Fairly certain that I am impacted by this. s-tui shows around 3600MHz sustained, which I can bump to ~3750 sustained if I undervolt a bit. I would happily swap to PTM but I am in no way comfortable screwing around with the liquid metal. I did the RTC replacement on my 11th gen boards and did the HDMI card rework on one of my two boards, so I have some comfort level mucking around. Liquid metal seems like a bridge too far - one slip-up and goodbye board.
I am not really interested in swapping out for a different board that still has the LM interface. If I go down the path of replacing the interface I am likely to ship the machine off to a qualified third party shop to do the work. This frustrates me. I really like all of my machines, but choosing to use LM feels like a miss, particularly since the whole heat mitigation solution does not seem to be working.
I may reach out to Support simply to be on record as having the issue. I would really like it if Framework would take back these systems and swap the LM on their dime and at their risk. It’s more about the risk than the cost.
This charger definitely does not provide 180w on a single port so they might complain a bit about that before paying attention to thermal throttling. It does look like it might be one of the reasons it’s not boosting to a higher wattage out of the gate but it’s hard to tell from the limited info.
In a CPU only Scenario the Charger does not impact the CPU Performance drasticly. He IS hitting Thermal Limit at under 40W TDP, So Yes the System is affected.
I Can run my System on my 90W PD Charger integrated in my Screens Connection and i am still hitting above 50W. There is no Peak above 70w when doing so i am peaking at 57W, but i am still running the maximum Performance set for 54w Slow PPT.
@Machine The 180w Charger is not needed if you are running a dGPUless System.
Okay. I was seeing limits on a 100w charger with only CPU load so I may have had something misconfigured at the time.
Also @Machine it definitely looks like it’s thermal throttling. I just thought the charger was causing it to be limited on the initial boost wattage which doesn’t really matter once heat is limiting it instead.
Well i am fully with you. It is reducing the max Performance. But as @Machine machine is already thermal throttling under 40w TDP the charger isn’t even a noticable change.
I am always using X86 Universal Tuning Utility to overide any PPT Settings by Framework or Windows.
And with the Extreme Preset i am getting the Values as mentioned before with the 90w Charger.
It’s worth it if you get lucky and don’t kill a board.
This is a 10 min run. It appeared to be holding a consistent 60w in hwinfo during the test (next time I need to try to reset it right as it starts to get an accurate average reading)
Just making an update after ~1.5 months that my performance also did drop. Went from 54W sustained to 48W, Delta went from 6C to 8C, and my score dropped from 15.7k to 14.8k, or a 11% decrease . I should try the repasting tips in this thread tbh.
For those undervolting the CPU, how do you do it? I have the 7940HS and couldn’t get it to cooperate with the Universal Tuning Utility, but maybe it’s been updated since I last tried. Or my 7940HS just doesn’t run well when undervolted.
I am on linux, using ryzenadj. I assume that it is working as I can see the operating frequency change when I change the value, and when I try to set it too aggressively, the machine restarts.
Yes. I am using X86 Universal Tuning Utility with per-core curve optimizations for undervolting. That won’t really change the thermal limits of the cooler, but it does increase the score in CB23 a little. I do also try to more aggressively tune the cores that are more hot than others to avoid overall thermal throttling at lower clocks.
Before the conductonaut though it did almost no good.
. I should try the repasting tips in this thread tbh.
