Is it possible to take a macro photo so that the resistor values ​​are visible?
And a screenshot from GPU-Z

shevalier wrote on 2025-06-30, 16:51:

Is it possible to take a macro photo so that the resistor values ​​are visible?
And a screenshot from GPU-Z

I'll try for the macro picture of those resistors, but the screenshot from GPU-Z is impossible in this state of the glitching videocard as it rarely makes it to win desktop before "black screen".

I have a theory: Asus intentionally kept the GPU core voltage at it's minimum (about 1.25 V) as the GPU's heatsink is pathetic. After all this is a cheaper variant of R9600NP.

And a second theory: these Hitano electrolytic caps are even cr*ppier than "normal" low quality junk as they degrade without any visible signs of leaking or bulging.

I just went into the basement and measured three cards:

HIS Radeon 9550 RV350 256MB: VDDC = 1.179V
HIS Radeon 9600 RV350 256MB: VDDC = 1.220V
Sapphire Radeon 9600 Pro Rv360 256MB DDR2: VDDC 1.291V

The voltage could be completely in range for a 9550.

All above mentioned cards work without issues.

analog_programmer wrote on 2025-06-30, 16:39:

So no difference, but obviously GPU voltage is not enough.

Judging by the diagram, it should be exactly 1.2V.
But on real cards I saw this more often.

tehsiggi wrote on 2025-06-30, 17:05:

HIS Radeon 9600 256MB: VDDC = 1.220V
Sapphire Radeon 9600 Pro 256MB DDR2: VDDC 1.291V

Replace the capacitors....

tehsiggi wrote on 2025-06-30, 17:05:

I just went into the basement and measured three cards: […]
Show full quote

I just went into the basement and measured three cards:

HIS Radeon 9550 256MB: VDDC = 1.179V
HIS Radeon 9600 256MB: VDDC = 1.220V
Sapphire Radeon 9600 Pro 256MB DDR2: VDDC 1.291V

The voltage could be completely in range for a 9550.

All above mentioned cards work without issues.

Ok, but what if my card has some momentary voltage drop caused by ripple which I can't measure with multimeter?

I think I'll just replace the two suspected caps and if there's no improvement - the card goes for scrap or spare parts.

BTW what do you think about the liner VRM voltage values? Are these OK?

analog_programmer wrote on 2025-06-30, 17:12:

Ok, but what if my card has some momentary voltage drop caused by ripple which I can't measure with multimeter? […]
Show full quote

tehsiggi wrote on 2025-06-30, 17:05:

I just went into the basement and measured three cards: […]
Show full quote

I just went into the basement and measured three cards:

HIS Radeon 9550 256MB: VDDC = 1.179V
HIS Radeon 9600 256MB: VDDC = 1.220V
Sapphire Radeon 9600 Pro 256MB DDR2: VDDC 1.291V

The voltage could be completely in range for a 9550.

All above mentioned cards work without issues.

Ok, but what if my card has some momentary voltage drop caused by ripple which I can't measure with multimeter?

I think I'll just replace the two suspected caps and if there's no improvement - the card goes for scrap or spare parts.

BTW what do you think about the liner VRM voltage values? Are these OK?

While you're at it, go for all four. The linear ones are thermally stressed.

And yes.. Caps can degrade without visible chnages to their appearance.

shevalier wrote on 2025-06-30, 17:10:

Replace the capacitors....

The cards work fine, the caps are within spec and have been already tested for ESR, i will not change them.

tehsiggi wrote on 2025-06-30, 17:15:

While you're at it, go for all four. The linear ones are thermally stressed.

I mean are the voltage values within specs?

As for these Hitano caps - I'll replace all of them with more descent (some low ESR Panasonic or at least Samxon).

The datasheet of your memory specs 200MHz / 400MHz effective at VDDQ and MVDDC of 2.5-2.7 volts. So voltage wise, they are fine.

tehsiggi wrote on 2025-06-30, 17:15:

The cards work fine, the caps are within spec and have been already tested for ESR, i will not change them.

I apologize if you took this advice as applying to you.

shevalier wrote on 2025-06-30, 17:36:

tehsiggi wrote on 2025-06-30, 17:15:

The cards work fine, the caps are within spec and have been already tested for ESR, i will not change them.

I apologize if you took this advice as applying to you.

The quote was on my answer, so I was a bit confused 😁

All good!

analog_programmer wrote on 2025-06-30, 17:12:

BTW what do you think about the liner VRM voltage values? Are these OK?

A linear regulator in which the regulating element (a transistor of any type, including a MOSFET) is connected in such a way that it has local and fast negative feedback "inside the transistor".
https://en.wikipedia.org/wiki/Common_drain
Therefore, it reacts faster to changes in load and is less dependent on the state of the output capacitor.
With equally lousy ESR of capacitors, the linear regulator will still remain some operational.
And yes, capacitors from the same batch, located on the same board, wear out more or less equally.
Especially after 20 years.

shevalier wrote on 2025-06-30, 16:51:

Is it possible to take a macro photo so that the resistor values ​​are visible?

Is this good enough as macro picture quality to find out what voltage Asus originally intended for GPU core?

shevalier wrote on 2025-06-30, 17:10:

tehsiggi wrote on 2025-06-30, 17:05:

HIS Radeon 9600 256MB: VDDC = 1.220V
Sapphire Radeon 9600 Pro 256MB DDR2: VDDC 1.291V

Replace the capacitors....

Ahaaa, I see. This was for me and you gave me the examples from tehsiggi's cards with GPU voltages more than 1.2 V.

Yeah, I'll definitely replace these two as a start. And if there's an improvement - all of these Hitano electrolytic caps.

P.S. Thank you very much for all the information you gave. And this is the simplest ans clearest explanation on capacitor's ESR meaning 😀

shevalier wrote on 2025-06-30, 13:34:

Ripple Voltage = Capacitor ESR * Current_Change

tehsiggi wrote on 2025-06-30, 17:35:

The datasheet of your memory specs 200MHz / 400MHz effective at VDDQ and MVDDC of 2.5-2.7 volts. So voltage wise, they are fine.

Thank you very much for this clarification and the measurements you took on your R9550 cards!

analog_programmer wrote on 2025-06-30, 18:22:

shevalier wrote on 2025-06-30, 16:51:

Is it possible to take a macro photo so that the resistor values ​​are visible?

Is this good enough as macro picture quality to find out what voltage Asus originally intended for GPU core?

https://www.hobby-hour.com/electronics/smdcalc.php
07B = 1.15k
(0.8V/2.2K) * (2.2k+1.15k) = 1.22V
Quite a good match.

What I would do personally
- yes, I would change all the capacitors
- if that doesn't help, replace 1.15k with 1.3k. +50mV usually adds a lot of stability.
- if that doesn't help, then flash the frequencies in BIOS, for example 100/100MHz, and use ATi tray tool to find stable ones.
- try BIOS for different memory.
- in my experience, Voltmod never works with memory on video cards of those years, so I personally don't see the point in it.

shevalier wrote on 2025-06-30, 19:17:

https://www.hobby-hour.com/electronics/smdcalc.php 07B = 1.15k (0.8V/2.2K) * (2.2k+1.15k) = 1.22V Quite a good match. […]
Show full quote

https://www.hobby-hour.com/electronics/smdcalc.php
07B = 1.15k
(0.8V/2.2K) * (2.2k+1.15k) = 1.22V
Quite a good match.

What I would do personally
- yes, I would change all the capacitors
- if that doesn't help, replace 1.15k with 1.3k. +50mV usually adds a lot of stability.
- if that doesn't help, then flash the frequencies in BIOS, for example 100/100MHz, and use ATi tray tool to find stable ones.
- try BIOS for different memory.
- in my experience, Voltmod never works with memory on video cards of those years, so I personally don't see the point in it.

Great tips, thanks!

I'll start with replacing the suspected bad caps. But before the GPU voltage-mod (resistor replacement) I will go with custom-made BIOS according to datasheet I found for Infineon GDDR memory chips on my card - I think, the memory timings settings according to memory frequency have to be written somewhere in the pdf.

If the recap is not helping and you suspect the memory, you can test it with the tool called R3MEMID under DOS. In case your card gets that far.

But one step at a time

tehsiggi wrote on 2025-06-30, 19:53:

If the recap is not helping and you suspect the memory, you can test it with the tool called R3MEMID under DOS. In case your card gets that far.

But one step at a time

Yep, one step at a time - from easiest and most reasonable to hardest and less reasonable 😉

Good to know that there is such a DOS tool. I've never heard of "R3MEMID" and here it popped-out in a vogons thread.

tehsiggi wrote on 2025-06-30, 19:53:

If the recap is not helping and you suspect the memory, you can test it with the tool called R3MEMID under DOS. In case your card gets that far.

I just read your threads on ATI R3x0 based videocards repairs and they're really interesting and useful. I got the R3MEMID utility, but I still can't understand how to decode which memory chip throws errors.

You're giving an example for memory chips IDs with 256 bit memory bus card here: Re: VGA Repair report collection But how to sort the things out with these memory chip bits, IDs, etc. if the card is with different memory configuration - 4 chips, 128 or 64 bit bus...? My failing Asus R9550 has 8 chips with markings U29 - U36.

P.S. I will not be able to use this utility on my Asus R9550 'til I fix the problem with card's stability, but I have some bad feeling that, if I manage to "cure" the card from switching off/hanging the system, there may be some problems with memory chips (because of the artifacting).

Perhaps I didn't read properly, does the issue also occur if you just idle around in BIOS etc?
Some cards with memory issues that I had in the past worked mostly "okay" until the driver has been loaded.

Regarding mapping your memory to the result of the tool:

You'll need a schematic of a 9550 or 9600, often the memory alignment (data bits to memory chips) is the same even if the rest of the layout is different.
There you'll see which data bit goes to which IC. They often also describe where the memory is located on the board.
The RV350/RV360 have two memory channels, A and B, each 64bits => 4 x 16 Bit ICs if TSOP-66 memory is used like on your card.

Fortunately board partners often just copied big chunks of the reference design, like the memory part. So you'll find the exact same memory arrangement and identifiers even if other parts are different. I attached the schematic of a Radeon 9600 from gigabyte for you. It's not exactly the same in terms of the power section, but memory looks good.