Robot Chassis (refresh to update)

now that's much much nicer and works much much better and much much longer good work Bill

hey this is a little off topic but where can i get a mongoose. i saw the website but is there a eeb site thta supplies it?

The Mongoose is too new yet, I've had to work on another project that will be posted on my site in the next couple of days.

The Mongoose chassis now on rev 11 is much better than before. The rear wheel plate (gold in the images) will have a few mount holes and a place for a power switch, it also is the front plate for sensors.
**broken link removed**
**broken link removed**

nice. your project is getting better every time.
but where do you get the materials needed to construct the chassis? (especially the "gold" ones). do you have a small workshop of your own to bend, cut and drill the parts or do you have a good hardware store?

instead of the double ball bearings, which look pretty expensive? how about a single piece of nylon stock, like 1/4 or 1/2" thick. nylon's not the right term for what I'm thinking of, it is that soft "greasy feeling" plastic commonly used in the diy hobby, and also the stuff they make the big white cutting boards out of.

it should slide around nicely on any smooth surface, especially if the wheels drag it rather than push it. plus it will work better on carpet or uneven surfaces, since it will want to ride above the carpet pile rather than sink down into it.

There is a metal shop in Mississauga Canada (Transduction) which I've worked with before. I'll have the parts made there with a clear anodize or coating. It's only two different parts and the BIG one is only 4.5" long! (The larger round "taillight" holes (fit LEDs just fine) are 0.25" dia
There are no PEM nuts to keep the cost as low as possible, perhaps a PEM nut blue anodize and signed limited edition someday.

As for the roller bearing they're pretty cheap and not hard to find (Tamiya) and about $7 CDn a pair.
And they are the rear drag wheels.
**broken link removed**

nice invention. i hope it'll get along just fine and make you a buck or 2...
if only i could find a place to get my hands on some parts I'm missing I'd go out and make my own bot to compete with you only my bots wheels are probably bigger than your bot...

PS i just reread the thread and found that you are using S sensitive magnet sensors? isn't it easier to use light instead?when positioned accurately you could even read the gaps between toothgears teeth also when using those you can sense the direction of movement

I think I've found a better way to sense the motor RPM. A reflective sensor and a couple of black strips on the crown gear.
**broken link removed**

OTS: sent you email. let me know if it didn't go through.

bloody-orc said:

OTS: sent you email. let me know if it didn't go through.

Click to expand...

Hmm, didn't notice it. Might have gotten filtered.

hmm... you have 2 users? ok... anyway sent to suport on blueroomelectronics.com

You have got those ideas. just thought I'd make them public

it's just that i have some ideas regarding to your robot (mongoose).
the rotation sensors you have on it are IMHO not very good... the sensing of B and W is not that accurate. especially on higher speeds. the sensors you need for this must be positioned hyper accurately over the area. they are sensitive only in one given range, that normally is 1.2mm +-0.2. and the area must be dust free at all times. i know, I have used them for a similar task.

instead you could use small magnets embedded to the wheel + a hall effect sensor or reed switch to count the rotations. or maybe use the similar concept that is used in PC mice (the older ones, with the ball under it). this way you can also see the direction of movement if needed.

Those things don't need so much maintenance.

---

Well i was thinking of using small magnets (2-3mm diameter) and drilling a hole in the tooth gear. just slip the magnet in and use some super glue to fix it. only problem would be obtaining sensitive enough hall effect sensors or reed switches that can detect the magnet fast enough...

Or maybe you cold mount the slotted disk from the PC mouse to the output shaft of your bot. then you could use the whole system from the mouse to calibrate the bots speed.

all just ideas though. (and it's getting late here. 00:28 in the night here)

Good luck

I've actually got the crown gear version with two tiny magnets epoxied into holes drilled through the gear. I like this solution BUT it's very easy to destroy the gear if you're not careful. I also like the opto just cause they're kinda old school and just a "Sharpie" on the crown gear does the job.

Using the comparators to sense the the optos / halls (hmm maybe I'll put the holes for both on the PCB).

I thought it would be nice to see if I could squeeze a L298 on the Mongoose. Might just fit and would give current sense resistor feedback for the motors.
**broken link removed****broken link removed**

Hey Bill! That looks great, so nice and clean.

Weren't you changing your mind and moving over to the SN754410 IC? What happened to that idea?

Anyways, it look excellent!
Omar

Actually yes, to replace the L293D, I just like the L298 for the extra current and the mounting tab. Those L293D and SN754410 run very hot even with only a few volts and a 500ma load. (Two motors @ 250ma each) I do like the current sense (using the PICs A/D channels)

I wish the L298 had the clamp diodes built in, and the 2W 0.51ohm resistors are a little large. I'll build the prototype around it first.
**broken link removed**

Ohh, I see now! It seems the reason why people go with the SN754410 is pretty much the fact that it has most things built in (including the clamping diodes).

But you're right. The 754410 doesn't have the current sense stuff.

Seriously Bill, I'm in awe of your drawing capabilities with Google Sketchup. I use it frequently, but let's just say I'm nowhere near as good as you are!

Omar

I've been using it for years. Data sheets and a micrometer, get the measurements right on all the parts, save as groups or components (components let you make changes to one part that affects all copies) The green chassis is a component, it's flipped but changes in one are mirrored on the other half. Most drawings start as a rectangle (I type in the dimentions to get them exact) and simply pull the surface (again typing in the height). Here's a drawing someone else did a couple of years back, I've cleaned it up (it may have been imported from another cad as it had a fair amount of extra lines in the drawing) and added the lights and tweeked other areas.
It's a complete mirror image except the stacks. I do have the experience now to draw such a beast, the treads would take some time but the rest would be measure / draw repeat.
**broken link removed**

WOW. A-m-a-z-i-n-g.

Funny... you make it sound so easy !
I understand the whole making a rectangle and pull it to make different shapes, but it takes so long! And anything that takes so long is hard. Plus my Sketchup drawings look like drawings done by wild chimpanzees. As well, I usually get tired of trying to input the measurements and start just wing the measurements.
Omar

someone's been busy I see

Omar.M Cad'ing is really fun. nice music, cool bear (or soda), shiny micrometer and a sensitive lazer mouse and a way you go . skill comes over time. I once couldn't make a AA battery, now I can make kinda everything (also tracks and chains )