The Life and Times of Leo

It's been awhile...well a long time

2010-10-29T03:44:00.000-07:00

I've spent the last year bouncing around like a bb in a runaway boxcar going through a mountain pass. Between two hospital stays and trying not to get too far behind on everything while working like a madman, there hasn't been much free time let alone time to breathe. So finally caught up with life, sorta(crosses fingers) and not planning on going anywhere for awhile I hope.

Tomorrow is UPS day

2010-01-14T00:42:00.000-08:00

I will finally get the picaxe usb cable I need to program the motor driver. I've been looking at the AtMega8s as the next step for the Motor Driver Board, it has plenty of PWM pins and such. It's typical of research, it starts off in one direction and then veers off to somewhere else. To fully utilize it, I'd have to use 12 digital I/O.

The new cable works great, lesson learned, get the real thing and don't get a clone. Ah well, I'll still get stuff at SF just not that lol.

Spending time researching

2010-01-12T21:09:00.000-08:00

If you haven't read

If you have built a few robots, you know the frustration of wanting it to be able to do different things, but don't have the knowledge to do it. This book goes into the decision making, types of behaviors and combing simple behaviors to make complex ones. Overall, it's a good informative read that would not put you to sleep.

In a few days, I'll be able to finish up the motor board. One of the reasons I'm not going to put a the stepper portion in is, because if I do that there isn't any room for the nap function or any of the error codes, function codes. I could crank things down a bit more than I have, but still would run into a memory problem. I could go with an ATmega 168, but that would be proc overkill. It'd give me all the PWMs, Analong pins and digital I/O I'd need to do it all from that chip, but not viable unless I can find a cheaper source for them.

Long Time No Blog

2010-01-08T18:47:00.000-08:00

Well, it's been two months to the day that I last posted a blog. A lot has been happening in my life in that time, just nothing in the realm of my hobby life of not has been happening. I have all the parts for the Motor Control Board(MCB) and it's partially assembled, just waiting on the programming cable to come in the mail next week. Yes, I had to relearn a lesson, don't try to cut corners with bad clones. Clones are like copies, neither are as good as the real thing.

One thing I did cut out of the MCB was the servo control part, because all that is needed to run servos is microcontroller itself. All the other chips would be a waste of space, power and money to control a bunch of servos. So, the MCB will be able to control 4 DC motors or two bipolar stepper motors. Because I'm use a PIC, it will only have two PWMs though, I haven't decided to use one PWM for each L293D or use both PWMs on both chips, probably the later. I'd thought about using two PICs, but that would be a bit of overkill and increase the cost and size of the board.

I have ordered a PicAxe 8M and protoboard along with the correct and non-clone programming cable.

Update on Leo

2009-11-07T05:03:00.000-08:00

Well got the rear bumper fixed, so Leo no longer has a ghetto bumper and won't be a hazard to knit shirts or other robots.

Leo's Ghetto Bumper or Most Embarrassing Pic for a Bot

Dayum, I'm good, only had to modify one hole in the bumper to get the bumper on. Actually, pure luck since at the time I drilled the holes to put on the aluminum plate I didn't have the part or know what I was going to use.

Just a bad pic of a decent job.

Monday I should be getting in the new motors for Leo, they can go up to 12V. I can now run 9v batteries all around if I want and probably will once Leo goes into UAV mode and can take care of his own charging duties.

Leo's Ghetto Bumpers

2009-11-04T09:33:00.000-08:00

Here's a pic of Leo's Ghetto Bumpers. One can improvise badly, yanno LOL

Leo's Ghetto Bumpers

Just laying out some thoughts

2009-11-02T17:58:00.000-08:00

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.

Finally better

2009-11-02T07:26:00.000-08:00

I am finally over the strep, man alive that is some bad bad mojo. I did get a chance to do quite a bit of reading between naps(about 20+) because my schedule was wake up, take antibiotic, take cough meds, pass out, repeat. I did get a chance to re-read a lot of the bot stuff though, and have decided to make a small change in plans. I was going to make a freeduiono board with screw terminal blocks becuase I get tired of when Leo hits a bump or rolls over that a wire somewhere on the arduino comes loose.

The change is that there is going to be two boards, one to control Leo's body and one to do mapping, read sensors, make decisions, etc. The first board will control the drive motors, the neck and head position motors, and monitor their position and state sensors. I should be getting the parts in the next few days for both boards. It will be fun putting them together and then getting them to talk to each other. Leo will go from one MCU to two and will have two L293D chips. I'll be ordering the better motors for him this Thursday. Only a few things left to get for Leo. Need to find a sonic ranger, a PIR sensor, a mini servo.

Been awhile

2009-10-28T16:16:00.000-07:00

It's been awhile since I've posted. Been dealing with a bad case of strep throat, and getting better now. Looked into getting the 6v motors for Leo, but it will have to wait till I want some other stuff from Polulu.com. If I order now, I get 4 motors for $20 and if I wait till next month I can order 15 motors for just over $30. They are inexpensive, but the S&H is not.

Going to order the replacement L293D for the drive board and a few other things tonight from BGMicro. I have a few ideas I'm working on. IF they pan out, I will blog about them, if they fail....well if it's spectaculiar, I'll publish the film and if not ah well, you will never know *wink*

Mech outta here

Current projects

2009-10-15T02:24:00.000-07:00

My current projects since Leo is almost physically complete are building up 4 arduino clones with a few personal additions and a few things like dual H-bridges with opto-couplers, IR sensor board(yes, I know you can buy sharp sensors ready to plug in to the MCU). I got the magnets for the magnetic encoders, just have to find some bipolar hall effect sensors such as the 1302, which is found in the tachometer gear of most cars. Found some on Ebay.com, but was hoping to support the local economy first before the world economy.

Leo is almost physically complete. Just need to wire up the IR sensors and get the hall effect sensors. Once those are installed the body part is done. After that I will switch out his motors for 6v ones and maybe add a head with some sensors on it.

Mesmic 2125

2009-10-13T00:33:00.000-07:00

Was looking around Radio Shack last Friday and found a Mesmic 2125 accelerometer in the sensor bin, I also found a PING module from parallax too. Got the accelerometer, but not the PING module. I like to stay within a budget and I think I can find the sonic module cheaper online with S&H included. Got it home, got it hooked up, wrote a sketch to look at the data stream. So may have to do some data sampling. Writing some procedures to do some data logging. I will find out tomorrow evening if they are worth the time it took to write them.

Video Clip of Leo

2009-10-08T03:39:00.000-07:00

Leo's Maiden Run. Here's the link as promised. After he rounded the corner he got stuck in one. Time to work on sensors and upping his battery supply. I may go ahead and put a low voltage sensor on his battery packs, so that I know when he's running low or just stuck.

Leo's First Run

2009-10-08T00:36:00.000-07:00

Well, he lucked out on his first run, only found half of the corners and got stuck in every one of them *chuckles*. Otherwise, it was a great and informative maiden run. Especially, considering that the only sensors he has right now are his bumper switches. The next project will be to adding some range finders to him so he can hopefully avoid obstacles altogether. Though adding another 7805 FET would probably be a good idea if I want to add some juice to his motors. I need to get more rechargeable batteries as well, it seems that most of mine are dead and I do mean dead as in won't take a charge dead.

I'll try to upload the video of Leo's first run tomorrow. The lighting is bad and there is no sound, but it's a cool video still.

Leo has front bumpers now

2009-10-07T14:50:00.000-07:00

Well, couldn't find the small hinges I needed for the bumpers, so I had to improvise. Seems that soda cans are useful for makeshift parts. I'd joked about using one for the hinges, but didn't think I was going to have to resort to such. Working on the rear bumper now and hopefully will get it done tonight so that I can do some more coding. Hopefully, tomorrow can work on the range finders. I need one more for the rear, but all is good.

Yay!

2009-10-07T02:59:00.000-07:00

Leo now has front bumper switches! Now just need to attach the bumpers and work on the rear bumper. I've even got the bumpers coded. I did check mouser.com for the bipolar hall effect sensors, and they don't keep them in stock. So, I'll have to scrounge ebay for them, yippee. After I get the bumpers completed, I will start on the IR rangers. That should keep me busy till the hall effect sensors come in.

Working on bumpers and redoing the design

2009-10-06T16:14:00.000-07:00

Well got the rocker switches for the bumpers. Just need to put a few pull down resistors, screw them in and attach the bumpers. I did find a Mesmic 2125 accelerometer at Radio Shack for $32.99 plus tax, but found a 3 axis one at adafruit.com for $20 plus S&H. The latter is out of stock though and has been for a couple of weeks now.

I did order the magnets for the tachometer part of Leo today as well. Using 1/4" rare earth magnets on the front wheels coupled with analog hall effect sensors. Will probably order them this week or find them on ebay.

Just need to mount the switches and redo the bumpers on Leo. I was thinking about putting the IR ranging sensors in the drive part of the body, but starting to run out of room in there with batteries, switches, the drive board and still need to put another or change out the battery packs altogether. A few different options, especially if I make Leo an autonomous being and therefore able to charge himself when he needs to. That is sometime down the road.

The drive unit will be using up 7 out of 11(I don't count the tx and rx pins) digital I/O pins and 5 analog pins. I have 3 clones that I bought from Dale and need to build them up, but need to get a lot more familiar with the arduino I already have.

Well off to do something, even if it's wrong!

Smokin'

2009-09-28T11:06:00.000-07:00

Well took the stuff off of the solderless breadboard and put on the drive board. Somehow fried the 7805 regulator, not sure about the L293D chip yet. I need to make an independent power supply for times like this, preferably switchable from 5v to 3.3 for electronics and maybe even 12v for motors. Just a bit frustrated, can't find my solder wick and thinking I've tossed it, because it was in a bin outside that got rained in and everything in it was ruined and smelled funky.

The plan was to supply 7.2 volts to the 7805 to drop it down to 5v so as to not over voltage the motors, yes would have a prob with some heat. Keeping a steady amount of voltage to the motors would have been worth it. Now have to switch to something else I think to at least get it going. May use 4 AA cells to supply 4.8v to the motors till I figure out what smoked the 7805. Just called radio shack and they don't have any desoldering wick, major bummer. Was hoping to take Leo to the RBNO at the Generator tomorrow in working order. Time to grab a wet paper towel and desolder the hard way. At least enough to get my 4 AA holder back and the bad reg out of the board. If the reg was totally blown and not shorted, I'd just bypass it with jumpers, but it's in circuit and I don't want to blow the L293D.

Old problem that is new to me

2009-09-24T02:56:00.000-07:00

Ok, I've gotten Leo to move some with his motors. Still working on the sensors and such. While working on the motors, one of the things that came to mind was the fact that any time his batteries are going dead, they have to be manually changed out. So if he's getting sluggish, then he has to be ripped open, partially gutted, sometime in the next few hours or even left that way for a day or three, then have freshly charged batteries shoved back in him, put back together and expected to perform just like he did before. Sounds violent and violating, I know so there is a much cleaner solution out there. There are seven batteries that have to be charged or changed out when he's go 'dead'. Yes this a problem for later, sense he can't even sense his environment yet, let alone navigate it. Suggestions are welcome.

I don't want Leo to be something I toy around with till I'm bored and then put on a shelf for who knows how long. Neither do I want him to be an electronic pet or pest. Like I said before, it's a problem for later. I believe that reaching any level of autonomy for a bot is a huge step. The roombas do their tasks, then return 'home', recharge and do their appointed task when their programming says to do so, so why shouldn't other bots do the same?

Got the motors running!!!

2009-09-23T06:05:00.000-07:00

The L293D does weird things if you have both Enable pins high and nothing connected to the one of the Half H-bridge. Got all the wires connected in the right spots and low and behold, it works even the PWM is going well. That means that I get to take from the solderless breadboard to the permanent breadboard. I'm going to add a few filtering caps first though on the 7805, batteries and wires pick up lots of RF and especially from the cheap DC motors that make up the muscle part of Leo's drive unit. The next step will be to work on his bumper switches and then work on encoding the front idlers with hall effect sensors and rare earth magnets.

Progress Report:

Body Modifications 33% (still have the shell and to remount the shovel to make it a neck)
Drive Train 75% (Motors mounted, L293D working just needs to be finished)
Sensors 0% (2 IR receivers and 8 IR LEDs bought, 3 switches that need mounted, etc)
Circuits built 10%(L293D x2, 38MHz IR circuit, bumper circuit, etc.)
Programing 1% (motor test program works)

What needs to be bought:

4 hall effect sensors $8
4 0.25" rare earth magnets $10
Sonar Sensor $25
PIR sensor $10
4 new motors $12
1 servo <$10
2 potentiometers <$5
Accelerometer $20
2 ardiuno chips $14

Total = $110 +/- $3 plus $20 in S&H

Motor probs

2009-09-23T04:13:00.000-07:00

Got the L293D hooked up, unsuccessfully. I can only get the right motor to turn in one direction. I don't know if I smoked the chip, will take the meter to it tomorrow. Both motors work fine and I have L293D hooked up as the spec sheet specifies. I probably need to add some 0.1uf caps to the 7805 regulator. Next time I find a bunch of them at a good price, I'm going to pick up a fistful of them. A bit worried that I smoked the chip and not in the good way you do ham and turkey. I can't seem to get the PWM on the arduino to work either, hopefully all these problems will be solved in the next few days. I'd like to add the bumper switches and probably an on/off switch so I don't have to chase him down.

For some reason, the right motor likes to stay in reverse.

Been busy and doing a lot of thinking

2009-09-21T20:36:00.000-07:00

I've been looking into behavioral programming and it seems like the way to go. As humans we don't have set paths we take and certainly don't have encoders on our ankles to tell use how many steps or 'clicks' we take. Still need a tachometer of some sort, just like our brain keeps track of how many steps we take from point a to point b, even if we aren't conscious of the count. Yes there are some blind people who can use a form of PID to get around quite well, but we don't use it that much. Yes, I know...we depend on our sense of sight to get around. Well, we have lots of sensors to choose from such as IR ranging sensors, sonic ranging and internal mapping.

Here's been my thought process of late. We have an internal tachometer(the soles of our feet), we have an inertial guidance system(our middle ears), object detection(our eyes), and internal mapping(our brain). So following that bit of logic:

Human Bot

Feet Wheels, tracks, GEV
soles Hall Effect sensors and rare earth magnets on the wheels
Mid Ear Accelerometer (preferably 3 axis)
Eyes CMOS camera
Brain High Powered MCU with lots of Memory

If you go along with behavioral programming, then you are taking low level actions to make higher level actions. Combine that with some neural programming in the decision making and threshold processing you have. If you think about, if a bot has it's right bumper triggered, it backs up the path it came till it can make a decision on how to get around what ever made the collision..

Just some thoughts

Tonight's Work

2009-09-05T22:35:00.000-07:00

Worked a bit more on Leo's body. All that is left after the glue dries is to do some trimming and the MCU shelf part is done. I've been trying to figure out what kind of hinges to use on the bumpers that won't stick. I had an epiphany while drinking a soda, reuse a soda can to make the hinges for the bumpers. It's quite fitting since Leo's body started off as a RC toy and has been upcycled to a bot body. I will be working on the front bumpers tomorrow afternoon.

My current dilemma is that I'm out of pins before I've begun really. I may put a pot or a resistor on the back IR distance sensor and make it digital, so I can have 2 analogs for the compass. Then I need to make the passive battery sensor only go low when it gets down to the point it needs to head 'home' to recharge. That would leave one digital pin unused, besides pins 0 and 1, which I already have plans for.

Progress Report:

Completed
Bot Body : Dozer Model(Modifications at 50%, Battery compartment and MCU shelf done)
MCU : AtMega328 Arduino

Incomplete
Sensors : 3 Bumper Sensors(3 Digital Pins)
5 IR Distance Sensors(5 Analog Pins)
Passive Battery Monitor(3 Digital Pins(75%,50%,Time to head home)
2 Encoders(2 Digital Pins)
Dual H : Still haven't ordered it(4 Digital Pins)
Compass: Still haven't decided which one (2 analog pins)

2009-09-04T17:28:00.000-07:00

This is is what it started off looking like:

From Leo the Bot

I took off the bucket assembly :

From Leo the Bot

Then made an addition to the body:

From Leo the Bot

I am making the shelf the MCU sits on a bit wider, the batteries and the L293D will be under that, to keep heat and the RF they might throw off to a minumun and keep the the center of gravity low

I'd like to enclose the MCU part, once I get a compass in it. Then reuse the bucket assembly to make a neck of sorts for a PIR module and for sonar to sit on. I know by then I'll be out of I/O on the AtMega328, so will have to get another one and figure out how to make them communicate with each other. I need to pick up some solder wick, since I've yet to find what I did have.

The home depot trip

2009-09-02T01:55:00.000-07:00

Got the wood needed for Leo's bod mods, got the ballast for the light in the kitchen though I think I'm going to take it back and get a new light though. Did some of the mods today. Really thinking of rethinking some of the mods. Don't want to run into weigt issues, or have too much time invested that I want to only use this bot body either.

Weekend is done, over, in the past

2009-08-31T17:50:00.000-07:00

Not a bot work this weekend, not a lot of sleep either, considering I put in a 40 hour weekend. It's done, it's over and I still want a nap.

Going to home depot tomorrow. Going to get the wood for the mods on the bot body. On the DPRG list there has been some concerns about the tracked body. Wheeled, tracked, legged or hovercraft type drives all have their issues. Most of them are enviromental though ie bumps in the carpet, wet spot on the hard floor, human picking up bot to examine it and setting it down somewhere else(course that could be me doing that lol) and whatever else that you can plan, program for and hope it works like you wanted it to.

I've been thinking of the navigation program and putting it somewhat on hold till the bot is actually moving and has working sensors. I figure the next few problems will be 'debouncing' the bump switches and adding on the IR range finders. Right now need to find a chip that is a volt meter to monitor the drive batteries. May have to go with a different rechargeable 9v battery than the lithium ion since they are $20 a piece and Ni-cd are around $5 to $10.

Goals for the end of September:

Modifying the body
Installing what sensors I have already(bumper switches and encoders)
Installing the L293D (yay for locomotion)
Installing the IR range finders

Tomorrow will be posting a progress report on Leo