Jump to content

980M has one less VRM than any previous nVidia MXM card back to the 260


Ethrem

Recommended Posts

Anyone else think this could cause failures when messing with voltage? The solder point is there but it has no VRM while the 970M has all 4.

http://rjtech.com/shop/images/detailed/DSC01680.jpg

Kind of worrying for me with all this sudden lock down of overclocking and such. Why would they remove a VRM?

Sent from my Nexus 6 using Tapatalk

Link to comment
Share on other sites

Anyone else think this could cause failures when messing with voltage? The solder point is there but it has no VRM while the 970M has all 4.

http://rjtech.com/shop/images/detailed/DSC01680.jpg

Kind of worrying for me with all this sudden lock down of overclocking and such. Why would they remove a VRM?

Sent from my Nexus 6 using Tapatalk

You Wana no what's funny. When I ordered my 980m from hidevolution and opened the card from it package I looked inside the package because I thought some of the chips fell off because there were blank spots on the card were a chip would go that weren't like that on my other cards.lol

Link to comment
Share on other sites

The chips you are referring to are high-side and low-side power FET combo chips. Usually one chip is just a low side-FET or a high-side FET, but Nvidia uses chips with a high-side and low-side combined on one chip. These are involved with voltage regulation, but they are not the VRM. The VRM is located on the back side of the card. Power FETs are responsible for powering the core, while the VRM controls the power FETs. All current to the core travels through the power FETs, which in the case of the 980m are 3 combo chips.

Now that that's cleared up...

The 980m has 3 power phases while the 680m, 780m, 880m, and 970m all have 2. On the 980m each phase has just one chip, while older cards and the 970m have 2 on each phase. This does mean that each chip must pass more current, but overall there is less strain on the voltage regulation. There are 2 reasons for this. One is there is a 3rd inductor. Inductors dump a lot of heat out onto the card because they are fairly resistive, and having a 3rd to split the current load reduces heat output significantly. Other reason why the 980m is better is because having 2 chips in parallel is not a purely positive thing. The 2nd chip on the phase means that the VRM must push and pull twice as much charge to turn the FETs in the chip on and off. You want to switch back and forth between high-side being on, and low-side being on, and vice-verse as fast as possible. The longer it takes means lower efficiency and more heat put out. The 2nd chip per phase slows down the switching speed.

So combining the positives of the 3rd inductor and faster switching speeds, the 980m seems to me to have better power delivery than the 970m or earlier cards.

  • Thumbs Up 13
Link to comment
Share on other sites

The chips you are referring to are high-side and low-side power FET combo chips. Usually one chip is just a low side-FET or a high-side FET, but Nvidia uses chips with a high-side and low-side combined on one chip. These are involved with voltage regulation, but they are not the VRM. The VRM is located on the back side of the card. Power FETs are responsible for powering the core, while the VRM controls the power FETs. All current to the core travels through the power FETs, which in the case of the 980m are 3 combo chips.

Now that that's cleared up...

The 980m has 3 power phases while the 680m, 780m, 880m, and 970m all have 2. On the 980m each phase has just one chip, while older cards and the 970m have 2 on each phase. This does mean that each chip must pass more current, but overall there is less strain on the voltage regulation. There are 2 reasons for this. One is there is a 3rd inductor. Inductors dump a lot of heat out onto the card because they are fairly resistive, and having a 3rd to split the current load reduces heat output significantly. Other reason why the 980m is better is because having 2 chips in parallel is not a purely positive thing. The 2nd chip on the phase means that the VRM must push and pull twice as much charge to turn the FETs in the chip on and off. You want to switch back and forth between high-side being on, and low-side being on, and vice-verse as fast as possible. The longer it takes means lower efficiency and more heat put out. The 2nd chip per phase slows down the switching speed.

So combining the positives of the 3rd inductor and faster switching speeds, the 980m seems to me to have better power delivery than the 970m or earlier cards.

Cheers Khenglish, that's very enlightening. How did you find out / determine all of that about the power delivery systems of the new Maxwell cards?

Looks like a lack of adequate power supply is not the reason for the ban on overclocking - because the power delivery systems of the new Maxwell cards are as good (970M) or better (980M) than the previous Kepler cards.

Link to comment
Share on other sites

Cheers Khenglish, that's very enlightening. How did you find out / determine all of that about the power delivery systems of the new Maxwell cards?

Looks like a lack of adequate power supply is not the reason for the ban on overclocking - because the power delivery systems of the new Maxwell cards are as good (970M) or better (980M) than the previous Kepler cards.

Power phases take up a large amount of pcb area and are very consistent between boards, so they're easy to spot from a picture. 1 inductor = 1 power phase.

Link to comment
Share on other sites

Power phases take up a large amount of pcb area and are very consistent between boards, so they're easy to spot from a picture. 1 inductor = 1 power phase.

Ah, ok, cool, I see what you've done, you've combined your knowledge of inductor and FET Combo chip arrangement with your observation that there is an extra inductor on the 980M card and a more efficient arrangement of FET Combo chips to marry with those inductors. You've used your engineering knowledge to come to that conclusion. (I think someone on NBR told me you're studying electrical engineering.)

Link to comment
Share on other sites

  • 4 weeks later...
After testing, adding in the VRMs does cut the card from dying at 1.2v with a memory over volt but is only required at the utter extreme end of benchmarking.

cool, so it is a fet issue and not an inductor one. hmm maybe I should put in stronger high-side fets on my 7970m... vid goes up to 1.225v but it crashes far below that

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.