Shop OBEX P1 Docs P2 Docs Learn Events
Hexapod / Humanoid bot servo position log? — Parallax Forums

Hexapod / Humanoid bot servo position log?

Bill HenningBill Henning Posts: 6,445
edited 2015-11-15 17:15 in Robotics
Hi guys,

I will hopefully be able to re-start work on my RoboPi / Pi based hexapod during the xmas break.

As some of you may recall, I ran into quite a bit of problems with jitter using the crappy MG 995 Pro servo. If not, here are some links:

http://www.mikronauts.com/robot-zoo/hexpi-hexapod-pi-robot/
http://www.mikronauts.com/the-better-mousetrap-servo-tester/

Given that HexPi is not walking yet, and given that I want to do some repeatable power / jitter tests, I was wondering if any of you had a servo log file you could share with me.

Something like:

servo N position
wait X

one per line... but any format is great, as log as I can tell how to parse it into something like above.

The more servos the merrier, so I'd prefer something from a hexapod.

The idea is for me to be able to replay a known-working hexapod gait, and monitor the power consumption and noise.

If possible, I'd like ~10 seconds of data.

Thanks,

Bill

Comments

  • WhitWhit Posts: 4,191
    Bill,

    Did you see this from erco? http://forums.parallax.com/discussion/162759/torroid-ferrite-core-eliminates-servo-jitter#latest

    Don't know if it would help or not...
  • Hi Bill,

    As you know I have a couple hexapods. I never log servo positions. All the positions are generated by algorithms. One algorithm generates the xyz coordinates of the hexapod's feet and another generates the angles of the servos based on some IK formulas. These values are generated at 50Hz.

    The power consumption is of course very dependent on the load to the servos so I'm not real sure how playing a sequence of motions will help you understand the power needs unless your robot is walking at the time.

    I know from talking with other hexapod builders trying to get enough power to the servos is often a challenge.

    I opted for the brute force approach and used a big LiPo and big regulators.
  • Hi Duane & Whit,

    Sorry, I forgot about this thread!

    Whit: I'll try bead cokes, however I am 99.99% certain the issue is very poor design of the servo controller board.

    Duane: I was thinking of using the log for repeatable power consumption & torque experiments
Sign In or Register to comment.