Shop OBEX P1 Docs P2 Docs Learn Events
stepper feedback & voltage measurement — Parallax Forums

stepper feedback & voltage measurement

kutalinelucaskutalinelucas Posts: 80
edited 2011-07-14 07:09 in BASIC Stamp
Hey guys.

I would like to get a little feedback from a stepper driven by a bs2 and L293d driver, but i only really need to know when theres too much load on the stepper rotor for it to continue turning...

Does anybody know of a sure-fire method of doing this? I'm acatualy just looking to get a positive logic signal to another IC when the stepper is halted

I haven't actually recieved my steppers yet and am still on LEDs, but i'm assuming that if the rotor shaft is held (against its will) the voltage will rise across the driver pins? I see experiment #28 in the paralax booklet uses an ADC0831 ADC chip to measure voltage across a variable resistor...could I just replace the vss & vdd pins with each side of one set of the driver pins? a schmitt trigger seems a bit overkill and my board is getting populated enough!

Comments

  • ercoerco Posts: 20,256
    edited 2011-07-12 14:42
    I think an encoder (optical or hall-effect) may be the only way to detect motion or lack thereof. A stepper won't necessarily pull any more current if the rotor stalls.

    You might be able to pick up some kind of transient signal from your non-energized coils as the magnetic rotor moves past them, but that would require some extra circuitry.
  • kutalinelucaskutalinelucas Posts: 80
    edited 2011-07-12 14:53
    I'm not too savy when it comes to analogue circuits, I'm more or a programmer, but how difficult would it be to implement hall-effect feedback? the steppers are to control the lower knuckles of a robotic hand and I'd really like to get some kind of real-time feedback, ie, if an object is detected (without triggering the tactile feedback on the finger-tips) I would like to know where the motor thinks it is...I know steppers aren't nearly tourquy enough for that application but i've built twist'actuated joints for the other two knuckles which compensate for this, but i wanted the precision of the steppers for the bottom knuckles. The steppers are directly connected to mitre gears, so 30 steps of the stepper equate to the 45 degree motion of the human finger

    Do you have any suggestions?
  • ercoerco Posts: 20,256
    edited 2011-07-12 15:18
    Once the stepper stalls, you lose all position info until you home position again. Seems like a strange choice for a robotic hand. Why not servos?

    Higher voltage & current will give you increased holding torque on a stepper, if you dare. Same with energizing more than one coil at a time.
  • kutalinelucaskutalinelucas Posts: 80
    edited 2011-07-12 15:27
    simply because servos are so bloody big, are only precise as the tendons joining them, and I needed something different for my masters in robotics dissertation! everybody seemed to go for vision systems or localisation but I thought i'd go old-school! I'm using a planitary dc motor using a twisted tendon to actuate the other 2 knuckles so tourque was never an issue, and I wanted some form of precision...which I do achieve until an object is grasped without touching the tips of the fingers, the steppers just hold their ground and make a nasty noise. Do you have ANY suggestions on how I could overcome this? pretty much everything else is working as planned...nut I could really do with some feedback from those steppers
  • ercoerco Posts: 20,256
    edited 2011-07-12 15:47
    PhiPi had an ingenious app for a dirt-cheap magnetic encoder. Might be right for your application. See http://forums.parallax.com/showthread.php?130799-Refrigerator-Magnet-Encoder&highlight=magnetic+encoder

    I'm still mad I didn't think of it!
  • kutalinelucaskutalinelucas Posts: 80
    edited 2011-07-12 16:01
    looks good...but i'm thinking I may use some form of tactile feedback on the inside of the finger...the stepper won't know its exact position, but as there's 30 steps/45 degree angle, it'll have a pretty good idea, and i'm sure I can skip over that during my presentations...what do you think?
  • ercoerco Posts: 20,256
    edited 2011-07-12 16:32
    You have multiple steppers, correct? Like one per finger? Much depends on your particulars and presentation level.

    Describe your setup more fully and attach some pics for better help.
  • kutalinelucaskutalinelucas Posts: 80
    edited 2011-07-12 16:43
    sorry, in bed now (in wales!) but will be back on in the morning...thanks for the help so far...
  • ercoerco Posts: 20,256
    edited 2011-07-12 16:54
    Some people have made tactile sensors out of conductive foam sandwiched between conductive plates. Squeezing it changes the resistance, and a Stamp could sense that. See Gordon McComb's new 4th edition book (plug, plug) "Robot Builder's Bonanza".

    Alternatively, you could have several contact switches inside the fingers, so that when one or more switches open or close, you deem contact has been made. You could connect multiple switches in series or parallel, depending on what you're grasping. You could get some inter-finger feedback from something like that. Once you grip something, you need to keep it squeezed to avoid dropping it, so you'll keep the last coil energized, correct?

    Finally, you could have a load-sensing motor mount. Mount your stepper motor so that it is free to rotate around its shaft. Spring-load it against a fixed stop. It will try to rotate against the spring force away from the fixed stop as the load increases, and it can trigger one single limit switch at a preset load before the motor stalls. You could also have a spring-loaded inline coupling between your motor and mechanism, but then you'll need slip rings or something.
  • kutalinelucaskutalinelucas Posts: 80
    edited 2011-07-14 04:52
    I actually have that book but have never picked it up! i've given it a quick read but it seems kinda flakey...I am thinking of giving piezo film a bash, I could be wrong but it doesn't look that tricky and I'll interface the feedback through my prop...I only need a threshold change in voltage not measurement...does anybody know of any deicent tutorials?
  • ercoerco Posts: 20,256
    edited 2011-07-14 07:09
    I'd go with the motor mount.
Sign In or Register to comment.