Modified overvolting vbios of GTX460m on G73SW.

Both are pretty good but the v1.00 seems to give better stability over v0.962 at high clocks.

 

With 1 V bios I maxed at 848/1696/1600 seems g53sw has binned gpus or me and kingpinzero are just unlucky lol

 

That's not too bad already!

 

Yeah I'm pretty happy of my oc, 2fps more wouldnt turn my gpu into a gtx 580.

 

Just an FYI folks. I don't know if these are binned or not..

I'm achieving 840/1680/1740 on stock voltage. .96 Volts Shuts computer down.

it isn't a gpu Temperature. and isn't current related.

.96V and .1V shuts off at approximatly the same amount of time. At the same overclock freq i have on stock voltage 840mhz stock voltage = fine 840mhz .96V = crash in 20 min. 840mhz @ 1V = crash in 20 min.

I am convinced its a VRM thermal protection on a piece that isn't properly sinked. Which one who knows.. I don't have an thermal scanner. but one person with one would show the culprit to over volting.

at stock voltage .91V i don't shut off at all no matter how long or how hard push it. as long as im at a stable clock.

btw with 1V bios i was close to 900mhz on core, but tried figuring out why it kept shutting off on me and decided to quit trying. I'd rather game at a safer voltage then have a vrm Meltdown. for those who don't know what a vrm meltdown is or havent seen one... google GTX590 puff puff pass. or watch this youtube video Geforce GTX 590 burns @ SweClockers.com - YouTube you'll notice the computer shuts off just as the laptop does. and poof.. there goes your 700 dollar card.

another thought, this card blew at +350mv i believe. while we are trying to push nearly 900mv example .91V-1.0V

 

The gtx590 issue isn't related at all with our overvolted cards.
In that review, the gtx590 blew up because of the drivers and the vbios.
Fermi have by default an overvolt protection limit enabled by driver in conjunction with vbios call.
What happened there basically was that the voltage limit was voided past the safe limit while trying to push it.
Occt gpu stress tests disable that feature thus throwing the card in the open without a way to protect it, that's why it blew up. It was a noobish mistake.

I know many and many gtx590 users with a good overclock and a voltage increase of 0,500mva, while staying stable, theyre not burning the cards.

Also our grx460 still relies on the old fermi architecture which seems to not have such problems as the voltage limit is built in in the vbios, rather than called by driver itself.

Have you ever wondered why even with high clocks your card doesn't crash in furmark after hours, but it does while gaming?

That's the same reason. After reaching 90c the vbios limits power output and capacity; in furmark this happens as soon as you start the stress tool, creating a "fake" stable stress session.
When gaming, however, the power limit would kick in only while reaching those temps, thus "really" testing if the oc is stable or not, since the voltage protection had to be triggered yet.

Some retail gtx460m (clevo iirc) have 0,962 or 0,95 by default. Our cards althought "gimped" are basically the very same cards therefore the overvolt is still safe.
As Saltius already explained, the 0,962 value is in the vbios as well, just not used, I guess by Asus request and decision.
Gtx560m uses 0,962 by default. This + the newer core with less power consumption and less heat + higher clocks makes break 900mhz barrier easily.

 

Hi Guys, just my own input as feedback !
I'm using lastest Nvidia 285.62 drivers.
I succefully overvolted to 0.962V my G53SW thanks to this thread.

I can reach 830/1510/1650(shader is linked anyways i don't have the exact value).

Thanks you all for this great mod!
I am just disapointed that my memory clock can't go beyond 1510 after it make glich and artifacts!!!! What a shame on this kind of laptop!
Have a nice day

2261 points at 3D mark 2011 ( with the G53 I5 version ! ) with TS

 

Hello Saltius, this is the Original Vbios of my G53JW-SX276V for a Overvolt
Vbios : G53JW-SX276V Vbios

The actually modified Vbios for G53JW don't work with me :

picture : http://img89.imageshack.us/img89/7029/imag0004xw.jpg

(sorry for my bad english :/)

 

All you need to do is change the sub system ID to 2043 in Nibitor and save it as a new vbios.

 

Tanks you ! Its work fine

http://img14.imageshack.us/img14/5940/imag0006bw.jpg

 

For some reason, if I edit my clocks in my VBIOS with Fermi Bios Editor, nvflash refuses to recognize the new VBIOS and says the file can't be found. Does anyone know how to get around this?

 

?? while im at .96v it will shut off the computer on both gaming or fur ??

During gaming im peaking ~78c although i had it over 90 before i repasted in fur.

MSi for another example uses .95v

So if our cards are essentially the same. why havent we tried tearing down the msi bios or clevo and applying it to our boards... or is there a reference design which is flawed in which we are using in the asus.

Who here has a clock above 840 mhz and a voltage of .96v and higher that doesn't have system shutdown after 20min into crysis 2 or hell even battlefield 3 now.

Even though battlefield 3 crashing doesnt nessisarily shut the laptop down its crashes to desktop before it gets a chance to shut down... crash happens prior to 20 min.

obviously there is something flawed... about the asus board if we are powering down at a reference board design's voltage thats in the clevo.

So whats our options. Continue overvolting thinking that its safe, yet only good for a benchmark or two.. " at that point whats the point of overclocking it then" or two, we can start disecting the bios and find out why its shutting down.. and prevent it. assuming that it isn't thermal protection on a vrm.

But know this. i clock at 840mhz on stock voltage. if i raise the voltage to .96v it shuts down.. even if i lower the clock to 800. gpu temps are way below 90c

 

OK So i bought 16gb of ram off newegg for 70 bux, during install I wanted to recheck the gpu board... It does seem like at the top of the board there are two large sized regulators that arnt sinked which are on the other side of the board??? why leave those unsinked yet sink the rest. I took the machine apart after gaming for awhile and they were still hot to touch 15-20min later. Doesn't seem like a good idea to me to leave them bare, unless they're trying to get you to purchase another in a couple years or so. I double checked the compound used on the ram as well.. Messy stuff. Like dried up buggers that somehow wipe away cleanly. LOL..... also staring at that warrenty sticker on the cpu... I said hell with it and took it off and repasted... if they have a problem with warrenty. I'll send a copy of my certification along with the laptop. I found TWO MAJOR problems with the CPU paste..

1. The sloppiest paste job i ever seen. the amount they used was enough to cover 3 1/2 cpus. they put a plastic square cutout on the cpu to keep the paste above the cpu. but they put so much on it, it found its way under the guard and practically out to where its pouring over the sides.

Not sure about how conductive thier compound is if any and if it has any capacitance value, but it completly covered the surface mount resistors on the top of the cpu. While they did cover the entire Die, it also created a thermal barrier around the rest.

2nd. The square plastic guard they applied overlapped the die and was sitting in between the sink and die. I'm thinking they didn't want you to remove that sticker so you couldn't see what a shotty job they do at building thier machines.

And finally my cpu temps were lowered 15C without any curing time on AS5.... I decided to pull that sticker since i was monitoring Battlefield 3 troubleshooting... Looking at the logs i was baffled and noticed my cpu temps were 10-20c higher then my gpu temps.. I thought NO WAY.... The cpu was peaking out 92C on small fft on prime95... way too hot. but suprisingly stable. I bet this chip with the right bios would clock insane. to bad asus limits these.

Back on to the GPU, the two regulators are really close to the heat sink, i wonder why they didn't go the extra and cover them as well.?????!?!?!?!?
I may take the back off here soon and use a thermal probe and see what they are reaching.

 

Alright running some benches. Found that those two regulators are too hot to hold your finger on them. They should be sinked for sure. but they are not... Why????? why has asus made their proprietary MXM card inferior to the reference design??

BTW those who are looking for a good cooler for their G73 i'm using the CoolMaster Strikeforce sf-19.

I have the gpu repasted. and clocked at 840/1670mhz on .91v and 1740 on the memory. I've run furmark for 10:00 tests using the cooling pad, and tried it with the bottem cover off as well.

Xtreme burn at 1920x1080 Max temps recorded at full on the pad were 78c and cover off was 75c Medium settings (to the point where LED's dont fluctuate "kinda quiet") i peaked 81c and 79c with the back cover off.

VS. GPU max temp of 87C sitting on the table by itself... Again 10"00 test.

The CPU on the other hand was running crazy hot from factory but explained that in the last post.

I tried touching the two regulators on the back of the card while laptop was sitting idle with an electronic thermometer and it shot up to 110F and started blinking Hi. so i need to get something a little more accurate. but touching it while it was running furmark, felt 3-4seconds and can't hold it on there any longer... i'd say 130f range maybe or a little higher. thats stock voltage.. its only a test away with a proper way of measuring temperature when i go and set .96V

I'm also going to lookup that part from a supplier and see what the threshold is on that particular regulator.. Thermal overload and runaway in a semiconductor starts at 160F if i remember correctly on a standard transistor. and lowers the internal resistance, and pop. instantanous bright light and a little puff of smoke.... Done it on an XFX 680i board pushing the vreg on the CPU to 1.6 something or other. q6600 vapochill days. they will pop. even sinked when pushed.

But where i'm standing right now and believe.. is this board is going into some thermal protection.. not the GPU. GPU is well under 80c during gaming, and its not current protection.. as i can draw 130W from the wall "Monitored and logged via Kill-a-watt" and the computer will shutdown at a sequence in crysis 2 where it was only pulling 110~111W depending on where you are in which particular levels etc, facing what direction grabbing x amount of framerate. it all plays a role. but i can expect that it will shut off at this given amount of time playing. Even sometimes if you try to restart the computer immediatly after it shuts off. it reboots itself over and over, until you give it a few seconds off to come down off what ever trip its gone on to restart normally.

 

btw cooling those two vreg with just fan cooling no sink, i was able to go from 840mhz on gpu to 855 on stock .91 voltage... just an fyi for everyone

still waiting on some hardware to get temps and a way to cool them.

 

Thank for your efford Jbmeth007, could you tell me more about the vreg on the backside of the card? Like where it's located, could the copper thermal blocks be the possible solutions to it? I've uploaded 2 screenshots of mine, would you mind circling it and reupload the picture, just want to make sure where those 2 regulators are at. Thanks.

 

Maybe it would be better off explaining what a vrm is.. composed of capacitor, which we should all know what they look like.. an Inductor, and a mosfet.

I don't know how many phases are on this particular card without looking at schematics so i generally explain them as one or 12 depending on how many i find on the card. VRM's are grouped in what they power.. one for cpu one for gpu etc etc... while the vrm can be explained as the total power supply on the card. litterally meaning true. I'm explaining the exposed section which you see in the upper left hand side of picture 1.

composed of mosfets and 2 larger inductors. The inductors is what is getting too hot to touch. i haven't stuck my fingers down into the mosfet pit of the vrm.... I was waiting for a tool so i can measure.

Removing the bottem cover they are the first things you see exposed of the GPU, and i've got the coolmaster sf-19 blowing on the bottem of the notebook. I've noticed i can clock 15mhz higher on it while the gpu temperature stay around the same idle.

 

also I believe this is the voltage control circuit, I wonder if its anything like a 7900gt.. if so a little conductive pen mod or something similair would be available for this card giving access to voltage control. or tweak.

 

I see, I think I'll get a few thermal blocks and paste on both of them to see if it helps.

 

I was thinking of doing the same thing, let me know if it works for you. I was planning a way to actively cool them somehow tieing it to the thermal plate of the gpu.

i've notice a stability increase with just a breeze across them. The fans in the Coolmaster sf-19 dont really push a ton of air, but they are loud of course.

 

I have a G73SW-XR1. I am only able to achieve 795/1590 1600 using nibitor. jbmeth - after following your posts I decided not to increase voltage seeing as you had possible thermal shutdowns. Let me know if you find an effective way to cool the gtx460m.

 

Could you modify my vbios please ?
Thanks man

 

Have you repasted? IC Diamond will do wonders for your temperatures.

 

Can anyone point out what I need to change in the GTX560M vBIOS to get the 1V?

 

Noticed nobody has given any updates on thermal solutions for the exposed parts of the graphics card. I took a closer look and was thinking I might just try finding some small heat spreaders to throw on there. Saw something like this used to cool a 6970 in modded JH models and figured it was worth a shot and a lot easier than a heatsink mod. Just have to locate spreaders that are the right size . . .

 

Did the mod on my G73SW 3DE. Ended up at 848/1696/1625 using MSI Afterburner 2.20 beta. Might be able to go faster on the memory since it's the max the software allows, but memory OCing hasn't seem to impacted performance much. I think it's the core/shaders holding things back at this point. I tried both 0,9625 and 1.0 V, and upping to 1 didn't impact the max stable OC any. At this point I'm not sure what's more stable of the two voltages. Leaving it at 1V for now, to see how the long term effects are.

 

I doubt it even changes anything besides the voltage label.

 

The VBIOS? We've gotten some pretty consistently documented increases in overclocking potential running the higher voltage . . .

 

He means the difference between 0,9625 and 1 V, given that the mod for the G73SW doesn't actually change the profile setting, but instead modifies other values. The gain from 0,9125 to 0,9625 is quite tangible however.

 

Yeah, I meant 0.9625 -> 1.0v.
With 0.9125 -> 0.9625, I've increased my stable overclock from 820 MHz to 880 MHz, so surely that works. However, the my laptop shuts down after 1min on 880 and after 10min on 820, so 0.9625 is not really usable for me.

Is there anyone here who doesn't get shutdowns with the 0.9625 vbios? What's your model (jw/sw) and VBIOS version?

 

My G73SW 3DE is stable and doesn't shutdown at 848 - I used a backup of my own BIOS and modified that, as there were too many differences to the ones posted here using a binary compare. I was worried it would screw up the 3D functionality. VBIOS version is 70.06.25.00.0C, using the latest Nvidia beta driver 295.51. Prior to the mod, the max I could OC was 816.

 

Same here, but I have gotten shutdowns even at 790 MHz and standard voltage. I'm 99% sure it's a hardware fault by now.

 

Just threw some copper RAM heatsinks on the two exposed inductors, could feel them giving off a good deal of heat while cooling down after shutting off the laptop, so I can only imagine how hot they might get during benmarks and gaming. Going to test as soon as I have time to see if I get a more stable overclock.

 

Saltius.. Tank you for posting the original g73sw vbios rom.. thank you !

 

There's nothing that could be changed in this vbios. It's already at the max possible value. There are only two entries in the voltage table.

 

As a matter of fact, 1V is really 1V ? I am still not sure because same issue previously has been discussed and it was not quite sure.
Ok, all bench is showing 1V but temperature mostly remains same and it is impossible. from 0.912 to 0.962 which I also use last 5 approx. temp goes up 5 deggress C almost so means ok.
Nevertheless, 855 core with 1625 memory " border on MSI Afterburner for mem " system stable.
BUT, in the main forum page someone get 1000 core ??? even with stock 0.912 ??? present system seems better than PS3 for sure already connect to 40"LCD and all is ok,
so if you really think that you have 1V so temperature should be really increased if not it may be seen only as 1V, as you know previously just for fun GTX 460M print has been replaced via bios as 560M or I remember such post here I didnt tried out but seems also possible.
If you have any 1V bios pls let me try so post it, I have already 2 different 1V but no temp increasing so I m using 0.962.
I have G53SW-IX023V, but seems 73sw models bios also possible to use.
hope someone reply with mod new bios or can make little explanatin about 1000 core ???

 

Sorry to wake up this thread, but is it possible to vBIOS the 460M in the Qosmio X505 line?

I ask because I'm curious to see what I can potentially get my GPU up to.

 

Might be possible, check out this thread on MVKtech.net. Saltius posted the MSI 460M 1v vbios there:

Need help tweaking the voltage of a GTX460M on G73JW

 

Anybody have a stock G73SW vBIOS?

And do driver installations change the vBIOS?

Need to send in my laptop to Asus

 

Did you just brick your card? And no, driver update does not touch the VBIOS. You can still blind flash if your graphic card is bricked.

There you go, original G73SW VBIOS.
View attachment G73SW.zip

 

Thanks! No my card is fine haha, but I wanna flash the vBIOS back to stock before I send it into Asus for other repairs and I dont want them blaming me for modding it The overvolt didnt actually help my OC'ing anyway.

 

You're sure? This is a VBIOS of an engineering sample card.

 

I remember getting this off someone that owns a G73SW on the forum, it's been quite a while now, that thread was about will swapping the vbios make your oc more stable, something like that.

 

I flashed the G73sw 0.96 vbios and everthing is ok. 3dmardk11 shows 2452 points running at 850/ 1700 - 1550 mhz, but when i try to play SWTOR the game hangs a lot, even with lower oc values that worked with stock vbios, can this be a driver problem? Im using the new beta drivers 301.24.

thanks

 

I tried using the version 296 driver without success. I don´ understand, with the stock VBIOS I can overclock to 823 Mhz but with the VBIOS running at 0.96v I cant use even this clock and the game keeps slowing the fps as if loading textures. I tried to run other games and can reach 865 MHz in BF3 without problems. I concluded that SWTOR somehow does not accept the overclocked voltage. If someone knows why this happen, tell me, im curious to understand... lol

 

since this place seems like the spot to ask, anyone mind posting an untouched copy of the g53sw 460m vbios?

either i fail, or there's only modded copies here, at mkv, and anything google finds/ originals for the 73/74/sx/j*. (never modded mine, need another clean copy for comparison to rule something out.).

thanks in advance.

 

GF106.bin

Is this what u are are looking for?

 

org.rom

can anyone change my gpu voltage to max for gtx460m? (i dont know how much exactly is that)

 

@shinobi. looks like it. thank you.

 

hi,

if someone read this, can he post stock bios from g73jw or show how to modify vbios to stock bios. please

thanks in advance