Let's Make Robots!

Leo

Goes bump in the night

Main Goal : To listen for a knock at the door and then notify the human occupant of the house.  Eventually will be an AUV, complete with charging station.

 Sensors not yet implemented :

 Voltage sensors

Hall  Effect Sensors on the front idlers (ordere)

Sonar Range Finder

IR Range Finder

PIR sensor 

 

Leo%20the%20Robot%20002.JPG
From Leo the Bot

 Ghetto Hinges or how to upcycle your soda can into bad bot parts

 

Got them replaced with regular hinges but still need to replace the back bumper, I'll post a pic of it if anybody is interested.

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Intereting :) More detailes?
I'll post more pics this weekend and a possible Walk Through on a problem I was having. Basicly, the solution is to be a smarter techie, and work things out ahead of time. Right now, writing some data logging code for the accelerometer, because right it's like I'm flying blind trying to get Leo to drive in a straight line.  

Well, got Leo back together completely and now have a short somewhere in the drive board. I unplugged the arduino and it's getting power still. I don't know if I bridged something in the drive board or what yet. I'm thinking that is where the problem is. Will find out later today.  

 

It might not be a short at all, are you still using L293Ds? As far as I can tell, they use CMOS buffered inputs, which can 'float' if left unconnected. That means the inputs won't return always to logic low if you remove the Arduino.

You can solve this by adding pull-down resistors (around 10-20kΩ or so) between the L293D inputs and ground.

I hooked it back and now it's doing nothing.  I've got the batteries on the charger, because the L293D needs at least 5v coming into it to power the motors.  I checked my board and it seems alright so far, will check things this evening, when I get a chance to look at it again.  

Thank you for reminding me about about the fact that it can 'float' if not grounded.  I learned about the 'float' on the L293D the hard way, when I first got it. I was very upset after connecting one motor, watching it oscillate, and not respond to the commands from the arduino, LOL. After connecting both motors it worked great, till now.  Hopefully I didn't smoke the chip.  If I did, I'll order a few this time and work on something else for awhile.  I hope I didn't, but if I did, it's a learning experience.  I think from now on though I'm going to run the outputs through diodes and setup a separate power source for the motor board and sensor suite, to help protect the main board from frying.

 

My next bot is going to be bigger, lol. Working on Leo is akin to surgery at times, trying to get everything routed and put it without killing something or my fingers is tough at times.  His insides are a bit bigger than an index card. 

 

Good news, the arduino seems ok, bad news is that the L293D smoked itself. Still trying to figure out how and why. Not finding any shorts on the board and the voltages were well within.
Check out Pololu's info on the motors in the track drive.  They stall at 2.1 A at a 3 volt level, will only be more at a higher voltage. The L293D is good to about 600 mA, continuous, 1 A peak. Also, running the motors at a higher voltage will burn them out quicker too.  Pololu has a couple replacements, here and a Solarbotics one.
It seems that the short is in the board. Either things got really hot and melted the plastic coating off the wires or the soldering iron got too close and heated them up.  Time to redesign that board and start again on the accelerometer coding.  I haven't checked the chip yet, but maybe it's still good.
I smoked it good and created a short from the inputs directly to the outputs via the power pin, ie you apply power and you get it out on the output pins. Good news is that more chips are on the way and I've had a lot of time to think of how to avoid this situation.  Don't route negative and positive wires over each other where things get hot, eventually they get too hot, the insulation melts and they short together and while draining the batteries they heat up the circuits big time!
Ok to keep Leo moving in the direction he's supposed to be I'm going to use:

A1302 Hall Effect Sensors on his front idlers coupled with magnets to give feedback on his track speed.

2125 Mesmic Accelerometer is used to detect and calculate drift on carpet, smooth floors, track skidding and bumps in the floor.

Even with those, there is still going to be a lot of error from track skidding, carpet nap, bumps in the floor, track wear, and from not being put back down exactly where he got picked up.

Going to use a PING(or other sonar type sensor) and a sharp or two map out his local environment to cross reference with his internal map. Basically, trying to program Leo in a fashion that biologicals think or at least my poor understanding of how I perceive my self and others do things.

Voltage monitors will be implemented as well so that Leo just doesn't 'die' in the middle of the room. He will know when to return to his base(home) to re-charge(to eat). So will have to figure out how to get a charging circuit into his little body as well as everything else.