Progress update December 16, 2006:

As you can see from these two images I have made great strides in building the mobile base.

Today I:

• Mounted the two lower decks where the casters mount
• Mounted the casters
• Machined some standoffs from acrylic rod (see the second image)
• Attached the motor mount assemblies with angle brackets (very stiff and perpendicular)
• Made a mount for the sonar (using 3mm PVC)
• Redid the sonar cabling (using 22ga stranded versus the solid, and following my wiring standard layout)
• Mounted the sonar

From the front:

Note: the front lower deck will also be the mount point for the line following sensors. You can see two of the angle brackets I'm using to shore up the motor mounts. At this time the bot weighs just over 2 pounds ... largely due tot he motors and encoders.

From the rear:

Note: I used the acrylic rod because I have 20 feet of it that a friend gave to me. I am considering lighting it with LEDs … but that's for a later time. This deck with also hold the Lipo pack.

To do:
It might help if I actually kept up with my todo lists but I use them as a brainstorm tool and right now my mind is racing with ideas.

1. Resolder the motor controller circuit
2. Rewire the encoders with longer leads using my wiring standard (which is to say the same as a servo lead) | GND | 5V | SIG1 | SIG2 …
3. Get this bot mobile with some simple code to count encoder pulses and drive a square
4. Add in PID to start using waypoints and let the "algorithm" decide how to get from A to B

Onwards,
Jay

December 11, 2006 progress,

I was on a roll this weekend and have some photos of my progress to share.

This first photos shows how I printed templates for my motor mounts. I first cut the mounts to their final width and height, glued them together with wood glue (tried spay adhesive first but it was too messy), glued the template on, marked the holes with a center punch, and drilled the mounting and clearance holes on my drill press. The photo below is my first design. Taking a page from my RC airplane days of constructing planes from foam, I added tabs and slots.

Note: I originally drilled the holes the same for both side, but due to the motor shaft being offset it was not symmetric like the encoders. Also, the mounting hole closest to the output shaft is too shallow for the screws I had. What is, was symmetric are the two mounting holes at 180 degrees which is what I am using.



The second image shows the motor and encoder mounted along with the gears and wheel. The motors are heavy and the assembly feels like it weighs half a pound or more (will measure it and update this post).



To add the mounts again I when with a tab-slot design. I printed a template for the base to locate the slots and used a razor knife to cut them out. I also had to make slots for the small portion of wheel that protrudes through the base.



Next steps are to:

• Strengthen and square the mounts using l-brackets and a cross bar of some sort. I have some acrylic rod I am going to try. Failing that, I’ll add a skin of 3mm PVC along the bottom to tie them together.
• I have decided to redo my motor controller board using the ribbon cable wires versus wire wrap. The insulation is better, and I had an issue with intermittent shorts when I tested it before.
• Create the cables for the encoders to the interface board
• Make some mounts for the caster wheels (it’s really back heavy right now so 1 may be enough)
• Write some code to drive in a square

Onward,
Jay

December 4, 2006 update:

Well, I received the gears and wheels that I ordered. The wheels from www.budgetrobotics.com are great.

I used to be fairly proficient with my Taig microlathe, but not Sunday. The gears are bored for .25” (1/4”) but the shaft diameter on the motors is .234”. I decide to make an adapter that would fit better on the shaft and integrate with the wheels and gears. I tried to use some Fortal aluminum, but it wasn’t cutting as well as I had remembered. It took my quite an effort to hacksaw a chunk to chuck in the lathe, and then it kept bogging down on anything but light cuts. It would have taken a day for one adapter to whittle 1.5” down to .375”. So I switched to nylon. All told it took 5 hours to crank out one adapter.

The good news is it works great; the bad news is I have 1 more to make. I’ll post a photo tonight.


Onwards,
Jay

11/28/2006 update:

This past weekend I did indeed get some “me” time to work on my bot. I spent most of Friday making the cables and soldering the interface board. I decided to use ribbon cable (old IDE cables) for the point to point wiring versus wire wrap. I really like the way it worked out and can recommend other give it a try too.

I decided against creating four 8-pin cables from the port headers on the MAVRIC to my interface board. Instead, I designed it so that I used all of PortB for one cable and a combination of PortF, PortE and PortD for the other. In other words, I went the wiring harness route.

From left to right, simple switcher power supply, MAVRIC, interface board, encoder, DC gearmotor

I also decided against reusing my tank base for this bot. The vacuum cleaner belts were starting to dry rot and I really wanted to redo the mechanics of the drive. I spent the rest of my weekend looking for solutions to couple the motors, wheels, and encoders. When I finally settled on gears I then began the search for affordable gears. www.servocity.com have a nice selection, but they are very proud of their gears. $4.95 for 1 gear and it doesn’t even have a set screw not is it bored for the ¼” shaft. Instead they want to sell you a hub for another $4.95. I have a micro lathe and can make my own hubs, but $4 is too much for a single gear IMHO. I found Stock Drive after trying www.servolink.com (great prices but $30 minimum order) and MSC (no details on the gears). They hit me with a $3.95 handling charge for using the online store and then another $6.18 for UPS for 4 freaking gears. The gears were reasonably priced at $2.10 a piece. If they had offered USPS I would have picked that. It’s a major pet peeve of mine getting overcharged for shipping. But I digress.

Oh, and I also ordered some wheels ($10 minimum order) from www.budgetrobotics.com . After deciding on the gears, I was able to knock up a model in CAD. The silver motors will be aligned with the center of the 9" diameter disk platform. The red encoders will be linked via the gears. The wheels will be attached to the gear on the motor shaft via 4 screws. I am planning for the gears to be a press fit, but if not, the hubs on the gears or thick enough to drill and tap for a set screw. Failing that, I'll bore out the hub and make a sleeve hub out of aluminum.

Onward,

Jay

This iso wireframe shows how the gears come into play. The base is made from 6mm pvc.

11/15/2006 progress:

What have I been up to? Well mainly I have been thinking about how to physically interface my sensors and peripherals to the MAVRIC. The MAVRIC has 6 8-pin headers (he offers screw terminals too) for each port of the ATMega128 cpu. I am being playing with my pin layout to consolidate as many pins on a port as I can. Some pins are too specialized for this (A2D, UART) and thus I have it down to 4 ports.

Also I have three serial devices but only two serial ports. I need to play with bit-banging to one of them (most likely the voice module).

I have not touched the bot in over a week, but I did place an order to www.sparkfun.com for some odds and ends. This weekend I am testing for level three in Krav and then I’ll work on a interface board. But then I’m still in a circle. Do I need an interface board, or do I create custom wiring harness with 5 8-pin connectors on one end and sensor connectors on the other. Or do I create 4 8-pin bus connectors to a connector board and tie the port pins to their connectors using wire wrapping or point to point soldering and then run another connector from the board to the sensor. I like the second method better but it’s more connectors and thus more to screw up.

Here is what my connector board looked like on a previous robot: http://www.geocities.com/jaydmdigital/ieee/jamaica.html 3rd picture.

After I get the whole connector thing decided, I can work on the PID, line sensors, IR sensor, etc… so you see it really is a key decision.

Hmmmmm…

Jay

11/2/2006 progress:

I tested my PWM code (and corrected a typo in my blog) tonight. I’m happy to report that locked anti-phase is working. The “sings” at the locked position but I can live with that (I think). Hopefully it won’t be louder than the voice chip speaker. Also, using my RC wattmeter my motor never goes above 1A even with a locked rotor (i.e. my stopping it by hand) with a 12V motor source.

So next on the agenda is PID testing. Then onward to line following.

Onward,

Jay

11/1/2006 update:

Things have slowed a bit. I decided that the best way to test the line following was to have a moving platform. Therefore I decided to rebuild my H-Bridge circuit. I decided to compare the circuit I built nearly a decade ago with the reference design. Wow, not sure what I was thinking but my component values for the RC blanking circuit were way not right, and my current sense resistor wasn’t even connected.

Locked anti-phase
Previously I was using PWM of the enable pin for speed control and the phase pin for direction. Recently I read about PWM of the phase call locked anti-phase control. At 50% the motor is locked and thus neither moves forward or backward. If the PWM is less than 50% then it moves backward, greater than 50% forward. How much greater or less than 50% controls the speed. So now you get speed and direction with one control line.

New circuit
I soldered up my new circuit with more correct values except the current sense resistor. RS doesn’t carry a .5ohm resistor and paying $5 for a resistor is nuts. Long story short 10ohms was too big and I ended up removing the current sensing and blanking components from the circuit to make it work. Also I need to get a smaller tip for my iron as I had a short (or two) that I had to trace down too. Once I’m sure on my circuits, drafting a PCB seems like a good idea.

Testing
To test my circuit I first tried it as a servo. I must have been high from solder fumes because I knew that PWM of a servo doesn’t use a duty cycle it uses pulse timing (1.5ms for center, 1ms for 0 degrees, 2ms for 180 degrees roughly). So as a quick and dirty I bit banged a 500ms 50% duty square wave. The motor was locked alright but it was singing too. So I tried smaller periods and settled on 50ms. I’m hoping that using the built in phase correct PWM of Timern that I can get a nice smooth 127 steps in both directions with little wasted power at “stop”. I am thinking that I’ll need to add a brake line for each motor to de-energize the motors at 0 speed (50% duty).

PWM
There are 4 timers on the ATMega128 Timer 0 and 2 are 8-bit and Timer 1 and 3 are 16 bit. All are capable of PWM using output compares. If I only need/want 127 steps then and 8 bit timer is enough, but I can see no penalty in using a 16 bit timer either. So here is the scenario. Using the phase-correct mode of operation the period of the waveform is set by the WGMn3:0 bits. Either 0x00FF, 0X01FF or 0x03FF. For 8-bit 0x00FF is what I’ll use. TCNTn will count up to 0x00FF and then count down to zero. If OCRnx == 0 then the output will be zero. Similarly if it is 0x00FF (or TOP) then output will be 1. So first step is to decide the actual period and set the prescaler to get close. The CPU clock is 16MHz. A prescaler of 64 will give us a 4us timer clock source. This means our period will be roughly 2ms (2 * 255 * 4us) or 500Hz. Now we need to configure the Output Compare bits in TCCRnA. We will configure them to 2 (10b) to clear the output bit on up counting match and set it on down counting match. Finally we need to set the DDRx bits for our output pins.

void timer1PWMInit (void)
{
// Set prescaler
TCCR1B = 0x03; // Timer clocked at F_CPU/64

// Init to 8bit phase correct mode 1
TCCR1A = _BV(WGMA0);

// Configure Output compares for A and B mode 2
// output bit will be cleared on match while up counting and set on match while down counting
TCCR1A = _BV(COMA1) _BV(COMB1); ;

// Set DDRB pins 5 and 6 for output
DDRB = _BV(OC1A) _BV(OC1B);

// set motors to 0 or 50% duty for locked anti-phase controller
OCR1A = 0x7F;
OCR1B = 0x7F;

// Enable timer1 overflows
TIMSK = _BV(TOIE1); // enable TCNT1 overflow

}

Onward,
Jay

10/25/2006 progress:

What a good day it turned out to be yesterday. I went home at lunch to fix the serial cable connector. I booted it up and tried the Soundgin software. YEEHA it works! I spent some time playing with the vocabulary by adding my own words to the listing. So now I move on to the line following sensor.

I have decided that each sensor will be individually mounted and wired. I will use 6’ servo leads. I am leaning towards the QRD 1114 part for ease of mounting. I will try analog first.

Still need to get some proto PCB. I’d rather use the solder mask plated thru hole material vesus the RS stuff, but it’s hard to tell what you get if you order it from a B&W photo. Anyone have any part numbers for me to use. I am thinking 4”x 6” or 2”x 3” .1” PTH with solder mask and mounting holes in each corner.

Onward,
Jay

10/24/2006 update:

I decided to take a break form programming and build my audio/voice circuit for the Soundgin. I spent about 2 hours Sunday laying out and soldering the circuit provided by the manufacturer. I didn’t get a chance to test it until that night and it didn’t work … well it did make noise, but not what I was expecting. So I had to wait till lunch Monday to do some debugging. Rather than look at the schematic and trace the line, I draw the schematic based on my soldered circuit. I used that link to creating your own graphing paper to generate PCB graph paper (i.e. dots versus lines) spaced at .2” so I could see what I had. Everything was correct but I had forgotten one connection from the low pass filter to ground. I added it after dinner last night and bingo, I get the demo with plays all the allophones to work.

This morning before leaving for work I added a header to allow me to jumper the demo input to ground, and either use 5V for the LM386 or a 9V lead if I need/want more volume. I also plan to play with adding capacitance to the gain pins.

I had hoped to test the Soundgin software with an RS232 to TTL cable I made ages ago for an HC11 Botboard. Unfortunately, I made a guess as to which pins should RX and TX when I soldered my circuit to the header and I guessed wrong ;) Oh well, with my new crimps I’ll be able to redo the cable in no time. You see, I tried to fix it in the field and ended up pulling one of the wires out ... doh!

Well, back to my list. I think I’ll work on #3 next. That’s the line sensor circuit. That will also require #2. Then I’ll tackle #5. I have 5 of the 3252SB chips in my inventory.

Onward,
Jay

10/21/2006 update:

Progress has been either stopped or else moving backwards. I’ve been down the very dark street of the TWI on an Atmel device. Long story short, I can’t trust the TWI not to freeze and thus block in my code. So I’m dumping it, the I2C LCD (which I think I may have shorted out too L ), and the last few days of my life. Yeah it’s been bad, but on the the good side, I now have a good handle on how to use the debugger.

So, now I’m dropping back to serial for my LCD (only 16x2 and 9600baud versus the 20x4 and 100KHz). I also grabbed a copy of FreeRTOS which has an AVR port though I’ve had to do some work to port it to mega128 versus the mega232. I need to get my srf04 and encoder code (I intend to pull it out of AVRLib to keep it simple) and get a basic system to have the display updating 5Hz, the sonar measuring (pinging) 10Hz using an external interrupt, and the encoder as an interrupt.

Oh, and all my parts for making wires came in as well as my sound chip and components to get that circuit built … too bad I fried my brain (and nerves) on the $^%^* I2C first.

Onward,
Jay

October 16, 2006 to-do list:

Things are starting to get a bit overwhelming for me. I tend to get ahead of myself and end up second-guessing, blue-skying, and just plain day-dreaming. In other words, I stall out until I can procrastinate no longer and I end up doing marathon sessions. In an effort to prevent that from occurring this time, I have decided to create a to-do list and try to link any dependencies so I can see the forest and the trees.

Note not listed by priority.

1. Build some cables and a break out board.
Status: The parts for the cables has been ordered and should be here by Wednesday. I have a PCB protoboard from Radio Shack but it may be too small.

2. Make a sample of the contest board and line for testing.
Status: Not started. Not a high priority but will need it when I test line sensors.

3. Make a line sensor and test different parts and configurations.
Status: I have QRD1114 and QRB1114 parts and can either feed them into the A2D or to digital I/O via an inverter. I have used both and am leaning towards the A2D.
Dependency: Maybe on wire making parts … but not required for prototypes

4. Make a terminal serial cable for debugging with AvrX
Status: Need to get a surplus serial cable. Then just connect the wires to the MAVRIC. I’ll search for one at lunch.

5. Rebuild/build H-bridge circuit.
Status: I have my original (5 years old now) Allegro Micro A3952SB circuit board and some of the newer A3959SB parts. I need to redo the circuit with cleaner wiring.
Dependency: Wire making parts

6. Design the shape and size of the robot.
Status: I found this link to a website that let’s you print your own graph paper. Will start sketching some designs for the wheel layout and base shapes

7. Design the motor mounts and encoder mounting
Status: Currently I have a tank-based mobile platform with the encoders serving as the other end of the tank tread. I will need to design a way to transfer the shaft motion to the encoders without interfering with the wheels. Also, I’d like the make the motor/encoder mount as a module that can be removed/installed as a component incase I ever want to change the base shape/layout.

8. Prototype a circuit that will decode the 1200Baud I/R signal.
Status: No started but do plan to use one of the UARTs on the MAVRIC. Have not decided on photodiode or phototransistor and will need to do some prototyping.

9. Add sound/speech.
Status: Have ordered components but am backordered on actual voice chip. No ETA. Will call manufacturer to get their opinion and may switch to competing product depending on timeline.

10. Build ball launcher.
Status: I have a pretty good idea for the design using a PVC elbow, a brushless DC motor (because I have one), and BLDC controller using a servo output. At this point I don’t have any real size constraints. The module should be sell contained and thus make it easy for me to change position and angle of it. At this point I am looking for the launcher to sit on one side of the robot.

Onward,
Jay

October 15, 2006 progress:

One step back, two steps forward. I became obvious to me once I started trying to put a few sensors together and control the rate at which they scanned and displayed their data that I need some sort of task scheduler or RTOS. A quick search on Google and it brought up one I had read about but never used AvrX.

So I have sent the last day reading about AvrX and experimenting with shoehorning my existing sensor drivers into it. After some initial issues with the makefile I was using I have successfully grafted the encoder, i2c, and matrixglcd functions onto some example code Larry provides. I love quick wins it just boosts my morale.

So now that the quick and dirty is out of the way, I am reading everything I can about how and when to use the AvrX api. I have joined the Yahoo Groups support group as well. I hope to have the SRF04 code added by tomorrow. However, it looks like I may want/need to make a serial cable to enable debugging first.

Onward,
Jay

October 12 progress:

I got the AVRLib i2c library test working and then molded in into what I needed. To do this, I created a i2cLCDsendbyte function to replace the uartsendbyte in rprintf. Now I’ll remount the processor, the new power supply, then new LCD (122x32 or 4x20 with the option of using a 240x64 model I have), the servos and an interface board (when my parts order arrive). If I have time tonight I will modify the SRF04 and encoder programs to use the i2cLCD functions.

Onward,
Jay

October 11, 2006 progress:

Microcontroller: The replacement parts have arrived from Brian but looking at how small they still are I’m a bit hesitant to try and solder them in just yet. Besides, it’s working fine without them. In truth I’m going to wait till this weekend when I have some free time at home.

As for programming, I now have quadrature encoding working flawlessly between ARVlib and my 1000PPR encoder. The nice thing about the encoder lib is that it keeps track of the position of the wheel (signed number of ticks) either forward or backward. Next step I2C to graphic LCD. That and I plan to modify my SRF04 code to be standalone from AVRLib for posting on AVRfreaks.net

Voice: I pulled the trigger and ordered a Soundgin (backordered according to the vendor I ordered it from) and the parts to build up a protoboard circuit (amplifier, low pass filter, etc…) most of the parts I found at www.hobbyengineering.com

Odds and ends: Connectors is always something in the back of my mind. Therefore I also made a order from www.action-electronics.com for some Molex KK shrouds, pins, and headers. I won’t have to worry about connectors ever again :)

Onward,
Jay

October 8, 2006 progress

Sunday was 4 hours of yard work and a very unhappy little boy so minimal progress was made. I decided to research speech chips and got off on all kinds of tangents. Makezine.com BLOG is addictive and I was up way too late Saturday night reading it.

So anyway, in a former life I was a midnight engineer for a buddy’s fledgling company. I had designed and programmed a device to translate the clock synchronization signal of one master clock to that of a tower clock controller. This was going to be a big product for his company so he had several PCB’s manufactured along with several sets of parts. Unfortunately the company went bust. I was contacted by the major investor and asked if I wanted the leftover electronics stock before it was “tossed”. Long story short I have a surplus of PCB’s, LCD’s, relays, etc… The reason I explain all this is because I had designed the system to use a switching power supply and had designed the 5V DC supply in one corner of the board. A few minutes with a hack saw and I had a nice PCB for a 5V DC switching PS. I also had all the parts to stuff it and as luck would have it, the diode bridge layout matched the .156” power header part. So, now I have a very stable high current PS (5A IIRC) that I can power from my stock of RC batteries (LiPo and NiMh up to 18VDC IIRC). I will go back any look at my notes to see what the design criteria were, but the system was capable for switching 4 relays at the same time if need be.

Back to the speech chips. Devantech’s is too expensive (at least for now) so it’s down to the SpeakJet or the new (and as far as I can tell not yet released) Savage Innovations Soundgin. Interestingly Magnevation which makes the SpeakJet is a partner in some fashion with Savage and both chips are very similar in that they use phonemes to generate speech. They are the same price but the Soundgin’s manual and software have given it a very strong lead on my score sheet. I’ll have to hack together a demo board with amp, but the schematic is presented. Since both use RS232 communications, I will most likely switch to a I2C LCD (a good excuse to play with I2C again) maybe even a graphic one I have laying around.

So, I think I will spend lunch making a parts order for a speech circuit J

Onward,
Jay

October 7, 2006 Updates

Wow what progress! Slowly but surely I was able to get the SRF04 sonar module working. Not without a lot of lost time on false starts, dead batteries and general ignorance on my part. I have created some notes for you to consider if you are going down the same path.

* The SRF04 outputs a pulse whose width in us = the distance. To scale it to cm divide by 58 and to get inches divide by 148 per the instructions.
* You can use an external interrupt to measure this pulse by storing the timer count on the rising edge and then subtracting the second time on the falling edge.
* External INT0-3 do not allow “Any Edge” triggers (i.e. both rising and falling edges trigger the interrupt) 4-7 do.
* A 16 bit timer has fewer overflows and helps limit erroneous readings
* The timer counts you use are in CPU ticks not us. Depending on how you configure your prescaler a tick can represent .0625us up to 64us. Since your measurement is in us you will need to convert your ticks to us.
* The SRF04 has a 3cm minimum distance (100us) and measurements start at the element, not the PCB


So using Pascal’s AVRLib I chose to use Counter/Timer1 which he configures with a prescale of 64 (each tick = 4us with a 16MHz clock). I also chose to use INT4 which is on PORTE pin4 configured for both rising and falling edges. IN my ISR I check to see if it is the rising edge (PINE PE4 =1). If so, I capture the current tick count (SonarDistance = TCNT1). Otherwise, I set the SonarDistance = TCNT1 – SonarDistance because the TCNT1 is set to count down.

Now all I need to do is write a function that will ping the sonar X times and return the average.

Depending on how much more time I’ll have today, I’ll start on the encoders.

Onward,
Jay

Update for October 05, 2006

Well on the physical front things aren’t so good. I have had chronic back issues (muscle spasms) ever since 7th grade (maybe even before) where I am somewhat debilitated and unable to stand. Yesterday I woke and was not able to stand up without a lot of pain. It’s hard to concentrate when you have a constant pain in your lower back so I stayed home from work.

On the robotic front I made huge strides. The spasms are weird in that the come and go in spurts (Ibuprofen helps to relax the muscles). When I felt a relaxation I went into my computer room to work on playing with AVRLib. Specifically, I wired up serial communication to a serial LCD and servo control.

Serial LCD:
This was cake to wire as the MAVRIC has screw headers from the Maxim 232 chip. The serial LCD from http://www.crystalfontz.com/ can take level shifted 232 or TTL 232. So I enabled the onboard RS232 pullups (via the dip switches) and gave it a try. It worked, but needed some tweaking to allow the LCD to init before sending any text. It would be nice to have a bigger LCD (4 line, 20 chars but anything is better than nothing).

Servos:
Now this was a PITA. Long story short, I already knew to change timer.h to timer128.h to accommodate the Mega128 device. However, I also had to modify the makefile to do the same thing. I had a 3 hour debug session wherein I dragged out my O-scope to watch the pin transition. After I got one working, adding more was cake. The AVRLib allows you to control up to 8 (any port, any pin) servos simultaneously by using the 16-bit timer/counter. Seemed a bit jerky, but that’s probably due to my adding printf statement to show the position of the servo before updating it to the next value.

I also made a mounting board for the LCD, MAVRIC, and servos. What I’ve learned thus far is I need to add two toggle switches. One for the power to the MAVRIC as there is no reset switch and 1 for the servo power. Come to think of it adding 1 more for the drive motors would be a good idea too.



You can see the MAVRIC, serial LCD, two micro servos,
my wire wrapping connections for the servos, oneof the RC battery packs, and the Ecro ICE Cube JTAG unit.


Onward,
Jay

Had a chance at lunch to play with the MAVRIC some more. Went through the intro document that comes with WinAVR. Was able to enter a debug session in AVRStudio and step, watch, modify register and program variables. One thing I learned is that everything slows way down from real time.

I then proceeded to debug the hello world program which uses interrupts. Ended up having to modify the program to get any realistic speed, but the program did step as it should. To improve the speed, I will play with breakpoints.

I then modified an AVRLib program using his library to do the same thing as Hello Wolrd. After I figured out the different places I need to check to make sure the right library and target chip is configured it works too. All in all a very productive day.

Next steps. Time to play with some hardware. I will add a serial LCD, a server, and a quadrature encoder. maybe even a push button ... I think I have a "test" board with some LEDs and buttons on it from a Rabbit2000 kit.

Onward,
Jay

Microcontroller board update

Good news, with Brian's (of www.bdmicro.com) patience I have successfully run a test program on the Mavric board. Of course had I done everything right last night I could have reached this step sooner. Here's what all I had to do to get this thing to blink the LED as it should:

1. After soldering the rest of the thru-hole components on the board and getting the go ahead from Brian about the missing (but in the mail) caps I decided on an RC reciever battery (5V DC) to connect the Vcc and GND pins bypassing the LDO regulator.
2. I installed AVRStudio 4 and it's service pack, WinAVR (except PN), and AVRLib
3. Following along with the ICE Cube quick start guide, I atteched it to the JTAG header, connected the battery, and ran AVRStudio.
4. Using the Tools menu I selected AVR Prog to loag the JTAG ICE image to the ICE CUBE. Worked perfectly
5. Using the command line ran make on the Hello World sample from Brian's site. Compiled fine.
6. Using the JTAGICE under Tools programmed the board with the .hex file generated. Programmed fine.
7. Nothing ... no blinking. After an hour or so, I decided to look at it with fresh eyes and went to bed
8. I decided to read the user help in AVRStudio (I know ... wierd isn't it) and read the following under known issues: "ATmega128 in ATmega103 compatibility mode: Not supported, but will be available in a later version of AVR Studio 4." I remembered seeing, reading, something about 103 compatibility mode and found it under the Fuses tab in the JTAG ICE screen. Cleared the fuse and now was seeing a very slow (say 8hz blinking LED). Ok, so it works ... kinda.
9. Read the ATMega datasheet and run ferquency calc in Excel to confirm register settings.
10. Brian's email telling me that the chip has not been programmed for the 16MHz external clock and that the fuse sertting are in the Mavric user manual. DOH! That's where I saw them.
11. Set the fuses per the manual, reset the code fro the Hello World same program ... viola!

Next step, figure out how to use JTAG debugging. Oh, and I've noticed what may be a solder bridge on two of PortB's pins that need to be inspected closer.

Onward,
Jay

Encoders and odometry

Along with knowing how many RPM my wheels will be spinning, I can calculate velocity. With the dimensions of my wheels I can calculate distance. And with the dimensions and layout of the wheels to the center of the bot, I can calculate angle. David Anderson has a great site to dicuss these and showcase his robot SRO4. Be sure to have a look at his writeups towards the end of the page.

Also, I have found a nice lab pdf that uses the AVRlib to do quadrature encoding.

Onward,
Jay

Coding 

On the trail of PID motor control I found this interesing library.
This may save me a lot of "reinventing the wheel" and allow me to
concentrate on high level tasks ... just as the author intended ;-)
http://seasight.dforge.cse.ucsc.edu/details/data_acquisition/ATmega128/avrlib/docs/html/index.html

The more examples the better.

Onward,
Jay

Microprocessor Update

Well I learned a few things this past weekend. First and foremost my SMD soldering skills are rusty. We perhaps I’m being too critical of myself. I have never actually tried to solder a discrete component smaller than 0805 and the bypass caps in this kit have to be 0402 or smaller. Smaller than a grain of rice for sure. Anyways, I got through 98% of the SMD components on Friday night but two of the .1uF bypass caps were lost as my tweezers shot them into wherever. I contact Brian to get some replacement parts so maybe those will get here today and I can knock them out later tonight.

As I couldn’t finish the board I decided to wit on the thru-hole parts too. To pass the time I started re-reading my Mobile Robotics book and looking for online content along the lines of PID motor control code using the ATMega128. I found a very promising site (still investigating) at www.drtak.org

I have also been playing with some driving strategies for the course which may end up affecting my base platform’s shape.

Onward,
Jay



 

Microcontroller

I have ordered a microcontroller kit from www.bdmicro.com to use for my robot. I have used HC11’s HC12’s, Basics Stamps II’s, and OOPics on former robots. In my engineering life I have played with PICs , AVRs, and Rabbit2000s. My favorite, the AVR. Thus when I found the ATMega128 kit I felt at home. I prefer C to ASM and Basic and the AVR is optimized for C. I also wanted the built in features the AVR offers (A2D, PWM, I2C, etc…). So my kit will be here today and I will have either a long night or a short weekend of SMD soldering and testing.

I also order the ICE from www.ecrostech.com as I only have a STK400 and don’t need want the STK500. I will be using WinAVR and it’s associated tools (though not PN, I use Ultredit32 and love it) and AVRStudio from Atmel for the actual flashing.

I know for certain that I will have a serial LCD (as I have a bunch of them I bought when a local business went out), will do line tracking, will use a sonar (I have one so why not), will try to incorporate a CMUcam (I have one and have always wanted to use it so…). I would like to implement speech/sounds but this is “bells and whistles” stuff so it’s on the back burner.

Onward,
Jay

Projectile motion

Well, I can't seem to get my mind off this thing. I mean it's really simple and yet I find myself considering and reconsidering materials, configurations, issues ... it's a flaw I have. I actually get more done when I don't have much time and I'm shooting from the hip. I can make things work versus engineering them to (that's my take anyway). So yesterday I spent my lunch rediscovering the formulas for projectile motion.

When a ball is launched at an angle (theta) the only force acting in the x direction (horizontal) is the initial force that propels the ball in the x direction. But the y direction has gravity that works against it. My thinking was that I know the height and range from the hoop. At what angle and what linear speed would the ball have to be shot to hit this point at it’s apex (upward velocity = 0)?

I created a spreadsheet to play with different angles to determine the velocity of the ball. Then using the calculated values, I plugged these back in and plotted the flight path with respect to time. I added a vertical line to show the hoop. In this way I can actually change the angle, initial velocity and launcher height to see the trajectory and time to target. Have a look below.



In addition, I calculated the RPM of the motor for the specified velocity and then reversed it to specify the motor speed to calculate the linear velocity so I could play with the angle. It’s quite useful, but we’ll see how reality plays out.

Onward,
Jay

Base

For this bot I plan to use PVC that my brother gave me. It's 1/4" thick and most likely the same stuff sold here http://www.budgetrobotics.com/shop/?shop=1&cat=103 I am able to cut it with a razor knife, but I have cut myself a few times in my excitement to get a base mocked up. I'll try my Dremmel to see how it goes.

I went to Radio Shack and picked up some discontinued VEX motors. In reality they are nothing more than Futaba S148 servos modified for continuous rotation. That and they use the VEX square drive shafts and clutchs. At $9.97 ea I thought I ws getting a good deal ... though I'm not so sure of this. Servos are slow and 4' x 6' is a rather big field ... I am seriously considering using my old tank base with DC motors and encoders already mounted.
In an effort to play with prototyping a new wheel layout though I mounted 1.5" strips of Velcro on my cutout base so I could play with motor/wheel spacings. This is on hold as I'm brainstorming my ball launcher too much.

Tank base

Prototype Base. 8"x8"

Onward,
Jay

Ball Launcher

I actually started the process back in August. The first thing I did was mock up a ball launcher. It is/was a simple affair. I used a surplus DC motor (I dismantle and keep any electronic and mechanical bits from old PC's, VCRS, and thinks to my dad photocopiers ... my wife calls it junk so it's my junk box 'o stuff) a scrap of PVC two sheets of paper, tape, a foam wheel (from my RC landing gear), a rubber band, and an old AT power supply (+12V and +5V). Using a 40mm ping pong ball as my guide I rolled the first sheep of paper into a tube and taped it. Then to stiffent the tube I added another sheet of paper. I mounted the wheel on the motor shaft using a piece of the rubber band to take of the slack ... it works well actually and if a trick from RC. Then I cut a notch out of the PVC a bit smaller (always cut smaller as you can shave off material easier than starting over ;) than the diameter of the motor case and wedged the motor in. I used an exacto knife to cut a slot in my tube long enough to accomodate the foam wheel. Then using the 5V I ran some tests. Conclusion: it works fine, and so I will continue to refine my design.

[Add photo here]

Onwards,
Jay

This is my first post to this blog. I'd thought I'd explain what my intentions are and what the content will contain.

Background: I have been playing with autonomous robotics since my sophomore year at college. The Southeast IEEE student chapter hosts an annual Hardware competition. It involves a team of student constructing an autonomous vehicle to navigate a "game" board to complete assigned tasks and compete head to head against other student teams in the same region. While a EE has the electronics understanding, practical knowledge (and I only speak for my school) was nil. Worse, mechanical and machining knowledge were not included in the EE curriculum (although helpful subjects like art appreciation and philosophy were :rolleyes:) So a lot of it is OJT (on the job training). I have attended 4 of the competitions and helped with a fifth. This year I am advising my alma mater's team while building my own bot to compete with the same rules but not to be entered. It's good to have a set of goals when you start something like this so you know where you're going and how far you've gotten.

Intent: My purpose for this blog is to keep anyone interested up to date on my progress as I construct and code this bot. Hopefully you can learn from my experiences and I can be “kept honest” by feeling the need to make frequent updates and progress.

Content: I will try to make this a media rich blog with progress pictures, maybe video, charts/graphs and links to interesting sites I’ve uncovered.

Rules: The rules to the competition are http://www.southeastcon.org/2007/doc/SoutheastCon_2007_Hardware_Rules.pdf Broken down the robot will need to navigate to a code box, received/decode the IR transmission, then navigate to the ball box and present the code to get 3 more shots. To me the most fun part of this is the shooting of the balls through the hoop. This is what I’ve been working on the past week or so.

Onward.

Jay