P751DM2-G: installed second M.2 drive, now usb is failing when using HTC vive.

Solved: check post #14

I'm experiencing an issue with my clevo and a high bandwidth usb devices (HTC vive and kinect) after installing a Samsung 960 M.2 EVO 250GB next to the already installed 512GB Samsung SM961 M.2. When using the device for a couple of minutes the usb seems to drop out, tries to reset or fails completely until the system is rebooted.

On some occasions I get this report in device manager: Windows has stopped this device because it has reported problems. (Code 43). A USB port reset request failed. When this happens all USB ports become unusable until I reboot.



On other occasions the vive disconnects without the error in device manager. When that happens I can software reset the vive, but the same problem occurs again after some time.

A few times the system also shuts down (without a BSOD). In the realiability monitor I get this report:


When I remove the 250gb M.2 drive from the system the problem doesn't occur and everything works fine.

The HTC vive is connected to the USB 2 port and HDMI. The problem also occurs on any other usb port.

So I wonder where to look and what to try in resolving this issue.

This is what I tried so far without result:
Formatted the 250gb m.2 drive
Swapped drives
Only the 512gb drive in other slot

Maybe the m.2 drive itself or the laptop has hardware issues, or some kind of conflict?
Or the 960 draws to much power?

 

It's all running through the chipset, if an odd conflict is going on it could cause issues. Have you tried taking an image backup of the 250GB drive, formatting it and then inserting it as blank?

 

I just reformatted the drive as NTFS and did a test, the problem still occurs.

 

Does it change anything if you swap the slot the 512GB is using?

 

Thanks for your suggestions so far!
I just swapped them, and tested, but the problem still occurred.

 

Now if you keep the 512GB in the same slot you moved it to does the problem still persist without the other drive?

 

That...is really strange. The USB and SATA/M.2 lanes only ever converge for the 3G/LTE M.2 slot and the WiFi M.2 slot.
You can try running a Linux ISO, plugging everything in and seeing if it all detects (use the "dmesg" command to see things like USB reset errors).

The only things I can think of from a hardware standpoint is if the PCH is faulty or power delivery is faulty (although it'd be unlikely for 1 M2 slot to work and the other to fail as they're both supplied by the same 3.3V line).

Have you tried swapping the M.2 SSDs around? Then you should be able to see if the same slot fails both time.

Alternatively, you may want try using a USB 3.1 Type-C -> USB 3.0 hub. The USB 3.1 Type-C ports are supplied by the Alpine Ridge chipset which run completely separate to the regular PCH USB controller.

 

Just tested, with only the 512gb in the other slot it works fine.

 

The devices show up fine in different operating systems, the problem starts to occur only when using the vive. I've tested with other high bandwith usb devices (kinect 2, webcams) but then it doesn't occur.

Yeah in different configurations, also with the 250gb formatted to ntfs. The problem doesn't occur when the 250gb drive is removed from the system, and the 512gb drive is in the first or second slot.

 

A small update to this, the problem doesn't only occur with the HTC vive. I just had a longer session 3D scanning with a kinect, and the same problem occurred after some time. When removing the second 250gb drive it didn't occur. It seems that any high bandwidth device triggers the usb to shut down and .reset

 

That is totally strange. My first thought would be power related but they use different voltage feeds and I'd expect everything to fail simultaneously.

Only other things I can think of would be checking for new drivers, new BIOS update (there's several for the P751), re-installing the OS and maybe contact Prema in regards to his BIOS mod (he may also know more about the issue itself).
If you could achieve the same load under Linux you may also get far better error logging as the windows events tend to be garbage for debugging.

Failing that, I suspect you're only left with RMA.

 

The VIVE isn't really hammering the USB connection in terms of data/bandwidth load, but it is continuous usage. Something like an external hard drive would utilise the USB much more - have you tried something like this as well, with the same problems? It's pretty mysterious, I haven't seen a report like this before.

 

It could be the southbridge itself has an issue, very rare but could happen.

 

Solved! My supplier suggested to try the samsung drivers for the NVME drive (I was on the stock windows drivers), and that solved my issue. Thanks for your replies!

You can get the drivers here: http://www.samsung.com/semiconductor/minisite/ssd/download/tools/

 

Thanks for that, must be a low level I/O conflict on the standard drivers. Good to know +rep.