It happened to me on my 3-rd card.
Suddenly after several days of working with USB ports, that includes connecting and disconnecting of devices, USB ports stopped detecting devices connection.
I do not see any new device under /dev/ upon connection any more.
The devices are working on same format on XU4 card without any problems, including connect and disconnect many times a day.
It looks like hardware problem, probably the card’s USB hub hw is not protected enough.
Which OS are you running ? Debian ?
Do you still have 5V on USB VBUS pin (1) ?
Is there any message in dmesg on device connection?
Hi @Loic,
I’m running Debian build 215.
No 5V on USB terminals. Both USB seems dead.
No error message on USB connection in dmesg.
From the schematics, there are per-port over-current protection thanks to NX5P1100UK chips. These chips are designed to isolate a VBUS voltage source from a VBUS interface pin automatically when a fault condition occurs. It would be weird that both of them are damaged. VREG_5V is used as input for these circuits, so maybe the problem is here. Could you check VREG_5V (+5V) on your board (available on Low speed connector, pin 37).
5V exists on this pin.
Some hardware between USB connector and the VREG_5V on low speed connector ?
The circuit schematics are here:
https://developer.qualcomm.com/download/db820c/dragonboard-820c-schematics.pdf
I had a similar issue here, with a brand new DB820c (it worked fine for a couple of hours). The USB 3.0 Hub Controller IC is overheating and no USB port is working anymore. It’s a hardware defect, in my case. It’s sad.
This is exactly what i have on my second card.
Same issue.
Yes, exactly. This IC (VL812) is getting very hot as soon as the DB820c is powered on, and the USB ports no longer work.
I remember reading a post somewhere, I think by ndec, that you should not switch S1 with any USB devices plugged in. Presumably powered ones. I will try to find the post.
It was in the installation documents, not a post: https://www.96boards.org/documentation/consumer/dragonboard820c/installation/board-recovery.md.html#using-usb-flashing-tools
Warning: putting the device into EDL mode while the device is powered or when USB is plugged can cause damage to the board, the actual root cause is under investigation, and a hardware fix will be implemented in newer board revision.
Hello, Loic.
I’m reading “MP25-P2214-2 REV A” , “25-P2214-P1 REV A” and “PWB, MAIN, P4, DB820C, SBC, APQ8096”. I’m not sure which of them means the hardware revision.
Thanks.
I do not have the problem described in this thread (yet 
), but I have always taken that warning seriously. But, just for comparison, this is what I have found on my board:
Stickers
Job# 4941523-H100
Rev: D
COO: US
N10N3DNRT
Silk Screen
PWB, MAIN, P4, DB820C, SBC, APQ8096
MP25-P2214-2 REV A
25-P2214-P1 REV A
OPC E-08 94VO
RL.36 16(03)
MADE IN CHINAIn my specific case, the DB820c worked for a while (a couple of hours) and suddenly its USB ports stopped working in normal use (not EDL mode). While investigating the issue, I noticed that the VL812 IC was overheating abnormally.
I do not think its is specifically EDL mode that causes the problem but any change of S1. For instance, did you try booting from an SD card?
No, I didn’t. I’ve used the fastboot method for flashing the DB820c.
The issues occurred much later. I do use USB devices such as USB wireless keyboard dongle and USB flash drive (thumb drive).
Anyway, If the DB820c is so sensitive to the S1 switch positions, I believe this should be fixed in the hardware design. User interface should never harm a product, I my opinion.
In my case:
Stickers
Job# 4941523-H100 **
** Rev: D
** COO: US**
Silk Screen:
PWB, MAIN, P4, DB820C, SBC, APQ8096
MP25-P2214-2 REV A
25-P2214 - P1 REV A
