tehsiggi wrote on 2025-09-23, 10:51:

Yep, you just ripped of it's dual FET (Q9) :D No more switching there, that's for sure. […]
Show full quote

Yep, you just ripped of it's dual FET (Q9) 😁 No more switching there, that's for sure.

While it's gone (lol) feel free to measure the resistance between MVDDQ and GND.

Here's btw. a picture of the SS pin from ISL6522 with over-current protection active:

The attachment Screenshot 2025-09-23 at 12.48.38.png is no longer available

Yellow is the waveform of interest. I think blue was one of the Gates, either high or low-side. In this case the FET for the low side had a disconnect on it's gate. Leading to the voltage not being generated correctly.

Well that’s the end of that 🤣. The FET refused to come off the heat sink so got some pliers and slowly twisted side to side, the IC eventually came off but unfortunately one of the legs broke off. The factory glued every chip under 🥲

RIP.

The FET is a STS8DNF3LL or comparable. Dual nFET with 30V rating and 8A. With good RDSon at 4.5V as well.

tehsiggi wrote on 2025-09-24, 07:06:

RIP.

The FET is a STS8DNF3LL or comparable. Dual nFET with 30V rating and 8A. With good RDSon at 4.5V as well.

https://www.digikey.com.au/en/products/detail … S8DNF3LL/603495 looks alright?

Yep, looks good. Did you get a reading on resistance between MVDDQ and GND?

tehsiggi wrote on 2025-09-24, 07:31:

Yep, looks good. Did you get a reading on resistance between MVDDQ and GND?

Yep it's 69.8 ohms

Don't want to interrupt you guys, but did you check that ferrite? Looks suspicious on the photo but could also just be some dirt. Also the tantalum(s) below have somewhat black marks.

ChrisK wrote on 2025-09-24, 08:00:

Don't want to interrupt you guys, but did you check that ferrite? Looks suspicious on the photo but could also just be some dirt. Also the tantalum(s) below have somewhat black marks.

That'd be the input ferrite for MVDDC or VTT, I suspect if it's a goner, we'd have a voltage missing here. (Un-)fortunately, those are there and look fine. The dirt dot looks suspicious, I agree.

zuldan wrote on 2025-09-24, 07:41:

Yep it's 69.8 ohms

That sounds good so far! Similar to my card, so no hard short at least.

ChrisK wrote on 2025-09-24, 08:00:

Don't want to interrupt you guys, but did you check that ferrite? Looks suspicious on the photo but could also just be some dirt. Also the tantalum(s) below have somewhat black marks.

Thanks for noticing that. I cleaned with IPA and it appears just to be dirt.

ChrisK wrote on 2025-09-24, 08:00:

Don't want to interrupt you guys, but did you check that ferrite? Looks suspicious on the photo but could also just be some dirt. Also the tantalum(s) below have somewhat black marks.

The attachment Ferrite.PNG is no longer available

Diode looks dusty or faulty too

Where do you see a diode on that picture?

Brain fart.
Tantrum caps.

tehsiggi wrote on 2025-09-23, 10:51:

Yep, you just ripped of it's dual FET (Q9) :D No more switching there, that's for sure. […]
Show full quote

Yep, you just ripped of it's dual FET (Q9) 😁 No more switching there, that's for sure.

While it's gone (lol) feel free to measure the resistance between MVDDQ and GND.

Here's btw. a picture of the SS pin from ISL6522 with over-current protection active:

The attachment Screenshot 2025-09-23 at 12.48.38.png is no longer available

Yellow is the waveform of interest. I think blue was one of the Gates, either high or low-side. In this case the FET for the low side had a disconnect on it's gate. Leading to the voltage not being generated correctly.

FET has arrived (https://www.st.com/content/ccc/resource/techn … .CD00002691.pdf). Any hints on what orientation it goes? A shame pin 1 isn't marked on the chip. Arrg!

If I'm not mistaken, the ST logo should be the "Pin1 mark" for it, and that's by also looking at the datasheet.

Oh, and it seems the Pin1 side has a slightly larger edge than the opposite one. But yeah, the ST logo should be your pin1 marker alright.

The datasheet you listed is from the STS9DNF3LL, your chip is an STS8DNF3LL. However, Pin 1 can easily be found:

The package as one side with an angled edge, you can see it easily when looking at the package from the side.

That let's you find pin1.

If I don't misjudge your picture, pin one is where the ST logo is.

tehsiggi wrote on 2025-10-12, 06:19:

The datasheet you listed is from the STS9DNF3LL, your chip is an STS8DNF3LL. However, Pin 1 can easily be found: […]
Show full quote

The datasheet you listed is from the STS9DNF3LL, your chip is an STS8DNF3LL. However, Pin 1 can easily be found:

The attachment Screenshot 2025-10-12 at 08.17.22.png is no longer available

The package as one side with an angled edge, you can see it easily when looking at the package from the side.

That let's you find pin1.

If I don't misjudge your picture, pin one is where the ST logo is.

Thanks PcBytes and tehsiggi, just to double confirm, I'll install the new chip in this orientation?

pin 5&6 and 7&8 connected internally, but not a 6&7
Why not just check the resistance with a multimeter before installation?

Yes. Your alignment is correct.

On your picture the footprint on the PCB has a corner on the top left, indicating pin 1.
Also, if assume pin 1 to be top left, you can see that it is connected to 5 and 6, exactly as in the schematic. So on the PCB the angled corner is 100% pin 1.

The STS8DNF3LL has the pin one at the top of the angled edge in your picture. That fits well.

The pin you marked with ground is PIN3, which is also connected to ground in the schematic.

Looks all good to me.

tehsiggi wrote on 2025-10-24, 08:46:

Yes. Your alignment is correct. […]
Show full quote

Yes. Your alignment is correct.

On your picture the footprint on the PCB has a corner on the top left, indicating pin 1.
Also, if assume pin 1 to be top left, you can see that it is connected to 5 and 6, exactly as in the schematic. So on the PCB the angled corner is 100% pin 1.

The STS8DNF3LL has the pin one at the top of the angled edge in your picture. That fits well.

The attachment Screenshot from 2025-10-24 10-46-27.png is no longer available

The pin you marked with ground is PIN3, which is also connected to ground in the schematic.

The attachment Screenshot from 2025-10-24 10-44-52.png is no longer available

Looks all good to me.

tehsiggi thank you so much for your help (again). You won't believe it, the FET (STS8DNF3LL) I ripped off was faulty and was the cause of the card not working. Once I installed the new FET, the memory came back online and the card immediately started working again. What luck!

1
2R3MEMID version 1.07, (c) Copyright ATI Technologies Inc, 2003
3Log file generation enabled to .\R3MEMID.LOG ...
4Reference data file (RDF) loading disabled ...
5[1  ] Fill                                                : PASS
6[2  ] Data line toggle (GUI)                              : PASS
7[1  ] Fill                                                : PASS
8[2  ] Data line toggle (GUI)                              : PASS
9[3  ] RW page (GUI)                                       : PASS
10[4  ] RW channel (GUI)                                    : PASS
11[5  ] Data mask                                           : PASS
12[6  ] Macro + Micro8x2 tiling read                        : PASS
13[7  ] Byte swap macro+micro8x2 read                       : PASS
14[8  ] HOST_DATA/<f,b>/M2L                                 : PASS
15[9  ] Line slopes/dir/P/solid                             : PASS
16TEST RESULT SUMMARY:
17====================
18R300 (0x4e44) detected.
19128M video memory.
20Test suite ran 1 of 1 times.
21Checking test status array ...
22[1  ] Fill                                                : PASS
23[2  ] Data line toggle (GUI)                              : PASS
24[3  ] RW page (GUI)                                       : PASS
25[4  ] RW channel (GUI)                                    : PASS
26[5  ] Data mask                                           : PASS
27[6  ] Macro + Micro8x2 tiling read                        : PASS
28[7  ] Byte swap macro+micro8x2 read                       : PASS
29[8  ] HOST_DATA/<f,b>/M2L                                 : PASS
30[9  ] Line slopes/dir/P/solid                             : PASS
31No failures detected.

Nice to see one ATi Radeon 9700 Pro back in life. I only have his younger and cooler brothers: Radeon 9600 pro and xt (no extra power needed)

zuldan wrote on Yesterday, 09:36:

tehsiggi thank you so much for your help (again). You won't believe it, the FET (STS8DNF3LL) I ripped off was faulty and was the cause of the card not working. Once I installed the new FET, the memory came back online and the card immediately started working again. What luck!

I directly believe you! Remember how we got here?

tehsiggi wrote on 2025-09-23, 10:51:

Here's btw. a picture of the SS pin from ISL6522 with over-current protection active: Screenshot 2025-09-23 at 12.48.38.png […]
Show full quote

Here's btw. a picture of the SS pin from ISL6522 with over-current protection active:
Screenshot 2025-09-23 at 12.48.38.png

Filename
Screenshot 2025-09-23 at 12.48.38.png
File size
381.16 KiB
Views
1009 views
File license
Public domain

Yellow is the waveform of interest. I think blue was one of the Gates, either high or low-side. In this case the FET for the low side had a disconnect on it's gate. Leading to the voltage not being generated correctly.

I had a card where one FET was slightly disconnected. Leading to a very similar pattern that you did see. Not the full voltage, while the rail looked not suspicious from a resistance point of view. I looked at the SS pin of the ISL and then saw we're in current limiting, though the rail itself was not the cause.

So I'm very glad that we lifted the cooling plate to check that ISL below 😁

Happy to see that card back in action, very glad it worked out for you!