VRM temperatures not showing.

[Brim]

Hello,

I am new here and i was told to use the GPUz to monitor my GPU as i am getting set to overclock it with a universal GPU water block and just a fan to blow on the VRM and memory to cool it. I am not using the NZXT G10 but the case i am using has a adjustable 140mm fan blowing on my GPU at the moment and the GPU cam with a heatsink on the VRM with the stock cooler. It is a Galaxy 660Ti GC.

I wanted to record a before and after temperature recording so i can know how much of an OC i can do without destroying my GPU. GPUz at the moment only shows the VDDC voltage and stops there. Is this fixable?

 

[buildzoid]

GPUz is probably not able to show more because the card doesn't have more sensors. But the GTX 660Ti is so low power you shouldn't worry plus you can't push more than 1.2V without modding the BIOS.

 

[Brim]

GPUz is probably not able to show more because the card doesn't have more sensors. But the GTX 650Ti is so low power you shouldn't worry plus you can't push more than 1.2V without modding the BIOS.

I do have the BIOS modded. I have the GPU frequency at 1267MHz, Voltage at 1.212Mhz and Memory at 7012Mhz and the temperature is no higher than 64C at ultra in CoD Ghosts. I can mod the BIOS even higher and i want to do so but i need more cooling performance and that is why i want to put on a universal GPU block as my card does not have a full cover block for it.

EDIT: here is proof. This is after 4 hours of gaming....

 

[buildzoid]

The GTX 660Ti pulls about 150W@1.175v at the connector so the VRM output for it should be 127A(with efficiency of 90%) by default. The VRM on your card seems to be a 6+1phase design with about 180A(or more) max out put so you would need around 1.6V at stock clocks to burn it out. So really you shouldn't worry much about it because at maximum power output that VRM will probably be throwing of about 32W assuming 90% efficiency. However I do recommend putting some heat sinks on the MOSFETS once you get a universal block just as a precaution and to marginally improve efficiency and power quality.

 

[Brim]

The GTX 660Ti pulls about 150W@1.175v at the connector so the VRM output for it should be 127A(with efficiency of 90%) by default. The VRM on your card seems to be a 6+1phase design with about 180A(or more) max out put so you would need around 1.6V at stock clocks to burn it out. So really you shouldn't worry much about it because at maximum power output that VRM will probably be throwing of about 32W assuming 90% efficiency. However I do recommend putting some heat sinks on the MOSFETS once you get a universal block just as a precaution and to marginally improve efficiency and power quality.

This is my exact card....http://www.galaxytechus.com/__US__/Product6/ProductDetail?proID=8&isStop=0 and it came with a long heatsink on the VRM. I plan on removing the stock air cooler and implement an EK universal GPU water block to cool the GPU and i will have a 140mm fan in my case blowing on the GPU to blow some air on it. I have the NZXT P630 and if you know the case you know where that fan is located. What i wanted to know or make sure is that i don't cross the "safe" zone for the GPUs VRM when i remove the stock cooler and implement the EK block.

 

[buildzoid]

This is my exact card....http://www.galaxytechus.com/__US__/Product6/ProductDetail?proID=8&isStop=0 and it came with a long heatsink on the VRM. I plan on removing the stock air cooler and implement an EK universal GPU water block to cool the GPU and i will have a 140mm fan in my case blowing on the GPU to blow some air on it. I have the NZXT P630 and if you know the case you know where that fan is located. What i wanted to know or make sure is that i don't cross the "safe" zone for the GPUs VRM when i remove the stock cooler and implement the EK block.

I referenced the PCB for that card so the numbers are sound. Also if the VRM already has a heatsink then just point a fan at it

 

[Brim]

I referenced the PCB for that card so the numbers are sound. Also if the VRM already has a heatsink then just point a fan at it

and this is my plan but i read an article on pcper about the NZXT G10 fan bracket for water cooling with certain AIO water coolers and even though the GPU temperatures were reduced by about 40C the vrm temperatures rose slightly so i just wanted to make sure i don't fry my GPU.

Thanks for the fast replies though.

 

[Brim]

I referenced the PCB for that card so the numbers are sound. Also if the VRM already has a heatsink then just point a fan at it

Got another question for you Sir. Does GPUz support kepler GPUs namely mine for reading the sensors?

 

[buildzoid]

Got another question for you Sir. Does GPUz support kepler GPUs namely mine for reading the sensors?

It will definetly support core temp frequencies and fan speed but voltages and VRM temps are dependant on the sensor choice of the manufacturer.

 

[Brim]

It will definetly support core temp frequencies and fan speed but voltages and VRM temps are dependant on the sensor choice of the manufacturer.

Well it showed the voltage but not the temperature of the vrm so in my understanding, it may not be a problem with GPUz or the GPU sensors but could be the Nvidia driver or some configuration in windows itself that may be causing the problem?

This has me very puzzled and i don't think i want to remove the stock cooler for the water block just yet. I really want to understand what is going on before i do. Galaxytech is no help as they do not answer e-mails from outside US & Canada or maybe they just do not like my e-mail so i doubt i can get help from them.

I used this...http://www.hwinfo.com/download.php and it to had no VRM tempertures but showed the voltage and every thing else for the GPU.....

 

[sneekypeet]

They have IR thermometers on Amazon for under $20 that will get you a good enough reading to solve the question of VRM temps.

 

[Brim]

They have IR thermometers on Amazon for under $20 that will get you a good enough reading to solve the question of VRM temps.

Yes i know about those but i would like to solve this problem before i resort to that.

 

[buildzoid]

VRM temp readings are on a separate sensor from voltage and many gpus don't have them because most VRMs can run fine at 100+C.
BTW I run a core only water cooled gtx 590 with no VRM heatsinks just a fan and it's perfectly ok.

 

[Brim]

VRM temp readings are on a separate sensor from voltage and many gpus don't have them because most VRMs can run fine at 100+C.
BTW I run a core only water cooled gtx 590 with no VRM heatsinks just a fan and it's perfectly ok.

This is very reassuring coming from you. I have gotten this same response by the fine people, namely Joe, at frozencpu.com where i get all my water cooling parts from and other experienced water cooling "gurus". So i will just go ahead with my plans.

Thanks @buildzoid and all the others for your help and views.