Building the Coat Hanger Walker

Here are a few things I noticed when building the "Coat Hanger Walker" from the Absolute Beginner's Guide to Building Robots -- see my last post -- that I think it would have been nice to know in advance. If you're not intending to build one in the near future, you should probably skip this post :-)

The original text is online here, so I will only mention the difference between my experience and the original author's.

• It's worth noting that the kit from Solarbotics does not include the 4cm gear described in the text; it perfectly clearly states what it does contain, and the gear is not in that list, but because all of the other components are included it's easy to assume that there is not one missing piece. Well, it was easy for me, anyway. I got a set of gears from Maplin that included one of an appropriate size -- "Spur Gear Set, part #917D/6", it says on the packet, though I can't find it on their website. Someone I spoke to at Solarbotics also mentioned that an old VCR could well contain an appropriate gear, along with lots of other useful stuff.
• Solarbotics have also switched the leg mounting pads that were included from being their LMP2 units (which are semi-triangular) to LMP1 (which are rectangular). I've suggested they switch back; the LMP2s require significant cutting with a rotary tool to fit the robot's plan, whereas the LMP1s required quite a lot less. This leads me on to....
• Before cutting the LMPs to fit the design, it's probably a good idea to make and test-fit the legs to them. This will make sure that you don't cut off bits that are required to hold the legs in place (cough).
• If you find yourself gluing the legs directly to the gears, having inexplicably found the leg mounting pads unusable, you will discover that five-minute epoxy is not strong enough. Indeed, five-minute epoxy isn't strong enough for any of the bonds for this robot, except perhaps the attachement of the battery packs to the motor (which is strengthened by a screw). After many unstickings and re-gluings I eventually got everything to hold together nicely with Araldite's 8-hour epoxy.
• Making the idler gear without the leg-mounting pad involved gluing the hub directly to the gear (having no LMP to glue it to). Again, I had glue problems; try as I might, I could not get the two to stick strongly enough. They would seem to hold for a while, but the bond was much too weak to stand up to the repeated motions of the gear. I eventually wound up supergluing a strengthening bit of plastic along the flat edge of the gear, on top of the hub.
• I found that when I connected the circuit to the servo, rather than moving (say) 30 degrees left, then 30 degrees right, it would move 30 one way and then 25 the other. This meant that after a while, it would rotate -- hardly the reciprocating movement required to make the walker walk! This may well have been caused by my distinctly non-l33t circuit-building skills, but I couldn't find anything obvious by poking around with a multimeter. So I used a mechanical solution -- the instructions recommend replacing the servo's final gear, which has a stop to prevent it from rotating more than a certain amount, with a gear without such a stop. Backing out that step and putting the stop back in there meant that I could arrest the extra 5 degrees leftward movement, giving an even 25-25.
• I can't emphasize how important the shape of walker's legs are. Once I had completed the build and started it up, it wobbled for a couple of seconds then fell over and twitched pathetically until I switched it off. My initial thought was that the frequency of the leg movements was too great, so I tried slowing it down using an elastic band to slow the movement of the front legs (as described here as Gait Dampening), but this didn't help. What did help was first carefully bending the legs to lower the centre of gravity (effectively by making it "crouch" more). Once I'd done that, it was able to stay in one place, with its legs slipping over the floor -- imagine a dog trying to run on an ice rink and you'll have the right idea. It was only when I tried flattening the "foot" ends of the hind legs that it started walking -- backwards, as I posted. I've been experimenting with different leg shapes, and will post anything interesting that I find.

If you've read this far, I really do hope you found it interesting :-)