Sensor usage with CM-5

[sdeba]

You mean documenting building this whole Sensor project?

I've made simple program in RoboPlus Task, just to test if it works with IMU. It was writing gyro values without problem, but for some time. After about 1 minute it started to show a connection error. Then I've unplugged and plugged the cable from IMU to CM-5, and then it started working again for another minute, then connection error. I think it is occuring because there is to many requests, do you have any ideas how to fix it?

[i-Bot]

Are you using the latest version of the code with twimaster.c ?

Please document and share anything you think is useful or of interest to others. Text, pics, video, anything.

[sdeba]

Yes, I've directly uploaded *.hex file from Debug directory in new version. I'll test if this problem occurs while using libbioloid, but RoboPlus Task will be preffered while I would like to use motion

I'll try to make a guide to do this sensor board, but don't know when. I'm an executive editor of one of the biggest Android websites in my country, so I have some ongoing duties

[i-Bot]

I just took a look at the asm listing of the 6DOFIMU project, and the I2C timeout value is way too long. The patch doc said a few ms, but it is a lot longer. Since the timeout function is blocking this is possible cause.

I will reduce timeout, and add a call to process dynamixel requests within the timeout.

I am surprised that Task does not retry. It should not need replugging to recover though if this is cause. Is Task OK with the BioloidIO code?

[sdeba]

Ok, so if you'll do so, give me a sign. Thanks for that

I wasn't testing a lot with IO code, but from what I remember there weren't any issues.

[i-Bot]

Sorry this is taking longer than I thought.

I fixed the IMU timeout, but when I checked with Task it did not fix the problem. When I tested with HUV Bioloid-IO, that also has same problem, so I now need to look at the HUV code, not my changes.

[i-Bot]

mmmh interesting. The problem appears to be in the bioloid packet handler. After a while the CM-5 sends a bad CRC (CRC = FF) for some reason. The Bioloid IO code tries to log the error and hangs. I commented out the CRC error logging and it works OK. I will take a look at the HUV logging functions.

[sdeba]

Thanks for sharing status. That's strange behavior for sure. Good luck and thanks for helping me

[i-Bot]

I put the updated projects in my file area

[sdeba]

The latest updated version works perfectly even with RoboPlus Task. Minor problem was that I realized that IMU adresses from here: http://www.bioloid.info/tiki/tiki-index ... umentation are reversed in my case. That means if 26 was responsible for Sideway Accel then it is in my case Pitch Rate But I've tested everything and it works ok. I've also updated my adress table and now everything is fine. Thanks for help. I'll have some more time during upcoming holidays so then I'll share my humanoid with IMU with you all

[i-Bot]

The allocation of X, Y, and Z and pitch/rate is dependant on the IMU orientation, and register allocation is easily changed in the code. In the same way gain adjustment, sign conversion, and filtering may be done within in the IMU if required.

Have fun over the Holidays

[wiweet]

Great thanks to you i-Bot, I've followed your guidance here, and I'm succesfully make a Dynamixel Device for IMU, the difference is I use Robotis Gyro and MMA7361 as the sensor, both of them are analog output, so my code is much simpler.

Thank you for sharing your expertise.

Regards,

wiweet