https://bugzilla.kernel.org/show_bug.cgi?id=204807 --- Comment #138 from Kamil Pietrzak (kpietrzak@xxxxxxxxxxx) --- (In reply to Kamil Pietrzak from comment #137) > With regard to current voltage scaling factors for nct6798d chip, they are > most likely not correct and probably will require future changes. For > example I can't see +12V and +5V is sensors output when using current > voltage scalling factors. Partially responding to my own comment here. Due to the lack of publicly available documentation for NCT6798D chip I checked docs for similar chips (NCT6791D, NCT6791D). Looks like voltages like +12V and +5V can be connected to any one of general purpose voltage inputs on the SuperIO chip. So on one motherboard +12V can be connected to pin VIN0, but on another one with same SuperIO chip it can possibly be connected to other general purpose voltage pin like VIN1, VIN2, VIN3 etc. In that case it will not be possible to properly scale these voltages without hardcoding motherboard models in module code, so scaling should take place in userspace apps like lm_sensors. The only voltages that can be safely scaled in module code are Vcore, AVSB, 3VCC, 3VSB, VBAT. Pins to which they are connected should not change between different motherboards. So it looks like current voltage scaling factors are as accurate as it can be without hardcoding motherboards models. However, I am still curious about Vcore voltage readings on my TUF Z490 board in BIOS and Asus software. According to docs Vcore should be calculated with formula Detected Voltage = Reading * 0.008 V but Asus in BIOS and in their software on Windows calculate it probably with some additional scaling factor, most likely something like Detected Voltage = Reading * 0.008 V * 1.11 The only reason that comes to my mind for calculating Vcore in that way is that they (Asus) wanted BIOS/software Vcore readings to be more accurate in relation to voltage readings using for example multimeter near the CPU socket. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
