Discussion in 'Dell' started by Spring1898, Apr 16, 2020.
More info on the tdp bug and possible fixes:
The problem for me doesn't seem to be the temps of the components, it's how hot the laptop itself gets.
i.e. earlier my CPU and GPU were both reporting below 60C (with turbo disabled) but the bottom of the laptop was extremely uncomfortable, almost burning my legs. But the air intake for the fans was ice cold (the fans were on)
I've had other laptops (even the G3) where the CPU and GPU reach even higher temps, up to 80C and it's nowhere near as uncomfortable.
It seems to be so inconsistent. There is a random piece of metal on the bottom of the back cover, anyone know that that' for? Seems like insulation would be better if that were removed.
Did you repaste your G5? Was there good die contact with the heatsink? Is the exhaust air not very hot even under load? There could be a defective heatpipe if that is the case.
It is possible there could be a software issue. Something could be holding the 5600M turned on even with no game running. The 5600M should appear grayed out for most sensor readings in HWInfo64 when it is powered off at idle. You would also notice this issue if the battery life is only 1-2 hours. Also if you drained your battery and are plugged in charging, the battery will get warm as it charges up to full. You can customize the charging method by using Dell Power Manager. It will definitely get hot if it is set to the ExpressCharge method.
On a side note, I got my machine swapped with one that does not have a missing heatsink standoff. It also has a bit different spec. I got the 68Wh battery and a 1TB SSD / 16GB RAM. Also, it has an AUO 144hz LCD panel which does not have PWM flicker (Dell part number K055G) which is great.
Strangely Dell has delisted the G5 SE from their site. It would be great if they end up releasing a new revision with improved thermals.
Yeah. Sadly re-pasting didn't help at all. I did tighten the heatsink down, but was scared to tighten it too much in case I cracked the motherboard or something.
In fact scores were lower in Cinebench R20 so maybe I did do a poor job. I did swap out the 2x 4GB 3200C22 RAM for 2x 8GB 2666C16 though, maybe this lowered scores? Scores went down about 1% and temps were the same
Maybe one day if I can be bothered I'll take it apart and see what the spread looks like on the heatsink.
I don't think the RX5600M is stuck running. I do wonder what the point of that metal is though... I was thinking about removing the mesh that reduces the air intake as well, maybe I'll remove that metal too
The heatsink screws should be tightened down until they stop. They have special spring loaded washers to prevent over tightening.
I do think you should have stuck with DDR4-3200 even if its CL22 vs CL16. The bandwidth is ~44000MB/s vs ~37000MB/s and also the infinity fabric clock can not go beyond 2666mhz as well. Also when gaming on the 5600M it copies the framebuffer over from the 5600M dedicated memory into the Vega IGP's area of system ram, just like Nvidia Optimus, so more bandwidth can positively effect this operation. It is part of the reason why the system performs so much better if you hook an external monitor to the HDMI or mDP, then the Vega IGP can just sleep instead of being used as the final output device.
I did repaste mine again with Thermalright TFX this time. It seems to have made my CPU temps better but my GPU temps worse. I may have done a poor job on the GPU die so I may have to take it apart again.
Separate names with a comma.