New laser cut robot chassis for GM series motors

Robot – work in progress

By Casca @ Sat, 2008-07-12 04:18

test bed for GM motors - currently sonar / avoidance

Cost to build:

$58

Time to build:

2 hours

Actuators / output devices: GM9 motors (pair) JR ST47 servo
Control method: autonomous
CPU: Picaxe 28x1
Power source: 6V
Programming language: Picaxe basic
Sensors / input devices: SRF05 Ultra Sound
Target environment: hard surfaces

Description:

Well, I got busy tonight and designed a chassis for the GM series of motors. The GM motors are really tops with price/performance... at $5.50 each, they are VERY hard to beat.

I thought you might like to see a video of the chassis being cut out, so it's included. The material is plexi .093 thick - it's clear, but has a blue protective plastic coating designed to protect it durring cutting and handeling - plexi scratches easily.

I designed this because there are absolutely no good chassis for the GM series motors, and that's a shame because they are so nice.

I've got to mount a few more parts, but it should be done over the weekend. Thanx to Fritz for sending me some code to build on :-)

Everything has a solid mounting point on the chassis (screws) and the servo is a fuill size JR ST47, mounted off center so the shaft is centerline preventing cone errors durring scanning.

I started out as a machinist so I try to make everything with extremely close fitting tolerances and built to last - the deck is screwed together, but the interlocking tabs would hold it together by friction alone. Motors are hard point mounted with screws and add to the rigidity of the chassis.

Just a few more tweaks and I gotta figure out the battery configuration...

By fritsl @ Sat, 2008-07-12 04:34

Very, very nice!!

By Casca @ Sat, 2008-07-12 04:38

Thanx, and also for the code

Thanx, and also for the code - it really helps.

By CaptainTuna @ Sat, 2008-07-12 05:03

the chassis is so nice! i

the chassis is so nice! i wish i had a laser machine :)

By Casca @ Mon, 2008-07-21 04:40

ThanxThe mechanics are like

Thanx

The mechanics are straight forward for me - it's the code I'm finding difficult -

By BaseOverApex @ Mon, 2008-07-21 17:23

PIece of cake

Given a good sound platform like this, the code should be a piece of cake. You're in the right place!!

By GroG @ Sat, 2008-07-12 05:05

Great! Are you going to go

Great! Are you going to go through a walk-through on how to build a CO2 IR cutting laser and 2 axis table?

By Casca @ Sat, 2008-07-12 05:30

:-) :-) :-) I can give you a

:-) :-) :-)

I can give you a walk through on a X Y plasma cutter - not really good for plexi though..

By Casca @ Mon, 2008-07-14 03:15

Added a video of the GM9

Added a video of the GM9 chassis working with the Parallax ping sonar - the code is still a bit buggy.... but it's getting there -

Chassis turned out to be really really rock solid and I'm proud of it.

By TheCowGod @ Mon, 2008-07-14 03:26

Laser cutters sure look like

Laser cutters sure look like useful things to have access to. What kinds of places tend to have those? I don't even know what to look up in the phone book. Provided I found one, what would a chassis like yours tend to cost, if I brought something like a CorelDraw file? Just curious whether this is something I should be looking into more seriously.

Dan

By Casca @ Mon, 2008-07-14 04:05

Might try sign shops - they

Might try sign shops - they tend to use laser cutters to cut out plexi letters.

I use my cutter/engraver to engrave marble and granite mostly - www.dragongraphicscorp.com my sight gives good examples of the type of work I usually do. I'm going to expand the site and offer cutting services for hobbyists :-) then it'll be a lot less like work to me!

The rest of the industry is high as a cats back - roughly $120 per hour, $60 minimum, but I say this chassis is about a $15 dollar item (materials included).

By TheCowGod @ Mon, 2008-07-14 05:09

Thanks for the info. Huh.

Thanks for the info. Huh. Yeah, $60 minimum is not worth it to build a simple robot for fun, but if you do start offering your services at rates closer to that $15 figure for work of this size and complexity, hell, I think just from this site alone you'd get a good bit of business :) I know I'd like something more permanent than plywood for my robots, and I'd probably make good use of something like this if I could spend $15-20 and end up with a chassis that fits exactly the way I want.

Or just for building brackets and other such structures, to mount servos and motors exactly where I need them.

Of course, I don't generally do too much planning before I start cutting. Maybe in my workflow, this king of thing would be better for a second step, after building the "first draft" in plywood. Either way, it's an interesting prospect.

Dan

By Casca @ Mon, 2008-07-14 11:29

First drafts are always good

First drafts are always good - saves money down the road. I'm personaly fond of card stock templates - nice cheap way to make sure everything fits. Also, the hobbyist's best friend is an inexpensive set of digital calipers - they're invaluable for layout. I took all of the measurements off the servos (including center points for screw holes) and ALL of the trick mounting points for the GM9 motors with calipers. Each individual component you wish to fit takes less than 5 minutes to fit properly the first time with proper measurements and the calipers are the right tool for the job.

Also, the chassis you make needs only the basics - don't worry about every little screw hole that you may ever need - plexi is easy to drill if you need to add a few more mount points.

By GobliZ @ Mon, 2008-07-14 09:11

Nice robot, and nice movie

Nice robot, and nice movie :)

I think you can get the navigation smoother. Instead of just driving until you hit something and then look left and right and pick the best way, you can constantly turn his head and measure. This gives way better results. Look into this for more info.

Nice work, and I'd like to use your laser cutting services too, if you are going to offer it for hobbyists. :)

By Casca @ Mon, 2008-07-14 11:39

I started working with that

I started working with that code last night - the Ping has turned out to be really quirky. all of the > < logic is reversed from what's written for the SRF05! The code from th second example drives the servo and pulses the Ping and produces good measurement results - now I just need to fix it for the Ping... or abandon ship entirely and order a SRF05.

I'm happy to offer the cutting services at hobbysts prices - let me see if I can come up with some basic "good practice" design rules and tips for layout.

By fritsl @ Mon, 2008-07-14 12:50

Regarding the cutting-shop:

Regarding the cutting-shop: That would be cool. If you want, we can give you a permanent link and your own page / paypal low tech "shop" here on LMR. PM me if you would like it :)

I think it would be cool for all!

By .TheWiz. @ Wed, 2008-07-16 00:37

Nice And Clean.!!!!!!!

Nice And Clean.!!!!!!! Exellelent!!!!

By Casca @ Wed, 2008-07-16 02:49

Thank you!

By Casca @ Fri, 2008-07-18 03:43

Finally!!! Ping code that works!

Played around a bit more with some code snippets and added some of my own - nothing fancy, but it navigates and pulses the ping correctly.

Goal is to make it as smooth as the "start here" project II video - it uses some code snippets from that - but cut WAY DOWN.

Added a new video of the robot with updated code -

I atached the code to this post - could someone with a Ping test it on their setup?

By CaptainTuna @ Sun, 2008-07-20 16:30

i will try it tomorrow!

Again great job!!! I like the chassis so much!

By Casca @ Fri, 2008-07-18 02:28

New parts!!!

I made a few new parts tonight...

Ping/SRF05/ sony IR servo bracket

and a full size to mini servo adaptor for the roundbot GM9 chassis - lets you use a mini servo in place of the standard -

By jklug80 @ Fri, 2008-07-18 03:10

You could make a mint

You could make a mint selling those brackets. My SRF05 was mounted using strips of aluminum and hot glue.

By Casca @ Fri, 2008-07-18 03:15

Well, the first bracket I

Well, the first bracket I bought will be the last - I'll never ever pay $5 for another one!

I figure these are $2 brackets -

By TheOther1 @ Sun, 2008-07-20 16:48

Great job on the entire

Great job on the entire thing! Looks awesome, very sturdy and well though out. You should be very proud.

BTW: If you are interested in selling a couple oif the SRF05 bracketrs, I would be interested in buying! ;)

By Casca @ Sun, 2008-07-20 17:30

Through Frits generosity,

Through Fritsl's generosity, I'll have a web shop hosted on this site very shortly - I should have the text done in a few days. In the mean time, feel free to PM me and I'll be happy to get you some brackets :-)

By BaseOverApex @ Sun, 2008-07-20 20:43

AMAZING

That is amazing news. Your workmanship looks superb! All I need now is a design...

By Casca @ Sun, 2008-07-20 21:15

Thanks so much for the

Thanks so much for the compliment. Although it won't happen overnight, I'll also post some general guidlines for design, and I think that should help with the learning curve. I'd also be happy to check over a design and try to make sure it'll fit together.

On a side note, I've been playing with google's "shetchup" over the weekend - there is a free version available, and it offers DWG (autocad) exports as well. The package is quite friendly and in just 2 days I'm starting to get the hang of it... for those who have invested time learning a new CAD package, this is a bold statement to be sure.

By CaptainTuna @ Mon, 2008-07-21 11:49

hey Casca, there's something

hey Casca, there's something i don't really understand about your .bas file. it says:

decision:
gosub ping
gosub ping2
if W0>W1 then
gosub motor_reverse
gosub body_lturn
else
gosub motor_reverse
gosub body_rturn
end if

return

---

won't the two gosubs (ping and ping2) give the same results as the head hasn't moved between the 2 gosubs?

By Casca @ Mon, 2008-07-21 17:31

Yep, your right. I need to

Yep, your right. I need to look at that a bit more...

I'll try and incorporate some of the code from Fritsl's "project II" robot dealing with the head position of the servo - that should straighten it out.

The current decision routine is rubbish.

By CaptainTuna @ Wed, 2008-07-23 07:55

what you need to do is just

what you need to do is just keep a record of the "ping" gosub you went through before. Example:

1- gosub ping

2- record the ping value

3- turn head

4- gosub ping

5- is this value bigger than the previous one? if yes we can just keep going this way!

6- record the ping value

7- etc etc..

but actually this code supposes that there is no or at least little movement in between those two gosubs.

By OptimusTronic @ Fri, 2008-08-01 17:44

Very good!

Very good. This project is great. It's very fast and the programation it's not too long.