Shop OBEX P1 Docs P2 Docs Learn Events
Very cheap robot chassis — Parallax Forums

Very cheap robot chassis

LeonLeon Posts: 7,620
edited 2012-04-28 17:30 in Robotics
I just got an email from Cool Components listing their new products, and this caught my eye:

http://www.coolcomponents.co.uk/catalog/magician-robot-chassis-p-975.html?utm_medium=email&utm_campaign=d1489db428-April_Newsletter_2012&utm_source=Newsletter+List

I've ordered one.

Comments

  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-04-23 08:48
    Hi Leon,

    SparkFun also sells the Magician chassis.

    These have been talked about here on the forum a bit. I think Rayman makes a board (Merlin IIRC) to connect with a QuickStart that makes driving the robot easier.

    I have one of these chassis. I broke one of the pieces trying to assemble it and I postponed building it until I find the acrylic glue I purchased not too long ago.

    One of these summers, I hope to hold a small robot summer school so I'm on the look out for inexpensive robots. I'm not sure if I'd want to use a kit like the Magician with motors to drive directly or to use CR servos.
  • LeonLeon Posts: 7,620
    edited 2012-04-23 09:06
    A friend of mine in Paris wants to do something similar for the kids at his daughter's school, and I've just told him about it.
  • RaymanRayman Posts: 14,662
    edited 2012-04-23 12:15
    Leon, I think Duane is right, it looks like the Sparkfun Magician chassis.

    "Cheap" is the right word to describe it... Maybe the biggest problem is that it doesn't aways exactly drive in straight lines...
    Still, for some people, price is everything and I think this is the lowest price robot out there...
  • LeonLeon Posts: 7,620
    edited 2012-04-23 12:42
    Yes, they are the same.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-04-23 13:00
    Leon,

    What are your plans for the robot?

    I've been putting Nordic nRF24L01+ modules in most of my robots to use them remotely. I'll probably add a Nordic module with a QuickStart to my Magician. Of course it will need some sensors. A robot's not a robot if it can't move on its own (and not just by radio control).

    I'm working on a remote for my various robots. It uses one of Rayman's 4.3" touchscreens with a PlayStation 2 controller.

    Just last night I learned how to read the pressure information on the PlayStation 2's buttons. So now, instead of just the four analog inputs from the joysticks, there will be a total of 16 analog inputs.

    BTW, while I'm rambling, I thought I'd mention, don't trust generic game controllers. The PlayStation 2 clone SparkFun sells only has 5-bits of precision on its joysticks (they should have 8-bits of precision).
  • LeonLeon Posts: 7,620
    edited 2012-04-23 13:46
    I haven't given it much thought. I'll probably control it with one of those new DIP28 LPC1114 ARM chips that I'm playing with.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-04-23 19:34
    This is a substandard chassis if it comes with the high-speed/low-torque motors. The SparkFun page for this product is filled with complaints about how the robot veers off course, and my own tests with two chassis confirms this. One of the developers at the company that puts this kit together (though he didn't design this one) says the chassis is supposed to come with low-speed/high-torque motors. He sent me some sample motors to test, and indeed that improved things, but two of the four motors arrived DOA. Apparently there can be an issue with how they're geared internally. I didn't take them apart to see if they're fixable.

    I'd say this is a decent chassis if you can get the right motors on it. You'll know what motors you got when you try to back drive them. The low-torque version offers little resistance. They're probably no more than about 50:1.

    -- Gordon

    PS: I forgot to mention Step 1 in building the Magician: Find the included screwdriver and throw it away. It'll chew up the heads of the screws.
  • LeonLeon Posts: 7,620
    edited 2012-04-23 19:47
    Thanks, Gordon. I'll probably replace the motors.
  • ercoerco Posts: 20,256
    edited 2012-04-23 20:00
    I have one of these still unassembled, and a variety of gearmotors. I like a good challenge and all the warning bells people are ringing make me want to make this dude perform. Who wants to join me and make this orphan drive straight to do a nice repeatable square pattern, and also a clean figure 8? IIRC, it even comes with slotted encoder wheels. :)
  • LeonLeon Posts: 7,620
    edited 2012-04-24 05:22
    My kit has just arrived.

    I saw the encoders, they are called "Speed board holders". If they are used I don't see why accurate control can't be achieved.
  • ercoerco Posts: 20,256
    edited 2012-04-24 08:35
    I'm with you, Leon. Wheel watchin' (monitoring wheel position & RPM for straight-line travel & odometry) requires constant attention and takes a lot of processor overhead (most of a BS2's resources, depending on RPM), but it can surely be done. Low-geared motors are still desirable for accurate braking, otherwise it is easy to coast past an intended stopping position.

    Makes me wonder if the same company sells the "speed boards". In that case, they would purposely use mismatched motors in their Magician kit to make sure you subsequently bought their speed boards to solve the problem and go straight!

    Heck, selling patches and upgrades has worked for Microsoft for many years now. When does Windows Ten come out? :)
  • Martin_HMartin_H Posts: 4,051
    edited 2012-04-24 08:59
    The notches in the "speed boards" are narrower than the surrounding tabs. So the high and low portions of the wave form won't be of the same width. I'm not sure if that matters for that level of accuracy or not.
  • LeonLeon Posts: 7,620
    edited 2012-04-24 09:39
    I checked the motors at 6V on my bench supply, and one takes a lot more current than the other, which is rather strange.

    Also, the speed board holders don't fit the shafts very well, and one is very sloppy. It is a good fit on the outer shaft, though! I might have to use some glue on them.

    One of the motor brackets was a lot thicker than the others, and wouldn't fit the holes, so I used it in an outer slot.

    It looks as though their manufacturing tolerances are very wide.
  • ercoerco Posts: 20,256
    edited 2012-04-24 10:05
    Leon wrote: »
    It looks as though their manufacturing tolerances are very wide.

    Tolerance is a wonderful thing! We even have a Museum of Tolerance here in Los Angeles. I should visit it to see if the Magician chassis is featured therein.

    http://www.museumoftolerance.com/site/c.tmL6KfNVLtH/b.4865925/k.83A7/Whats_Happening_at_the_MOT.htm
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-04-24 10:22
    Wrong Tolerance. Tolerance is a town between Redondo Beach and Carson. Right along the 405. Can't miss it.

    On those speed thingies: Being red, and not even 100% opaque to white light, they are fundamentally transparent to near IR. If you want to use them as encoders you'll need to apply several coats of heavy black or other IR-opaque paint to them. While not symmetrical, when counted as high- or low-going pulses and averaged out over a half-second or second, they'll do okay as distance counters.

    Because the speed differences between the motors can be so sudden and drastic any encoder system will have to work overtime to keep the thing going straight.

    -- Gordon
  • ercoerco Posts: 20,256
    edited 2012-04-26 01:43
    @Gordon: Heck, I live in Tolerance! Or is that Intolerance...?

    I assembled my Magician chassis tonight and I was pleasantly surprised how well the laser-cut acrylic pieces fit together. The motor plates & screws were snug. The only questionable/sloppy fits were the battery holder screws. My main dislike is that cheap steel caster ball. Erratic friction induces random drifting. I don't know why some people think they're a plus. CBA's drawer pull (dragging on the ground) would offer more consistent performance.

    Motors are definitely geared too high, but with some attention to detail, I got it driving in a nice straight line on 3V. The wheels can rub the chassis, which will cause erratic turning. I'll switch to much lower geared motors at first opportunity, the manufacturer should follow suit.
  • ercoerco Posts: 20,256
    edited 2012-04-26 11:00
    The stock motors are geared 1:48. I scored two different pairs of 1:87 gearmotors, but neither has an output shaft on both sides, so I won't be using the stock encoder chopper. It was too low-res anyway (sour grapes). From experience, wheel encoders need 16 slots or 32 stripes.

    One pair have similar motors as the stock motor and operate predictably at nearly half the speed of the stock motors (good). The other pair have different motors and rotate MUCH slower with gobs of torque (great).
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-04-26 11:16
    The manufacturer has known about the under-torqued motors for months, but it may be a case that SparkFun bought a gazillion of them and is working on moving out older stock. What's in the pipeline is all the high-speed motors. Plus there's the issues you noted with the transfer ball (it's the kind that is only made to be used ball-up) and the wheel rims that chaff against the chassis. If you work hard enough at it you can often get something out of the kit, so for the price it's worth a shot. Fifteen bucks is pretty good for a pair of motors and wheels and a nice battery holder with 2.1mm barrel connector on it.

    -- Gordon
  • ercoerco Posts: 20,256
    edited 2012-04-26 11:20
    If you PWM the stock motors at high voltage (6V or higher) and low duty cycle (50% or lower) , they might work OK. But I hate patching poor mechanicals with electronics. Too many mediocre examples out there!
  • ercoerco Posts: 20,256
    edited 2012-04-27 03:40
    New gearmotor is noticeably slower in RPM, with corresponding increase in torque. One stock gearmotor 1:48 motor and one 1:87 gearmotor with different motor winding shown here, both operating off 4.8V.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-04-27 19:46
    Think of it this way: your bot is doing one half of a figure-8 course. You're almost there already!

    -- Gordon
  • ercoerco Posts: 20,256
    edited 2012-04-28 17:30
    Now that's makin' lemonade out of lemons...
Sign In or Register to comment.