<< return to Pixycam.com

Does pixy has a whoAmI I2C register?

I just upgraded several pixies to the latest firmware (2.0.5) and im running into a software/hardware issue which was not present in a previous version of firmware. Upon power up, sometimes Pixy is not getting initiated properly. I have Pixy talking to arduino DUE via I2C (and i have several other devices with different addresses on the bus as well). This happens approximately every 1 out of 8 times, the indicator led either lights up to a bright blue state, or barely lights up at all. As a result, pixy is not sending any data back to micro controller. Also, the LED wont change its state as the camera is pointed to known signatures. Im guessing that this is due to some failure to start i2c communication with pixy. Is there a register that i can write to and get some kind of hex number back or whatever so that I know that there is i2C connection between pixy and arduino? If not, as a temporarily solution i can just turn off the power to pixy with a relay and restart it until it works.

Thanks !

Best,
Vadim.

Hmm, it would be good to try to figure out what’s going wrong (instead of just patching the issue with a reset scheme). Can you try some testing with just 1 Pixy attached via i2c and see if you can reproduce the issue? Also, how are the Pixys getting power?

thanks!

I have a USB male connector with just VCC and GND powering pixy (no data lines). The 5v power itself comes from a voltage regulator step up chip, https://www.pololu.com/product/2115
The chip gets power from a 3.7V lipo battery. The voltage regulator can produce LC voltage spikes but i believe that I got that filtered out. Ive been using pixy on my robot for about 5-6 month with the old firmware without this issue. I have 3 devices connected to my i2c bus. I have sparkfun IMU https://www.sparkfun.com/products/12636, Pixy, and arduino FIO. SDA and SCL wires leading to FIO are pulled up with 3.3k resistors.

I have pixy #3, and I will install new firmwhare on it and connected it to DUE and see what happens. Im also wondering what would happen if I unsolder my 3.3k pullup resistors…
I also know that my I2c Lines are not locked up. When the pixy fails to initiate properly, my robot is able to run around and read data from the IMU just fine.

Hey Rich!
The problem is still here, although it happens less frequently now (1 out of 50-60 times), I have uploaded a video on youtube here

Have you tried using a different power supply (for testing purposes)?

So i tried using my 3.7V battery with a voltage step up described above as a power supply, as well as just USB power from arduinos Vin Pin. I have not tried it with a bench power supply,
I think this issue is related to I2C bus itself. I think that something is funny happening with clock or data signals, and I will try to hook it up to scope and see what im getting.

Its been a few days and I still cant figure it out. I think that I will be reverting back to the previous version of firmware. Is there a chance that I could get a modified version of a previous firmware? All I want is just disabled auto white balance so that when i turn power to pixy off and on, it comes back to seeing the same color signatures.

HI Vadim,
You and this gentleman: http://cmucam.org/boards/9/topics/5078?r=5135 are seeing similar things.

The same issue came up a while back:

http://cmucam.org/boards/8/topics/4056?r=4401#message-4401

We removed the flash integrity checking and it seemed everyone was happy. We added what we thought was a more robust check (less likely to result in false positives for flash corruption) in the 2.0 firmware, but we clearly need to review this. We’ll have a follow-on release next week and make sure we do our best to fix this.

thanks!

Thank you Rich!

I switched to previous version of pixy firmwhare/pixy mon/library and problem described in this post is no longer present. I have been using pixy in current configuration for over 3 month daily so I knew immediately that something was not right with the current update.

Also, I remember having signatures disappear on me, and I was able to figure out the steps to make it happen in my setup [[http://cmucam.org/boards/8/topics/4056?r=4401#message-4401]] . The problem magically went away when I stopped using white power connector and provided power to USB connector instead.
I was providing 5V and it was probably not enough cause there is a voltage regulator which needs >5V, so its quite possible that there was a lot of quick powering on/ powering off stuff happening due to power interruption or something. Looking forward to test out your new version of firmware as soon as it becomes available! Feel free to notify me as soon as it becomes available (either by posting here or emailing me, you have my email because I emailed you guys before)