Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
PCH temps on Giga GA-H61MA-D3V
I have a Gigabyte GA-H61MA-D3V - an H61 motherboard, as the name suggests, with the stock F1 BIOS, and am using HWInfo 3.95-1600.

The line for the PCH is blank, but the next lines refer to an IT8728F. I'm not sure if this is a bug, or just the way the temps for the PCH are reported.

Any advice would be great, I've had quite a problem with the PCH heatsink on this board, and although it's fixed now, I'd still like to see the PCH temps.

The attached rar file is the html report


Attached Files
.rar   KESTREL.rar (Size: 12.08 KB / Downloads: 2)
Seeing the Intel PCH sensor without any values means, that the sensor was detected by HWiNFO, but it doesn't report any valid values.
In fact, for the 6- and 7-series chipsets I haven't yet seen a board where the PCH sensor would report values. It's strange, because the PCH itself features a thermal sensor and it should report it's temperature, but it doesn't. Usually the PCH temperature is pushed to the SIO as well, so it might be available from there.
I also have an Intel DH67GD board with i3 CPU used as a media PC. The attachment is HWInfo 3.95-1600, same as the Giga board.


It shows quite a different set of figures, including those from a Nuvoton chip showing a temp for the PCH. The Nuvoton chip itself is described as a 'Legacy IO' device, and includes silicon for a CIR infrared receiver, one of very few boards that do, and a little useless because getting the IR hardware to use with it is nearly impossible. Some ASRock boards use the same chip I believe, so they may also show the PCH measurement. This Intel board definitely shows the PCH temp though, as does the Intel Desktop Utilities which unfortunately won't run on non-Intel boards.

The 63 degrees it's showing isn't great though. The photo of its heatsink is what I found from the factory. I refreshed its heatsink compound and put the stock hs back on, but it seems that has not gone well over time.


There's still no sign of a temp reading from the chipset itself though.

My concern is that if the PCH was actually not cooled properly or at all (talking about the Gigabyte baord) that it'll shorten its lifespan. No big deal, I guess, for a $65 board, but not reassuring. Do the top end UD5 or UD7 boards get the same treatment? I'm not picking on Gigabyte, this just happens to be the oddest case I've seen, in a practice that seems commonplace among manufacturers. And seriously, what percentage of customers remove chipset heatsinks for inspection or upgrading?
In case of the DH67GD board, the Nuvoton chip reports the PCH temperature, because this temperature is routed to this chip and I know which exact temperature register is mapped to it. However, the layout of sensor inputs is vendor specific and every mainboard does this in a different way depending on how particular engineers decide.
That's the reason of such big variance in reporting of sensor values. The problem is that only very few vendors publish this information and for the rest of them, one needs to guess/evaluate/try to determine what's the real sensor layout. Now imagine how many different mainboard models are there and how tough it is to support them without information from vendors.
So in the case of Gigabyte GA-H61MA-D3V board, the PCH temperature might or might not be reported by the SIO/LPC chip and only the manufacturer knows this for sure. Usually if a manufacturer's tool for monitoring and the BIOS doesn't report such temperature, then it's not supported on the mainboard.
You're very quick to answer, Martin, thank you. And thank you for the great utility.

Yet more shoddy behaviour by motherboard manufacturers, not publicising the details you need? I thought Gigabyte might have been different, at least they show a block diagram of their mbs. Crude, but better than nothing.

As an experiment, I copied a few gig of data between USB devices as this raised the temperature of the PCH on the Intel board, moreso than SATA devices, but nothing changed on HWInfo.

That's truly bad form from the manufacturers, I think. Especially when their own efforts look like a 4 year old on LSD designed them.
They usually publish such information only if it would bring them some benefit. It would require to allocate additional resources (people + time) to take care of it... And that means additional effort = money... The same applies to any kind of additional information and nearly any vendor, where a need to know must be justified, and that in most cases = money...
Yep, I hear you. People tell me I'm naive, but I believe good will is worth a lot, and colours a company or an individual. Surely pushing existing design diagrams onto a website wouldn't require a taskforce, though I've worked for a very large company here in Aus, and could just imagine the meetings needed to decide such a thing.

Forum Jump:

Users browsing this thread: 1 Guest(s)