a driver update running on my 240W.
Of course i i could have killed that card during all the swapping to find out what was wrong later but I was able to change VBIOS oddly enough. (like you)

Your problem as I see it is that the one card does not boot alone in the computer since its a type that should. Just like mine.

You got it new or you got a used one?

I really hope im wrong.

Do you have other things running that you can stop?
External disks? Overclocking of CPU in BIOS? Fantray from USB? Etc.....

 

This sounds like the issue I'd been having with the card from Riri-Fifi, with the exception that I hadn't been able to bench with it at all, I could boot it in secondary slot without drivers (Standard VGA Graphics in Device Manager) but installing drivers would blackscreen freeze in windows and on restart would freeze during loading screen. I sent it back and bought another card and it worked.

Have you tried doing the 30-second power drain on the machine? Remove battery, hold down power button for 30 seconds. I'd even recommend taking out the cmos battery for 5 minutes and trying again.

But, and this is most important:

Check the card and heatsink seating and make sure all relevant VRM's (little black chips between and around the big inductors) - as well as the big inductors, ram chips and core - make good contact with the heatsink. Very important. I'd had non-booting cards (7970m included) that refused to boot unless properly cooled.

 

It's not a power issue. Chances of dead card, unfortunately.

 

No it was a brand new card from Eurocom. I don't have anything running. Even tried turning devices off in the bios

W does one find out for sure?

 

If it were a power issue, the machine would still boot correctly, you'd be able to use the card, and if power consumption went overboard, the machine would shut down and the light on the PSU go out immediately.

 

Yeah it shouldn't be due to excessive power draw, it should be in 2D mode anyway.

Maybe try the vBIOS that came with the card. Try a few different vBIOS's and drivers, if it still doesn't work you may have to return it. Sine you can flash it, it doesn't seem to be bricked, just incompatible.

 

The way I understood it, it worked originally, but blackscreened during a 3dmark bench, and then stopped working.
Could have been driver corruption, but since it doesn't appear to be the case, sounds like brick to me.

 

Sorry I mean dead

 

Hi. My 5 cents now

On the photo attached ealier i could see that you have those 4 small pads glued to the x-bracket.

Remove them completely


I was having issues with them, they would make load of pressure on the card when i was pushing it down to tighten the 2 screws for the final mounting on the mobo.

GPU PCB can bend, and the connection with the MXM slot is not 100% complete,

Also, tightening the screws of the heatsink too much when using thick thermal pads can bend your PCB even more, and you will end up on having card like this:



SO make sure your cards PCB is flat and NOT BENT at all, and Remove those Foam Pads from the bottom !

Cheers and good luck !

 

Thanks for the idea. I really wish it was something that simple but no luck

That was what I was and still am afraid of. Wish I knew what I did to get down this road. This is turning into one expensive headache

 

I think flingin's post is really something to take note of. If the PCB was bent by the foam pads and/or over tightened screws the card could have been damaged when it heated up possibly expanding during your 3Dmark run.

I really hope its not that though. CLEVO cards aren't like Dell cards since if you use the CLEVO x-bracket you have to use the CLEVO screws and there's no c-clips to prevent you from over tightening the screws.

That reminds me, Aristotles 8970M also stopped working during loading the GPU after the first install and it didn't work again. He couldn't even flash the card though.

 

sorry, i did not mean to discount his suggestion. I tried it right away and had actually cut the pads in half earlier thinking the same thing. and I have the heatsink screws on with maybe a turn right now. Earlier they were tighter so it does worry me that i might have cooked it during the 3d mark run. I feel like its starting to be more and more likely

 

After everything fails, Contact Eurocom to replace the the card. tell them it stopped working after a week and you just have no POST with the card.

They will replace it for you.

Cheers !

 

Well i might be wrong.

Bricked?
If he can load new VBIOS then it might not be bricked? Mine did the same. Isnt bicked when VBIOS dont want any updates?

Incompatible?
Why does it stop sometimes in early boot and move past other times, again stopping when drivers are loaded the few times it moves past if its incompatible?
During safe mode the drivers are not loaded and if you move past POST it will always start windows since the working card is not loading the system much.

During the Boot, when all cards and fans are powered in POST we have high consumtion. (Power On Self Test)
If your PSU is cold or warm it can move past or not depending on how much higher it gets, also warmer PSU will faster go into thermal shut down if power is to high.
Bad card does not mean a short, there could be resistors restricting power to 100W when GPU has full opening, reducing top load to 100W. (If gpu is dead or power parts of card)
With the not working card as only card it will always stop during POST, but not due to power lacking, more due to the only card not working.

My card did the same after 11 months use. I could also load new VBIOS.
That card was not incompatible. Could be another issue here, but same issues so strange if its not the same.

I will test my 240 again at some point, but what i remember from last time was that it did not shut down, it just frooze during heavy load.
My 240 can run my crossfire if its undervolted. But has frozen at times during heavy load during benchmarks or gaming. Anyone else seen this?

I think the different vendors of PSU has different protection. I guess some shut down to self protect and some has other protection, (current limiting) causing different indications.

But again, i might be wrong.

One of my 2 cents is that newer drivers made card use more power (more effective drivers) moving it past the limits.
Also when moving back and forth, u do it faster and faster, perhaps leaving battery in since you dont have to remove power button hinge cover every time since its already off for these fast tests and the sloppyness could have killed the card.

The other part of my 2 cent is that something did not have proper contact with heat sink during high load and died on the card. (memory?)
This component could be consuming high current and moving the PSU past the limits.

But at the end of the day, i agree with flingin.

 

BTW
I have Dell cards.
Plain screws going into the X from the top via top of the heatsink via the spring "wings" where screws go, no retention clips.

I bottom all my screws but very carefully and slowly going in sequence, but im quite sure they are all the way in bending the wings down to the board.
My first card that died had great heat leading pads, but i tried 0.5 mm thick pads on memory, it seemed they had good contact. Perhaps not? After that i have had thicker pads and no problems after I got my new card. Using 1mm and still have great cooling on the GPU. 62 deg during bench with 975 / 1440 could be a good indicator? Having such good cooling makes me not want to reduce to 0.5 again.
(leading me to think my 2'nd of the 2 cents were most likely)

 

did you try my vbios i sent you?

 

One of my friends was changing the mxm card in his laptop. I was watching him closely. He forgot to take the battery out before putting the card IN the mam slot. We both saw a small arc that jumped between the card mxm pins and laptop mobo when the card was approx 5 mm from the mxm connector. I looked at him. He looked at me. We knew what happened. He flipped the laptop to check for the battery which was of course in. He said "what the % $#@!".

The card got fried no POST at all. He was Lucky mobo was ok.

After I saw that I knew I have to remain calm during every upgrade, so far so good....

Sent from my SM-N9005 using Tapatalk

 

Yeah he did, no dice

... that's hectic... but you cant disassemble to install the card without taking out the battery first, how is that even possible???

 

Easy. Ive almost done it myself when moving back and forth a bit for testing quckly.

You leave your hingecover and keyboard loose and connected enter battery and cable, then do test.

Remove cable, remove hinge cover and keyboard, exit GPU. zzzzzzzz fried.

 

POST in BOOT should be when it detects all cards etc. and when computer discover that the board is fried normally but a full short is probably not moving BOOT that far either.


All depending on where error occure on the board and perhaps where the arch hits it can move further and shorter into the boot sequence.
Perhaps when moving the board back to get it out of the slot connector componets might hit the frame?
As a second card as here it seems it also can get started all the way.

 

When I'm test mode with the hingecover loose the bottom is always open so I'm only running on AC. I hope he gets it sorted though and it isn't a compatibility issue cos if one of my cards decide to die, finding an earlier revision CLEVO card is going to be a mission.

 

Bro that was not R2 it was Medion laptop. You could take the bottom cover out and leave battery in.
Sent from my SM-N9005 using Tapatalk

When I was reading that post and started to read "zzzzzz" i could almost feel electricity going trough my body LOL

 

well thats whats in my mind too..

i have both my cards litecoin mining 24/7 since i bought them and i wonder where i will find a replacementif one of them dies...

 

Ah, okay

Is there any progress on the HeatSpring group buy? I need to repaste my chip soon as I'm hitting 100C on 2 cores when gaming with 80 TDP.

 

My comments bolded below

Aye, have to be very careful not to forget the battery, make sure no static electricity occurs, not to short anything etc. etc.

Let me guess, good old 16F1?

Still need like 2 more people. I've ordered some Coolaboratory Liquid MetalPad (rated at around 36W/mk) to see how it behaves.

 

I'm wondering if you didn't damage the MXM slot. Leaving those pads intact puts a lot of pressure on not only the card, but also the MXM slot - in the upward direction. I almost made that mistake installing mine.

It probably would have taken 15 - 20 lbs of force to get those pads to depress enough to fully seat the screws holding the card in place. The MXM connector gets hot with ~100+ Watts flowing through it and then it's attached to a hot PCB (video card). Thermal stress combined with the insane pressure could have weakened the integrity some of the pins connecting to the motherboard on the MXM connector.

Have you tried other known working cards in the slot? I wouldn't be surprised if you have a dead card and/or a dead MXM slot.

One way to check and see if it was the pads applying pressure to the card that killed it would be to attempt to reflow the solder on the card. A warped card combined with ~100C benching temperatures can weaken the solder joints on the various BGA components on the card. I wouldn't be surprised if the card started to work again (at least for a little while) if you did a reflow.

If you don't have access to a reflow oven, any cooking oven will work. Suspend the card on a tray by it's corners so that none of the components under the card are making contact with anything. With the card in the oven bring the temperature up to ~415 degrees F, hold it there for ~8min and then slowly let it cool down back to room temp. If you can, put a 2-3 lb weight on the gpu die.

I've repaired countless cards this way showing similar symptoms as yours. Although I've never got them to last more than a few weeks before the process needs to be done again. If I had a proper reflow oven, then maybe it could be a permanent fix.

 

Close ! X8711 Barebone

Tried that already and i was not impressed at all. Had a hard time to get that thing melted/sticked between CPU DIE and heatsink.

Good luck anyways bro

 

TY Bro. Mind me asking, I'd heard it needs to be burned at at least 80C for a sustained period. How did it perform/what happened with it?
I plan to crank the OC up on that, using like 3 stacked layers.

 

Bro ask me anything

I had problems to reach desired temp to melt it. If i would put any load on cpu the temp would go straight away to 100+. Idle was about 60+. If you overheat the cpu too much you can damage the memory controller early. ..

Sent from my SM-N9005 using Tapatalk

 

Ah, it sounds like you didn't get proper contact. Even unmelted, the thing should have performed adequately enough for you to put moderate load on it at lower TDP.
Consider that I managed to use a fat bit of thermal pad for cooling with no problems.

You sure you didn't just burn in the metalpad into the cpu core without much contact to the heatsink itself?

 

My PSU thing over here is not trying to explain why the card doesnt work. It is also not trying to identify if his PSU is good enough.
This is more why I think it sometimes moves past POST, and almost never past loading of drivers and most times just halts during POST.
(like i had)

The card he has worked presumably, and now it dont. Removing the possibility of incompatible card, and the fact that it was bricked i guess.
This card needs to be replaced. (like mine)

But if the card is dead/fried and consumes more current, there is no telling how much more power it takes and how it behaves since we dont know what is dead.

When going through POST and the card has a problem, noone knows how much more power the self test takes, because noone knows what is killed if anything. If it was same behavior all the time it should either never move past POST or always i guess, not just some times. And for me it was during post it stopped most times.


I also never have had the light disappear on my 240 PSU, but still i know for 100% certain that my 240W PSU has given my computer issues when running since the computer freezes up, and more problems the higher GPU voltage i use. Almost always stopping when using stock voltage. These are problems i dont see now with my 330. Could be due to noicy voltages towards the top. I dont know.

 

The Flextronics 240 watt PSU's do emit a lot of noise, so I wouldn't be surprised if that was contributing to your issues.

I don't think its an issue of the card consuming more power and that's what's causing it to fail due to a lack of current from the PSU. If that was the case, the PSU would go into overprotection, and the card would most likely fry itself to the point of being 100% dead - no start/no video.

A bad bga connection can behave very inconsistently. This is something I've experienced a lot. In most cases all the balls are still making contact, but it's simply not seated nearly as well as it can/should be on the pad. This results in dirty signal transmission. Sometimes they're clean enough to not be a problem, and sometimes they're indistinguishable between a high and a low.

I've found this tends to get worse with increased gpu load (and higher clocks) probably because the increased load results in icreased current draw which reduces the core voltage for a brief period of time before the VRM's can properly compensate. This couple of mV drop is often the difference between a solid high or a solid low on a already degraded signal due to a poor BGA seat. And then when you combine that with higher core frequencies the signal has less time to stabilize before it needs to be interpreted.

And yes, a noisy PSU would only further hurt an already damaged card.

I had two faulty 6990m's that I was reflowing every other week to keep them stable. They would experience intermittent failures which would get more frequent and then result in an inability to boot into windows after post due to the cards failing when the graphics drivers were initialized - which results in load on the cards. How far they got into the windows boot sequence was always random - Sometimes they'd actually get to the desktop and then fail, but usually they'd fail to even get past the black screen at the start of the windows boot process.

The card is shot either-way, so it won't hurt attempting a reflow. The way in which is was physically strained before the failure could definitely damage the integrity of the bga solder joints. My bet is it will be fully revived for at least a little while so long as his MXM slot is undamaged.

 

hi i got a alienware m17x r2 my prosseor its Intel Core i7-820QM , 8gb ram and now i got 2x4870 grafic card, i am thinking to update grafics 7970 will work on my laptop. if its works in future i can make crossfire?? tenks

 

Bro. Read PM

Also @nitsun69

Flash the both cards with @Nospheratu vbios. I think slv7 vbios is making all this.

Report asap please
Sent from my SM-N9005 using Tapatalk

 

Thanks for your time and ideas flingin

 

If anybody is still looking for a CLEVO card just saw this listed on ebay for $450. They are new cards and it the seller has very good feedback. Free shipping in the US but they also ship internationally. You would have to contact them for shipping prices if you're outside the US.

 

Anyone tried the 7970 VBIOS 024 for Clevo yet in our X-fire setups? (or single "older" card also)

Is it with the stock 1000 -1400 or is it overdrive possible?

 

I will share some bad/good news with you regarding PSU mod.

The chinese auction for the R2 Jack

NEW DELL ALLIENWARE M17X R1 R2 DC JACK POWER PLUG IN PORT CONNECTOR SOCKET | eBay

First time i got R2 jacks from this seller i was thinking...Looks ok but a bit small for R2 jack....but ok it works.

I ordered 2 More jacks 2 weeks later because i was thinking that i could use them when modding PSU for other NBR members, and since delivery takes 3 weeks, then why not order them now...

Yesterday R2 Jacks arrived again, To my eyes, those are different !!!

Those are Bigger, Wider, have different mounting type, and feel much more...oryginal Yolo

Look at the photos bros, just to inform, i am using the Smaller one ( SIC! ), and now i regret a bit that i got those small ones in the first place. LOL







I am wondering now...how this small thing is gonna handle long term High power draw.....

 

I wouldn't use those flingin, the original R2 jack is the larger one with those funky shaped holes at the front and should have Foxconn branding on the metal shroud. If you ordered from the seller from your link you can claim that the item is not as described from the picture and get a refund.

The smaller version looks cheaply made and could have been made for any laptop that takes our PSU connector tip which could be dangerous.

@Trome71, I was meaning to try the 024 vBIOS but I'm thinking it might be a bad idea in case the card gets bricked and I have no programmer to recover. It is a hybrid Enduro/Non-Enduro vBIOS so its a bit risky.

 

Share the 024 vbios pls
@Flingin: That looks like a chinese knockoff to me, I have huge doubts that it will last while pulling hundreds of watts. May melt. You should open a case with paypal/ebay and ask for a refund. Submit the case with pics of the original connector side-by-side with the mini-chino.

 

From Nospheratu on page 183 or something in this thread:


m17x R2 Clevo 7970m BIOS Version 015.024.000.002.043595 (BR43595.003)

 

Hello guys,
There is a question!
Method of mod 330watt psu which offers Nospheratu identical as imsolidstate a method?
I know that the imsolidstate method gives over 330watt 33% (something near 440watt)