Printable View
hi there,
i wanted to volt mod my sparkle 8800gt but while comparing it to the g92 standard pcb design i doubt whether it's possibile to mod it at all.
Can some1 with a sparkle 8800GT show me the measuring points of vGpu and where to put the resistor??
thanks
How about posting some close up pictures so we know what we dealing with! :)
Hi,
yea sorry I forgot to take some :D
Here are 2 pics of my Sparkle 8800GT with (slightly modified) Artic Cooling Accelero S1- had to make it fit into my case:
http://img169.imageshack.us/img169/7...9579rb8.th.jpg
http://img512.imageshack.us/img512/6...4360jw8.th.jpg
The original Cooler was this one:
http://www.rueducommerce.fr/composan...GT512D3-HP.jpg
Background is I want to overclock the card as much as possible,
with BIOS vmod i only could change the vGPU Load to 1.1V which allowed me to set the GPU to 680Mhz.
1Mhz more and i have artifacts.
With the accelero (and 120mm fan blowing) the temperature doesnt exceed 50°C- so i have much space for higher voltages
First it would be nice, if someone could show me the + measure Point of vGPU Voltage. That i could check the actual vGPU (must be 1.1V) and see if the BIOS accepts higher voltages than 1.1V (i dont think so, because with 1.17V set to BIOS i didnt achieved better OC results.. so it must still stay to 1.1V)
BIG THX and sorry for my bad english :)
vGPU can be found from these red points - possibly also vMEM (which would be the one farthest from PCIe connector) but I can't be sure since the HS completely blocks the frontside pic.
Attachment 73661
The highlighted chip, what's the label on it?
Hi,
the chip's label should be
L6713A
990TS VH
MYS 99 729
(but u can read it now on the pics below)
Here are some more detailed pics of the front with heatsink removed:
http://img396.imageshack.us/img396/8006/frontdj0.th.jpg
http://img201.imageshack.us/img201/2...1117zz8.th.jpghttp://img372.imageshack.us/img372/8...7948oz0.th.jpghttp://img387.imageshack.us/img387/7...8157pg7.th.jpg
And here of the back, especially from the areas that may have sth to do with voltage
http://img254.imageshack.us/img254/6263/back1mi8.th.jpg
http://img142.imageshack.us/img142/3900/back2ma6.th.jpg
http://img142.imageshack.us/img142/5910/back3ry2.th.jpg
Sorry for the big filesize, but i didnt change the resolution because I think the more information u get the better u can help me
I tried to measure the vGPU with the red dots u pointed to- well i got the result of 1.8V (!) -
Did i do sth wrong´?! i connected the black cable of my multimeter to the case (which should be ground?!) and then put the red on the dot on the picture..
(I get this 1.8 Value on every big red Pin u showed me..)(
Something funny going on with the measurements. They can't all measure the same... (unless your vGPU is @ 1.8V that is)
After seeing the frontside I can confirm that the legs of the large black block-like part (inductor) that reads "1R5M" is vMEM, it's the one farthest from the PCIe connector. Legs of "R33M" is vGPU.
1.8V sond like the volt mod is done! He He! :rofl:Quote:
Originally Posted by largon
Hi, well it was my mistake :D
I measured the voltage of the part (looks like capacitor) below 1R5M and there was ~1,8V.
R33M is 1,1V in IDLE and 1,07V in LOAD
1R5M is the same- either vMEM is the same or 1R5M is also responsible for vGPU
(But vMem isnt interesting for me anyway- i can clock the Mem to 1030 without errors, that's not the problem..RAM doesnt improve performance that hard)
Well good to measure the vGPU.
I set the vGPU to 1,17 with nibitor- so it's correct that u can't set a higher value than 1,1V with BIOS.
And that my card comes with ~1,05V by standard
But the 680Mhz with 1,1V arent that good- so wouldn't it be nice to take the next step:
increasing the vGPU by resistor
It would be nice if someone helped me
Attachment 73870
- measure resistance between either of these points marked in red AND GROUND
- multiply the measurement result by 20
- use a VR rated at about the final result (measured_resistance × 20)
will try that later today , also i need to find the vgpu read-out point .
if some1 give more in put on the card's lay-out :D
@largon
Do I have to measure the resistante between those two points or of each point to ground?
If it is between each point and ground and the result differs from each other, which variable resistor do i have to take then?
And where to put the resistor between? (for increasing vGPU)
Thank you
Whoops, I forgot to type "AND GROUND" in the end of the line
It doesn't matter which point - they're optional mod soldering pointsQuote:
- measure resistance between either of these points marked in red
largon are u sure that the mod will work ?
use a 100k vr from the lower red point to ground and u got ure self a gpu mod :D tested and its good ;)
Hi
the result was around 1840-1850 Ohms from both of these Points to Ground.
So i have to find a ~37kOhm VR and solder it between one of these red points and ground- Is this correct?
Lowering the resistance will have the effect of rising vGPU?
Hmm is this safe? i don't wanna destroy my new card :D
Maybe some other people can confirm this mod?
Another big problem is the soldering act- These Points are damn small and i need a very fine SMD soldering iron i think :/
Hmm ok.. 100k seems more safe?!Quote:
Originally Posted by blind_ripper
How did u solder it? i Mean what for a soldering iron do u use
Heinzkopf,
Based on your measurement a VR rated 50k would be the optimal choice. 100k works too but it will be a little rougher to tune compared 50k. Solder it from either of the red points to ground and you're done.
the 100k set to 67k ohms gave me a start voltage off 1.088vgpu idle so with 50j omhs that will be 1.15vgpu ;).
greetz blind
okay,
thank you both.
currently i have 1.1Vgpu load with bios, i will use 50k ohms from lower red dot to ground to start with 1.15vgpu load then :)
But first i have to look after a thin soldering iron
now can someone find the vmem mod on this little pup ?
Yes that was cool, too.
Beside this another question- how much voltage would u declare safe for 24/7 usage? i mean i have a good heatsink, but i'm not sure..
Whats on this chipQuote:
Originally Posted by blind_ripper
http://img142.imageshack.us/img142/3900/back2ma6.th.jpg
@Rol-Co
Which chip do u mean exactly?
@largon and blind_ripper
it works perfectly with 50kOhm resistance for me.
I start off with ~1.135v IDLE and 1.105v LOAD
btw is it allowed to change the VR while computer is running or do u have to switch off every time u change the resistor?
Always tune the VR while computer is powered on and continuosly measure the voltage you're tuning.
Yeah, where to do the vmem mod? It seems that there is Hynix memory chips on these cards, atleast it says so on the chips :)
I got the same card. Thought it was nvidia reference so i got myself a 500 ohm VR before i knew it, and now i found this thread.
Thou my card doesn't have the exakt same Phase controller. Mine is named L6713A too, but
990B2 VH
MYS 99 751
instead of
990TS VH
MYS 99 729
as the one in topic has.
My sparkle 8800GT 512MB is revision S1.2
My measurements on those red dots are 1634 Ohms on both of them.
What kind of VR do i need? 1634 * 20 = 32680. So, will a 50k VR be fine for me too?
vMEM would be pin#4 on APW7067N to ground through a VR rated 20x higher than the resistance between pin#4 and ground.
micheskillz,
50kΩ works for you too.
Great :) Big thanks to you largon!
Edit:
vMEM: 2.05v !? wtf :)
vGPU: 1.10v idle / 1.07 load
EDIT2: vMEM mod
http://geting.se/image.php/96036-vMEMsparkle.jpg
Nice to hear from another owner of a sparkle 8800gt ;)
As i see u also have a vGPU drop while running under LOAD. I still ask myself why our Sparkle Cards have this issue.
btw what are u going to set ur max. vGPU at? Here i have 1.18v idle and 1.15V load right now
Quote:
Originally Posted by Heinzkopf
Hehe, indeed :)
Same revision too ;)
Thou i got a different "stock" HSF on my newest sparkle-card.. (this is my second sparkle 8800GT as the other one crapped out and only ran in 8x PCI-E mode)
Check this out
http://geting.se/image.php/96061-DSC05750.JPG
There is heatpipes going throught it and ALOT of fins under the cover.
Looks like an Accelero S1, doesn't it? Great temps with it, but i changed it to
a HR-03 GT anyways, because of the vGPU-mod i'm going to do asap. And first of all because of the noise that fan made even on 25%.
Shouldn't it always be some kind of drop between idle and load? Like CPU's, it's a built in "function" and that it should be that way? Maybe not as much as 0.03v, but i don't see it as a problem really :) As i don't want the voltage going up when i don't want it too :)
Btw, Is there any special about the hynix memory that these cards use? Usualy there is Qimonda and Samsung, but what about Hynix?
I'm going to use between 1.25 to 1.30v for vGPU and 2.15v for vMEM (for 24/7) if that'll be enough for me to reach my first goal of an OC up to 800Mhz core and 2000 shader 2200+ memory.
Also gonna try up to 1.4v when doing benchmarks later on, if that will help the OC even more :)
What's youre max OC on stock-voltage btw? Got any pics of your mod?
My max OC was 748 / 1728 / 2206 on my first sparkle-card, this new card craps out at wimpy 100mhz less on everything :( So that's another reason why i'm doing a vGPU/vMEM mod, because that previous OC gave me alot.
EDIT: My mod so far, for vGPU and vMEM reading, waiting for my VR's.
http://geting.se/image.php/96068-DSC05882.JPG
http://geting.se/image.php/96064-DSC05883.JPG
http://geting.se/image.php/96066-DSC05885.JPG
http://geting.se/image.php/96067-DSC05888.JPG
Easy access for vGPU/vMEM-reading :)
http://geting.se/image.php/96069-DSC05900R.JPG
Hi,
yes i also don't care about that load drop because i also had this on my p35 board before mod.
before doing this mod i just could reach about 650mhz clock with default voltage (1.05v--- with 1,1 I could set the clock to 680)
that was the reason for asking for a voltmod here.
ram i always can set to 1030 without voltage increase.
I'm still not sure whether I should do a RAM voltage mod too.. Because I dont wanna create more heat on my card :p it's still hot enough. And i think a higher ramclock doesnt have that much effect on performance..
Yea i could take some pics but there is nothing special to see. i just soldered the VR between those points and put it on my card with some glue :p
Ok!Quote:
Originally Posted by Heinzkopf
I don't think memory will increase performance that much either. But fun to have both.
Between those points!? You mean that you soldered it onto both of those spots?
No i just soldered one cable from the upper spot to VR and then to ground ;)Quote:
Originally Posted by micheskillz
U're right, i think ill do the ram mod today in the evening and see what freq i can reach then.
micheskillz, what VR did u use for vMem mod?
http://geting.se/image.php/96036-vMEMsparkle.jpg
I measure between this point and ground 650 Ohms and so i used a 13K Ohms VR.
But unfortunately i can't change my vMem.. Even if i put the Vr to nearly 0 Ohms it doesnt has an effect of the vMem Voltage.
The 2 Measurement Points for vMem always display 1,86V and 1,89V/1,91V (under Load).
The 2 Points I used are these:
http://img80.imageshack.us/img80/9157/vmemhh2.jpg
What am i doing wrong? wrong VR value? bad measurement points?
thanks
You are meassurig correctly on the 1R5M thingy. So yeah your only meassuring vMEM.Quote:
Originally Posted by Heinzkopf
Must be the wrong points for vMEM-mod if so, thou largon said Pin #4, and usualy the #1 is the one besides the dot on the chip. I could be wrong :S I haven't done the vMEM mod yet, as i've said, waiting for my VR's.
I got 618 Ohms between pin #4 and ground, compared to your 650. So i'd guess that it should be the correct pin!? As i have 2.05v as standard vMEM and you only have 1.89.
EDIT: It should be teh correct dot on my pic.
EDIT2: i was tired when i wrote this, so i edited it now.
Just to clarify:
The red dot in pic below is pin#4 so that's where the VR is supposed to be soldered.
http://geting.se/image.php/96036-vMEMsparkle.jpg
Hmm yes that was the pic that micheskillz posted first.
I chose that marked pin and soldered the VR between it and ground.
But something is still wrong- measuring points of vMem and VR value seem to be right- where is the mistake then?
edit:
i'm sorry it's working now :p
i found the mistake- a bad soldering spot at screwhole (for ground).
thank you both again.
Now im trying to find a good 24/7 value for vMem.
Right now im at 2V (risen from 1.9V) and can set 1080mhz vmem
But i'm not sure what i can set for a max vdimm for safe usage
Quote:
Originally Posted by Heinzkopf
Nice :)
I think you could use 2.05v without any problems on vMEM. As i got 2.05v stock @ vMEM without any vMEM-mod.
Thou i wonder what's a safe max vGPU-voltage. Running at 1.25v atm. No heatproblems at all, stays under 50C load.
Here's how it looks like after i did the vGPU-mod. If someone wants to see.
http://geting.se/image.php/96659-DSC03346.JPG
http://geting.se/image.php/96661-DSC03349.JPG
Didn't solder the GND, used a nut + screw into one of the stock mounting holes for the HSF.
Hi,
yes I think 1.25V gpu idle (here ~1.21V load) is a good setting for our cards.
I can reach 760mhz GPU with that.
Will also go to 2.05V vMem, because when this is default at ur card should be okay for me too.
Here a pic of my card, but as i said nothing special. Didn't wire the measurement points cause i can easily access them with my multimeter.
Used the VR i had to stock here.
http://img135.imageshack.us/img135/1385/gpuqs4.th.jpg
I get 1.238V at load, when using 1.250v. Gonna try ~1.3v in an hour or so to see if i can reach 800 / 2000.
Yeah, you should be fine with 2.05v vMem. I think if you'd have my BIOS you'd get 2.05v without doing any hardmod.
Always fun to see pics :) How hot is your heatsink to the right(on the mosfets) on the pic?
"Ouch i can't hardly touch it hot" or more like bodytemperature? hehe
EDIT: It took me 1.35v to reach 800Mhz & 2000Mhz, and it was all fine. All componentes stayed cool too! The Core-temperature @ load went up to 49C. 1C more than in sign.
Hmm the Heatsink gets pretty hot. U can still touch it but after a few seconds (unter load) u want to release ur fingers off the sink.
So i think it's between 70 and 100 degree?
What about ur heatsinks, do they get that hot too?
My coretemp is also no problem stays around 50°C.
My mosfet-HS don't get as hot as yours :/ It feels like it's about 45°, i glued my sinks onto the card thou, superglue+thermalright paste.
hmm i used a special 2 components heat conductance glue.
I mean the sign that my heatsinks get hot is on the one hand a good one because it shows me that the heat conductance works between mosfet and heatsink but on the other hand the temperature seems to be to high anyway.
For a test i set all voltages to standard and looked what temperature do the mosfets have then in load.
And i must say they are nearly as hot as without voltmods.
Regarding this I think these temperatures are normal and no danger for my card.
I set the vDimm now to 2.05V and I could improve my clock by a huge amount- 1115 mhz right now and still no errors.
Aha :) That's great!
I got alot of fans in my case so i'm pretty sure that helps alot for me to keep the card cooled, (7x120mm casefans) hehe.
And no, i don't think it's dangerous for your card with that temp, thou it shouldn't be too hot either. Normally the mosfets gets cooled by the stock-HSF. Thou they can't be that with our layout on the PCB.
Nice OC on the mems! That's 115 more than me :p
Reached 785mhz gpu and 1962mhz shader now with 1,325V idle and 1,28v load now.
but im not sure if i should keep this settings since the power consumption and heat generation have increased a lot.
what programs do u use for testing the stability?
That's what i reach with 1.325 too. So we seem to have exactly the same initial at different vGPU-voltages.
I use ATI-tool at first to see that i won't get any yellow dots or pink thingys at all and then i use 3Dmark06, not really hardcore-stresstesting any with like aquamark or such, as i am a 3dmark06-point-hunter mostly :)
Gonna test my cpu @ 3.3GHz and GPU @ 800+ 2000+ and 2100+ now to see if i can brake my old max-score.
12489p
CPU 3.3Ghz
Mem 1033Mhz 4-4-412
GPU
Core 800
Shader 2000
Memory 2106
*Note to self*
Hi,
yes seems like we have very equal settings.
I also have a C2D running at 3,3Ghz.
For comparison, i get 13100 3dmarks with 174.74 and vista 32 with my last graphics card settings.
edit:
system is e6320@3.3Ghz, 4x1GB A-Data @944mhz, asus p5k @fsb 472
But for me I'm off now, i'll return to lower voltages.
It was exciting to see what clockrates you can reach but now it's enough for me :)
Reason is simple: if u draw a chart with 2 curves, one is the overall performance of the system and the second is the power consumption and heat then you will see that at a certain point the second one rises exponential.
E.g. this means when u want to gain 1% more performance u must pay 5% more power consumption & heat.
And so it's not a good idea to let the system run with these high voltages 24 hours a day.. especially because i hate the big noise level.
Hehe, i see :) i only had around 12300 before vmod, i'm only getting 12500 now with vista 32bits too and 174.74 drivers. So yeah, very equal, except score :) win XP won't give me higher score either, i get lower score in XP actually..
Ey, what the heck, power consumption is not that important for me, i can use as much as i want for a specific price per month. :)
I wanna brake the 13k 3dmark06-limit with my 60$ CPU :(
EDIT: 12520p is highest so far :( http://service.futuremark.com/compare?3dm06=6239919
GPU-Z link: http://www.techpowerup.com/gpuz/8vndc/ ( Core 820 Shader 2043 Mem 2106 )
hehe
with 3mb more l2 cache u would have definitly broken trough the 13k barrier:)
Well my problem n1 is the noise level. I wanna have a silent system that is really quite. On the other side I also want to be able to play the latest games.
It's hard to find a good compromise between high performance and silence.
Higher voltages mean more heat, and my power supply unit gets really hot then. My load wattage is 300W. Then the fan has to turn up and generates a damn loud noise.
So my goal is now to find the compromise ;)
For CPU ive already found it. Running now at 429FSB which results in 3Ghz CPU. still a good push from 1,86 to 3Ghz. and the big advantage is it runs stable with only 1,27Vcore. 3,3Ghz needed 1,41v. In numbers that means my Core Temp sank to max 57°C from 70°C.
Now i have to find the max core of my graphics card with 1,1V gpu and 1,95v ram. should be around 680 , 1040
And then ill run 3dmark again and see what score I still have.. I suggest not much less then 11000.
Edit:
I found the max. stock clock for my graphics card now too: 700mhz gpu-1750 shader- 1055 ram
I did some power consumption measuring before and after voltage decreasing:
1,41Vcore - 2,15Vdimm | 1,325Vgpu - 2,05Vmem
IDLE: 110W
LOAD: 306W
1,27Vcore - 1,85Vdimm | 1,1Vgpu - 1,95Vmem
IDLE: 80W
LOAD: 200W
And->I still get 12012 3dMarks with these new settings...
Isnt that quite impressive?
My computer needed more than 34% more energy for gaining 8% more performance....
Now with only 200W the power supply's fan is much more silent as before.
Now i'm really happy :D
Hehe probably :)
Well with soundcompromise i know what you mean, i bought myself the perfect case, since i've been buying alot of cases in the last 2 years, Cooler Master Stacker 832 seems to be the perfect choise for me.
Running all my fans at low voltage wich results in low noise even thou the case is like a swiss cheese i have as much as 9x120mm fans when playing with these OC's to drop temps with about 5C, but i cut down to 6x120mm fans when i'm done playing. It gets so quiet that i can barely hear the system, and it usualy stays in a reasonable temperature-limit. And i've also switched out the PSU-fan to a very quiet one ;)
Wow, that's alot of change in power consumption i must say :) Cool to see. What are you meassuring with?
12k with the lill OC? i get 11k with stock @ GPU and CPU @ 3.2.
But yeah, impressive :D
Hi,
i took a small digital device from our local power station.
I have an old chase that came with small air openings. So i had to make an blowhole into the bottom of the chase.
The only noise I hear now is the sound of my harddisk (wich is decoupled though) and of my PSU.
I also replaced the ADDA 120mm fan of the Seasonic S12II 380 with a Slipstream 1200.
The 800er doenst run with 4,2Volts so i couldn't choose it.
Maybe another fan would be better with less noise?
hi guy's
im in deep sh** and some off u can help me!!
i desoldert these 2 smd's and i dont know the value's if some one can messeure these out for me ?
http://i3.photobucket.com/albums/y90...tb/8800smd.png
these are the 2 smd's , need help guy's !!!
Quote:
Originally Posted by blind_ripper
I'd be glad to help you out on that, but AFAIK you have to desolder them before meassuring, and i'd rather not desolder any of them :/
can any one get me a read-out ?
Btw does anyone know if these Hynix chips is good or not? the memory chips.
It's:
Hynix KOR
HY5RS123235B
FP-1 752A
VTFF480903
And it seems to be Hynix 1ns chips. But does anyone else know anything more about them?
EDIT: Found out that these chips are specified at 1000Mhz and are high-voltage chips. So therefor my vmem is high (2.05v), but they're specified at 2.2v @ 1000Mhz. And mine does 1050 @ 2.05v with ease.
@micheskillz
is this the specific model of your card? SF-PX88GT512D3-HP
hi guys i have a big problem. I have a non reference board v1.2 sparkle as the one in this topic, but i cannot overvolt the gpu core. I soldered one 50 kohm vr as described above to the ST L6713A and ground, soldering pin as described before but whei i try to increase voltage, it drops to 0.000 volt after it reaches 1.108. During test i tryed to bypass the OVP protecition of ST L6713A but without success. Is there a known workaround?? what i can do??
Tnx a lot
Quote:
Originally Posted by metallus84
Post a picture of the volt mode! Maybe you did a mistake! Let us take a look on it! :up:
ok i will post as i make pics
ok i made pics, i hope are detalied enough, i made only vmod for gpu core not mem
http://img516.imageshack.us/img516/8...1609ft6.th.jpg
http://img141.imageshack.us/img141/8...4163fs3.th.jpg
original size pics are here, 2 mb each (copy and paste to browser because direct link not works)
http://death84.altervista.org/img/1.JPG
http://death84.altervista.org/img/2.JPG
metallus84,
Your problem is the "ground" point - I mean it's not actually ground. That point is directly connected to a leg of a vGPU inductor meaning the solder point is at live vGPU voltage.
My search skills suck apparently. I need help to mod this card. :D
http://lakesidepc.com/100_3967.JPG
http://lakesidepc.com/100_3973.JPG
http://lakesidepc.com/100_3976.JPG
http://lakesidepc.com/100_3977.JPG
Thanks for any help.
No one? :confused:Quote:
Originally Posted by loonym
TBH, your pics are not very helpful. I can't read a single chip name. If you can give us the names of the voltage regulators, then we can help. ;)
I'm not sure. Is this them?
http://lakesidepc.com/s98.JPG
If I find the time, I'll mark all the interesting chips. ;) RT9259A doesn't look bad, though.
http://tw.richtek.com/download.jsp?t...RTOTI1OUEtMDJD
Could be Vmem. I don't think that it's responsible for VGPU.
Thanks, I truly appreciate any help and I can make better photos.
The 2 chips marked in red could be interesting:
http://www.hs-augsburg.de/~ingmar_k/100_3977.jpg
And please take a pic of the card's frontside, too. Perhaps the VGPU controller is mounted there. All the chips that I have seen so far don't seem like the typical VGPU controller. ;) Far too simple, IMHO. I have a feeling that there's one on the frontside of the PCB.
BTW, I would guess that the IC marked in red, on the left (the one with 8 pins) is the EEPROM, but I could be wrong.
Sorry my cam is not better. I see nothing typical about this when I look at it beside my 8800s.
http://lakesidepc.com/100_4008.JPG
http://lakesidepc.com/100_4028.JPG
http://lakesidepc.com/100_4016.JPG
http://lakesidepc.com/100_4017.JPG
http://lakesidepc.com/100_4015.JPG
One tip for the future: If you want to take a photo of an IC, with the markings readable, then you have to use flash and take the shot from an angle. If that combination doesn't do the trick, for example because the markings are barely visibly at all, then rubbing some white thermal compound on the markings, most likely helps.
Looking through the pics right now. The two chips that I had marked in red, are an EEPROM and a Quad-NAND, so nothing interesting in terms of vmodding. Looking through the rest of the pics now.
Edit: OK, seen all pics, but the only real controller chip remains the RT9259A. Seems strange. Not even an OPAMP there.
BTW; is there a chance that there is a chip under that huge white label sticker?
I peeled that, just some smd under there.
I really though that the RT9259A would control Vmem, but without another controller, it nearly has to control VGPU. Will have another thorough look at all your pics this weekend. Perhaps I missed something.
What you could do anyway, is to measure voltages on the RT9259A.
Of interest are voltage and resistance between the following pins:
- #4(FB) and #7(GND)
- #6(FBL) and #7(GND)
Resistance 4 to gnd=.626
6 to gnd=.785
voltages for 4 and 6 are .792 and .789 respectively.
I believe that "AP-AB" pictured in post #64 is vGPU ctrlrl, the look-alike chip above that, the fet driver.
RT9259 would be just vMEM.
That would make sense, but unfortunately I couldn't find any info on it. Have you successfully found anything? :confused:Quote:
Originally Posted by largon
I've seen a few of these mysteriously marked (eg. AP-AA, AP-AG, and now AP-BA ) 16pin QFN buck chips with very inconsistent pinouts. Some have FB pin in pin16, some pin14, some pin8...
:shrug:
Strange. Perhaps someone else can dig up some more info or even already knows more details. I'll have to give up on the VGPU mod here. For one because I don't have much spare time ATM and because I just don't know where to look for more details.
Seems like recently a lot of new small semiconductor companies from Asia entered the market, so it's quite easy to miss something. And some of them seem to lack good documentation, too. At least in english, that is.
Thanks for looking, it's appreciated. They were pretty cheap so I'm going to probe around a little and maybe do an experiment :D
For Vmem, you probably already know. Take the resistance between pin "FB" and GND on the RT9259A, multiply that resistance by ~20 and that's the value for the trimmer that you should use. Should work just fine. Trimmer also between FB and GND.
HAH, you'd be surprised how little I know about electronics but I'm slowly learning. I think the mem will be nice with a bit more volts so it's a good place to start. If I have a 'breakthrough' I'll be back. :DQuote:
Originally Posted by celemine1Gig
And don't forget to tune the trimmer to max resistance before soldering it in between FB and GND. That's the easiest way.