LPC4337 USB0 failure: suddenly disable

[lpcware]

Content originally posted in LPCWare by andres.cass on Tue May 17 09:54:18 MST 2016
Hello.
I am working with LPC4337 trying to make a CAN-USB translater. I use as base the vcom example, configuring the LPC4337 USB0 as virtual serial port. It was working very well but suddenly the PC stop recognicing the virtual COM. I am still able to connect to the LPC4337 through JTAG or SWD (LPC-Link II), download a firmware and debugging but the PC still not recognicing it as a virtual COM, not even as an "Unrecognized device". It is like there is nothing connected to the PC.

It's happened with two LPC4337, I want to know what could caused it and if there is any way to solve it.

Thank you.

[andrescass]

Hi. I have tried the software in an eval board and it is working fiine. I think I will try to replace the mcu part. I can't figure out anything else could happened

[andrescass]

Hi. Thank for the answer.

I will sumarize the problem.

I am workinig in a board designed by my self with the LPC4337JBD144 uC using USB0 port.

First I put series 33ohm resistors in DM and DP with 15Kohm pulldown. With that setup the USB was working fine with the example and with my own code, wich use the cdc example. But suddenly while I was debugging LPCXpresso threw an error about some register and stop the debugging session. After that USB never works again. Windows doesn't even recognice a device attached. Nothing.

Then I read the user manual and see that for USB0 it is not necessary tu put the series resistors so I took them off but the USB0 is not working.

I tried also the cdc example again but it doesn't work.

The other peripherials of the microcontroller are working fine. In the project I am using a CAN port and it works well. But the USB doesn't work.

It is posible that the USB0 phy could be damaged and the rest of the microcontroller still works good?

[brendonslade]

It is possible that a device can be damaged in a way that a single peripheral dies but other parts work (e.g. if an internal bond wire gets burnt out).

Its very hard to diagnose what might be wrong given that it is your software and hardware design. Do you have an eval board to try and in order to eliminate a PC or LPCXpresso IDE issue?

[brendonslade]

Hi Andre,

You should see Windows enumerate several devices when the board is plugged in via the J5 USB connector. One of these will be a VCOM serial port for the debug probe (i.e. the LPC4322 device on the board); if you have the correct drivers installed this will show up under Ports (COM & LPT) > LPC-LinkII UCom Port (COMxx) when xx will be whatever number Windows decided to use. When you run the cdc examples a different driver will be picked up, but this is reasonable sanity check.

How about trying using a power USB hub? The LPC43xx devices need a fair amount of current when running at full speed, and with the debug probe also running some laptops wont provide this. If you don't have one at hand you can try only plugging in the target connection (J4) and not the debug probe (J5) after flashing the code, thereby removing the latter from the power drain... if the demo starts to work then that is almost certainly your issue.

I did just try my LPCXpresso4337 and the example works for me, with both drivers appearing. I'm using a docking station.

regards

Brendon

[andrescass]

I write again with the hope of a solution.

It's all equal than before, but I've noticed that I had put 33 ohm series resistance with DM and DP and today reading the user manual I found that USB0 doesn't need the series resistances, so I took them of and connected directly DM and DP to the connector. But nothing's changed. The computer still not recognicing anything, it's just as if nothing is connected.

Is there the posibility that the phy has broke but the microcontroller still working?

The problem start when I was debugging the application, sudenlly LPCXpresso trhow an error pop-pu saying aomething about a register and then never connected again.

Thank you

[lpcware]

bump

[lpcware]

Content originally posted in LPCWare by andres.cass on Tue May 24 13:24:54 MST 2016
No. It was just a USB-CAN transeiver.
But I've tried the cdc_vcom example directly from LPCOpen and it acts equally. Not only nothing is happening in the PC, no device recognized as connected, the current consuption get increased to 200 mA.
First I thought that the PLL could be failing, but it worked well for a while (and the "while" was about four hours) and then stop working while I was debuging. It is any posibility that the USB phy get broken and the mcu still working?

[lpcware]

Content originally posted in LPCWare by bavarian on Tue May 24 09:13:05 MST 2016
I can't think of a good reason why it should work for some time, then stops and never works again. I could think of a temperature effect which lets the ref clock drift away, but this would mean that it starts working again when cooling it down.

Any possibility that your code is doing something strange after a while? Switch off clocks, change frequency settings, go to power down etc?

Regards,
NXP Support Team.

[lpcware]

Content originally posted in LPCWare by andres.cass on Tue May 24 07:04:38 MST 2016
The prblem is that the PC doesn't even recognize a device connected. It does not appears as an unrecogniced device or anything. Is like it is not attached to the PC. I tried another board with the same code and it worked well until it stoped worked too. That's why I think is a LPC4337 problem

[lpcware]

Content originally posted in LPCWare by bavarian on Mon May 23 02:22:46 MST 2016
This is most likely not related to the LPC4337 hardware but to the driver on the PC side. Make a crosscheck with another PC which has not been connected before.
There are a few versions of the Windows drivers, partly as install packages but also included in installations of LPCXpresso and LPCScrypt, so it's hard to say which is the right one.
Getting rid of a faulty VCOM detection on my PC by removing inf files etc resulted more than one time in a blue screen at PC startup and a time consuming reconstruction of the configuration, so better be careful with that, in case this is your PC with all things you need for you daily work ;-)

Regards,
NXP Support Team