Acer Aspire 4830t, help with diagnostics PLTRST impedance

[Yousuf]

Hi All,
I have interesting diagnostics experience, with Aspire 4830t motherboard. All the voltage rails are up (except for integrated graphics, which I think happens later in the boot sequence). The PCH_PWROK and DRAM_PWROK are asserted. I can also see PCH reading the BIOS and fetching (and executing) the ME_Firmware. But I don't see PLT_RST# being deasserted (I always see it driven logic low). All conditions seems to be correct for platform reset but that signal is never driven high. Out of desperation, I changed the CPU and still see the same issue.

Out of curiosity, I measured the DC resistance of PLT_RST to ground and I see low values like 5ohms. The specification says, it can sink/source about 2mA or thereabouts. So it looks like a short (or close) and started removing the buffers and even the SIO on that net. But still no luck, it is low. What do you think ? Do you have any measurements off hand or can you do one with the board that you are currently debugging ? Thanks for all your tips.

[Yousuf]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

Uploading the schematic.

[Sephir0th]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

It is a clear short to GND, so basically pretty simple to solve. If injecting doesn't help you can of course start removing component by component. You even shared the schematics so I really wonder where you're stuck. I had cases where PLT_RST# was shorted caused by a bad Thunderbolt controller for example.

So yes, you have also to rule out the bigger ICs, like the USB-Controller U35 and also the EC and of course everything else which is connected to this logic line.

Good luck.

[Yousuf]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

Thanks Sephir0th. I just wanted some confirmation, before I start removing more ICs on that net. I can see an ethernet controller, USB3 controller on the net which I have still not removed. I will do that now and see if that helps. Thanks a lot for sharing your experience, gives me more reason to remove the components

[Sephir0th]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

Don't worry I've asked the same question before. Luckily I was bored and started the task and was willing to proceed till I would have reached the CPU. Luckily it was the last IC to desolder, the thunderbolt controller...

Please report your findings

[Yousuf]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

HI,
Thanks, I did try and removed the controllers one by one from the PLT_RST rail and of course, it is always the last one that was bad. In my case, it was the ethernet controller which was given PLT_RST, but luckily there was a jumper (0-ohm resistor) that I could remove, and bang PLT_RST is high impedance. So after soldering back Super I/O (yes I could have removed it last, but I didn't), I could see the board booting, PLT_RST going high, and DDR3_RST also going high, but not a whole lot of progress after that.

I changed to a different processor, but still the same. I could try updating the BIOS and removing the ethernet controller and see if that helps, but otherwise, I am clueless about how to proceed.

[Sephir0th]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

The board went through the whole sequence. If it still doesn't P.O.S.T it is hopefully a BIOS issue then. Otherwise there are only internal failures of the BGAs left which are almost impossible to diagnose.

So what I guess we're still at no picture? Check with external keyboard whether CapsLock is responding and remove the LAN Controller as a last resort, then start reprogramming the BIOS. Try and report please.

[piernov]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

If Ethernet controller is bad there is no reason to keep it on the board. It can still interfere with the PCIe bus.

[Yousuf]

Re: Acer Aspire 4830t, help with diagnostics PLTRST impedance

[The rabbit hole deepens]

Hi All,
Apologies for getting back late. Thanks to Piernov and Sephir0th, I have removed the ethernet controller but the behaviour is the same. I went back and updated with a couple of different BIOSes available on the Acer website. But the behavior is the same, - ` no screen (no backlight) and no beeps.

I went a little deeper and looked at the spiflash transactions. I can see that the CPU is executing the ME code and then follows the reset vector and executes the various executables present in the firmware volume. By comparing the bios address from SPI bus to the module address provided by UEFITool, I can see that the code executes till MemoryInit. In some cases, the last thing I see on SPI is access to PchSmbArpDisabled module. Even though the different BIOSes have tried, shows different sequences, it more or less stops while executing MemoryInit or PchSmbArpDisabled. Could it be that the software assumes the presence of (the removed) ethernet controller and probably crashes because it couldn't find? But thinking from the other side, I will assume that the code has to enumerate this PCI device and do operations only if it is present.

Yet another observation is that I see DDR3_RESET going high (out of reset) event before PLT_RST going high. It seems weird because I thought the PLTRST would go off and then ddr3 will come out of reset as part of CPU reset. I cannot confirm this with any of the CPU documentation. Do you have any idea what is the expected sequence here?

If anyone of you is interested, I can share the bios trace, Thanks for the help.