The issue was not occurred when I used the HDMI at first. as long as I connected the monitor through a USB-C to DP adapter (straight SS/DP lane adapter, no other USB/TB3 device connected on same port), the issue starting surface. The DP cable is the cable from previous PC with a DP1.2 output and screen only at 1920x1200@59/60Hz. tried another brand new DP cable, same result.
When it's happen, the screen just go black for 1-2 sec, then come back. all icon position remain same. So, I assume the Monitor is still there just lost sync to DP Video signal briefly.
I read a lot of article online, most of then seems point to adapter, DP cable, and USB-C/TB hub firmware. but this is a simple lane adapter, should not has that FW issue. According to my observation form past few days. the issue is not really random. It's seems link to USB activities most of time some how. it's happens most of time when I plug/unplug USB device from front USB-A port, enable/disable BT, even when disabled/enabled the built in digital Mic etc. I even tried set the TB3 port to DP++ sink mode, does not make much difference.
Maybe this issue is the USB-C MUX between those 4 SS and DP lanes related to other USB Hub activities? Does anyone has same issue? Thanks.
I am actually a little surprised by what you are saying; I am used to seeing folks running into this problem when using the HDMI interface, not the DisplayPort (DP) interface.
I am not sure your conclusion makes sense, however. The TBT3 USB-C connector consists of three independent sections, (1) 4-lane PCIe 3.0 interface, (2) 8-lane DP 1.2 interface and (3) USB 3.1 interface. Again, these are supposed to be completely independent.
Have you looked closely at your cables? The so-called "Standard" DP Cable may support HBR2 (i.e. DP 1.2) or it may only support HBR (DP 1.0). Make sure you are purchasing cables that specifically say that they support HBR2 (or DP 1.2). Now, if the connected monitor is only 2K capable, this is moot; the necessary bandwidth was there even in DP 1.0.
Finally, regarding the USB-C-to-DP dongle, well, you get what you pay for. I have seen good ones and I have seen awful ones (it miffs me to see this when the prices are so closely aligned). I would try another one before coming to any conclusions. You could also try using a USB-C-to-HDMI adapter. There are good ones that can support 4K@60Hz (but don't expect any capabilities that aren't expressly claimed and especially watch for adapters that require DP 1.4 support before they support 4K@60Hz.
Have you tried running the latest Beta graphics driver? That is, version 188.8.131.5225 , which is available here: https://downloadcenter.intel.com/download/29679/Intel-Graphics-BETA-Windows-10-DCH-Drivers.
Thanks for your response.
I think for a USB-C to HDMI, it need active component in adapter which may cause FW issue etc. USB-C to DP should be a most passive bidirectional connection.
I am definitely not expert on this. from my reading USB-C has 4 SS lanes and 1 USB2.0 lane for USB function. 4 SS lanes can be all or partially used by DP (through MUX). not sure if USB-C support PCI-E. TB does of course. A individual lane can be only use by DP or SSRX/SSTX data at a given time, not both, I assume that what Alt means.
That cable never had problem in pass few years between another intel DP1.2 iGPU to same monitor. So I suspect cable may not the issue here. again, my monitor only run 1920x1200@60Hz.
I don't want exclude the factor of adapter, but the timing of it occurs every time is too coincidence with the USB device plug/unplug event to ignore. I say 85% of it. that's the reason I suspect the MUX. Just a guess.
I am still doubtful of your conclusion. This same set of symptoms seems to come up with some regularity and the DP (USB-C) interface is most often not even in the picture. Is there any way you could video this that would show this coordination between the graphics outages and the USB activity?
Regardless, one of the Intel Customer Support engineers will need to step in here and investigate. Because I know that they will ask, let's save some time. Please download and run the Intel System Support Utility for Windows and have it save the report to a file. Then, using the Drag and drop here or browse files to attach dialog below the Body edit box of your response post, upload and attach this file to the response post.
In another conversation, it was suggested that there could be an issue related to USB power management that can cause similar symptoms on the display. You could try following the "solution" suggested here: https://community.intel.com/t5/Intel-NUCs/NUC8I5BEH-random-system-halt-with-flicker-of-screen-ocassi... and see if it makes a difference.
Ok, replaced with brand new DELL monitor, a new DP to mDP cable, then replaced the adapter and DP cable altogether with a "plugable" integrated USB-C to DP cable. No difference.
Only one thing works. as scott said, I disabled the " allow the computer to turn off this device to save power" option on one of the 'Intel(R) USB 3.1 eXtensible Host Controller - 1.10(Microsoft)' (assume it's the one which control to the rare TB/USB-C port).
So seems like windows is try to turn off that USB-C controller to save power while there is USB activities on any USB port regardless which controller it connected to... the workaround does solved the issue. but this can't be normal behaver right?
Update, looks like don't need disable PM on that USB 3.1 controller. disabled the PM on the "USB Root Hub (USB 3.0)" which connected to that controller is also works.
Yes, sorry, I *am* wrong (too used to looking at the previous generation PCHs
The USB 3.0 Root Hub is below the USB 3.2 xHCI Controller. [Aside: Why this USB 3.2 xHCI controller manifests as a USB 3.1 Root Hub (not a USB 3.2 Root Hub) is an issue for another day.]
USB IF love rebranding USB standard. USB 3.2 2x1 for this assume.
back to the topic, I suspect the issue maybe to do with TPS65982 USB PD controller. Tried a slight newer(few days) driver from Lenovo, does not made any difference.
Well. as you said, only Intel engineer knows where the issue lies.