View Full Version : [Vcc mod] KFA2 Geforce GTX 560 TI
03-03-2011, 11:54 AM
[ *] Graphics Card: KFA2 Geforce GTX 560 TI
[ *] Volterra VT1185MF
[ *] Need FB-pin mod, total OCP, channel OCP, and if possible VID mod + 300mv
Help would be greatly appriciated .. cause Datasheet is not available :(
Or if you like to share the Datasheet ... would be great too, that way i should be able to figure things out myself.
He-he, ME4ME :) One of rare cards I want :D
Need FB-pin mod
There is no direct FB on Volterra's. Why not using Afterburner or similar software to controls? Also you can use EVBot to control stuff if you have it.
total OCP, channel OCP
Measure resistance to gnd on pins 46, 45, 44, 43, 42, 41, 40 please.
possible VID mod + 300mv
Mmm, why want that? :)
03-04-2011, 12:48 AM
GPU-tool and NVinspector can set voltage, but only up to 1.15v even though GPU-tool slider could go further. Not sure why... Also tried Afterburner, but it doesn't support this card.
Reason why i would like +300mv is to have a stable voltage level when running on LN2. In my oppinion things seem to run better when running constant voltage.. and not going back to 1v when crash or reboot. Keeping the GPU hot seems to run smoother. :)
Ill messure resistans on those pins when i get back home this afternoon. :)
Hope these pins are correct :)
03-05-2011, 02:25 PM
stupid me... totally missed the obvious...
backside of VT1185MF:
looks like VIDs.. so now VID-table would come handy :D
been thinking about normal vmod too.
From what i understood, VT1185 can be manipulated by messing with vsense+. (http://kingpincooling.com/forum/showpost.php?p=9178&postcount=18)
So looked up some pictures from your old EVGA Untouchables thread TiN, and found you used VR between resistors 17 and 18 (http://www.xdevs.com/kb/epower/ucard_rig1.jpg) to control things. Found even a better pic on xdevs (http://www.xdevs.com/kb/epower/ecard_vtmf.jpg)
in other words .. should boot up the system, check which points are connected to vGPU and ground ... messure resistance of the two resistors in case, and start messing with those? :D
just doing as much research as i can ..but really hard to find stuff :(
I would not recommend that, that VR on UT was already burnt and stay there just because I was lazy to remove it :D I used digital controls for UT, it have current DACs there to adjust voltage :)
You can go this way too, if you have EVBot.
Or, you can go VID-mod :)
For that u need to check voltages on pins 2,3,4,5,6,7,8,9 with powered card, or draw those resistor pairs to see what VID setting there.
Actually you may use 1115 DS, it can be found on baidu.com search engine ;)
03-05-2011, 03:39 PM
Don't have any EVBot, so VID mod sound like a good option.
Then +300 or something like that could be set, and do final finetuning by software +-150mv. Any idea which side is what @ the VIDs.. and which resistors to move? hard to know what to do, without VID-table.. :o
And what about OCP? :)
EDIT: ok thanks for the tip, will check 1115 DS!
Change resistor connected to pin #41 from 4.41K to 16.9k. This will boost your OCP limit almost twice from what I know.
Also measure resistance #38. There are few tables for VID.
03-05-2011, 05:16 PM
Thanks! .. will try OCP mod tomorrow.
Have been messure'ing the vids a bit:
so it's: 0111000 (without vid7 counted) ... pin 38 is 334 ohm.
so page 27 in datasheet right? ... It's kinda weird.. 1.50625v and 1.48125v matches too. Either way... thats way to high. Can it be that they've set that high, and adjusted by vsense to get normal ~1v?
cause stock VDDC for the card in windows is 1.012v and max any software can set is 1.150v.
Can it be that they've set that high, and adjusted by vsense to get normal ~1v?
I don't think so.
So, VID table for 1185 is different :D
03-06-2011, 04:15 AM
Pin41 from 4.42K to 16.9K .. that one you got from the VT1165M DS right?
cause 16.9K is not in the 1115... so it made me a bit suspious..¨
anyway.. about the VIDs... how should i handle this safest. Just go nuts, and move 1 vid at a time, and see what happens? Maybe change VID3 from 0 to 1, cause VID3 always seem to be +200mv?
oh and for fun, the rFREQ resistor is 26K ... means 800khz switching frequency ... nice, me like :D
. that one you got from the VT1165M DS right?
Nope, that's from my experimentation :D
As for VID's, if VRM is in SVID mode, moving that resistor will not do anything to output voltage :]
03-06-2011, 04:27 AM
So I should change VID mode by adjusting R_Sel3 first, and THEN adjust vids? :p
edit: ..just saw that R_SEL3 is the pin 38 which i messured yesterday, which was 334ohm. According to 1165M (and 1115) that should be VRM10.x mode
edit2: looking at this picture (http://www.xdevs.com/kb/epower/ecard_vtmf.jpg) ... any idea what the stock voltage/vid is of the Untouchables?
if higher then what i currently have on this card, i may just copy the vid positions from UT
03-06-2011, 09:27 AM
OK, Didn't have the patience to digg further, so i went for a VID mod.
changed VID =1 to =0.
turn out quite nice, maybe a bit overkill though:
now idle in windows voltage is 1.47v ... load with furmark about 1.57v(!)
Software is still working though, so i have about 190mv to adjust. It can only go -40mv ... so about 1.43 idle / 1. 53 load ... still quite much. But im happy to see it working :)
TiN, any idea why there's such a gap between idle and load? I know it's quite common, but 0.1v is quite much..
That VID table is invalid. You did only that resistor change?
03-06-2011, 11:05 AM
yes, only resistor change.
and yea... that table is useless ... but had nothing else to go after ...
will now do vmem and vpll mod, and give it a go at LN2 tomorrow... ocp mod will be for later, first i want to see if its needed :)
edit: vmem mod done ... skipping the pll mod for now.
only thing now needed to force EN-pin high for going cold.
Looks like thats the OE-pin @ 1115 .. so, now need to find out how to mod that one
03-07-2011, 07:14 AM
Just did a 10-min LN2 run with this card. It clocks nice! :D
but terrible booup problems when cold ... like expected.
need to do some mod...
this is how you mean, right TiN? :)
remove that wtf :)
03-08-2011, 11:14 AM
a little update:
modded the OE-pin like on the picture above. All is fine ... but still the coldboot issue remains. ... not sure what to do now, but i may just give the card a go, and hope it doesnt do a full crash so the system powers down.
one thing is for sure though ... even with this midrange card, the SLIM pot is struggeling. For next test the FAT one will be used for sure :)
03-29-2011, 12:23 PM
Hey, just before this thread gets forgotten and lost, i wanna say a last "thanks" to Tin for the support! Here are some results:
Vantage GT2 (http://www.nordichardware.se/images/labswedish/artiklar/NHOC/KFA2.560ti/fullimages/560_VANTAGE.jpg)
clocks could be a lot higher, but im running into trouble when going colder than -100c. Oh well.. still had fun with this card. :)
03-29-2011, 05:23 PM
Looks like a fun card :up: Nice PCB!
03-30-2011, 12:57 AM
Yeah it looks like you still had alot of fun :)
vBulletin® v3.8.3, Copyright ©2000-2013, Jelsoft Enterprises Ltd.