Hello all,

I have a GT1030 (GDDR5 version)

How much GPU usage %, you get when gaming.
Use MSI afterburner for the collection of statistics.

 

I get 99% on the core and 95-100% video memory usage. The memory controller is also almost always at 70% or greater load due to having to constantly swap textures to and from the low amount of vram. Temperatures never exceed ~70C.

BTW I would recommend HWInfo64 for in game stats, it has much more info than MSI afterburner and it can also hook into RTSS. (Also lower CPU/memory overhead in my experience)

 

I get 99% on the core and 95-100% video memory usage. The memory controller is also almost always at 70% or greater load due to having to constantly swap textures to and from the low amount of vram. Temperatures never exceed ~70C.

You will not like the news, but shunt-mod this is not for you, +3.5% at FPS cannot help anyone, when GPU is all ready at 100%.
The solution to your problem, this is my GPU on your system, and I have no intention to share it.
I am aware that you are determined to push your luck, but its pointless, there is no benefit.
You might burn your motherboard ....
What game are you playing? What are the minimum VGA specifications ?

 

accidentally replied twice

 

wouldn't shunt mod a card without 6/8-pin power. people avoid the bottom shunt by the PCIe slot on cards with 6/8-pin power for a reason.

not that long ago AMD had to release a fix for RX 400 series drawing >75W from PCIe and causing problems.

 

wouldn't shunt mod a card without 6/8-pin power. people avoid the bottom shunt by the PCIe slot on cards with 6/8-pin power for a reason.

not that long ago AMD had to release a fix for RX 400 series drawing >75W from PCIe and causing problems.

I understand the inherent risk in a shunt mod for the PCI-E slot, however the cards stock power limit is 30W, so even if I doubled the power limit with a shunt mod I would still be well within the 75W PCI-E safety limit.
I know that GPUs still can spike above their power limit for a millisecond or two, so that would be my only concern about exceeding 75W, however I don’t plan to push the power usage much farther than 45-55W based on the cards current thermal performance, so I don’t anticipate that even a power usage spike would be dangerous here, as I would likely run into thermal issues far before any sort of Power overdraw event could occur.

 

Shunt resistors are actually in the range of milliohms because they need to pass a lot of current with minimal voltage drop. Judging from this thread and others, I think the BIOS power limit is enforced through a different method than shunt resistors and the white SMD is actually a 10A fast blow fuse to minimize damage to the PCIe slot/mobo in case of a short. You can check by removing that SMD and cleaning the pads with solder wick to determine if there are multiple pads.

 

Shunt resistors are actually in the range of milliohms because they need to pass a lot of current with minimal voltage drop. Judging from this thread and others, I think the BIOS power limit is enforced through a different method than shunt resistors and the white SMD is actually a 10A fast blow fuse to minimize damage to the PCIe slot/mobo in case of a short. You can check by removing that SMD and cleaning the pads with solder wick to determine if there are multiple pads.

I have read that thread, however the PCB shown there has quite a few differences from my PCB. It looks to be a DDR4 variant, which is a few years newer than my GDDR5 (2017) variant.
Judging from that forum post, it doesn’t seem like anyone is quite sure what controls the power limit for these cards, except perhaps the people on HWbot who are running the gpu at 2700mhz as thats just not possible on 30W without maybe removing/disabling the vram.
Yeah I figured that 10 marked component wasn’t it as none of BuildZoids videos show any shunt resistors as white. I may reach out to him to ask

 

I suggest tweeting at buildzoid, he likely will reply

 

Not that I’m against making a Twitter account, but do you know if he have an active account on here that I can reach out to?

 

I've gone through the PCB and marked up all the resistors on the board:
Front: One marked in Red is 1.8 ohms (Code = 1R8)

Back: The one in green is 15ohm (Code = 150) and ones in blue are 2.2 ohm (Code = 2R2)

Is there any way to test via a multimeter if these resistors are used to calculate power draw?

 

Even if you didn't physically desolder the component and check to see if it is connected to multiple pads like my first comment said, I believe a multimeter can identify if the component is a shunt resistor by probing out what it is connected to. You can then use your multimeter + the pinout and/or functional diagram of your GPU's switching controller to determine which passive components are hooked up to the current monitoring sense pins and compare your results to see if the suspected shunt resistor is connected to the current monitoring circuitry. Having a full datasheet of the controller and the boardview of your PCB makes this arduous task vastly easier; but reverse engineering of the PCB layout is technically possible if you have lots of time and patience.

Be warned that without the pinout/functional diagram telling you what pin you are probing, you will need to start altering the characteristics of the passive components (attaching parallel resistances, replacing resistors with potentiometers for small value adjustments, etc) and looking for changes in voltage/current draw/card behaviour, etc to tell you how the component influences the switching controller. It goes without saying that physically modifying your card is always risky; but this haphazard reverse engineering of the switching controller is at extreme risk of killing something before you even figure out what stuff has to be modified for the current sense circuit to register lower than actual current draw!

In any case, what is the switching controller for your GPU? Posting a clearer picture of this component might help us identify what company makes the controller and potentially find a datasheet/pinout for it.

 

Even if you didn't physically desolder the component and check to see if it is connected to multiple pads like my first comment said, I believe a multimeter can identify if the component is a shunt resistor by probing out what it is connected to. You can then use your multimeter + the pinout and/or functional diagram of your GPU's switching controller to determine which passive components are hooked up to the current monitoring sense pins and compare your results to see if the suspected shunt resistor is connected to the current monitoring circuitry. Having a full datasheet of the controller and the boardview of your PCB makes this arduous task vastly easier; but reverse engineering of the PCB layout is technically possible if you have lots of time and patience.

Be warned that without the pinout/functional diagram telling you what pin you are probing, you will need to start altering the characteristics of the passive components (attaching parallel resistances, replacing resistors with potentiometers for small value adjustments, etc) and looking for changes in voltage/current draw/card behaviour, etc to tell you how the component influences the switching controller. It goes without saying that physically modifying your card is always risky; but this haphazard reverse engineering of the switching controller is at extreme risk of killing something before you even figure out what stuff has to be modified for the current sense circuit to register lower than actual current draw!

In any case, what is the switching controller for your GPU? Posting a clearer picture of this component might help us identify what company makes the controller and potentially find a datasheet/pinout for it.

Thank you for the well thought out answer. Unfortunately the component is too small for any of my cameras to focus on, however this is what is written on the chip:
Top Line: vP16660 (Possibly uP16660)
Bottom Line: PGA69T

Judging from google search, PGA69T is the name of the socket that the chip is attached to, but I could be wrong.

A quick google search shows that uP16660 is part of a family of GPU power controllers beginning with uP1666. Here is the data sheet for a similarly named model:

http://www.gdcy.com/download.php?id=2190767&pdfid=878A5A0918BEEAECEE88A4126A3C7D20&file=0596\up1666qqkf_9057604.pdf

A visual comparison shows that the two models at least have the same pin count: 20 with 5 on each side. That’s all I’ve gathered from quickly looking at the data sheet.

Edit: Upon closer inspection, it looks like the top line actually says uP1666Q, which matches up perfectly with the chip family in the datasheet.

Edit 2: Judging by this post on R/Overclocking, this chip runs the memory controller:
database/geforce20/2060 - overclocking (reddit.com)
If you CTRL-F the uP1666Q, every instance of this chip on a card is being used to control memory voltage, rather than regulate total power draw which I assume is what we are looking for here.

 

While your at it find out what voltage controller it got and find the feedback pin on the controller. Slap you a 100k vr on that to ground

 

The voltage controllers on this card are all made by uPI, there is a separate controller for the core and for the memory and I’ve linked the data sheet for the memory voltage controller in an earlier reply in this discussion.

 

the GT1030 isn't even requiring a 6-pin external power connector..so how this could be a BIG help??

 

Because it’s power limited at 30W, so there is still plenty of performance on the table due to the PCI-E slots 75W capability as well as the cards low temps.

 

I absolutely agree though sell the 1030 and pick up a new gpu. You can get 780tis/7970s for 100 bucks.

 

Thanks for the suggestion, I just checked eBay and was blown away that the 780ti is so cheap, I might buy a few just to test this Core2Quad as you have, this card must be bad for mining haha.
I think you are misunderstanding the situation, I’m not looking to make the most economical choice here (which I agree would be to sell the 1030 and buy a 780ti) I’m just looking to push this card as far as possible for fun, if it burns out I do not care.

 

Ah your pushing it ok thats better. Volt mod it slap a good cooler on it and change the shunts. Idk about that card ik that 1080s etc dont reapond to higher voltage and are more temp limited until you go sub ambient.

 

Ok, the uP1666Q datasheet I am looking at shows it is a single Vout controller; meaning it can only control the core with it's two phase output. The datasheet mentions that it uses MOSFET Rds(on) to balance the current each phase handles internally (with no dedicated sense pins for a shunt resistor); so maybe you can probe out the circuitry attached to the MOSFETs for the core and see if you can identify any current sensing ICs like IR25750 (they likely have 3 pins or more). It is possible that such a strict power limit is being enforced by a physical circuit comparing the output of a current sensing IC to a preset limit instead of the BIOS telling the controller how much current is allowed to flow; although I don't have enough technical knowledge to understand how that limit could be translated to the controller.

 

Thank you for reading through that! I had a look into the Bios with a hex editor and it does look like the power limit is at least stated in the Bios, so I am not fully convinced that the Bios isn’t at play here.
My process for determining whether the power limit was stated in the bios was to look into the Bioses of the gtx 1050 and 1070 where it is well documented what part of the Bios controls the power. Looking at the same offset for the GT1030 shows a hex code with the decimal value of 30,000 stated twice, which I assume is the power target and power limit stated on milliwatts.
I have purchased an EEPROM programmer and will attempt to modify the power limit in the Bios and report how that goes. If the card boots and runs fine with a modded vBios but the power limit is still locked at 30W then I think it’s safe to say that your assumption is correct that the vBios is not the only thing controlling power draw.

 

For anyone who may be interested, here is the guide I am following for correcting the checksum of the modded vBios:

[Tutorial] How to fix a bios checksum.

When a new graphics card has just been released there are often no tools available to properly modify the bios of the card. It's often fairly easy to find the clock speeds and voltages by comparing the bios to one on an older card, so that's not the problem here. The one big problem you will...

www.xtremesystems.org

 

Okay, I received my EEPROM programmer in the mail today and here is what I was able to accomplish:

I located the 25 series EEPROM on the back of the card, and attached it to the reader which I plugged into a separate PC.

Using the software ASProgrammer, I was able to read the vBios off the GT1030 and confirmed that it was the same as the one I backed up with NVFlash via a hex editor.

I located the Power Limit and Power Target hex bytes which are in row 30860, and changed them both to represent the decimal 65000 instead of 30000 (Power in milliwatts)

In order to correct the checksum, I found the difference between the new Bioses 32 bit checksum and the original Bios 32 bit checksum, and then manipulated garbage bytes at the end of the modded vBios so that the checksum matched the original per the guide I linked earlier in the thread.

From there I erased the old vBios from the EEPROM and wrote the modded vBios into it. I then reread the EEPROM once more to make sure the vBios was written correctly.

Then I put the card back in the system and turned the system on. Unfortunately on boot I got a flashing blue screen followed quickly by a static black screen.

I then took the card out of the system and reconnected to the EEPROM reader to the card. After making sure I didn’t accidentally forget to match the checksums, I decided to reflash them original vBios to the card, which resulted in the card working fine again once put into the system.

At this point I was hopeful that perhaps the bytes I had modified the first time were the reason the flash failed to boot. So I reattached the card to the reader and just modified a few garbage bytes of the original vBios at the end of vBios, still making sure that the 32 bit checksum matches.

Unfortunately, this flash also failed to boot, showing the same display glitches as before. One more reflash of the original vBios and the card is working just fine.

I’m thinking that the desktop pascal/turing/ampere cards truly cannot be bios modded without Nvidia’s signature, as I don’t seem to be able to modify even a single byte without having the card fail to boot, despite matching the checksum to the original.

I’m going to continue probing SMD’s with my multimeter/looking through data sheets to try and formulate a hardware mod, but unless anyone has any ideas how to progress with a vBios mod (Any examples of success on a desktop Pascal card, doesn’t have to be a 1030) I think I will stop trying to modify the cards vBios for now.

 

Well first of all, this isn't true otherwise the Mobile Pascal Bios Editior wouldn't work for anyone and I wouldn't be running with a 230W GTX 1070 (230W directly through the MXM slot) on my MSI Laptop.

What were the very first two bytes of your edited BIOS file?
And what were the first two bytes of the original BIOS file?
And more importantly, what were the first two bytes of the modded flashed Bios when read from the chip?

One thing I saw awhile back on the notebookreview forums in the Pascal Editor thread was those first two bytes changing when flashed to the card, from the original values to two other values.
This despite the modded BIOS having the correct values. When they were flashed with a programmer, the values changed. I think from VERY FAINT memory, that one problem on 1070's were the first two bytes changing from "55 00" to "4E 56", just by flashing the bios with the programmer.
And then the card would of course fail to boot.
I don't remember at all what people had to do to stop that from happening. Someone hinted at an option in the programmer software. Someone mentioned that a Skypro programmer prevents that from happening. Another person said if it does happen with a Skypro, you need to read the original raw file (modded), then flash, then immediately read the flash contents to see if they match.

The very last time I remember discussion about this problem, someone mentioned something about the 1.8v adapter being the culprit for this, and they (FROM FAINT MEMORY) tried flashing directly with 3.5v (which could fry the chip and DID fry some people's chips over there) and then the flash worked and the bytes did not change and the card booted.

I never had this issue with my skypro flashing a modded 1070 with my 1.8v adapter + Skypro. I also faintly remember someone mentioning a Coright software update "possibly" dealing with something like this?

You're on your own here but:

TechnologyGuide

Thank you for visiting the TechnologyGuide network. Unfortunately, these forums are no longer active. We extend a heartfelt thank you to the entire community for their steadfast support—it is really you, our readers, that drove

forum.notebookreview.com

 

I have found a resistor on the back of the card that is directly in-line with the 12V rails of the PCI-E slot via multimeter testing, however it is unmarked. Has anyone ever seen an unmarked shunt resistor on a card? I'm trying to test its resistance with my Multimeter but it is showing zero resistance when using the lowest scale on my multimeter (x10). Is this perhaps because this is the shunt resistor and its resistance is so low (in the range of milliohms) that it effectively looks to have zero resistance on my cheap multimeter?

I have also determined that this resistor is in line with the pins 5 & 6 of a MOSFET on the front side of the card labeled M3202S. There is very little information on this mosfet, except for on this vietnamese page where you can buy them:
M3202S: 380kHz, 2A Asynchronous DC-DC Buck Converter | linhkiendientu.vn - ***IDI

Edit: Here is the resistor in question marked in green with indications showing its apparently unimpeded connection to the 12v pins of the PCI-E slot:

 

Something somewhere on this card must be measuring power draw, otherwise the 30W limit couldn't be enforced. I assume this is either an IC on the PCB or the GPU itself measuring the voltage drop across either a resistor on the PCB or an internal resistor within the IC/GPU to calculate current and thus power draw.

My assumption is that whatever this device may be, it must be in line with the +12V pins of PCI-E slot, so I've probed every device on the front and back of the PCB to figure out what devices are in-line with +12V. Attached below are images highlighting which devices have a 0 ohm connection directly to +12V:
PCB Front:

PCB Back:

Also attached in this reply are my annotations for the front and back devices with names of the SMD's.

I have gone through datasheets of each of the highlighted SMD's and found nothing that indicates that they are to be used expressly for measuring current. A few of the datasheets mention the monitoring of an external resistor divider, which sounds something like what I am looking for, but the rest of those datasheets did not elaborate on that.

I assume there are 4 possibilities for how this card is measuring current:
The GPU is measuring voltage drop across an internal resistance
The GPU is measuring voltage drop across a PCB mounted shunt resistor
An IC is measuring voltage drop across an internal resistance
An IC is measuring voltage drop across a PCB mounted shunt resistor

From that there are 2 methods for success and one for failure:
If the device (IC or GPU) is measuring the voltage drop over a shunt's resistance then I assume its easy: Just drop another shunt on top in parallel or swap out the original shunt for a lower resistance shunt.

If an IC is measuring the voltage drop over an internal resistance, then I assume I cannot modify that internal resistance and would thus need to modify the IC's output pin to send a different output.

If the GPU is measuring an internal resistance itself, then there is nothing that can be done.


Does anyone have any idea what I am looking for in these datasheets to figure out what is monitoring voltage drop (current)? I apologize but my knowledge of electronics is very basic & thus much of the info in those datasheets does not make sense to me.

 

Something somewhere on this card must be measuring power draw, otherwise the 30W limit couldn't be enforced. I assume this is either an IC on the PCB or the GPU itself measuring the voltage drop across either a resistor on the PCB or an internal resistor within the IC/GPU to calculate current and thus power draw.

My assumption is that whatever this device may be, it must be in line with the +12V pins of PCI-E slot, so I've probed every device on the front and back of the PCB to figure out what devices are in-line with +12V. Attached below are images highlighting which devices have a 0 ohm connection directly to +12V:
PCB Front:
View attachment 2525599
PCB Back:
View attachment 2525600
Also attached in this reply are my annotations for the front and back devices with names of the SMD's.

I have gone through datasheets of each of the highlighted SMD's and found nothing that indicates that they are to be used expressly for measuring current. A few of the datasheets mention the monitoring of an external resistor divider, which sounds something like what I am looking for, but the rest of those datasheets did not elaborate on that.

I assume there are 4 possibilities for how this card is measuring current:
The GPU is measuring voltage drop across an internal resistance
The GPU is measuring voltage drop across a PCB mounted shunt resistor
An IC is measuring voltage drop across an internal resistance
An IC is measuring voltage drop across a PCB mounted shunt resistor

From that there are 2 methods for success and one for failure:
If the device (IC or GPU) is measuring the voltage drop over a shunt's resistance then I assume its easy: Just drop another shunt on top in parallel or swap out the original shunt for a lower resistance shunt.

If an IC is measuring the voltage drop over an internal resistance, then I assume I cannot modify that internal resistance and would thus need to modify the IC's output pin to send a different output.

If the GPU is measuring an internal resistance itself, then there is nothing that can be done.


Does anyone have any idea what I am looking for in these datasheets to figure out what is monitoring voltage drop (current)? I apologize but my knowledge of electronics is very basic & thus much of the info in those datasheets does not make sense to me.

All I can say is when I see 6 SOP8/SOIC8 chips on a board that all look like bios chips, I know it's time to give up and buy a 3090.

--------------------

Pretty sure C83 is some sort of shunt. I think I saw something similar like this in one of the shunt mod threads on NBR that were NOT in the main pascal editor thread?
In that thread, someone had something that did not look like any normal 2512 or 1206 shunt and they just desoldered it and put on a "real" shunt to mod their laptop card.

Also are R516 and R565 = 15 mOhm shunts?

 

I checked carefully the IC marking you posted above.
There are no current monitoring IC in the list, not totally sure, because sometimes it is hard to find the exact SMD marking.

Maybe it resembles to the 1050Ti, with no additional 12v power connector, which does not show evident current shunts.
The current is managed by the gpu itself i suppose, the gpu ask for a determined voltage and the VRM controller answer, with the desired gpu voltage.

So, to me, either there is another kind of implementation of the current monitoring at the 12v pci-e, that does not use a discrete monitoring IC.
Either the gpu sense the current at the output VRM, limiting its own current consumption.

If the current sensing is not using a discrete IC, a similar circuit with OpAmp would look like this.

We found that the LM358G is a dual OpAmps, maybe it is used to build up a current monitoring circuit for the gpu, memory.
So you have to retro engineer the components and the traces that come out from it, with your multimerter.

Or either begin to probe the output VRM, checking for traces or components that are directly linked to the gpu, like some kind of shunt.

 

There are no current monitoring IC in the list, not totally sure, because sometimes it is hard to find the exact SMD marking.

That’s what I figured as well. I think one or two of the data sheets mentioned current monitoring, however none of those data sheets had any resemblance to the INA3221 data sheet which is the current sense IC used on higher end GPUs.

I don’t think I’m understanding what to look for or what I would need to change. I agree/understand that the GPU die is likely monitoring current itself due to the absence of a INA3221-equivalent, however I don’t know anything about how these IC’s work, and certainly not how they would work with each other.

My basic understanding is that whatever is monitoring current (be it the gpu or an IC), has to have two traces/electrical connections: One to each side of a known resistance to measure a voltage drop across said resistance. Are these assumptions correct?

The reasoning above is why I proved devices to determine what was on the +12V plane, because every example of a power mod online illustrates the IC/GPU measuring the voltage drop of the +12V line.

Based on your previous answer, it sounds like the shunts for this card may not be on the 12V power plane, so I proved the rest of the PCB and found two shunt resistors directly behind the GPU die pictured here circled in faint green:

Although these black SMD’s report no resistance, I believe they are shunts as my multi is not precise enough to measure milliohms. I don’t think these are jump-resistors, as the actual zero-resistance resistors on the rest of the board are marked with a “0”, even ones much smaller than these two.

Both of these resistors measured as having approximately 70 ohms resistance to the 12V power plan, so I assumed they must be connected to an IC somewhere on the board which is stepping down the voltage.

I next probed these potential shunts to the pins of IC’s across the board. I found that these are both in line with a device called ICE5 ( I think?).

Not sure if that IC is meaningful at all, or if these shunts are even worth probing. I will keep looking across the board.

 

Since you mentioned that you are using a cheap multimeter, I assume that it came with a standard probe set. If you haven't already, I highly recommend that you pick up some 'needle' or 'fine' tip multimeter probes to ease the process of probing tiny components and staying on target while using the second probe on the other side of the PCB. The cheap 'gold' plated ones like these have worked good enough for me while tracing out component paths from individual voltage regulator pins and are sharp enough to stay put on a solder pad; just be careful not to bend/break the tips with too much pressure since they are obviously thinner/weaker than the standard probes.

 

Now that I know that SMD's marked with an "R" on the PCB are resistors, I probed every resistor on both sides of the PCB to figure out which ones might be current sensing (effectively 0 ohms on my multimeter). Below are the highlighted resistors:
Front:

Back:

The reason I did this was so I knew which points to test continuity to from LM358G's input pins so that I could retro-engineer the input circuits of LM358G.

I decided to first attempt to retro-engineer to the circuit for Input 2. When probing across the board looking for current sense resistors, I was most interested by R533 & R534 on the back of the PCB. These resistors were both measuring near zero resistance, and they were both also in-line with ground, similar to the diagram you replied with earlier:

But most interesting was the fact that they were next to R537 & R538, which were both 2 ohms, similar to "R1" in the example which were both the same resistance. Each "shunt" is also directly in line with the 2 ohm resistor to the side of it, however it is not in-line with a 2 0hm diagonal to it. For example, R534 is in line with R538, however it has ~4 ohms resistance to R537.

Next I took another look at the circuitry surrounding LM358G, and found that both the input 2 pins were in-line with polar SMD's which each had 45K ohms in the direction of LM358G, and 200 ohms going away from LM358.

I next probed the points marked 1 & 2 to R5XX area, and found that both points are in line with R537, and thus have only 2 ohms resistance to R533/534 (the shunts).

Is this a valid circuit to the model of an OpAmp you provided? Or are could it be that R4058 & C816 are only in line with R533/534 because they both run to ground?

Are there any probe tests i can do to eliminate possible candidates for current sensing shunts? Is it a valid assumption that any "shunt" not in line with ground is not a shunt given that the current sense resistor in your diagram went to ground? Or was that just an example and not the rule?

 

Well the rule should be.

ANY low value resistor that is in line with the 12v power plane. => big, high current shunts resistors.
ANY low value resistor that goes to ground from the 12v power plane or from the output of the vrm to ground. => smaller, lower current shunt resistors?

AND it should also trace back to the OpAmps, to the gpu or to some kind of monitoring chip.

 

So far I have probed for pin 5 (easiest to hold the multimeter probe to) of LM358G to every contact on the entire board twice and the only point of zero resistance is the 20ohm resistor next to it, which runs immediately to ground afterwards.

Am I possibly doing something wrong here? I have probed the entire board for connections with pin 5, once with the black probe on pin 5 and once with the red probe on pin 5 just to be sure, but there shouldn’t be any difference because I’m just checking for continuity right?

I have also probed most of the contact points on the board for continuity to pins 2,3 & 6 and also found no points of continuity.

Is it possible that all of these inputs are just running to the pins beneath the GPU? It really doesn’t seem as if the inputs of LM358G have any points of continuity to the rest of the board.

 

I actually found an example of some sort of power mod done to my exact model of GT1030 on HWBot:

It looks like this guy has uP1666Q hooked up to some blue device in the top left? Does any one have any idea what that blue device is and know what is going on here?

 

thats a trimmer pot. Idk the value of it though

 

I ended up paying $9, but I was able to get the GT1030 Schematic and Boardview for the MSI model, see attached. Also if there are any GPUs anyone would like schematics for just ask me, I get 20 downloads but only for a day.

Unfortunately I cannot share the boardview as Overclock.com doesnt allow sharing of .cad files. If you know of any other way to share I can do so

 

Sir you made my day, i know that being praised by a random on the internet is not worth much, but hats off to you, for dedication to a project. 👒 🎓🎩🧢👑
No kidding, you managed to restore some faith to an old tech/gaming passionate, thank you a lot.

I will start to work on it right the way and report back, the Broadview files are not stricly needed, it is just an easier and faster way to deal with the schematics, when one is doing repairs.

 

I believe those voltage sense lines are used to measure the on-die voltage so the voltage controller can compensate for the voltage drop across the PCB; not necessarily for the current monitoring of the VRM output.

 

So unfortunately, there is no real shunt resistor, as we found in the upper hand cards, as we suspected.
The gpu itself seems to sense and/or control the VRM output voltage, the gpu core is directly linked to pins that dictate the voltage output of the VRM controller.

And sure enough, where the PS_NVDD_VSNS comes from, you guessed, the output of the VRM.

Tho, i realized that these schematics don't match completely the card we are working on, i can't find the U505 for example, in the basic GT1030 schematics.
U505 is the OpAmps we were working on in the previous post, I will check the other Galax file, very curious.



Yep, i'm not sure how really the circuit works, since the gpu core should be able to control the VID voltage, through the VRM controller VID pin.
Nevertheless, the gpu core is also tied to the VRM controller voltage control circuit, for the voltage drop on the power plane as you said.

The gpu core have 2 sense lines, so it can monitor its own core voltage and do the power consumption calculation based on theses sensed values, i suppose?
Which means that the power consumption is totally managed by software, it would explain why the HWBot user could volt-mod the card, without the need of a shunt mod?

 

Here is what I've determined for the feedback circuit of uP1666Q:

The Feedback pin (pin 11) is in line with R37. R37 is in the exact spot where the HWBot user soldered the potentiometer.

Help me understand something, why is it that when I run the probes in one direction, I get 120 ohms for R37, and when I run the probes the other way, I get 850 ohms? I thought with resistors it didn't matter which probe was held to which pad?

Either way, I'm fairly certain that R37 is the key SMD in the Feedback circuit, as it's right pin has only a 6 ohms resistance going back to the Feedback return pin, which I assume is the resistance of whatever the feedback circuit runs through (the GPU core?)

Edit:
I found R37 in the CAD file of the MSI board, and it seems to function similar to the Gigabyte model, as it is positioned near U5 as well. This SMD appears to connected to voltage related parts of the MSI model, so very similar function it appears. It also reports a value of 1000 ohms on the MSI model, so 850 might make more sense for the Gigabyte model then.

 

Here is what I've determined for the feedback circuit of uP1666Q:
View attachment 2526322
The Feedback pin (pin 11) is in line with R37. R37 is in the exact spot where the HWBot user soldered the potentiometer.

Help me understand something, why is it that when I run the probes in one direction, I get 120 ohms for R37, and when I run the probes the other way, I get 850 ohms? I thought with resistors it didn't matter which probe was held to which pad?

Either way, I'm fairly certain that R37 is the key SMD in the Feedback circuit, as it's right pin has only a 6 ohms resistance going back to the Feedback return pin, which I assume is the resistance of whatever the feedback circuit runs through (the GPU core?)

Edit:
I found R37 in the CAD file of the MSI board, and it seems to function similar to the Gigabyte model, as it is positioned near U5 as well. This SMD appears to connected to voltage related parts of the MSI model, so very similar function it appears. It also reports a value of 1000 ohms on the MSI model, so 850 might make more sense for the Gigabyte model then.
View attachment 2526324

Here is what I've determined for the feedback circuit of uP1666Q:
View attachment 2526322
The Feedback pin (pin 11) is in line with R37. R37 is in the exact spot where the HWBot user soldered the potentiometer.

Help me understand something, why is it that when I run the probes in one direction, I get 120 ohms for R37, and when I run the probes the other way, I get 850 ohms? I thought with resistors it didn't matter which probe was held to which pad?

Either way, I'm fairly certain that R37 is the key SMD in the Feedback circuit, as it's right pin has only a 6 ohms resistance going back to the Feedback return pin, which I assume is the resistance of whatever the feedback circuit runs through (the GPU core?)

Edit:
I found R37 in the CAD file of the MSI board, and it seems to function similar to the Gigabyte model, as it is positioned near U5 as well. This SMD appears to connected to voltage related parts of the MSI model, so very similar function it appears. It also reports a value of 1000 ohms on the MSI model, so 850 might make more sense for the Gigabyte model then.
View attachment 2526324

As i said, if you put the probes on the resistors pads, you read the resistor value.
So not sure how you could read different values, R37 should be 1K Ohms +/-, dunno if they used the same resistor tho, as you pointed out.
If you read the resistance from the FB to the R37, keep in mind that you have traces that come in, in between.

Just beware reading randomly, as i said, there are trances popping in and out everywhere in the circuit.

If you want to do the @mllrkllr88 mod, you need to probe from the VSENSE+/FB to ground, to be able to read the whole circuit resistance value.
In this case, the VSENSE+/FB are the same thing, so from the FB side pad of the R37, to ground.

That's why i would rather look at the REFIN part of the circuit and simply change the REFIN voltage, like i showed you above.
If REFIN is 0.65v instead of 0.6v, your final voltage with be skewed by 0.05v.

As @o1dschoo1 said, these are trimmer pots, also called variables resistors, the Bourns brand is the best.