MC68HC11F1

[Jon Lawton]

Okay... I want to use this Motorola MCU for my $100 grey-box embedded system. The only problem is... Digi-Key is fresh out, and Future FAI will only sell in quantities of 18 (seemingly, I'm going to call and beg tommarow)! I've checked Newark, Mouser, Jameco, just about everyone. Does anyone have some creative idea of how to get ahold of one of these? If Future can't help me tommarow, then I'm going to have to go beg Motorola, or search the deep dark corners of the Internet to find one!

...or are there 17 teams out there that *also* want one of these $10 SiO2 wonders?...

[bigqueue]

Originally posted by Jon Lawton
Okay... I want to use this Motorola MCU for my $100 grey-box embedded system. The only problem is... Digi-Key is fresh out, and Future FAI will only sell in quantities of 18 (seemingly, I'm going to call and beg tommarow)! I've checked Newark, Mouser, Jameco, just about everyone. Does anyone have some creative idea of how to get ahold of one of these? If Future can't help me tommarow, then I'm going to have to go beg Motorola, or search the deep dark corners of the Internet to find one!

...or are there 17 teams out there that *also* want one of these $10 SiO2 wonders?...

There is a autonomous robotics website that sells 68HC11 boards and chips that seems to advertise them.

Look at:
http://www.mrrobot.com/prices.html

Toshiba is also a 2nd source of 68HC11 processors. You might check out:
http://www.toshiba.com/taec/cgi-bin/...ategoryID=6964

I wish you luck....tell us if you find a source.

-Quentin

[dense]

alright! go for it, i'm just curious to know what you're gonna use the MC68HC for? The stamp controls provided will more than suffice the controls for the robots.

I was skeptical about the stamp at first and still am, but it's got some nice features, basic is easier to learn than the assembler in a crammed 6 week period.

We haven't been able to max out all 48 in/out ports, yet.

g'luck

[Joe Johnson]

You could buy a C library and use it to do some real math that would be impossible with the time constraints and programming limitations of the Stamp2/RC controller.

I don't know what you would do exactly, but you may put some signal through a FFT and do something clever with the frequency content you discover.

I don't know.

Another possibility would be to make the slave CPU (the Stamp2) just a pass through micro. It would be possible to have all the PWM values & Relay outputs controlled via the HC11. Again, what do you do with it? Not sure, perhaps not be tied to the 40 Hz Loop time... How about something that has a real time clock associated with it?

I would like to get into doing something like this, but in the end, the Stamp2 is going to have to do.

So many cool things to try, so little time...

Joe J.

[Jon Lawton]

Well... the idea is to have the system gather power usage statistics from each motor, and possibly speed of each wheel (Stored on EEPROM). It can help keep track of when that breaker is going to pop! Yes, I know the stamp-based controller can do this as well... BUT... I'd rather muck around with a $10 68HC11 than a $1,000,000 Robot Controller!

As for programming... assembler is just fine for me, thanks. I still don't understand why everyone is so scared of it! If you wrap common logic structures with macros, it's even pretty clear to read. I'm sure there are C compilers with a back-end for the 68HC11... (Does gcc have one?)... if not, writing one would make for an intresting project.

What else can you do with 4MHz of Lean-Mean-MCU-Machine? Who knows? But I'd bet at some point somone will think of a sweet use for it. I've heard of teams out-programming the stamp... this would take care of it! Just a thought. And HEY! I've been looking for an excuse to build an embedded system for a while now!

[bigqueue]

Originally posted by Jon Lawton
Well... the idea is to have the system gather power usage statistics from each motor, and possibly speed of each wheel (Stored on EEPROM). It can help keep track of when that breaker is going to pop! Yes, I know the stamp-based controller can do this as well... BUT... I'd rather muck around with a $10 68HC11 than a $1,000,000 Robot Controller!

I think we are looking at doing something like this with the dashboard program. We will see if this makes it into the final code....but it is a good idea. (have to keep track of different batteries as they do get changed through the day)

[JoeJ]

HC11F1 is a nice controller but you need to build a fairly complex board to hold it. An E9 would make more sense because you could run it single chip mode.

We are using a PIC16F877A. This is a flash part with 8K of program memory on chip. 38 I/O pins and a very easy assembly language (35 instructions). Also the HC11F1 is a PLCC which is hard to build on protoboard. PIC comes in DIP. Programing tools are a free download. For $100 from digikey the 16F877 has an in-circuit debugger.

Our system is running the PID control loop for the GYRO and sending GYRO controlled steer commands to the control system. The PIC communicates with the control system with a TI DAC and some digital I/O pins.

Our system also monitors motor current and provides a circuit breaker overload timer. This will allow short periods of burst power that are computer controlled so the driver doesn't have to guess.

[bigqueue]

Originally posted by JoeJ

We are using a PIC16F877A. This is a flash part with 8K of program memory on chip. 38 I/O pins and a very easy assembly language (35 instructions). Also the HC11F1 is a PLCC which is hard to build on protoboard. PIC comes in DIP. Programing tools are a free download. For $100 from digikey the 16F877 has an in-circuit debugger.



Our system also monitors motor current and provides a circuit breaker overload timer. This will allow short periods of burst power that are computer controlled so the driver doesn't have to guess.

Sounds kind of interesting.....I agree about the Microchip device. Nice parts for hobbiests because the development tools are so cheap!

It also sounds like you are doing something similar to us with the motor current monitoring (integration) and circuit breaker timer.

Our custom circuit only measures current, and the heavy math/calculations are done by our Dashboard program on the laptop. This way, we can do things like keep logs of our matches as well as wan our driver of impending doom of the breaker!

We also wanted to keep computers out of the decision making process because the risk/benefit of "pushing" the breaker changes with the number of seconds left in the match. (and also, we didn't want a system shutdown should the current measurement curcuit go belly-up.

-Quentin

[JoeJ]

We listen to the commands from the second cpu but we can always bypass the unit. That is just a matter of using some dashboard switches to bypass the commands from the second cpu. Very simple stuff.

The small circuit breakers are the only concern for us. The large breaker is VERY tolerant. The small breakers pop quickly so there is no time for the driver to make the decisions.

Our robot is very fast so we are only planning on overload conditions during acceleration. Influencing other robots will be handled by ramming the goal or running away. The only way to pop the main breaker is to have high loads on more than two drive breakers.

The heavy math is required for PID control loops on the Gyro and the optical autopilot for a goal follower. Performing the filters on the Gyro data will consume more memory and performance than is available on the robot controller.

[Jon Lawton]

Originally posted by JoeJ
HC11F1 is a nice controller but you need to build a fairly complex board to hold it. An E9 would make more sense because you could run it single chip mode.

We are using a PIC16F877A. This is a flash part with 8K of program memory on chip. 38 I/O pins and a very easy assembly language (35 instructions). Also the HC11F1 is a PLCC which is hard to build on protoboard. PIC comes in DIP. Programing tools are a free download. For $100 from digikey the 16F877 has an in-circuit debugger.

The HC11F1 has a "Single-Chip" mode, too. I am not running it in this. Guess how much "glue-logic" I need? One hex inverter, and one RS-232 trasciever (MAX-232, if that counts as glue). I'll post pictures and schematics soon.

I do agree that the PIC is king of simplicity. If the PIC does everything you ever need, then indeed, it is superior. I went with the HC11 becasue I don't want to find myself at the compition going: "WOAH, that other team did somthing REALLY COOL with their system. Too bad I don't have the [cpu power/eeprom space/whatever] to do it." I really almost went with some x86 system, but the glue-logic starts to get heavy real fast there. Not to mention I love the flexibility of the HC11 address space.

You can get PLCC->DIP adapters for about $20. Can't beat that PIC ICE for $100, though. That's good. If you run the HC11F1 in single-chip mode, you have I/O lines up the wazoo! In fact, you can use unused control lines *even in other chip modes* as general purpose I/O. Programming tools for the HC11 are, as far as I know, free. Does the PIC have on-die A/D conversion? 8 Channels of it?

Either way, for the kind of work we are asking these systems to do, I think either chip is a good choice. I just wish there were more teenaged hardware geeks around!

Good luck with your 'bot!

[JoeJ]

I can understand wanting to use F1's. I used to write quite a bit of HC11 code. I even used F1 for one project. I still have a few laying around. I have also used the free tools for the HC11. I don't plan on going back anytime soon.

Of course, '877 runs up to 20Mhz (5mips) and has 4x the memory in single chip mode. It also has 8 10bit ADC's, 1 UART, 2 PWM's, etc... In many ways it is very similar to F1 in single chip mode only faster with more memory. And 2 chips cost the same from digikey as the PLCC to DIP adapter. The only glue logic required to get it running is a crystal or ceramic resonator. If that is too expensive then a resistor and capacitor will also work for the clock. Programing is in circuit via the ICE. In fact the demo board that comes with the ICE has enough proto area to accomplish most of what people want with the second circuit board.


Of course I still can't figure out how you were going to build a minimal x86 board for < $100 and still have enough money to interface it to the rest of your system. Especially not with digikey prices.

[Jon Lawton]

Originally posted by JoeJ

Of course, '877 runs up to 20Mhz (5mips) and has 4x the memory in single chip mode. It also has 8 10bit ADC's, 1 UART, 2 PWM's, etc... In many ways it is very similar to F1 in single chip mode only faster with more memory. And 2 chips cost the same from digikey as the PLCC to DIP adapter. The only glue logic required to get it running is a crystal or ceramic resonator. If that is too expensive then a resistor and capacitor will also work for the clock. Programing is in circuit via the ICE. In fact the demo board that comes with the ICE has enough proto area to accomplish most of what people want with the second circuit board.


Of course I still can't figure out how you were going to build a minimal x86 board for < $100 and still have enough money to interface it to the rest of your system. Especially not with digikey prices.

Yow! Those PICs must have come a long way since I last looked at them. The 16F84 was the last PIC I used, so perhaps I shouldn't have been so quick to dismiss them. Oh well, I'm happy with what I have now, and that's what matters. I'll definately take a closer look at them for the next board! (Is your team sponsored by Microchip? They wouldn't happen to have any neat-o toys to give to us to, uh, sway us their way, would they? )

Eh? You don't think I can make a x86 board for <= $100 (DigiKey)? I'm thinking I accept that challange as soon as the 6 weeks ends!

[JoeJ]

Actually our team is sponsered by Microchip and I am an applications engineer for Microchip.

I did a quick price estimate and you could build a 16 bit bus 386ex for a hair under $100. approximately $66 is required for CPU, FLASH & RAM. The remaining $36 will go fast for interfaces, glue, passive components and a power supply. Just a CPU card is not too bad. Making it do something useful will take a few extra devices.

I am building a PPC 403 computer. The parts cost is around $150 but is has LCD/CRT, audio, a PIC comms processor and a few other things. But it is not for the robot.

Our robot computer is fairly simple. It has a quad DAC, 2 current monitor IC's, rs-232 transciever and a 16F877. Current shunts are on a second PCB under the controller PCB. They stack into the box neatly.

[dense]

Wow, can you lead me on the path to learning PIC or atmel hardware and programming. The motorola microcontrollers sound cool too. There's not alot of microcontroller adepts around where i live, even some freinds who go to college don't do alot of it and highshoolers arn't much help either.

The most I've done is mess around with a non-flash PIC 17 or so, trying to build a serial interface, with some eeprom.

Do have any suggestions of books or websites for a beginner?

[JoeJ]

Your best step to get started is to purchase a Picstart plus.
This is a basic programer that covers all of the devices up to 40 pins in DIP packages. Next, get a few 16F628's. These 18pin devices include a serial port, internal clock, internal reset and a PWM. There are many projects that can be easily built with this start.

For books, there are a variety of introductory pic books but the best start may be the appnotes that are available on the microchip website. These are tutorials and miniprojects on using PIC's in a variety of ways.