[Guide] Installing AMD Radeon Mobility 6970 / 6990 in your M17x R2 (Single GPU and CrossfireX)

Would connecting ID cable from the original 240 brick to the m18x 330 brick work? Like having the 240 connected to wall just to keep the end of that ID cable active and using it as a replacement ID on the end cable of 330 ....seems like a easy temporary fix till BIOS2 is done if it would happen to work that is.
i'm just wondering...

 

Cutting the ID cable will result in the powersupply not being recognized at all and no power will be transmitted to the laptop.

The idea of using the ID cable from the 240 watt PSU would be tricky, but it might work. . . .

 

i sketched something up just to make it more clear what i mean to people ...whats the worst that could happen...if a bios fix is released then cables can easily be connected back XD this could serve as a quick easy fix to people's thirst ..ill try it when my brick arrives.

 

Ok, so I'm a bit confused. How would disconnecting the ID cable from the m18x pdu (330w) and connecting the id cable from the m17x pdu (240w), be any different then simply turning off the notification in the bios? If the identification cable is truly the only difference? Or does turning the notification off in the bios simply remove the warning and not actually do anything else?

 

The ID cable is not only used by the bios exclusively. I tried disconnecting the ID cable on one of the 240 watt PSU's and windows would not recognize it.

Also, I think the two PSU's would have to share common ground.

 

That is correct, turning it off won't do more than what it says , the BIOS won't allow the PSU to power the system.
What I'm wondering is if the circuits can be isolated and replaced, take the one from the 240W PSU and solder it on the 330W PSU PCB replacing the original

Thanks for trying aarpcard, even though I was on holiday I was following the thread.

 

Nice work Aarpcard! Congratulations!

Strange to hear about the psu's not sharing the load 100%. Do you think this may be attributed to slight variances in voltages?

 

I don't know much but I would have thought that it would draw more power from the psu with the least resistance? Is the the psu that plugs into your adapter board usually the coolest, the extra resistance from the plug may be enough? just an idea.
My project is going to end up looking something like this (if any one is interested) only with a circuit breaker wired in

 

Thanks!

I'm not exactly sure why the PSU's aren't sharing the loads equally. My best guess is, one provides a slightly higher voltage than the other, and overpowers the other PSU until the second PSU's breaker flips (which I have observed happening) and then power draw switches to the second PSU once it is reset. I'm going to try including a diode, to see if that makes any difference.

 

Hmm, under what conditions have you seen the PSU trip?

 

Really intense loads. And only 1 will trip. I think it's because 1 of the PSU's has a slightly higher output voltage than the other and therefore dominates the power draw sent to the the computer, and then at the same time forces current back into the other PSU. . . . I do think diodes will fix this which I will try in about an hour.

 

Sweet, can't wait to hear your progress. Your definately the pioneer on this front. Who's efforts I'm hoping to benefit from. Especially since I'm going ot have a pair of 6990's soon I plan to undervolt them and run them at stock clocks for everyday use, but in benching, dual PSU's would be amazing. Especially if the power draw is shared.

 

Ok just put the diodes in, everything is working fine so far. Going to do some gaming and test the loads on the PSU (as best as I can)

 

Awesome...pretty exciting stuff you got going on

 

Nice aarpcard, if You have 3DMark11 installed, can you do a run of the combined test and see if either of the PSUs trip? Do it with the highest overclock you can manage, and if you can, a CPU OC as well.

 

Ah, well here is something interesting.

I only had access to 3amp diodes. So I put 4 of them in parallel to theoretically achieve a rating of 12amps. I know it's not a good idea to put diodes in parallel, but it was the only thing I could do at the moment until I order a 15 amp diode. Since all of mine are the same part# and specifications, the risks are far less likely. Here's the interesting thing though:

I plugged both supplies in and ran metro with full gpu and cpu overclock. Everything was fine, except only 1 PSU got hot. I touched the diodes, expecting them to be a bit warm, and they weren't. With Metro still running, I unplugged the hot PSU. The laptop stayed on AC power, so now obviously it was running on the second PSU.

I played a bit more of Metro on only the second PSU to see if it would get warm. 10 minutes later, I checked and it was still quite cool. I touched the diodes, and they were scalding hot. So I quit the game and unplugged the PSU to let them cool. I tested everything out, and the diodes still work, so nothing got fried.

Neither PSU's cut out during this testing period.

What this does is confirm, my suspicion that only 1 PSU is powering the laptop even with both plugged in. If both were powering the laptop, the diodes would not have been cold when I touched them the first time.

I do have two more diodes back home that I'll chain onto the bunch to make a theoretical rating of 18 amps. It is possible that much more than 12 amps were being drawn by the laptop when they got nearly lethally hot and that is what caused the overload.

That might fix only 1 PSU running at a time, but I don't think so. I'm going to experiment playing around with the data lines from both PSUs to see if disabling 1 or the other makes any changes. I'm also going to buy more diodes and put them on the second line as well, incase that is messing something up.

There's something special going on with these PSU's that's alluding me at the moment. Other powersupplies can be put in series no problem and deliver their combined amps without event. There is something different about these, but I'm confident it can be worked out.

 

Hey, why don't you keep only ONE data cable?
I bet the one that doesn't have it( the data cable) will still be used.
We basically need only one that tells the systitem "ok,green let it through".

Try that if you have time and see what you get.

Thanks

 

I actually tried that last night. I didn't test it extensively, but things seemed to work fine (fine in terms of no changes). I'm going to try it again, this time with a few more combinations once I get the diode issue straightened out. I'll be able to mess around with this again more shortly after 5pm.

 

Gonna go get some more parts. I think I know now what the problem is.

 

awesome...looking forward to your results

 

Yep my idea worked. The PSU's are now sharing equal parts of the load. However, the Diodes are still overheating like crazy - even under minimal load situations.

I'm going to go to Queens tomorrow after work to hopefully pick up some 15amp diodes.

 

Ok, so it is to do with a higher voltage on one adapter causing almost a short like situation to the other adapter. This would explain why the diodes are getting so hot. Beefing them up should help, but I wonder what rating would be effective?

Should we be also taking in to account a forward voltage drop when the Diodes are introduced. I think it is a drop of like 0.3 to 0.7 volts...

You could be making the situation worse by placing the diodes on one positive, as this may make the voltage gap even bigger. However if you place it on both positives, the overall voltage will be reduced...

 

I'm using schottky diodes (with very low forward voltage drops) on both positives. Forward voltage drops are minimal - about 0.3 volts so the total output voltage is about 19v. Neither PSU is putting out exactly 19.5v as stated on the case.

The short situation is what I experienced, but fixed by using diodes on both positives - however, it wasn't a direct short, most likely it was just low internal resistance. If it was a direct short, the laptop would not have stayed on AC power.

The new diodes I hope to get, (might not find them in Queens) will have a 16 amp rating and a forward voltage drop of 350mv - all other specs fit the application.

What is happening is just as I described in an earlier post. It's not a good idea running diodes in parallel. You'll more likely have success in doing so if they are identical diodes, but it's still not a good idea. However, since I didn't have access to the ideal diodes for this application, I had to use these.

What's happening is, one of the diodes (of the 4 in parallel) has a slightly lower forward voltage drop than the others. Therefore, more current goes through that diode than the others. As more current passes through, the diode gets hotter, further lowering it's internal resistance (silicon will become less resistive with heat) and as a result further lowering the forward voltage drop (Vf) which then forces more current through the diode, and a circular chain reaction occurs until something overheats and fries.

You can put very small resistors in series with the diodes to prevent this, but then you're even further reducing the final output voltage, so that's not an option here. Once I get these 16 amp diodes, to replace the parallel arrays, everything should be fine as I now have the PSU's equally sharing the load.

 

Awesome Aarpcard

I was going also going to suggest only using 1 larger diode on each positive but you are already all over it!!

Keep us posted!

 

Yeah directing power properly from two sources can be a nightmare.

 

Not really, it depends on your diodes, like aarpcard said.
@aarpcard I recommend you use an already made diode bridge (at least you're sure they are the same and have insignificant loss) rated @20A.

In the past i found that it's better than going thru the hassle of matching them myself

 

I was thinking about doing that, but I wanted to find a correctly rated schottky diode if I could as that would be the best solution - which I did, but not in a brick and mortar store. I just bought them off ebay.

70amps at 30 volts constant DC forward. 0.3volt VF at 70amps and constant 30 volts DC reverse blocking voltage.

 

GUYS my idea works HAHAHAHAHAAHAHAHAHAHA sry im happy im no genuis when it comes to eletric stuff but read back threads for info. ill take pics...so for those people who are willing to settle for 330 watts and fairly simple mod m18x adapter was about $130 i got from dell. best thing is if bios mod fixes this i wont need both bricks ill just simply swirl wires together back to place and warp with electric tape. so far it charges system and doesn't nerf clocks on GPUS

 

So are you using a 330W PSU? I may have jumped the gun in buying a second 240W. Although yours forces you to use both a 240W and a 330W. Interesting though. I may need to open a 240W just to get a better idea of whats going on inside.

Although in both cases, yours and aarpcard, your still using two PSU's.

 

Well, assuming all is working like it looks it is...it makes me wonder if simply modding the bios to allow the m18x adapter should work...

 

yes i am and best part is its so ing simple lolol here look this is the m18x adapter right?

 

idk if the ID cables are even well connected let me unpower the 240watt and see if the 330 keeps working with a cut ID cable

 

Thanks Stealth55...that would be REALLY interesting to see

 

WOW 330 adapter seems to be working by just simply cutting ID cable hold on let me restart laptop to make sure

 

Nearly positive that won't work. I've tried doing that on a few PSU's I've tried. EDIT: If it does work, make sure the cpu multipliers are correct.

Did you ever try using the 330 watt PSU with it's own data cable?

 

would be huge if it works without the ID cable, good job

 

Is there still a concern, even if the 330W adapter does work in the above scenario that it could pull too much power? or potentially damage the components? Sorry electrical work isn't my specialty, so don't know the ramifications of doing something like this.

 

YEA guys it works with just cutting the ID cable i cut mine all way infront right before adapter plug thing tht connects to laptop.....bummer tho the end of my adapter looks ugly now Q.Q....gotta patch that up with tape now x.x well thats 1 way to solve 330 watt PSU problem... just be careful cutting tht black rubber guys...now i wait for my 6990 they on a plane on way here atm...OHH BTW the ID cable on 330watt psu is GREEN

 

Do you have a watt reader?

Would be nice to see some power draws on the 330 watt psu

Also, is the psu check in BIOS enabled or disabled?

 

interesting...I could have sworn that someone had tried this with the 330W adapter already. Now the real question..and I'm sure we tried this already. But has this been tested as is? I've seen people say it doesn't work, but have we confirmed it? I'm certain we have...I'm just amazed to be honest if it's really the darn id cable that's preventing it from working.

On the plus side, this would be a really easy mod.

Although I still give kudos to aarpcard for all his work

Will be curious how this all turns out

*EDIT* Hey Douse, have you got shipping confirmation on your 6990's yet? My replacements are supposedly shipping tonight (Friday AM China time)

 

i cut cable all way infront maybe they didnt....but yes aarpcard's PSU is a awesome one to have for extreme overclock

 

Also which bios are you using? Have you confirmed that the cpu multiplier is correct? Also try booting the computer without the battery plugged in.

If this is really true, then there is something really strange going on here - I wonder if it can be replicated. Cutting the data cable on the 240watt PSU results in it not being recognized even with the bios check turned off - although when I get home I'm going to confirm that again - maybe we will be surprised.

Very strange indeed.

 

my multiplier is at 24 with 5% oc on bios i have the V10 mod bios ill boot it when i come back i gtg somewhere real quick

 

Greywolf, mine are held up because my international transfer is sitting in limbo somewhere.....really, really frustrating...

I am working with pkhetan so hopefully it is resolved soon..

 

I wonder if that might be the difference. The 240 watt adapter was not recognized by the A10 (unmodded) bios when I cut the cable. I did just install the A10 modded bios a few hours ago. I wonder if it will work with that. It'd be amazing if it did.

Imagine if the Wiz unknowingly managed to tweak the PSU support in the first release of the modded bios! That would be ridiculous! **** lol I can't wait to get home from work in 45 minutes and try it.

 

That would be pretty amazing! Anyone know where I can source a m18x 330watt psu? eBay has some up for heaps of $$...must be cheaper somewhere else...

 

You might be able to get one from Dell

 

Douse, thanks for asking the question, I had the same one.

I don't have the modded bios installed, but will install it if it allows for the 330W PSU with no modifications (or simply the id power cable removed).

Mmmm, maybe I should have gone with the 6990's afterall... lol

Oh well, I'll be happy with the 6970's

 

the 330 watt wont work without the cable cut i tried it before i cut lol

 

lol...good to know, although if it works fine WITH the cable cut, that's just as good for me