PSC twitching

daveedavee Posts: 35
edited 2004-08-12 - 03:00:05 in Robotics
Antbody else getting servo twitching with the PSC? I get almost constant twitching even though I've routed the servo wires up and away from the pcb as much as I could. I heard a rumor about parallax adding capacitors to solve this problem.confused.gif

Comments

  • John BJohn B Posts: 82
    edited 2004-07-28 - 21:59:35
    Hi there,
    We've heard some people have this trouble. So far, what works best is lowering the voltage supplied to the servos to 5 VDC. We've added a few different values to caps on a few different boards to experiment with, but it did not make much difference. Lowering the voltage made everything solid. I heard that older servos will tolerate the higher voltages.

    John B.
  • daveedavee Posts: 35
    edited 2004-07-29 - 12:21:40
    I'm usinsg a 4 pack NmH d cell battery pack, I think the MOST it can do is barely 5 volt. It's rated 4.8 but on a full charge it is only 5.2 and when it gets down to 4.6 the arm quits workng as the low voltage is not enough.

    That by the way is a VERY narrow voltage range. Especially for thos wanting to use it off a battery pack.

    I'm beginning to see why people are suggesting the use of a voltage regulator, even on battery packs. Maybe parralax should put a regulator on board to solve the problem for everyone instead of basically requiring everyone to put a voltage regulator on it themselves.

    Or maybe even make it work with the BOE, like all he other popular serial servo controllers can. It's amazing that 4.8 volt 4500 mAh battery will only power the PSC for an our or two, even with small servos. It seems very power hungry compared to other SSCs (though I do like the ramping)

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Dave Evartt

    FUSION robotics
    What's in YOUR robot?
    http://wehali.com
  • John BJohn B Posts: 82
    edited 2004-07-29 - 16:04:58
    Well that shoots a big hole in our current theory. We will have to start over [noparse][[/noparse]troubleshooting] with your twitch issue. What kind of servos are you using? What are they attached to? I understand that you have a ~5V battery supplying the servos, but what are you using to supply the 5VDC to the PSC and the BASIC Stamp(R) module?

    I rather like the voltage regulator idea; it could certainly help with any noise issue. The only problem is that most shunt regulators supply 5 Amps at best, so with all sixteen servos running, it would be best to use two regulators. The addition of two regulators would add significant size and cost to the PSC. We'll have to think on this a bit more.

    John B
  • daveedavee Posts: 35
    edited 2004-07-29 - 17:19:29
    Here is what I have:

    Crustcrawler SG5 robotic arm as it came from Crustcrawler (minus battery packs and wiring harness)

    Controller:
    BS2p20/40 demo board
    BS2P/20
    7.2 volt 3000 mAh NmH RC car battery pack

    PSC:
    4.8 volt 4500 mAh NmH (4 D cell)
    5 volt regulated DC comes off the demo baord to power the PSC electronics.

    Arm:
    3 hitech 475
    3 hitech 645

    Twitching occurs regardless of the position of the arms so it doesn't seem to be an issue of load current as even when it is straight up, there is no load on any individual servo but it still does happen on a regular basis. It's also interesting to note how one can tell the PSC Servo power is low: The PSC simply stops telling the arm where to go, in essence, it just suddenly goes limp until fresh batteries arrive.

    I guess I 'll just have to build a battery monitor to add to the stamp baord so I can tell the stamp when the power gets below a threshold and it can park the arm and refuse to move it until the voltage comes up.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Dave Evartt

    FUSION robotics
    What's in YOUR robot?
    http://wehali.com
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2004-08-08 - 14:48:41
    It sounds to me like the quality of the servo electronics as they are matched to the mechanica setup has something to do with the twitching -- similar to 'auto-pilot dither' of the old auto-pilots which were used on commercial fishing boats.

    It occurs when the rate of seeking to readjust direction is not well coordinated with the mechanical factors involved.

    What happens is the seek function overshoots the correction, then has to go back, and try again, and again, and again.
    Some fishermen just give up and accept the constant racket of the auto-pilot and the zig-zag.

    I suppose the easiest thing to do is to try the controller (assumed to be a BasicStamp) with another brand of servo. If the problem goes away, it is obviously somewhere internally located in the servos configuration.
  • AlexAlex Posts: 20
    edited 2004-08-12 - 03:00:05
    Hello Everyone:

    The twitching has nothing to do with the PSC controller. The twitching is due to unregulated battery power going into the PSC controller. We have tested this extensively at CrustCrawler and have found that using a regulated 5v power supply or a 5v voltage regulator with the battery results in zero twitching of the servos. We get rock solid performance using either configuration. We are going to post a simple 5v regulator circuit on our web site this week for those who want to use it.

    Thanks,

    Alex Dirks

    CrustCrawler Robotics
Sign In or Register to comment.