Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
Do I see this correctly? You removed the SMB-DAT pad entirely? Or is it just such color of ‘nail polish’ over it?

Anyway, this proves that there is a problem with some SMBus device on the card. The high fan speed is probably a consequence of the fact that no IC on the card can be accessed via SMBus…
There are no SMBUS conflicts using this card within the iMac12,2 - there is only the nasty sleep/wake issue.
 
  • Like
Reactions: highvoltage12v
Do I see this correctly? You removed the SMB-DAT pad entirely? Or is it just such color of ‘nail polish’ over it?

Anyway, this proves that there is a problem with some SMBus device on the card. The high fan speed is probably a consequence of the fact that no IC on the card can be accessed via SMBus…

AMD FirePro W6170M - Update iMac 27" 2009


Hello 👋 dfranetic...

The SMB-DAT Line 32 (pin #16 on front side) I've removed completely 😉 ...shown
on MXM sheet the SMB-DAT goes to 100k Resistor (PTC / NTC) - so what kind of
other information except the thermal Ohmic resistance should the video card
send to the system ? There is an adress conflict and the video card send wrong
priority values (HWSensor 129° Celsius) on same adress as CPU thermal sensor.

I think about Ohmic resistance and Resistance parallel connection calculator...

Parallel resistance calculator

For some people a Video card is a miracle - but on the other side it's an
electronical circuit with logical function, so no wonder except the function
is possible... ;)

👍👍👍 Thank you 👍👍👍
 
Last edited:
  • Like
Reactions: highvoltage12v

AMD FirePro W6170M - Update iMac 27" 2009


Hello 👋 dfranetic...

The SMB-DAT Line 32 (pin #16 on front side) I've removed completely 😉 ...shown
on MXM sheet the SMB-DAT goes to 100k Resistor (PTC / NTC) - so what kind of
other information except the thermal Ohmic resistance should the video card
send to the system ? There is an adress conflict and the video card send wrong
priority values (HWSensor 129° Celsius) on same adress as CPU thermal sensor.

I think about Ohmic resistance and Resistance parallel connection calculator...

Parallel resistance calculator

For some people a Video card is a miracle - but on the other side it's an
electronical circuit with logical function, so no wonder except the function
is possible... ;)

👍👍👍 Thank you 👍👍👍
No, it is not that simple. That 100k resistor is probably there to act as a weak pull-up (or pull-down) as a safety measure (in case the connector on the other end does not respect the MXM standard). The card can send all sorts of information via SMBus (if it is ‘asked’ to). However, iMac’s SMBus is now unable to access anything on the GPU, and that’s the most likely reason iMac’s SMC drives all fans to full RPM.
 
  • Like
Reactions: Ausdauersportler

AMD FirePro W6170M - Update iMac 27" 2009


Hello 👋 dfranetic...

The SMB-DAT Line 32 (pin #16 on front side) I've removed completely 😉 ...shown
on MXM sheet the SMB-DAT goes to 100k Resistor (PTC / NTC) - so what kind of
other information except the thermal Ohmic resistance should the video card
send to the system ? There is an adress conflict and the video card send wrong
priority values (HWSensor 129° Celsius) on same adress as CPU thermal sensor.

I think about Ohmic resistance and Resistance parallel connection calculator...

Parallel resistance calculator

For some people a Video card is a miracle - but on the other side it's an
electronical circuit with logical function, so no wonder except the function
is possible... ;)

👍👍👍 Thank you 👍👍👍
Hi!

Please run Macs Fan Control (using the automatic mode for all fans) and HW Monitor and post some screen shots about the temp data you are getting after the SMBUS has been disabled.

Thanks in advance
 
Sure, but he has this card in an iMac11,1 (not the one in his signature). He now has no conflict, but also no data from the card via SMBus.
Just wanted to remind all of you. Some brave users put that card into the iMac12,2 ignoring of the sleep/wake story and might be confused about this discussion.

The sleep/wake issue with GCN cards is only related to the use of PC GOP vBIOS versions, it is not related in any way to this particular (iMac11,1 and iMac11,3) SMBUS problem.
 

AMD FirePro W6170M - Update iMac 27" 2009


W6170.png


Hehehe... I think there is nothing to analyze - ROTFL 😄😄😄


Please note that "some thermal resistors" are now on other positions inside
of this 12 y.o. iMac ;) - good display, but a real trashcan inside 😄😄😄

Attn: Ausdauersportler 👋
I NEVER use MacsFanControl on automatic mode - not on STEAM and
not on Valley Benchs at 3rd party video cards (Nvidia Quadro / AMD) -
I think about my three defected Quadro K2000M in three month...

On automatic mode with FullHD Valley Bench the Mac switch off after
scene 5 or 6 - the peak of the Video Card channel is about 75 - 93watts
on HWMonitor ... I don't use automatic mode anymore, unnecessary if
SMB-DAT is disabled (ATI) or not (Quadro video cards) - Everybody
need constant cooled GPU & CPU, I think the automatic mode increase
the oldering process of the components...

*** I have a modified Three-pipe-heatsink ***

👋👋👋
 
Last edited:
  • Like
Reactions: Ausdauersportler

AMD FirePro W6170M - Update iMac 27" 2009


Attn: Ausdauersportler 👋
I NEVER use MacsFanControl on automatic mode - not on STEAM and
not on Valley Benchs at 3rd party video cards (Nvidia Quadro / AMD) -
I think about my three defected Quadro K2000M in three month...
Usually I have replaced the ODD sensor to get back SMC automatic mode. It kicks in at 55C on the sink.
*** I have a modified Three-pipe-heatsink ***
Same heat sink here, same card, do not get confused about the M380, just playing with OpenCore. But I have the same weird CPU core 0 and CPU environment readings at 129C even without cutting the SMBUS lane.
Something is still wrong with your experiment...
W6170M SMBUS.png
 
  • Like
Reactions: Pacific1972
Usually I have replaced the ODD sensor to get back SMC automatic mode. It kicks in at 55C on the sink.

Same heat sink here, same card, do not get confused about the M380, just playing with OpenCore. But I have the same weird CPU core 0 and CPU environment readings at 129C even without cutting the SMBUS lane.
Something is still wrong with your experiment...
View attachment 1927148

:cool:;) I have a look on it :cool:;)

ATI FirePro W6170M​

W6170 Test.jpg


But if you have parallel switched two resistors,
for example

R1= 100k Ohm - GPU
R2= 100k Ohm - CPU

R = 50K Ohm - All (Thermal information at same-lane SMBus)

Two same resistors = 1/2 Ohm

and that is not the referred thermal info in °Celsius because of parallel switching on same lane !
(HWsensor shows 129°C, irrelevant if data comes from GPU or CPU) - but perhaps my CPU
sensor also is broken... I don't know ! 😄
 
Last edited:
  • Like
Reactions: highvoltage12v

AMD FirePro W6170M - Update iMac 27" 2009


View attachment 1927128

Hehehe... I think there is nothing to analyze - ROTFL 😄😄😄


Please note that "some thermal resistors" are now on other positions inside
of this 12 y.o. iMac ;) - good display, but a real trashcan inside 😄😄😄

Attn: Ausdauersportler 👋
I NEVER use MacsFanControl on automatic mode - not on STEAM and
not on Valley Benchs at 3rd party video cards (Nvidia Quadro / AMD) -
I think about my three defected Quadro K2000M in three month...

On automatic mode with FullHD Valley Bench the Mac switch off after
scene 5 or 6 - the peak of the Video Card channel is about 75 - 93watts
on HWMonitor ... I don't use automatic mode anymore, unnecessary if
SMB-DAT is disabled (ATI) or not (Quadro video cards) - Everybody
need constant cooled GPU & CPU, I think the automatic mode increase
the oldering process of the components...

*** I have a modified Three-pipe-heatsink ***

👋👋👋
Same thing for me on my late 2009 with RX480 graphics card.
Only difference seems to be that you didn't short circuit the HDD thermal sensor wires (or used some method of providing a separate HDD thermal sensor). The HDD is not an original one I suppose.
 

AMD FirePro W6170M - Update iMac 27" 2009

The SMB-DAT Line 32 (pin #16 on front side) I've removed completely 😉 ...shown
on MXM sheet the SMB-DAT goes to 100k Resistor (PTC / NTC) - so what kind of
other information except the thermal Ohmic resistance should the video card
send to the system ? There is an adress conflict and the video card send wrong
priority values (HWSensor 129° Celsius) on same adress as CPU thermal sensor.
Well, removing completely the SMB-DAT pin is not what I meant by "easily reversible" experiment 😂
Anyway, great it worked out for you 😄
 
  • Like
Reactions: Pacific1972

AMD FirePro W6170M - Update iMac 27" 2009


After a weekend of Tests and Blackouts (of me and the video card 😄😄 )
I have edit the MXM port of the card and disable the SMB-DAT Line 32 -
(on the Top of the Video card its Pin #16) - for disabling SMBus
thermal problems with same-adress resistor CPU on iMac A1312 (2009):

View attachment 1927095
SMB-DAT disable to avoid same adress problems with Thermal sensors
of GPU and CPU - explore MXM sheet for 100K resistor on MXM-line #32.

Preparation of W6170M:

View attachment 1927096


After a new install of High Sierra I got some first Benchmarks...

(use HWMonitor / MacsFanControl):

View attachment 1927098
Benchmark 1920x1080 - compareable with GTX860M ?


View attachment 1927099
Benchmark 2560x1440


...comparing these Video Card with old HD6750 it's nearly 3 times
faster than ATI HD6750 😄😄😄 - MacsFan was a little bit louder...

MacsFan: ODD 3500 / HDD 3500 / CPU 2000

Info only -
please make a second install with Big Sur & OCLP
on a separate SSD - the OCLP often only show Bootscreen if you
have something to pick ;) ... !!!

Attn: perhaps I edit this post later... the Metal Monster runs...
:eek:

You could simply desolder one jumper resistor. Even cheap iron solderer could be used for this....


also M6100 aka FirePro W6170M is using smbus adress 4C, which is same as native GPUs


ADM1032ARMZ−REEL 8-lead MSOP RM−8 T1J 4C

The ADM1032 and the ADM1032−1 are available with one
SMBus address, which is Hex 4C (1001 100). The
 
Last edited:
No!
You need a 2011 heat sink and need to modify it!
Read the first post and take a look a the pictures attached there. You certainly missed this post on your read through all posts.
Thanks for your answer!
I received the GTX780M and placed it into the Mac to see if it works.
I flashed it in bootcamp and everything works like a charm!!
I'm on Mojave now.
Now I'm on the original heatsink (2 pipes) and the temp of the GPU is around the 60 degrees C.
Now I know the card is working, I want to go for the 3 pipes heatsink.
My question is: does the 2011 heatsink fit in the 2010?
 
Thanks for your answer!
I received the GTX780M and placed it into the Mac to see if it works.
I flashed it in bootcamp and everything works like a charm!!
I'm on Mojave now.
Now I'm on the original heatsink (2 pipes) and the temp of the GPU is around the 60 degrees C.
Now I know the card is working, I want to go for the 3 pipes heatsink.
My question is: does the 2011 heatsink fit in the 2010?
Yes, it does. Is there any evidence it would not since we strongly recommend it?
 
:eek:

You could simply desolder one jumper resistor. Even cheap iron solderer could be used for this....


also M6100 aka FirePro W6170M is using smbus adress 4C, which is same as native GPUs


ADM1032ARMZ−REEL 8-lead MSOP RM−8 T1J 4C

The ADM1032 and the ADM1032−1 are available with one
SMBus address, which is Hex 4C (1001 100). The

🤣 Can you translate that for me ? I don't speak in HEX or assembly language... 🤣

I can't know or find every component of the video card, I've searched for R507
(MXM sheet I posted a few days before) but I don't find R507 on the video card,
not at the front, not at the back - so I must make another solution...

😇
 
Last edited:
??? what ???

🤣 Can you translate that for me ? I don't speak in HEX or assembly language... 🤣

I can't know or find every component of the video card, I've searched for R507
(MXM sheet I posted a few days before) but I don't find R507 on the video card,
not at the front, not at the back - so I must make another solution...

😇
It's probably R220 - the lower resistor/jumper circled in lime green in this @nikey22 post (scroll down a bit). (Different card, but I suspect they have the same PCB layout.)
 
  • Like
Reactions: Pacific1972
It's probably R220 - the lower resistor/jumper circled in lime green in this @nikey22 post (scroll down a bit). (Different card, but I suspect they have the same PCB layout.)
PCB Design is same and I see T1J Chip. If I desolder R220, I think in case
I can resolder it with wire bridge. R220 have 0 Ohm, only a layout resistor...

Thank you, I can cut the line there and don't need to deactivate pin #16
(that procedure have the same consequence but looks better on the circuit)

I try the Pros & Cons... ;)
 
  • Like
Reactions: edwardgeo
Here my latest GOP (this means you'll need OpenCore in order to have a boot picker) VBIOSes for all AMD Polaris cards.
Those VBIOSes have been tested for a lot of time now, with the exception of the VRAM configurations, which appeared only recently and were collected by @Ausdauersportler and @edwardgeo. (That means that effectively only the WX4130 and RX480 VBIOSes have been tested in their final state!)
They should allow for proper boosting and throttling, before and after sleep, as well as in Windows. However, boosting and throttling may behave differently after the first time waking your iMac from sleep.

The WX4130 and both WX4150 types differ in VRAM configuration only. Shaders are unlocked on the WX4130 so it should perform the same as the WX4150.

The WX4170 VBIOSes differ from the WX4150 ones only in terms of higher GPU clock - but the WX4170 has more shaders available in hardware, so you may use the WX4150 VBIOSes on your WX4170 if you want it to stay cooler (you'll not loose the additional shaders).

The _ALT_VRAM variants have the VRAM configuration from the VBIOS provided by @edwardgeo - you should try those if you experience issues with the "normal" one.

Please report your experiences with the attached VBIOSes!
For an official release we should sort out whether or not we need both VRAM configurations for the WX4150 and WX4170.
Thank you so much, internetzel!

My chronic WX7100 boost problems (See Post 6743) are gone since flashing your new GOP ROM!
The Card now accelerates as it should...

By contrast, as I still do not reach the desired Valley performance, I realized that my CPU throttles all the time.
Unfortunately I cannot tell since when, as i swapped an i7-2600 and a Xeon 1290 forth and back again.

I tried virtually everything I can imagine:
- installed OWC In-line Digital Thermal Sensor Cable for upgraded Internal HDD together with SSD
- checked every sensor cable at least 20 Times
- performed SMC Reset multiple times
- checked the two ferrite beads near the ambient temperature sensor connector (L5500 & L5501) -> no apparent prob
- Macs Fan control gives reasonable temp values for all sensors I am familiar with
- downloaded AHT via this Github PAge and installed it on OC USB Stick, ran it, found 4MOT/4/40000003: HDD-1301 -> changed the HDD fan -> no change at all
- in despair, I compiled a SSDT-PM.aml according to dortania -> no change

CPU is stuck at 1.8 GHz, so helpfull suggestions would be warmly appreciated...


Best regards,

roscho
 
Thank you so much, internetzel!

My chronic WX7100 boost problems (See Post 6743) are gone since flashing your new GOP ROM!
The Card now accelerates as it should...

By contrast, as I still do not reach the desired Valley performance, I realized that my CPU throttles all the time.
Unfortunately I cannot tell since when, as i swapped an i7-2600 and a Xeon 1290 forth and back again.

I tried virtually everything I can imagine:
- installed OWC In-line Digital Thermal Sensor Cable for upgraded Internal HDD together with SSD
- checked every sensor cable at least 20 Times
- performed SMC Reset multiple times
- checked the two ferrite beads near the ambient temperature sensor connector (L5500 & L5501) -> no apparent prob
- Macs Fan control gives reasonable temp values for all sensors I am familiar with
- downloaded AHT via this Github PAge and installed it on OC USB Stick, ran it, found 4MOT/4/40000003: HDD-1301 -> changed the HDD fan -> no change at all
- in despair, I compiled a SSDT-PM.aml according to dortania -> no change

CPU is stuck at 1.8 GHz, so helpfull suggestions would be warmly appreciated...


Best regards,

roscho
Run HW Monitor (and only this tool) to check the temp readings of all sensors. If you see more than the GPU and GPU environment on 129C you found the culprit.

Running ASD may be another option. Download links are hard to find…
 
Run HW Monitor (and only this tool) to check the temp readings of all sensors. If you see more than the GPU and GPU environment on 129C you found the culprit.

Running ASD may be another option. Download links are hard to find…
Option-D after the chime, ethernet boot, downloads and runs AHT.
Cheers
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.