Overclock.net banner

1 - 20 of 54 Posts

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #1 (Edited)
Credit to R1fast on reddit for finding the correct resistors. I was able to find them on Mouser: Panasonic ERJ-M1WSF5M0U

It was a slightly messy soldering job but that is much better than getting the board too hot or getting solder where it shouldn't be. A small bead of solder was placed on each end of the original shunt resistor and then the new resistor was held in place with tweezers while heating one end at a time with the soldering iron until the bead under it melted. The original resistor did need to be pretty warm before the solder would stick to it.



It worked absolutely perfectly. Power use is reported as exactly 50% of what it used to with the same load. Real world performance improvements are minimal but the power limit never triggers anymore. :D
 

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #3
Thanks! This certainly feels a lot more permanent and precise. I also mount this GPU vertically so liquid metal would be even more dangerous than it usually is.

I did try with a silver pen first but it didn't have any impact on measured power use. :eek:

I was pleased with how easy it was too, it wasn't much more difficult than applying liquid metal and a barrier to protect the rest of the board. No worry about metal incompatibilities or it moving/dripping. I also still know my real power draw, simply double what is reported. :)
 

·
Registered
Joined
·
858 Posts
So is this just the exact same OEM resistor stacked ontop of the existing one? What is the reason for doing so if I may ask? Does this method have some parallel do-hicky magic to lower the power limit? I am confuzzled :thinking:
 

·
Packet Pusher
Joined
·
2,317 Posts
So is this just the exact same OEM resistor stacked ontop of the existing one? What is the reason for doing so if I may ask? Does this method have some parallel do-hicky magic to lower the power limit? I am confuzzled :thinking:
2 resistors of the same value attached in parallel gives you half of their individual resistance.

2 8 ohms in parallel = 4 ohms.

Math checks out as he's seeing the load at 50% less than before.
 

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #6
So is this just the exact same OEM resistor stacked ontop of the existing one? What is the reason for doing so if I may ask? Does this method have some parallel do-hicky magic to lower the power limit? I am confuzzled :thinking:
Exactly, as already mentioned adding another resistor in parallel halves the resistance. The GPU measures power use based on the voltage drop. Using V = IR half the resistance means half the voltage drop so the GPU thinks it is using half the power it really is. :D
 

·
Registered
Joined
·
858 Posts
Very cool, I just ordered a set of resistors and will solder them on during my next maintenance break down. Should help hitting the PL with my new 3440 x 1440 display. :thumb:
 

·
Not a Fan
Joined
·
2,648 Posts

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #9
Very nice, I am glad to see another proper power limit mod. You seem to have been able to solder them on with a bit less solder added too. :)

Did you have trouble getting the solder to stick to the original resistor? I had to turn up the temp on my soldering iron and heat it up a bit more than I wanted to before getting good adhesion. :eek:
 

·
Not a Fan
Joined
·
2,648 Posts
I tinned the new resistor, spread flux on the solder pads of the existing resistor, and soldered using some quad eutectic solder and a small chisel tip. I solder at 750 °F for most things, get in and out without staying too long on the part. I don't remember it being too tough, just nerves over trashing a $1200 card.

If I had it to do over, I'd use 2M0 instead, but it's under a water block in a full loop and I'm not pulling it out to change it.
 

·
Registered
Joined
·
858 Posts
I tinned the new resistor, spread flux on the solder pads of the existing resistor, and soldered using some quad eutectic solder and a small chisel tip. I solder at 750 °F for most things, get in and out without staying too long on the part. I don't remember it being too tough, just nerves over trashing a $1200 card.

If I had it to do over, I'd use 2M0 instead, but it's under a water block in a full loop and I'm not pulling it out to change it.

What's the reasoning for 2M0?
Edit - ordered the 5M0 resistors last week and they just arrived :)
 

·
Not a Fan
Joined
·
2,648 Posts

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #13
Did you test power use on a repeatable workflow before the mod? Are you sure you got good solder connections everywhere? I can see that being an issue when not tinning the original resistor given how hot I needed to get it before the solder would actually stick/flow correctly.

I ask because you should be able to use twice the power before hitting the limit, I measured exactly that. Nothing I can run on it has triggered the power limit even for an instant, its max value in afterburner stays at 0. I have not and will not test something like Furmark.
 

·
Not a Fan
Joined
·
2,648 Posts
I didn't do any scientific before/after. I ran some benches and compared scores before/after, the after were definitely higher.
 

·
Overclocker in training
Joined
·
13,946 Posts

Attachments

·
9 Cans of Ravioli
Joined
·
19,816 Posts
Hi,
For the record you guys are talking about these objects ?
They look different than the link in the op pointed to in the link
The things with 5M0 printed on them are the shunts you need to "mod" (either with LM or solder).





IIRC - the bottom most one can be skipped as it controls power through the PCI-E slot and you don't want it taking more than 75w from the slot. But don't quote me on that.

I did the shunt mod for my 1060 but took it off after a few weeks because it didn't do much for performance and there are some horror stories about LM eating away the shunts and them falling off.
 

·
Overclocker in training
Joined
·
13,946 Posts

·
Meep
Joined
·
5,592 Posts
ordered the 5M0 resistors last week and they just arrived :)
If you do end up wanting more current, but don't want to wait on different resistors. You can stack 2 to get effectively a 2.5mΩ resistor.

You might want to make sure it still clears your cooler/waterblock if you do this tho.

Edit: post was a little older than I thought, probably not that useful of a reply, I'll leave it as it might be useful to somebody else.
 

·
Robotic Chemist
Joined
·
3,010 Posts
Discussion Starter #19
Okay yeah it was confusing see the image from this
https://www.mouser.com/ProductDetai...=sGAEpiMZZMtlleCFQhR/zWEAYIYjGcyDOAOWVexWG0k=

And nothing on the card show shows it lol :confused:
Yes, that was confusing to me too. The image is from one of the tiny versions of this resistor from the same line, the giant ones I ordered look exactly like the ones on the card, as you can see in my photos. I did order from that link with the wrong image. I almost ordered other ones that had a more similar image but I am glad I didn't.

They do say "Images are for reference only See Product Specifications" and they mean don't use the image for reference beyond what type of component this is. :p

If you do end up wanting more current, but don't want to wait on different resistors. You can stack 2 to get effectively a 2.5mΩ resistor.
I actually bought 2mΩ resistors as well, which is about as low as you can go without triggering a low power safety check, but decided to keep it at a limit of double the power in case someone did run something like furmark on it.
 

·
Overclocker in training
Joined
·
13,946 Posts
Hi,
Thank you both for clearing up my confusion
Interesting process for sure would these be okay for a 1080ti ftw3 :)
 
1 - 20 of 54 Posts
Top