Garmin ANT USB-m support

[Nick Stone]

I recently purchased an ANT+ stick which I intend to use with my Garmin Cadence / Speed sensor for training over the winter. I have downloaded GC 3 on my work laptop (windows) and the ant stick works fine and I can see the sensors etc.

However my home laptop is Linux based (Mint 17) and I seem to have a problem with GC seeing the sensors. If I run lsusb I get the device showing up OK.

Bus 004 Device 002: ID 0fcf:1009 Dynastream Innovations, Inc.

Running dmesg shows the usbserial module loading but I receive a message at the end saying:

usb_serial_simple 2-1:1.0: suunto converter detected
usb 2-1: suunto converter now attached to ttyUSB0

I have come across another forum: https://groups.google.com/forum/#!msg/wattzap-users/fTL3VMNES74/vrWXDHnr_psJ where it appears that the latest drivers may mistakenly be assuming this is a suunto device attached and thus adding driver support incorrectly. The forum suggests that a change was required in the app to fix this.

Does anyone else have this problem / know how to fix it?

[Jon Escombe]

Hi,

What *should* happen under Linux is that GC detaches the usbserial
kernel driver from the stick, and then uses libusb to talk to it natively..

Does the ttyUSB0 entry disappear once you've tried to access the stick
in train mode?

I guess other things to check would be that your GC build includes
libusb support, and that you have sufficient permissions to access the
device.

Regards,
Jon

[Nick Stone]

So the ttyUSB0 device remains visible when in train mode and I have checked the permissions and have added my user to the device group as per some other forum entries

The install I used was from apt so was pre-built. Is there any easy way to find out if libusb support was compiled in?

Thanks

Nick

[Jon Escombe]

It should say in Help -> About -> Version

Cheers,
Jon

[Nick Stone]

I just checked and the about GC has the build information:

libusb was built in

[Jon Escombe]

Hmm that's a puzzle then, I only have the previous (0fcf:1008) stick to
test with, but they should be treated the same way..

Couple of other thoughts;

Does it work if you manually rmmod the usbserial module before running GC?

Do you see any error message if you start GC from a console?

Oh, and worth checking your GC version I suppose. The code to unload the
serial driver is about a year old, so any build from 3.0 SP1 should
include it.

Regards,
Jon

[Nick Stone]

Think you might be on to something - looks like the permissions are the problem - I receive an error from the console:

usb set_configuration Error: could not set config 1: Operation not permitted
check permissions on: "/dev/bus/usb/004/002"

I will check my udev rule and make sure I have got it right

[Nick Stone]

OK permissions were the issue - I have manually gone in and over-ridden them and it all works now

many many thanks for the help....

[Jon Escombe]

Cool, glad it's working..

[Nick Stone]

The issue is that I copied a string from a forum that was incorrect. The start of the entry in udev/rules.d should read:  SUBSYSTEMS=="usb" and on the forum it showed the entry as SUBSYSTEMS="usb" with the result that the permissions were incorrectly set when the key was plugged in

Thanks again