X700pro volt mods?

Printable View

Show 100 post(s) from this thread on one page

03-07-2005, 12:48 PM
crotale

Quote:

Originally Posted by Steven

wat is ground on that card?or just to an molex ?

http://img120.exs.cx/img120/3400/vgpuarea8py.jpg
You can use the pin just below the yellow one, pin 3.
But the pins look small so I strongly suggest finding another ground, should be plenty of them on the card. Just measure on some of the larger spots.
03-07-2005, 03:40 PM
CaTalyst.X

I would reccomend using the grounded solder point of one of the electrolic capacitors. They are thru hole caps, just probe between the leg and GND on the resistance setting and see if it "beeps" if you dont have the beep function on your dmm, it should read like .002ohms resistance

-CaT
03-07-2005, 04:10 PM
Malves

Quote:

Originally Posted by crotale

Nice work CaTalyst.X, but don't you think we should advice the first person who would like to try it to start with 100k VRs instead?
It's better to be on the safe side. Either way, be sure to have 20 turn VRs and turn them slowly and check voltages carefully.

No need for that. The 20k trimmer should be enough for this mod, even a 10k will be ok. This is the old ISL6522 that ATI has been using for their cards for Vmem, including the 9500, 9700 and 9800 series.
03-07-2005, 04:39 PM
CaTalyst.X

Yeah but the 20K is needed IMHO. Just go with 20K and work from there. On the SP61 chip youre guess is as good as mine.

-CaT
03-07-2005, 06:03 PM
crotale

Quote:

Originally Posted by Malves

No need for that. The 20k trimmer should be enough for this mod, even a 10k will be ok. This is the old ISL6522 that ATI has been using for their cards for Vmem, including the 9500, 9700 and 9800 series.

I'm just saying that it might be a good idea for the first person to try a larger one to be on the safe side.

I also think that the appropriate range for "our" VR is more dependent on how ATI (or the manufacturer) has chosen to implement the feedback circuit around the IC.
03-08-2005, 05:39 AM
Malves

If it makes you feel better, go ahead. But I guarantee it is not needed.
03-08-2005, 07:52 AM
Snowman89

is there no other vmod out there? think it is pretty hard to solder on pins so much easier to use other spots like solder spots and so on, really not complaining on you guys great job that you found a vmod but ust wondering if theres no other way to do it on? like a pencil way or a easier one?
03-08-2005, 09:29 AM
persivore

Quote:

Originally Posted by Snowman89

is there no other vmod out there? think it is pretty hard to solder on pins so much easier to use other spots like solder spots and so on, really not complaining on you guys great job that you found a vmod but ust wondering if theres no other way to do it on? like a pencil way or a easier one?

I'll look over the pictures again later and see if I can find an alternative solder point for the Vgpu mod.
If someone can get a resistance reading between pin4 on the SP61 and ground, and pin4 and the Vgpu read point, I will work out the exact value VR you need to use for the GPU mod.

Edit: pin4 (feedback) appears to disapper to another layer of the PCB, so I can't see where it is connected to :(
03-08-2005, 05:29 PM
CaTalyst.X

Yeah PIN 4 Disapeears into another layer, however you may be able to set DMM to continuity and probe between pin 4 and points that are close to it on the board to see if there is an alternate mod point. LOL bro thats pretty cool if you can work out that VR from just that info :) How are you gonna do that?

-CaT
03-09-2005, 02:59 AM
persivore

Quote:

Originally Posted by CaTalyst.X

Yeah PIN 4 Disapeears into another layer, however you may be able to set DMM to continuity and probe between pin 4 and points that are close to it on the board to see if there is an alternate mod point. LOL bro thats pretty cool if you can work out that VR from just that info :) How are you gonna do that?

-CaT

I can use the formula for a non-inverting amplifier (which regulator basically is), 1+(Vin*(Rfb/Rg)) to work out the gain that the regulator has with different resistors and thus how much voltage the regulator will give out, where Rfb is the resistor between the Vgpu supply and the feedback pin, and Rg is the resistor between feedback and ground.
03-09-2005, 03:44 PM
CaTalyst.X

HEHE Thats over my head bro, do you have a background in Electrical or Computer Engineering?

-CaT
03-10-2005, 08:53 AM
FireDragon

That is also over my head i am in a Tech Elec class at school (it is my whole school day)...we are just getting into making amps and so on...i will take a good look and find if i can't find any other points to boot....
03-10-2005, 09:27 AM
persivore

Quote:

Originally Posted by CaTalyst.X

HEHE Thats over my head bro, do you have a background in Electrical or Computer Engineering?

-CaT

I'm studying electronics at college atm, and I'm going on to study electronic engineering at uni after that.
03-11-2005, 09:26 AM
the_countess

the amount of current that will be running from pin4 on the SP61 to the 20k trimmer wil be pretty minimal right?
so a single hair-thin (copper) wire solderd on the 4th pin should be enough right? (possibly secured with some epoxy)

edit : if someone dusnt beat me too it i'll messure the resistance as persivore asked when i get my multimeter but that could be a while yet(week/week and a half.)
03-11-2005, 10:58 AM
persivore

Quote:

Originally Posted by the_countess

the amount of current that will be running from pin4 on the SP61 to the 20k trimmer wil be pretty minimal right?
so a single hair-thin (copper) wire solderd on the 4th pin should be enough right? (possibly secured with some epoxy)

edit : if someone dusnt beat me too it i'll messure the resistance as persivore asked when i get my multimeter but that could be a while yet(week/week and a half.)

Welcome to the forum :)

The current flowing through the pot/wire shouldn't be much more than 5ma with the pot at quite a low resistance, so a single strand of wire will be fine. I've found that thin enameled copper wire works well when trying to solder to very small areas like pin4 on the Sp61.
03-11-2005, 06:07 PM
CaTalyst.X

U should strip a IDE cable :)

-CaT
03-14-2005, 05:24 AM
MikeMurphy

So whats the status on this now?
03-14-2005, 01:53 PM
the_countess

well im waiting for my multimeter to be deliverd

but if you have one i'll be happy to let you have the honor of being the first to do it. :)

Quote:

I've found that thin enameled copper wire works well when trying to solder to very small areas like pin4 on the Sp61.

exectly what i had in mind and the reason i asked. and using a ide kabel is a good idea, then it will be insulated right away
thx for the welcome btw :)
03-14-2005, 10:46 PM
Snowman89

so for the vmem mod i need to solder on one of the pins in the picture and then to ground?
whats the max safe voltage to run?
what is this ide cable thing your talking about?
03-15-2005, 03:48 AM
the_countess

Quote:

so for the vmem mod i need to solder on one of the pins in the picture and then to ground?

NOOOO!
well yes, but you NEED to put a (ajustable) resistor between the pin and the ground. 20k atleast to be safe.
if you just connect it to ground you'll blow up your card most likely.

the idea is that you bleed just a bit of power from the sence line so the voltage controler thinks its not giveing the memory enough. the resistor is there to make sure only a little is bled.
and you want it ajustable so you can decide what you want on your mem.

the idea of the IDE cable is that its a nice cable to strip and use. you'll have a thin copper cable thats shielded (to prevent shorts) thats ideals for soldering on small pins.
03-15-2005, 07:06 AM
Snowman89

Quote:

Originally Posted by the_countess

NOOOO!
well yes, but you NEED to put a (ajustable) resistor between the pin and the ground. 20k atleast to be safe.
if you just connect it to ground you'll blow up your card most likely.

the idea is that you bleed just a bit of power from the sence line so the voltage controler thinks its not giveing the memory enough. the resistor is there to make sure only a little is bled.
and you want it ajustable so you can decide what you want on your mem.

the idea of the IDE cable is that its a nice cable to strip and use. you'll have a thin copper cable thats shielded (to prevent shorts) thats ideals for soldering on small pins.

yes of course a resistor between :p: forgot to write it ust :cool: so this guide works and theres no risk of destroing the card if i ust do it right, like you maybe selected the wrong pin and so?
whats the max volt i should run on the memoryies?
03-15-2005, 09:43 AM
MikeMurphy

http://www.samsung.com/Products/Semi...K4J55323QF.htm

The max spec to put at these is 1.91V. 1.95v sounds good ;)
03-15-2005, 10:37 PM
Snowman89

Quote:

Originally Posted by MikeMurphy

http://www.samsung.com/Products/Semi...K4J55323QF.htm

The max spec to put at these is 1.91V. 1.95v sounds good ;)

lol, then its no need to vmod because stock voltage is around 2.1v
03-16-2005, 11:25 AM
the_countess

Quote:

Originally Posted by persivore

I'll look over the pictures again later and see if I can find an alternative solder point for the Vgpu mod.
If someone can get a resistance reading between pin4 on the SP61 and ground, and pin4 and the Vgpu read point, I will work out the exact value VR you need to use for the GPU mod.

ok presuming i had the correct ground (the minus side of a condensor, had the same reading as the plate at the front, you know the one with the vga connector and stuff.) then i got :
pin4 to ground as 739ohm
and pin4 to vgpu read point as 743 or 742ohm
(measured with the card OUT of the computer)

would that make any sence?
im a bit wurryed and i dunu if its correct because if i turn probes around (put red where i had black befor and visaversa) i get a ohm reading starting somewhere at 1200ohm and increasing all the time
or would that make sence too? (new DMM so just making sure)

edit2 : saw in the spec's pin3 is ground too and used it to check my chosen ground, and it checked out. so above reading should be correct.

p.s. i read in the specs that if the Vfb (aka pin4) "drops 0.25volt below the positive reference" it asumes there is a short and enters "huccup mode"
http://www.sipex.com/products/pdf/sp6132.pdf (page 3)

dus that mean a max overvolt of "just" 0.25v? (probably less to be safe)
03-16-2005, 01:17 PM
persivore

Quote:

Originally Posted by the_countess

ok presuming i had the correct ground (the minus side of a condensor, had the same reading as the plate at the front, you know the one with the vga connector and stuff.) then i got :
pin4 to ground as 739ohm
and pin4 to vgpu read point as 743 or 742ohm
(measured with the card OUT of the computer)

would that make any sence?
im a bit wurryed and i dunu if its correct because if i turn probes around (put red where i had black befor and visaversa) i get a ohm reading starting somewhere at 1200ohm and increasing all the time
or would that make sence too? (new DMM so just making sure)

edit2 : saw in the spec's pin3 is ground too and used it to check my chosen ground, and it checked out. so above reading should be correct.

p.s. i read in the specs that if the Vfb (aka pin4) "drops 0.25volt below the positive reference" it asumes there is a short and enters "huccup mode"
http://www.sipex.com/products/pdf/sp6132.pdf (page 3)

dus that mean a max overvolt of "just" 0.25v? (probably less to be safe)

Thanks for the readings. The different values with the probes connected differently are probably due to capacitors on the card charging slightly from the voltage the DMM probes give out.

A 20k variable resistor between pin4 and ground should be fine, and should give a 1.96% voltage increase above stock voltage at maximum resistance.

I'm not totally sure about the short circuit protection, but I don't think there will be a problem setting more than 0.25v over stock. I think that the regulator compares its internal reference voltage of 0.8v to the feedback pin, which will be fixed at 0.8v in normal operation by the regulator, and should only fall below that level if there is a short.

Show 100 post(s) from this thread on one page