ATI 4850/4870 voltmod thread

[quake6]

@wesleyclark: Just one question for vmem. Middle pin is ground?

[Clemmaster]

What kind of increase did you guy's see when going from 2.0 to 2.1v on the memory chips ? In the past the clocks went down when increasing the voltage. Because the boards where already running the chips to the max. Lowering the voltage sometimes gave a increase.

I'm thinking about v-modding my HD4850. But i'm also thinking of buying a HD4870. Those things have 1.26v on the core stock. And run like 790 MHz (max CCC) without a problem. I dont see my HD4850 with running 800 MHz with just 1.26v.

Nothing for me

@Quake6: yes, the you can solder it to another ground if you want, this one is directly on the chip but it's not mandatory

[chri5g]

Damn.. testing atm - my card wont even do 825 @ 1,45v...

1.417V in 3D
870Mhz stable in Crossfire

[Miwo]

Here are the values of the metal can caps if anyone is interested in seeing if adding/replacing caps will make a difference

Top Right Area of the 4850 starting from Left Side:

4x 820uF 2.5v
1x 470uF 25v Sanyo
3x 47uF 16v
1x 100uF 16v Sanyo

Bottom Right Area Starting from Left Side
1x Sanyo 100uF 16v
2x 820uF 2.5v
3x 47uF 16v

I am assuming that I am reading these ratings correctly. The Polymer Solid Cans have markings that go like this: These two are the most common:

ULR (manufacturer?)
16v (voltage)
470 (this should be 47 * 10^0 uF)
XI ( type?)

833 (type, or maybe week/batch code?)
821 (likewise this has to be 82 * 10^1 , because 821uF is not a common rating)
2.5v (voltage)

The main power filtering cap is the Black Sanyo I am guessing since its the largest? 25v 470uF

[adpr_02]

Damn.. testing atm - my card wont even do 825 @ 1,45v...

Does the screen go black after a short while?

Just got 15433 marks in 3dmark06 in vista 800/1100. Time to push higher

[WheelMan]

Read my posts further down.

Wouldnt do 1GHz 1,63v, trying to get 900-950 stable

[JamesBong420]

if any of you guys have GRID, can you test your overclock with it? I tried ATI tool stress test, and it was fine, i am retrying crysis to see if it loads the card more, but no matter what i do when i get into GRID the game shuts off my video card! I did flash to get 750 mhz but shouldnt it be stable at 1.2v????? unless im not getting 1.2 from the bios flash. I tried 725mhz in GRID and it too failed. The only thing i have overclocked is the core and its cooled by an S1, the ram dont have heatsinks on them yet but will very soon. Before i flashed, it would do it at 700mhz too, so im going to test it out at 700 see if it crashes. This sucks ass!

grid works flawlessly at 910 mhz core on mine.. haven't tried lower.. man that game looks amazing at 1920x1200 4xaa i so badly need to find an IDE drive in storage.. so i can actually get some real bench scores im getting passed up by peeps with 700 mhz core..

[zaraza]

So it is my core not getting enough power. Damn! ill keep it stock till later today i should be getting heatsinks and a multimeter. After that i have 27 days left to return it to bestbuy if it breaks lol. I dont like how hot it runs on stock cooling.

[JamesBong420]

if your using the fan header from the card then i hope you have come across this.
http://www.xtremesystems.org/forums/...d.php?t=192932

[jason4207]

What kind of increase did you guy's see when going from 2.0 to 2.1v on the memory chips ? In the past the clocks went down when increasing the voltage. Because the boards where already running the chips to the max. Lowering the voltage sometimes gave a increase.

I'm thinking about v-modding my HD4850. But i'm also thinking of buying a HD4870. Those things have 1.26v on the core stock. And run like 790 MHz (max CCC) without a problem. I dont see my HD4850 with running 800 MHz with just 1.26v.

See my quote from earlier in this thread:

You are correct about the Samsung chips. My GTS w/ Samsung loves voltage. It was theorized that it wasn't the Qimonda chips themselves, but the memory controller that was causing issues on the cards that had Qimonda memory. The reasoning being that turning down the voltage actually helped one obtain higher clocks, and at the brink of stability the timings on the Qimonda chips could still be tightened w/o having any adverse effect on stability. It seemed as though something else must be at fault. The memory controller was spec'd at 1.8v, but going from 1.2ns to 1.0ns Qimonda RAM only involved pumping up the voltage from 1.8v to 2.0v on the RAM (the controller had to come along for the ride since they shared the same rail). The Qimonda RAM did fine w/ the extra voltage, but the controller wasn't stable anymore. Iirc the spec sheet on the Qimonda has them rated up to 2.5v!

Hopefully, the Qimonda RAM on the 4850 will respond well to extra voltage.

I didn't get to do any extensive testing b/c the card I did was for someone else, but I can tell you that at stock voltage w/ stock fan speed 1200 didn't pass the CCC test, and w/ 2.09v, and the fan at 100% 1200 did pass the CCC test.


If the 4870 can easily do 790 w/ 1.26v, then why wouldn't the 4850 (uses the same core as 4870) be able to do ~800 at 1.26v? I know the 4870 has a better power delivery system, but they should still be close until you get up much further than that.

[zaraza]

will this multimeter do the job?

http://shop4.frys.com/product/542743...H:MAIN_RSLT_PG

[Specialfighter]

my 4850 is doing only 850@ 1,45V

its a bad core i think

on stock only 670 oO

my ram does 1175 @ stock

but for extreme oc on ram i need water on it

i have tested 1300Mhz @ 2,4V on passiv heatsinks and i think its to hot i get extreme artifacts





greetz from germany

[wesleyclark]

will this multimeter do the job?

http://shop4.frys.com/product/542743...H:MAIN_RSLT_PG

That multimeter looks like will work good. And it takes a 9v battery, or 1.5v, which is good, as opposed to my stupid MM needing a 12v battery. yay

[wesleyclark]

I May have the V-mod for a 4870 to test, someone who owns a 4870 and a MM please PM me.

jimmyz, could you post the proposed mod, even though though it isn't confirmed so others can look at it?

[jimmyz]

jimmyz, could you post the proposed mod, even though though it isn't confirmed so others can look at it?

nope. Im not being a jerk I just don't want someone frying a card and saying they saw my mod and tried it. as soon as someone confirms it ( I have at least one tester) It will join the other mods in the first post if K404 wishes.

[darkskypoet]

@wesleyclark: Just so you know... the 12v batt I got, which yours should take as well, will resemble :

Duracell 21/23 - 12V MB21/A23 From their Security line... (as in like Car alarm Key FOBS)...

I am not sure if any of the model numbers given are Duracell specific. Got it in Walmart.. Was just on their battery shelf can't be rare at all. But there is probably a few less components required if it runs off 12v rather then 9v or 1.5v (AA) - At least I'd guess that, because our DMMs are of the dirt cheap variety (at least mine was, and yours is probably similar) Nexxtech I am guessing?

[Booj]

I got my HR03 GT today, and did the pencil mod on my 4850. Currently stable @ 800mhz 1.31v, 1130 on the ram unmodded. Hoping to get 900mhz stable. I love this card.

[wesleyclark]

My DMM is radioshack branded. I went to 2 radioshack stored to find the battery, as it said it was available there, and NEITHER had the battery. So i am going to radioshack to return it in about 10 minutes. It was $20. Then i will be off to FRYS electronics to get theirs, which is half the price, @ $10, and uses a 9 volt or a 1.5v. So. Im happy.

[wesleyclark]

nope. Im not being a jerk I just don't want someone frying a card and saying they saw my mod and tried it. as soon as someone confirms it ( I have at least one tester) It will join the other mods in the first post if K404 wishes.

That makes sense. Ill wait


Alright, went to Frys. Got 2x5k, 2x10k, 2x20k, and 2x100k Trimpots. If i needed a 1k for something, i could replace the 1k with a 5k correct? I would just need to turn it down a lot more before it would get effective right?
I also replaced my solder. So ill be doing a little bit of work tonight. But no DMM. The cheap ones looked pretty unattractive, so i will go back tomorrow with more money and get a nice one for about $45. That should work good. Also got some shrink wrap, and electrical tape. looks like I'm good to go.

[darkskypoet]

My DMM is radioshack branded. I went to 2 radioshack stored to find the battery, as it said it was available there, and NEITHER had the battery. So i am going to radioshack to return it in about 10 minutes. It was $20. Then i will be off to FRYS electronics to get theirs, which is half the price, @ $10, and uses a 9 volt or a 1.5v. So. Im happy.

nexxtech is the radio shack brand... Strange they didn't have the batt when walmart did... We don't have fry's up here... So couldn't take advantage of that... this one was like 20$ too. Well hey Good to hear!

[865pe]

Guys,

4870 diffrent PCB where GPU measure point, memory measure poit. and vmod point?

[Theta]

Guys,

4870 diffrent PCB where GPU measure point, memory measure poit. and vmod point?

As per W1zzard and jimmyz, here is the 4870 measuring info:

on your dmm put the black probe to ground (like case metal or the metal piece on the graphics card) and the red one to what you want to measure

yellow area and green area should be either mvddc and mvddq memory voltage (dont know which is one because i have no board). any of the pads in one color will work fine, it is safe to short between same colored pads

the blue area is vddc (gpu core voltage). as you can see two solder pads go together (white boxes on the pcb), one is ground, the other one will be vddc. test both, the one with the 0 volt reading is the gnd one. dont create shorts between those pads.

if you have questions ask before causing more damage

[JamesBong420]

im thinking we need a shot of the front for the vmod...

[Theta]

im thinking we need a shot of the front for the vmod...

Nope... The back is easier at this point.

[JamesBong420]

Nope... The back is easier at this point.

what?
enlighten me then please.....