Quadruped Robot Chassis Opinion

Hello everyone! It's been a while... I can't seem to figure out how to set up a poll on the new forums (well, not so new, but, again, it's been a while) -- so if you guys could give me an opinion just as a reply, that would be great...

I'm looking to develop an affordable, totally open hardware, expandable quadruped walker robot over this summer. I have found that this sort of chassis is rarely found commercially available, I only know of about two examples, both of which are somewhat costly. My question comes down to, if you would be in the market for this sort of chassis, would you rather go for something a bit heftier, costing around $800, or perhaps something smaller, lighter, and weaker, around $400-500? This is hardware only - 4 legs, a body, and all 12 servos (3 DOF each leg). I ask because I can take two routes with the design but I don't know which would really be more popular (I could always develop both, but in that case it comes down to which one to work on first).

I realize the details of the difference between the two possibilities are rather vague here, ask if you have any questions that might help you form an opinion, I'll try to answer them, I just don't know what else to say at the moment.

Thanks in advance!

I'd rather see something at a sub-$100.00 USD price-point; I think it could be done, too - if you used the small 9-g servos, along with PCB material for the leg pieces. In fact, if you took this route, the entire "kit" could consist of a bunch of "break-apart" pieces on a single "board" shipped flat, with a kit of component parts (electronics and other hardware) to solder on the "body"; if you set it up right, you might even be able to have traces along the leg pieces to small right-angle mounted push-buttons at the "feet" end, to sense "pressure" and relay it back (or something similar).

I'm not sure if $100.00 would be the manufactured cost, or after markup, but even if it were the manufactured cost, you could bump the SRP to $150.00 and it would still be a worthwhile kit (as long as all parts - on board stand-alone arduino controller, servos, etc - were all included).

I'm looking to develop an affordable, totally open hardware, expandable quadruped walker robot over this summer. I have found that this sort of chassis is rarely found commercially available, I only know of about two examples, both of which are somewhat costly.

I suggest you make a working quadrapod prototype. It could be something simple first made with small servos, craft sticks, and hot glue. Also study other's projects like below.

http://www.lynxmotion.net/viewtopic.php?f=8&t=7632

cr0sh:
I'd rather see something at a sub-$100.00 USD price-point; I think it could be done, too - if you used the small 9-g servos, along with PCB material for the leg pieces. In fact, if you took this route, the entire "kit" could consist of a bunch of "break-apart" pieces on a single "board" shipped flat, with a kit of component parts (electronics and other hardware) to solder on the "body"; if you set it up right, you might even be able to have traces along the leg pieces to small right-angle mounted push-buttons at the "feet" end, to sense "pressure" and relay it back (or something similar).

I'm not sure if $100.00 would be the manufactured cost, or after markup, but even if it were the manufactured cost, you could bump the SRP to $150.00 and it would still be a worthwhile kit (as long as all parts - on board stand-alone arduino controller, servos, etc - were all included).

Interesting, I could sure try. Something that cheap would take a bit of effort - hardware itself is very cheap (laser cut plastic chassis, some screws/nuts, and a couple 3D printed brackets), it's always the servos that bump the price up. I could go for some micro servos, those go for around $15 minimum though... that brings price up to at least $180 for servos alone. Of course, mass production lets you drop the price per servo when buying in bulk so technically that could be dropped to maybe $140 or less. I would argue something closer to $200 would be possible at a minimum -- do you think that would still be attractive, cr0sh?

I suggest you make a working quadrapod prototype. It could be something simple first made with small servos, craft sticks, and hot glue. Also study other's projects like below.

craft sticks

Um... We're talking about a commercially viable design here.

He said make a prototype yourself with the craft sticks and small things before trying to make somethign bigger. Spend $50 and get the thing working right, and then large scale it. Perhaps as you play around with your prototype, you might make some videos, and get some feedback from people about things they might want.
Though I would agree with Cr0sh. Something small, like made from a small fiberglass board that you breakaway, then a bit of assembly. That would be somethign that could be sold pretty easy (I'd buy one). Maybe give the thing a few sensors with it for people to play around with and a sample code for them to update.

Um... We're talking about a commercially viable design here.

Then you give him the quick answers. :wink:

jezuz:
Interesting, I could sure try. Something that cheap would take a bit of effort - hardware itself is very cheap (laser cut plastic chassis, some screws/nuts, and a couple 3D printed brackets), it's always the servos that bump the price up.

I still say you should go with PCB FR4 material or something like that; it's fiberglass, and very strong. Your PCB routing (cutouts) and circuitboards could all be done by the same company; if you laid it out right, you could do a slot-tab construction with solder joints to hold the whole thing together (maybe with some extra wire bracing for certain areas). That way, you could make the whole chassis "3D" for future prototyping (add proto-holes to the legs, for instance). It should be real cheap to do; much cheaper than acrylic and assorted hardware, probably.

jezuz:
I could go for some micro servos, those go for around $15 minimum though... that brings price up to at least $180 for servos alone. Of course, mass production lets you drop the price per servo when buying in bulk so technically that could be dropped to maybe $140 or less. I would argue something closer to $200 would be possible at a minimum -- do you think that would still be attractive, cr0sh?

If you're going to do this, you'll likely have it all fabbed in China somewhere; go on ebay, look at the small 9g servos on there. I pulled up a page where they were under $3.00 each - that's the price point you'll want to shoot for.

As far as a prototype is concerned, you'll want to do this. You could probably do one in acrylic, first, with an eye to translating the design to PCB material; work out the design kinks, etc - then you'll have a viable design for sale.

The below servo seems to have good reviews for $1.95.