Some help please - smcFanControl issue

I could use some help. I have a first generation Retina MacBook Pro, 15" 2.7 GHz and I'm in a dispute with Apple.

If you could run Macs Fan Control and tell me what the minimum and maximum speeds of your left and right fan speed _default_ values are, I'd really appreciate it.

Also keep an eye on the CPU 1 and CPU proximity sensors and see if they vary much. I see up to a 30ºC difference.

Many thanks,
Mark

 

Long story. I've been using smcFanControl for years along with Intel Power Gadget, the temperatures closely agreeing. My screen started to go and I noticed discrepancies between the two temperatures of up to 30ºC. The monitor became unreadable and I sent the Mac in for repair.

When the Mac came back with only the clamshell replaced and 20º to 30º discrepancies between the CPU temperature sensor and the Proximity sensor remaining. That's when I started using Macs Fan Control. Left side fan min 2160, max 5940. Right side fan min 2000, max 5499.

I figured the symptom had been cured without the underlying problem being fixed because of the discrepancy between the two CPU temperatures remaining where prior to the problem they agreed within ~5ºC.

After a LOT of arguments between second tier tech support who conferred with engineering, I was told not to worry about the temperature discrepancy with no explanation for it.

I complained to Customer Service and sent in the computer again. According to the paperwork the logic board was replaced, the SSD was replaced, both fans were replaced, Cable I/O Flex was replaced, Cable I/O Right was replaced and the heatsink was replaced.

When I got the computer back I saw precisely the same fan speeds, both minimum and maximum, and the same discrepancy (20º to 30ºC under load) between the CPU temperature and the CPU proximity sensor. In other words no difference. Even the fan speeds were the same to four significant digits.

I suspect all they did was erase my SSD and send me back the computer.

 

They didn't replace the logic board when the display went, just the clamshell.

Prior to the display going bad, smcFanControl and Intel Power Gadget agreed within a few degrees. After the display failure I started seeing a difference of 20ºC to 30ºC. Operations that prior to the repair would trigger the fans, would no longer trigger the fans.

For instance, running Geekbench 3 would show a temperature of 95ºC with the fans going at a high speed a few months ago, now I see a temperature of 58ºC with no increase in fan speed. To add insult to injury, I get a lower Geekbench score because the i7 detects it's overheating and throttles the clock speed down.

That's why I started using Macs Fan Control, since it would show both the proximity sensor and the CPU sensor (and allow me to choose which sensor to use to trigger the fans).

 

The CPU sensors on the die have nothing to do with Apple, they're used by the CPU for its own purposes including to prevent itself from overheating by slowing the clock. The CPU Proximity sensor is the one supplied by Apple and the one the SMC uses to control the fans.

I suspect this sensor of failing.

Prior to the display failing the two sensors were usually pretty close, after the display failed (before and after the repair) the proximity sensor would show so low that the fans would not engage under load.

 

Here's a comparison showing Macs Fan Control vs just the OS X default using the CPU Proximity sensor fan control while running 10 iterations of the Geekbench stress test:

Note that the final Geekbench score is in the range of a dual core i7 MacBook Air rather than a quad core i7 rMBP and took over 7 minutes longer to finish when the CPU depends on the SMC for cooling. It really looks like a bad CPU proximity sensor not causing the fans to spin up to speed and relying on the i7's thermal self protection to drop the clock speed to keep from overheating.

Mark