STM CircleOS forum / Unstability of the USB connection

Francis · 2009-02-20 11:19:35

Sjoerd,
Your evolution of CircleOS is very nice.
I tried this version on two different Primers, and I have an unstability. It seems that the USB connection is somehow unstable, and I see (under Windows) the status of the icon 'toggling'. Sometimes, a second explorer is opened... I was wondering wether the problem is caused by the switch mechanism we added on the USB (copied from the ST schematic of their own evaluation board). Or would it be software ? I will try to uinderstand.

Francis · 2009-02-20 11:30:52

The management of the disconnection looks good. It could be the quality of the routing that is not perfect for USB (the signals go from the bottom of the board to the top...).

sjoerd · 2009-02-20 12:06:26

I've tried it on two Primer2 units and it seemed to work fine.

I didn't do any stress testing and I haven't checked how well debouncing of the signals is handled.

Francis · 2009-02-20 13:13:41

I tried to increase the cpu clock before launching USB-MS, but it does not change anything. I also tried to use an intermediate hub, but I have still the problem.
I use your CircleOS V3.6. Typically, when connected, a new explorer window is opened every 30s... But when I use one of these windows (file copy/delete/move,...) everything works fine (you implemented a very useful feature).
Moreover (not critical) I have seen that the menu management is more or less broken when USB-MS has been launched. I will check this.

Francis · 2009-02-26 08:12:00

Sjoerd,

I tried to disable the timer2 interrupt, and it works... The Windows Explorer is now stable. The Timer2 interrupt is used to get thwe MEMS information (and to communication with the audio chip through the I2C port). We set this irq at the highest priority level because the I2C communication is quite delicate (it fails when it is interrupted). But USB seems delicate too...
The solution consists in disabling the Timer2 interrupt (audio and MEMS) whilst the MS feature is enabled. This is acceptable...

Francis

sjoerd · 2009-02-26 08:39:28

Francis,

Good to hear you've found the cause. It's sounds a bit strange that the two interrupts can't work simultaneously. I can imagine that the I2C communication shouldn't be interrupted but in case of USB audio things might get a bit tricky.
For now I agree that disabling Timer2 interrupt when USB MS is active is a acceptable workaround.

# 1 2009-02-20 11:19:35 Unstability of the USB connection (project: USB Mass storage 1.1)

Unstability of the USB connection (project: USB Mass storage 1.1)

# 2 2009-02-20 11:30:52 Unstability of the USB connection (project: USB Mass storage 1.1)

Re: Unstability of the USB connection (project: USB Mass storage 1.1)

# 3 2009-02-20 12:06:26 Unstability of the USB connection (project: USB Mass storage 1.1)

Re: Unstability of the USB connection (project: USB Mass storage 1.1)

# 4 2009-02-20 13:13:41 Unstability of the USB connection (project: USB Mass storage 1.1)

Re: Unstability of the USB connection (project: USB Mass storage 1.1)

# 5 2009-02-26 08:12:00 Unstability of the USB connection (project: USB Mass storage 1.1)

Re: Unstability of the USB connection (project: USB Mass storage 1.1)

# 6 2009-02-26 08:39:28 Unstability of the USB connection (project: USB Mass storage 1.1)

Re: Unstability of the USB connection (project: USB Mass storage 1.1)

Board footer