RivaTuner plugin supporting GT2xx and HD4870 Voltage Tuning

finally here is a first release of my RivaTuner plugin wich allows voltage control on vga cards equipped with a Voltera VT1165 VRM controller.

as far as i know this are: GTX260 65nm, GTX280, GTX295, HD4870 and HD4870x2

this VT1165 plugin is an enhanced version of the VT1103 plugin already included with RivaTuner and supports the previously available monitoring features plus some added functionality:

already included in VT1103 plugin

• vrm output voltage
• vrm output current
• vrm temperature for up to 5 phases

added the functionality

• change vrm output voltage on the fly
• monitor status of OCP
• monitor status of OVP

monitoring of OVP and OCP might be useful for hardmodded cards to verify if either of this safety features is activated or not.

to integrate this in a RivaTuner plugin has the advantage of utilizing the available RivaTuner functions to access the VT1165 directly. there is no need to manually scan for a device, write down register addresses and enter this values somewhere else, besides in RivaTuners scheduler to apply the settings at startup.

also you don't need to care about which one of the registers is used in which mode, as the plugin will detect the currently used register itself and apply the settings to this register. so to completely setup your cards all you need to do is to bring the card in the mode you want to change (idle, 2d, low 3d or performance 3d) or what ever it is called and change the voltage as required. there are not more than 4 modes available. most cards don't use all of them. my gtx295 is using only one mode. usually i would assume putting the card under load (atitool 3dview would be fine for this) and change the voltage then. i cant see a reason to change any other modes, but its up to everybody to do so.


you do all of this at your own risk! don't blame me if you overvolted your card too much and thus burnt it. this is a very powerful and dangerous tool!


also i would not have been able to create this without the help of Unwinder, the creator of RivaTuner. although he did not take part in the devolpment, he put some valueable info in the "evga voltage tuner" thread here in the forums and also kindly answered all my boring questions!


that said, lets roll ...

the plugin can be found here: vt1165.zip

as the plugin is created with visual studio 2008 you need to download and install microsoft vs2008 runtime components. this is just about 1.7MB and can be found here:
Microsoft Visual C++ 2008 Redistributable Package (x86)

here is a step by step guide on how to install and use my plugin:

experienced RivaTuner users wil most likely know almost all of this already and just look at the few points below and at some screenshots at the end:

• download the plugin vt1165.zip
• download and install the Microsoft Visual C++ 2008 Redistributable Package (x86)
• download and install the latest version of RivaTuner
• start up RivaTuner once to let it configure its database and exit RivaTuner again (rivaTuner might get started directly after ths installation. if so, just let it configure its database and manually exit again.
• copy VT1165.dll into RivaTuners plugin folder. this is PlugIns\Monitoring within RivaTuners installation folder. on my system this is: C:\Program Files\RivaTuner v2.22\PlugIns\Monitoring
• run VT1165.reg found within vt1165.zip. this changes a single registry key to activate the plugin. don't do this in the gui as there is a bug within my code that might crash RivaTuner. i have not yet sorted this out. use VT1165.reg as a workarounf in the meantime to activate the plugin for the first time!

finally start up RivaTuner again.

the plugin checks upon startup if a card is supported and pops up the following window for each card unsupported card:

http://fg000005.host.inode.at/pub/NoVT1165.gif

so if you see this, you are out of luck!

if everything went right and your card is supported the following window should pop up:

click on the small button besides Customize...1 to open up the menu and then select Hardware monitoring 2 to bring up the Hardware monitoring window.

http://fg000005.host.inode.at/pub/p1.jpg

depending on the version of RivaTuner you are running, a warning might come up first:

http://fg000005.host.inode.at/pub/p2.jpg

clicking on the [UP] button will increment the voltage. if you are running on two identical GPUs and have selected Apply settings to both GPUs at the beginning, voltage for both GPUs will be increased simultaneously, as you can see in the following screenshot. clicking the [DOWN] button will decrease voltage respectively. the currently selected voltage is displayed in the VRM Voltage Control window immediately as you click the button. the graph might follow with a delay depending on the update rate specified in Hardware monitoring setup, usually 1 second.

if you want to make the settings to be applied automatically on startup, simply use the values displayed in the [B]VRM Voltage Control[B] window besides Parameters: when setting up a scheduler entry in RivaTuner under the Scheduler tab.

running on a system with either a single GPU installed or if there are two GPUs of different type installed, you should have selected Apply settings to first GPU only at the beginning. this results in changing the voltage for the first GPU only. in this situation, there is currently no way to change the voltage for the second GPU. as soon as RivaTuner V2.23 gets released, this restriction will be lifted, and all GPU's can be modified individually!

if you selected the wrong mode on startup, just close the Hardware monitoring window by clicking the [x] button and reopen it again. you will be prompted to select the mode on every startup.

http://fg000005.host.inode.at/pub/p3.jpg

finally, two new graphs are available that might be of interest for some hard core overvolters using hard modded cards. this two graphs show the state of OVP and OCP. if OVP or OCP is inactiv (normal) the graph shows a steady line at 0. if OVP or OCP gets activated the graph jumps up to 1 signalling OVP or OCP condition.

http://fg000005.host.inode.at/pub/p4.jpg

as i have not hard modded my card i was not able to verify if this works as expected. so if any hardcore overclockers are out there and suspect OVP or OCP is kicking in on their cards, please use this plugin and report back. also i had just one gtx295 at hand so i was not able to verify the functionalty on different cards. please report back if you run into any problems or have any further questions or feature requests. but keep in mind, my programming knowledge is somehow limited and dont expect miracles.

once again: remember this tool allows you to change voltages to very dangerous levels for you GPU, so be careful and don't blame me if you burn your card. you do all of this at your own risk!

Nice! Was waiting for someone with the coding skills to do this :up:

Thanks a bunch bro!! :D

Sorry but the more option is not available to me;):(

Wow, cool plugin :up:

No go on 64 bit vista + visioktek 4850 512MB crossfire. No error message, and no plug in shows up either.

would bios volt modding work?

Quote:

---

Originally Posted by DMH

Sorry but the more option is not available to me;):(

---

sorry my fault, mixed up the screenshots when merged them during editing! you need to rightclick on one of the output voltage graphs to get to the correct setup menu.

updated the screenshot in my post.

Quote:

---

Originally Posted by afireinside

No go on 64 bit vista + visioktek 4850 512MB crossfire. No error message, and no plug in shows up either.

---

just got a reply from Unwinder, 4850 is using a different VRM controller, thus not supported!

as mentioned in my posting, i just tested this with my own gtx295 and got the list of supported devices from various other posts. unfortunately, the 4850 series cards are not supported!

Quote:

---

Originally Posted by Boogerlad

would bios volt modding work?

---

may be, have not tried it so i cant tell.

but when you are talking about tools like nibitor, most likely not as you can change only the selection of preset voltages, but not the voltages itself as the plugin does.

modding the bios with a hex editor might work, but thats a different story and beyond my knowledge.

Quote:

---

Originally Posted by afireinside

No go on 64 bit vista + visioktek 4850 512MB crossfire. No error message, and no plug in shows up either.

---

I haven't heard of anyone that got HD4850 Softmode work with rivatuner :shrug:

If you need someone with x64 to recompile and check the source just shoot me a message. Probably just needs a few adjustments made for 64bit pointers, though RT is 32bit so that shouldn't be an issue itself.

Are you using RT API to write the registers?

BOOL WriteI2CEx(BYTE *BUS, BYTE, *DID, BYTE *REGISTER, LPCSTR *g_complex[gpu_id].m_dev) I believe

There is also ReadI2CEx(...)

the Ex functions are to access any GPU device, the non-Ex are only for primary GPU.

Quote:

---

Originally Posted by fgw

sorry my fault, mixed up the screenshots when merged them during editing! you need to rightclick on one of the output voltage graphs to get to the correct setup menu.

updated the screenshot in my post.

---

Nope still not working man:(

Quote:

---

Originally Posted by DMH

Nope still not working man:(

---

what card are you using?

Works nicely on my 4870x2. Thanks for releasing this.

Quote:

---

Originally Posted by mikeyakame

If you need someone with x64 to recompile and check the source just shoot me a message. Probably just needs a few adjustments made for 64bit pointers, though RT is 32bit so that shouldn't be an issue itself.

---

guess the problem is he is using 4850 cards. unfortunately i originally thought the 4850 is using the VT1165 but i was wrong! updated my post and removed the 4850 from the list.

Quote:

---

Are you using RT API to write the registers?

BOOL WriteI2CEx(BYTE *BUS, BYTE, *DID, BYTE *REGISTER, LPCSTR *g_complex[gpu_id].m_dev) I believe

There is also ReadI2CEx(...)

the Ex functions are to access any GPU device, the non-Ex are only for primary GPU.

---

yes i know and use this functions.

have you tried it on your gtx280 under 64bit?

Neh. It's only 9.30am here :D I'll grab it now and take a look.

Ok first thing to fix.

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Unwinder\R ivaTuner\Monitoring\Plugins]
"VT1165.dll"=dword:00000001

Reg key needs to be like that for 64bit OS.

So just make an extra registry key called vt1165_x64.reg and tell people with 64bit OS to run that.

32bit registry keys are stored in there.

And it loads once I do that but throws an exception about unreferenced memory address and takes out Rivatuner.

When I restart RT it seems to load ok, but it has no values in your message box and they are wrong!

GPU=0, VID=CBh, Voltage=0.000v...

It also randomly selects VID register offsets heh. It needs to be linked against newer platform sdk and i'd say pointers aren't being handled correctly for some functions. Shoot me the source and I'll see if I can find the problem and fix it.

Works fine in Vista x64 here.

Quote:

---

Originally Posted by Solus Corvus

Works fine in Vista x64 here.

---

thanks a lot, for confirmation, solus!

Mind you I'm booted on Windows 7 x64 at the moment, and don't really want to boot back to Vista SP2 unnecessarily.

Works great for me vista x64 CF 4870's
Thanks a lot for the post...hehe time to go clock them up now :D

Edit: Can you also give more instructions on how to load the settings on start-up.

XP Pro, SP3.

Does not work on my Asus 4870 1 GB Dark Knight card. Gives me the red error message up above.

Quote:

---

Originally Posted by fgw

what card are you using?

---

GTX260 216SP Gainward,Vista Ultimate 64-bit

Quote:

---

Originally Posted by DMH

GTX260 216SP Gainward,Vista Ultimate 64-bit

---

I think that specidfic gainward card does not have the reference nvidia design, so it might not have the Volterra 1165 voltage regulator, that's why it isn't working, filho... :p:

Even tho' i don't have the hardware to take advantage of your plugin, i have to thank you big time for that very nice contrubution to the OC community. You sir represent the true XS spirit :up:

Again, thank you very much.

Very nice. Anyone with a 4870X2 checked whether OVP kicks in above 1,37V (as we know it does), or if the plugin can display that? Unfortunately I am away from my Mainrig right now.

Quote:

---

Originally Posted by mikeyakame

GPU=0, VID=CBh, Voltage=0.000v...

---

Same problem here.

Quote:

---

Originally Posted by jcool

Very nice. Anyone with a 4870X2 checked whether OVP kicks in above 1,37V (as we know it does), or if the plugin can display that? Unfortunately I am away from my Mainrig right now.

---

Well it says that OVP stays off all the way up to 1.5v (as far as I dared so far). I don't have a way to verify that the voltages are actually what they say they are or to check OVP so I only have what rivatuner reports at the moment.

Quote:

---

Originally Posted by Cold Fussion

Same problem here.

---

I can find and fix the problem but that's dependent on whether the author will release his source to me to make changes to!

It's hard to troubleshoot and fix problems that you can't replicate yourself I know that much from time programming.

Monitoring the current and power draw of the card/computer, we already know that it won't give more than 1,37V to the 4870X2's GPU core by software. So it just doesn't display it right, much like w/o the plugin.

wow man good work

:up::up::up::up::up:

Quote:

---

Originally Posted by Luka_Aveiro

I think that specidfic gainward card does not have the reference nvidia design, so it might not have the Volterra 1165 voltage regulator, that's why it isn't working, filho... :p:

---

No it does;)
It's straight reference,hell all GTX260/280/285 and 295 for sale are reference design,and it works with AwardFabrik and that other guy voltage vmod,just not with this plugin but hey I got up from 712/1512/1100 to 734/1566/1100 with 1,08V(0,02V over default):p:

Quote:

---

Originally Posted by DMH

No it does;)
It's straight reference,hell all GTX260/280/285 and 295 for sale are reference design,and it works with AwardFabrik and that other guy voltage vmod,just not with this plugin but hey I got up from 712/1512/1100 to 734/1566/1100 with 1,08V(0,02V over default):p:

---

Do me favor if you have Everest installed. Right click on bottom of window, goto Video Debug -> ATI SMBUS Dump, save it to a file and post it so I can take a look at it.

Would someone with a GTX 285 care to do the same as well for me.

Thanks.

:mad:

It doesn't work on my 4870 :censored::censored::censored:

Dam, prob cause its a non reference PCB :(

Tested on my BFG GTX 295 in Vista X64, working fine :up:

Quote:

---

Originally Posted by fatguy1992

:mad:

It doesn't work on my 4870 :censored::censored::censored:

Dam, prob cause its a non reference PCB :(

---

More than likely, most, if not all non-reference 4870's do not utilize the reference Voltera controller, as the majority of non-reference cards use analogue PWM.

Working great on Windows 7 x64 with EVGA GTX 295, THANKS!!!!
(181.22 Drivers)

thank you for this tool:up:

I wonder if anyone will ever see "1" appear as a OCP -state...
But atleast it is absolutely certain OVP will never, ever trigger and give "1" when using this plugin. Really, OVP monitoring has no purpose at all.

Guys I have a problem...

http://www.shrani.si/t/5/9L/AfQ5ltp/voltage-problem.jpg

Vista 64bit, and CF sapphire 4870 (one with blue, and one with red PCB), RivaTuner never asked me if I want to change only one or multiple GPU-s...

mikeyakame
have seen posts here were people are able to use the plugin successfully on w7 x64!

anyway, in your RivaTuner installdirectory, where you copied the plugin into \PlugIns\Monitoring there is a logfile VT1165.log.
close Hardware monitoring window, delete the logfile and restart Hardware monitoring again. try to change the voltage to get some information into the logfile, upload the logfile somewhere and post the link. pls don't post the logfile here directly as this would put a lot of mess in this thread! anyway, would be interesting to figure out whats going on there.

i also added a tread on my forum, you can upload the logfile to this thread so i have all the information in one place.
but as its a forum, you have to register in order to be able to upload ...

Cold Fussion
pls do the same and get me a logfile!

M-XXXX
looks like you are using the wrong plugin! look on the source setup window (the lowest window on your screenshot), you can find the name of the plugin at the top besides Data Provider. you are using the VT1103 plugin!
click on [Setup] in the graph window and then on the [Plugins] button. here you can uncheck VT1103.dll and check VT1165.dll instead!

regarding the discussion on OVP and OCP:

first of all, i still dont have a datasheet for the vt1165! all my knowledge is based on the information in various forums, mainly the posts from Unwinder and some private conversation with Unwinder.

so my understanding of this is:

OVP is triggered as soon as the output voltage exceeds the programmed voltage plus a certain, via resistors, predefined margin.
if its working this way, we will never see OVP kicking in when using the registers to change output voltage, as output voltage can hardly exceed the voltage we set by registers.

e.g. if the register defines 1.15V and the margin is set to 0.2V then OVP would kick in when output voltage exceeds 1.35V

if we now change the register to a reflect voltage of 1.30V and add our margin of 0.2V we get also a changed value of 1.50V where OVP would kick in! so, no matter what voltage is defined via the register, OVP would always kick in if output voltage exceeds this value by the margin, which simply would not happen.

this is totally different when using a hard mod: here the voltage set by register remains unchanged, thus the base for OVP. a hardmod forces the voltage increase by changing some resistors externally, thus output voltage plus margin is likely to exceed the unchanged predefined voltage in the register.
so using the example above, no matter what output voltage you set via a hardmod, the OVP reference level remains unchanged at 1.35V (1.15V + 0.2V) and thus let OVP kick in as soon as output voltage exceeds 1.35V.

OCP is a different story, as there is no register involved here. OCP will kick in when a certain current draw is exceeded! i have no idea at what point this will be. if my plugin works correctly we probably will see this sooner or later.

implementing OVP and OCP status in the plugin was no big deal and was, the above said in mind, targeted to hard mods mainly, tough OCP might be seen even when soft modding.

my vrm's die at anythign over stock on 4870 lols they go like 114 degrees then CRASH BANG WALLOP

Quote:

---

M-XXXX
looks like you are using the wrong plugin! look on the source setup window (the lowest window on your screenshot), you can find the name of the plugin at the top besides Data Provider. you are using the VT1103 plugin!
click on [Setup] in the graph window and then on the [Plugins] button. here you can uncheck VT1103.dll and check VT1165.dll instead!

---

Tnx, its working:D

Botton line:
Sapphire 4870 (one RED and one BLUE pcb) on vista 64bit, works like a charm:D

fgw,

Shall do that mate and get back to you.

I just finished reading and re-reading the Intersil ISL6327CRZ multiphase closed-loop feedback and compensation PWM Voltage Reg datasheet, the VR used by NV in the die shrink G200 reference cards, GTX 260 55nm and GTX 285 55nm.
Seems NV screwed you guys over with VR cost cutting this time around, mind you the IC itself only costs $5usd or so.

It's a VR designed with cost savings and affordability in mind so it lacks any kind of external low-pin count bus interface accessibility and relies on the pcb designer to set up all the necessary operational values in design and testing phase. There are 8 VID pins which when pulled up the base VID is read by which pins are 0 or 1 and converted to digital signal by DAC. '

Voltage adjustment is done through current offset via the OFS pin, which is polarity dependant for negative or positive offset, switch to Vss (gnd) for neg, switch to Vcc (vin) for pos. Offset is configured through external resistors connected between the DAC and REF pins, the impedance dynamically controls the current which in turn determines the Vout (voltage out).

Long story short, unless NV have some kind of low pin count programmable interface to adjust a pull up or pull down resistor connected between DAC digital VID signal pin and REF voltage reference pin there is absolutely NO CHANCE AT ALL to control Voltage Adjustments via software on either the GTX 260 55nm or GTX 285 55nm. The only way would be to change the resistor between DAC and REF pins, to give a greater current offset to drive Voltage out higher.

If EVGA is saying it is possible on their cards, short of physical changes to the NV reference cards to add this basic functionality I can't see how it is possible at all. I read over the datasheet a good 4 or 5 times just to make sure I completely understood the electrical design and don't see any way to do it.

fgw,
My point about OCP was that if it triggers, Rivatuner will not catch it because once OCP activates, the PWM is immediately clamped or reversed to a safe mode -> crash -> soft reset.

OVP should have 2 scenarios for taking place.

1) If Vout reaches OVP Cut Off VID, which is probably between 1.5-2.0v.
2) If Vout - VID > Vout_overshoot offset.

If only controlling VID through I2C bus registers then scenario 1 OVP is the only case one will encounter.
If controlling VID through resistor impedance for hardward mod, scenario 2 OVP is likely if the resistance pushes Vout past overshoot offset from programmed VID.

If using combination of hardware trickery mod and software programmed VID, then scenario 2 should be avoidable if you can determine the offset allowed for Vovershoot.

Quote:

---

Originally Posted by mikeyakame

1) If Vout reaches OVP Cut Off VID (...)

---

How do you know an OVP of this type exists?

Well I've seen similar in other VR datasheets, and like anything that regulates voltage and current there is always a safety cut off point incase of malfunction. It would be in this case either an OCP or OVP hard set value just incase something goes terribly wrong and Vout overshoots uncontrollably, a hard OVP limit controlled by something like a closed loop slew rate feedback circuit would catch this hopefully before electricals get fried.

Tried in Vista64 and XP does not show any voltages for card, or increase voltage. Single GTX 260:(

http://img.photobucket.com/albums/v2...h_RTPlugin.jpg

mikeyakame,
I haven't seen a single VID controller that has a fixed OVP. And I can't see how it would even make sense to have VIDs that are higher than OVP trigger voltage?

i got a green pcb 4870 1gb card with black cover and heathpipe system.
"unable to detect VT1165 ...."

Quote:

---

Originally Posted by Hardass

Tried in Vista64 and XP does not show any voltages for card, or increase voltage. Single GTX 260:(

http://img.photobucket.com/albums/v2...h_RTPlugin.jpg

---

I have exactly the same problem on a SAPPRIHE 4870 512MB. Any one know how to fix it?

Quote:

---

Originally Posted by largon

mikeyakame,
I haven't seen a single VID controller that has a fixed OVP. And I can't see how it would even make sense to have VIDs that are higher than OVP trigger voltage?

---

Intersil ISC6327 VR used on 260/285 shrinks uses this kind of OVP as safety. It has a preset of 2.0v if using recommended resistance, but it is upto the vendor who implements it to set it differently if they want. It also uses offset value above VID of +125mV as soft OVP with recommended design.

Quote:

---

Originally Posted by sch3ll3p3ll3

I have exactly the same problem on a SAPPRIHE 4870 512MB. Any one know how to fix it?

---

I'm helping fgw look into its since I have same problem. He hasn't got a 64bit OS on his system to troubleshoot unfortunately. But rest assured its being looked into and fixed ;)

Once its fixed there will be an updated dll available, things like this happen when API libraries get updated and small changes break previous usage :)

Tried this on my EVGA gtx260 55nm....didnt work....

Time to break out the solder iron. :up:

Quote:

---

Originally Posted by mikeyakame

I'm helping fgw look into its since I have same problem. He hasn't got a 64bit OS on his system to troubleshoot unfortunately. But rest assured its being looked into and fixed ;)

Once its fixed there will be an updated dll available :)

---

Thanks, but mine does not work on XP 32bit either.:confused:

Quote:

---

Originally Posted by [XC] Hicks121

Tried this on my EVGA gtx260 55nm....didnt work....

Time to break out the solder iron. :up:

---

Even though it's been said on the current thread page and many times in the past...

260 55nm and 285 55nm shrink were fitted with a cheaper 4-phase InterSil VR that doesn't have programmability like the Volterra VT1000 series. They saved NV money and cost the customers full software programmable voltage adjustment.

There is a soft OVP in the datasheet mentioned of 125mV or 0.125v, whether NV stuck to this design value or used a larger VID offset for OVP it's unknown. Let us know if you reach OVP and at what point if using an adjustable pot.

Quote:

---

Originally Posted by Hardass

Thanks, but mine does not work on XP 32bit either.:confused:

---

In Rivatuner 2.22\Monitoring\Plugins, you will find a VT1165.log file, paste me the contents in a private message if you like so I can see the debug info.

Quote:

---

Originally Posted by mikeyakame

In Rivatuner 2.22\Monitoring\Plugins, you will find a VT1165.log file, paste me the contents in a private message if you like so I can see the debug info.

---

Sent. And Thanks.:up:

Quote:

---

Originally Posted by mikeyakame

Intersil ISC6327 VR used on 260/285 shrinks uses this kind of OVP as safety. It has a preset of 2.0v if using recommended resistance, but it is upto the vendor who implements it to set it differently if they want. It also uses offset value above VID of +125mV as soft OVP with recommended design.

---

I can't seem to find any reference to such 2.0V limit.
But even if there was such a limit, it wouldn't matter as max VID is 1.6V.
And I don't even see the point such 2.0V limit as the offset voltage trigger (VID+175mV) will always cause OVP event before the 2.0V trigger.

guys is it possible to change GFX card RAM voltage with this software

Quote:

---

Originally Posted by largon

I can't seem to find any reference to such 2.0V limit.
But even if there was such a limit, it wouldn't matter as max VID is 1.6V.
And I don't even see the point such 2.0V limit as the offset voltage trigger (VID+175mV) will always cause OVP event before the 2.0V trigger.

---

IOUT - IOUT has the same output as IDROOP with
additional OCP adjustment function. In actual application, a
resistor needs to be placed between IOUT and GND to
ensure the proper operation. The voltage at IOUT pin will be
proportional to the load current. If the voltage is higher than
2V, ISL6327 will go into the OCP mode, this means it will
shut down first and then hiccup. The additional OCP trip
level can be adjusted by changing the resistor value

It's a form of OCP based OVP.

Quote:

---

Originally Posted by dinos22

guys is it possible to change RAM voltage on these

---

Don't believe so mate. Old fashioned way of increasing Vddr seems the most plausible method.

I'm pretty certain Unwinder has said there is no programmable interface to GDDR VR that he knows of.

Thank you for sharing the great tool :up:

Quote:

---

Originally Posted by Hardass

Tried in Vista64 and XP does not show any voltages for card, or increase voltage. Single GTX 260:(

http://img.photobucket.com/albums/v2...h_RTPlugin.jpg

---

me niether :(

is there not a way of getting this to run on x64 (need idiots guide)

Quote:

---

Originally Posted by mikeyakame

IOUT - IOUT has the same output as IDROOP with
additional OCP adjustment function. In actual application, a
resistor needs to be placed between IOUT and GND to
ensure the proper operation. The voltage at IOUT pin will be
proportional to the load current. If the voltage is higher than
2V, ISL6327 will go into the OCP mode, this means it will
shut down first and then hiccup. The additional OCP trip
level can be adjusted by changing the resistor value

It's a form of OCP based OVP.

---

That is OCP. Absolutely nothing to do with OVP, mosfet output voltage, sensing nor limiting it.

Quote:

---

Originally Posted by mikeyakame

Do me favor if you have Everest installed.
...
Would someone with a GTX 285 care to do the same as well for me.

Thanks.

---

Here is the dump file from GTX 285. If you can do something you will be the best.

Quote:

---

Originally Posted by mikeyakame

Even though it's been said on the current thread page and many times in the past...

260 55nm and 285 55nm shrink were fitted with a cheaper 4-phase InterSil VR that doesn't have programmability like the Volterra VT1000 series. They saved NV money and cost the customers full software programmable voltage adjustment.

There is a soft OVP in the datasheet mentioned of 125mV or 0.125v, whether NV stuck to this design value or used a larger VID offset for OVP it's unknown. Let us know if you reach OVP and at what point if using an adjustable pot.

---

1.26vcore - os boot
1.27vcore - no os boot

http://img244.imageshack.us/img244/2594/vmodph8.jpg

most im able to get via the vid pins is high 1.25vcore low 1vcore

Quote:

---

Originally Posted by mikeyakame

fgw,

Shall do that mate and get back to you.

I just finished reading and re-reading the Intersil ISL6327CRZ multiphase closed-loop feedback and compensation PWM Voltage Reg datasheet, the VR used by NV in the die shrink G200 reference cards, GTX 260 55nm and GTX 285 55nm.
Seems NV screwed you guys over with VR cost cutting this time around, mind you the IC itself only costs $5usd or so.

It's a VR designed with cost savings and affordability in mind so it lacks any kind of external low-pin count bus interface accessibility and relies on the pcb designer to set up all the necessary operational values in design and testing phase. There are 8 VID pins which when pulled up the base VID is read by which pins are 0 or 1 and converted to digital signal by DAC. '

Voltage adjustment is done through current offset via the OFS pin, which is polarity dependant for negative or positive offset, switch to Vss (gnd) for neg, switch to Vcc (vin) for pos. Offset is configured through external resistors connected between the DAC and REF pins, the impedance dynamically controls the current which in turn determines the Vout (voltage out).

Long story short, unless NV have some kind of low pin count programmable interface to adjust a pull up or pull down resistor connected between DAC digital VID signal pin and REF voltage reference pin there is absolutely NO CHANCE AT ALL to control Voltage Adjustments via software on either the GTX 260 55nm or GTX 285 55nm. The only way would be to change the resistor between DAC and REF pins, to give a greater current offset to drive Voltage out higher.

If EVGA is saying it is possible on their cards, short of physical changes to the NV reference cards to add this basic functionality I can't see how it is possible at all. I read over the datasheet a good 4 or 5 times just to make sure I completely understood the electrical design and don't see any way to do it.

---

anyone pin point out the exact resistor?

ill hunt it down

alright what resistance would that resistor be ?

1k ?
26.6k ?

^^^^And thus this topic has officially gone WAY over my head. :rofl: :ROTF:

Quote:

---

Originally Posted by dinos22

guys is it possible to change GFX card RAM voltage with this software

---

afaik, there is no way to change RAM voltage by software. voltage for RAM is at least not controlled by the vt1165!

to everybody out there having problems with my plugin on single gpu setups:

i developed this software on a system with a gtx295 and think i'm handling single gpu setups incorrectly!

give me a short break to verify this. if my assumptions are correct i might come out with a fix by tomorrow!

WTG m8 ;)

pleas epleas eplease can you do a x64 fix please :D

This is pretty awsome! Thanks for the massive effort here bro. I'm not having any luck with this on my 4870 either but I'm sure you'll get figured out soon.

Quote:

---

Originally Posted by mikeyakame

fgw,

Shall do that mate and get back to you.

I just finished reading and re-reading the Intersil ISL6327CRZ multiphase closed-loop feedback and compensation PWM Voltage Reg datasheet, the VR used by NV in the die shrink G200 reference cards, GTX 260 55nm and GTX 285 55nm.
Seems NV screwed you guys over with VR cost cutting this time around, mind you the IC itself only costs $5usd or so.

It's a VR designed with cost savings and affordability in mind so it lacks any kind of external low-pin count bus interface accessibility and relies on the pcb designer to set up all the necessary operational values in design and testing phase. There are 8 VID pins which when pulled up the base VID is read by which pins are 0 or 1 and converted to digital signal by DAC. '

Voltage adjustment is done through current offset via the OFS pin, which is polarity dependant for negative or positive offset, switch to Vss (gnd) for neg, switch to Vcc (vin) for pos. Offset is configured through external resistors connected between the DAC and REF pins, the impedance dynamically controls the current which in turn determines the Vout (voltage out).

Long story short, unless NV have some kind of low pin count programmable interface to adjust a pull up or pull down resistor connected between DAC digital VID signal pin and REF voltage reference pin there is absolutely NO CHANCE AT ALL to control Voltage Adjustments via software on either the GTX 260 55nm or GTX 285 55nm. The only way would be to change the resistor between DAC and REF pins, to give a greater current offset to drive Voltage out higher.

If EVGA is saying it is possible on their cards, short of physical changes to the NV reference cards to add this basic functionality I can't see how it is possible at all. I read over the datasheet a good 4 or 5 times just to make sure I completely understood the electrical design and don't see any way to do it.

---

lower resistance @ the 1kohm resistor dac/ref might lead to the back of the card lowers vcore.. higher resistance doesnt allow os boot @ 1.33vcore

suggestions?

Quote:

---

Originally Posted by swale

pleas epleas eplease can you do a x64 fix please :D

---

The plugin should work on w7 x64 and a gtx295. Are you using a gtx295?

fgw,

Once I finish writing up my helper code in C for managing the I2C device lists , transactions and debugging you can add it in if you want. Since no point both of us writing same plugin differently :)

mikeyakame,

I tried using that new evga voltage tuner, it was an early patched version that worked with all cards, turns out it was buggy and when I pressed the default settings button it put my gpu voltage to 2.0v..

I did notice straight away as I was running rivatuner's monitoring graph and restarted my pc.. It rebooted fine everything seems ok but do you think even a few seconds at 2.0v would have damaged the card's circuitry? I mean could it have caused physical damage which could get worse If I tried overclocking at all in the future? I knew I should have stuck to rivatuner..lol

Thats the problem with simple slider tools like EVGA are offering. Rivatuner method is a bit more work, but at least you know what its doing.

Personally I think that with this of feature a lot of checking needs to be done, and some user monitoring. It becomes easier when you see this to understand Unwinders decision to not include GUI functionality for such a feature. He honestly doesn't want users who don't know what are doing frying their expensive hardware, as he knows they will do it and will feel responsible because he made it easily available.

Ok so I got it working...but whats the max safe voltage for my gtx295???? for 24/7 stock cooling!

For the people with x64 where it hasn't came up automaticaly all I Done was followed the guide but I had to manualy allow the plugin by going into rivatuner hardware monitor, setup, plugins find vt1165.dll in the list and click the tick all will then become apparent.

to everybody who had problems with the plugin on single gpu systems:

uploaded a new fixed version of the plugin!

this version fixes a bug in my code which prevented the use of the plugin on systems with a single gpu!

everybody using a single gpu should download this fixed version!

download location is unchanged and can also be found in the first post of this thread!

pls report back!

working mate. ;)

i'll throw you a line once I finish up all the accessor structs, custom malloc handler, pci vendor id checks and all that.

It's working for me to (sapphire 4870 512MB & vista x64). Thank u for the effort.

It's working on my 4870 1gb now.

I also noticed if you want to change 3D voltage you have to put your card in 3D mode before you up the voltage. I used ATI tool artifact scanner to force 3D mode then up the voltage and after you exit 3D mode your voltage should return to default levels (1.2v 2D) while keeping your new voltage in 3D mode.

Working on single GTX 260 now. Thank you very much.:up:

FGW,

Great Program, Thanks for takeing the time to do all this.
:up::up::up::up::up::up:

Do you know how much extra voltage is considered ''safe---ish :)'' for 24/7 general usage with a bit extra uummpphh??? :D

Quote:

---

Originally Posted by Mysterfix

I also noticed if you want to change 3D voltage you have to put your card in 3D mode before you up the voltage. I used ATI tool artifact scanner to force 3D mode then up the voltage and after you exit 3D mode your voltage should return to default levels (1.2v 2D) while keeping your new voltage in 3D mode.

---

yes, thats the way it works.

it changes the voltage for the mode you are currently in!

good to mention it once more for everybody who missed this in my first post!

Quote:

---

Originally Posted by swale

Do you know how much extra voltage is considered ''safe---ish '' for 24/7 general usage with a bit extra uummpphh???

---

no, unfortunately i have no idea as nvidia is not supplying specs for their processors. so we are at our own and need to wait for experiences of other users.

The "more" option in the monitor is grey for me. I have a 4870 512 and Vista 64.

Quote:

---

Originally Posted by Nizzen

The "more" option in the monitor is grey for me. I have a 4870 512 and Vista 64.

---

screenshot please ...

Quote:

---

Originally Posted by Nizzen

The "more" option in the monitor is grey for me. I have a 4870 512 and Vista 64.

---

You may have to manually load the VT1165 plugin, I did. The VT1103 plugin will display some similar graphs but the more button is greyed out.

FGW,

While testing an overclock I'm getting a regular voltage spike according to your vt1165 plugin but rivatuner's vt1103 plugin is not detecting it, any ideas?

http://img12.imageshack.us/img12/494/spikesat9.th.jpg

Quote:

---

Originally Posted by M-XXXX

Tnx, its working:D

Botton line:
Sapphire 4870 (one RED and one BLUE pcb) on vista 64bit, works like a charm:D

---

Thanx! Then it is working :)

Remember to put this to the "hard" thing to the first page :up:

I'm using this card:

ASUS Radeon HD 4870 EAH4870 DK/HTDI/1GD5

And I'm getting the Unable to Detect VT1165 error. As far as I know this is a reference board other than the aftermarket cooler.

Any ideas?

Hum no still isn't working for me:(

OMG, I am soooo happy to see this!

will make it much easier to push the envelope with OCP/OVP!

No working for my Gainward 4870x2 R700 GS GLH
It says Data Provider VT1103.dll:0:0 instead of VT1165

Quote:

---

Originally Posted by Protocol_48

FGW,

While testing an overclock I'm getting a regular voltage spike according to your vt1165 plugin but rivatuner's vt1103 plugin is not detecting it, any idea

---

as i can see in the screenshot, you have both plugins (vt1103 and vt1165) active at the same time.

this is from RivaTuners V2.10 Revision history:

Please take a note that Volterra voltage regulators are rather sensitive to frequent polling and may return false data under heavy load, so it is not recommended to use VRM monitoring in daily monitoring sessions.

as regular polling of the vt1165 might already return false date, i can imagine, have it polled by two plugins wont help much! i assume this is simply one of this false results returned by the vt1165 under heavy load.

try to disable the vt1103 plugin and see if the spikes go away!

you can further reduce polling rate by also disabling all datasources supplied by the vt1165 plugin you are not really interested in.

interested in your results!

Quote:

---

Originally Posted by Telperion

I'm using this card:

ASUS Radeon HD 4870 EAH4870 DK/HTDI/1GD5

And I'm getting the Unable to Detect VT1165 error. As far as I know this is a reference board other than the aftermarket cooler.

Any ideas?

---

if you get this, you are out of luck! the plugin was not able to detect a vt1165! assume, this card is using a different controller ...



DMH

are you using the updated plugin? pls pm me your logfile: YOUR_RIVATUNER_INSTALL_DIRECTORY\PlugIns\Monitorin g\vt1165.log



initialised

go to setup -> plugins and disable the vt1103 plugin and enable the vt1165 plugin!