Issues with Windows SDK v0.9.49 ?

[Francis.Jambon]

I have some issues with the Windows SDK v0.9.49 (with Windows 7 Pro 64) :
- most of the time, the calibration process do not end: the calibration stops at the last point
- the eye-tracker disconnects after about 10-30 minutes

I downgrade to version 0.9.41 and everything is back in order. I also performed tests with another Eye-Tracker, with another USB 3.0 chipset (Fresco Logic FL1100 and NEC uPD200201) with similar results.

Does anyone run into the same problem?

Francis

[Martin]

Hi Francis,

Could you please past the output of the log located in
C:\Users\MYUSERNAME\AppData\Local\EyeTribe

I have some good news on the stream-stopping on older USB 3.0 controllers. There's been a lot of work on the firmware to accommodate these. We now have a first alpha of a firmware that keeps the stream running indefinitely. In our internal testing it has been tested with the NEC / Renesas controllers that has been very problematic in the past and it seems to just work. We are preparing the firmware and flashing tool for release. I will send you a link once it has reached a beta stage.

Is anyone else experiencing that the calibration hangs with the 0.9.49? I will double check internally to see if there's been any reports on it.

[Francis.Jambon]

Hi Martin,

Thank you for you reply.

I performed more tests today.

About disconnections, the issue seems significantly less frequent with another USB 3.0 chipset (NEC UPD200202 chipset PCIe card with up to date firmware and driver) but it is still present. At the time of disconnection, the logs usually indicates "tetrode:2254". I send you by PM the logs file.

The problem with calibration that do not ends seem to append when the calibration process is difficult (for instance with poor lightning or dirty glasses), but there is no clear indication of any problem in the logs. The symptom is that the circle in the center of the cross stops changing size. The calibration could be ended withy the escape key with no problem. In the logs I send you, there is three calibration attempts. I think this is not related to USB 3.0 issues, but linked to something like uncached error situations in the calibration algorithm.

I am ready to test the new firmware once it is released !

Sincerely,

Francis

[Rafael]

I have some issues with the Windows SDK v0.9.49 (with Windows 7 Pro 64) :
- most of the time, the calibration process do not end: the calibration stops at the last point
- the eye-tracker disconnects after about 10-30 minutes

Hi,

We have the same setup here (SDK 0.9.49 and Win 7 Pro 64), and the system has a Renesas USB 3.0 controller. Calibration runs fine, but the unit is also disconnecting after about a few minutes (sometimes 2 minutes, sometimes it lasts up to 60 minutes).

Could we have access to the beta firmware for testing purposes, please? Also, Francis mentioned the problem disappeared downgrading to SDK 0.9.41. Could we have access to that version, please?

Thanks!

[Martin]

Hi Rafael,

Check your email for a beta

We're preparing a general release for the firmware updater with a new firmware version. Seems to do the trick on NEC controllers.

[JeffKang]

Calibration stopping at the last point

The calibration stopping at the last point just happened to me a couple times today.
The freeze was during 2 attempts to send feedback data, which I had successfully done several days ago.
It has gone away for now though.

USB 3 error

Also today, I got a “The Eye Tribe Tracker must be connected to a USB 3 host” error, which is remedied with a device manager uninstallation of the unknown/unrecognized device, and a replugging of the USB cable.
This is an old error that I don’t think people get anymore, and I only get this on occasion.
I don’t think this has anything to do with the stopping of the calibration process.
Although, when I pressed escape to exit out of one of the 2 stuck calibrations, I do remember that the Eye Tribe UI animation eyes were missing, and there was a message that said either that the device is not connected, or not working.
In a log file of today’s “The Eye Tribe Tracker must be connected to a USB 3 host” error, there are entries of “The Eye Tribe Tracker is not working”, and “The tracker device has been connected but is not working”.
But again, I don’t know if my USB issue from today has anything to do with the freezing calibration from today.

Glasses

The problem with calibration that do not ends seem to append when the calibration process is difficult (for instance with poor lightning or dirty glasses), but there is no clear indication of any problem in the logs. The symptom is that the circle in the center of the cross stops changing size. The calibration could be ended withy the escape key with no problem.

It’s interesting that you mentioned that because it’s only this week that I tried sending feedback with my glasses on.
I’m currently at 2 successful attempts for 4 total attempts (2 calibrations that stopped) for sending feedback data with my glasses on.

[Francis.Jambon]

An update from my 20 January 2015 post,

With strictly the same hardware and software configuration (and same guy and glasses ), I was not able to reproduce the USB 3.0 issue with SDK v0.9.49. Now the devices (Firmware 293 as well as Firmware 203_2) work without disconnections. I test them individually more than 4 hours continuously. The computer was not modified except automatic Windows updates (it is used only for eye-tracking) . Sometimes, the calibration takes more time than expected, but the last point small inner circle never stops blinking.

Good news! The new Firmware 203_2 solve the problem for the 203 firmware, but the issue, which was also present with the firmware 293, disappear for it and I don't know exactly why.

[Martin]

Cool

If there's anything else on the new firmware or the flasher tool, please use this thread.