Society of Robots - Robot Forum
Electronics => Electronics => Topic started by: Iron Man on August 17, 2008, 12:05:11 AM
-
Hey Guys,
I was just wondering what your dream microcontroller board would be.
Mine would be somewhere along these lines:
Tons of pwms
ADC
SPI
I2C
UART
Low Power
Capacitive Touch
Line Reflectometry? *not sure that's what you call it, but it's a way of measuring the lengths of wires and checking for breaks in connections.
CAN- extensive sensor networks
Op Amps - use for development of your own sensors.
Tons of programming memory
The reason I ask is that I may or may not be designing a microcontroller board during the next two semesters. But i'd like to get a good idea of what to shoot for if i do.
For your help, if i do make the controller board, i'll see if i can get a couple to add into the prize pot for the robot winners, and i'm sure admin can have one.
Bribery never hurts right? :P
Thanks for the help guys.
-
From what I've been seeing the new AVR family support many of these features alone in the chip...
Some peripherals are needed for sure in a board... but what's the point after all...
You can always build some small external boards for more advanced peripherals...
Anyway, I think you forgot some basics like temp sensors and a clock with back up power supply...
Maybe some 7seg, a buzzer and even a card reader (or memory card holder, not the electronics....)
An LCD holder... some push buttons or a button matrix better....
That's for me....
-
built in motor controllers
-
Might not be important to others, but I would love to see one with built in wifi.
-
I'd like to see a cross between the Axon and the Parallax Propeller. The main feature of the Propeller is the 8 parallel processors. Not having to mess around with interrupts is a beautiful thing. It does have shortcomings, though - no analog I/O, no "real" C compiler (yet).
-
Do you have any links or information on this Line Reflectometry?
-
@trickynekro
I know some of them are supported on chips already, but the board would be an easy access to the PIC that i would be using, so chosing the appropriate PIC is important. Also what do you mean by 7seg?
because it's a PIC i would eventually include the emulation circuitry, by that i mean i would take one of microchip's
programmer tools and integrate that into the board.
either way you won't have to buy a programmer to program it, but i'm sure the debugging/emulation circuitry would be nice.
I'd like to keep the board down to the size 3" x 6"
Just to sum what's been thrown out there so far:
LCD
button matrix
motor controllers
Wifi - could use zigbee or MiWi for reduced size and energy consumption could create a mesh network with these
swar
No interrupts..... idk about that one.
temp sensors
and a clock... depending on which chip i chose this may already be included, but still a very good point.
a buzzer... maybe just some flashing LED's.
card reader
@Jesse Welling
http://en.wikipedia.org/wiki/Time-domain_reflectometer
I called it the wrong thing, sorry.
-
Forgive my ignorance, but is this a type of project that you could work with Admin on? Something like an Axon 2 or Axon Deluxe? For a beginner like me, the Axon is such a nice cicuit and a lot of great features. Seems like there is a solid foundation to work from that these extra features could be added onto.
Just a thought
-
I don't want to be a spoil sport but TDR doesn't seem very viable for a single micro controller system. It appears to be more of a 'testing' solution in the lab. I would suggest double redundant sensors (meaning 3 sensors of the same kind measuring for each input that is deemed safety critical) and using a simple threshold voter to arbitrate the discrepancies or who to trust. If it is really really important that you have fail operational systems that is. A fail silent system could make due with just single redundancy...
-
@ Jesse Welling
the TDR is possible with only one chip. Many of the new 16bit PIC's have a CTMU component *charge time measurement* This unit allows for TDR and capacitive touch. it can detect a change in time down to a nanosecond. May not be fast enough for smaller robots, but might be helpful in larger projects, like lets say your creating an automated system to check the like 6000 miles of cable between your room and the laundry room in your dorm basement. There is also a proximity ability with the ctmu, it can detect small changes in the magnetic fields in wires... or something like that, i just saw a demo of it and was like woah.... that wire can tell when i'm getting close to it.... that's trippy.
What do you mean by a fail silent system?
@ Mark Brown
I would most definetly appreciate Admin's help. But as is right now this is nothing definite, still planning, brainstorming.
also it would have to be a PIC, so it'd be different than the Axon in that major respect.
-
Well I meant more "Fail Safe" which means that your processor knows something has failed and can go to a safe state.
A good example is quadrature. If you are using PID to regulate a quadrature measured speed, if the quadrature sensor fails, your processor will think it needs a rather large P gain element no mater what you do, so you will get this huge command even though your end actuator is probably doing just fine, So you'll get a robot that will go full bore for the wall. You could play some tricks if just one channel went down so it is kind of semi redundant....
Now if you had two quadrature sensors you would be able to tell that one of them failed, your controller could under stand that your system is one failure away from going crazy and put itself into a safe state so it would not hurt itself or any thing. It's easy to do this with machinery, but not medical equipment because if you stop doing something on with a machine it's not likely to kill any one. But with a medical device, you don't exactly want to stop what you are doing in all cases, so the safe state isn't always 'do nothing'
Fail Silent in this context means to not do anything so as not to cause damage. You want Park Breaks to fail silent in a sense that if they fail you don't want them to activate while some one is driving at 80mph. Like wise you want speed control to fail silent and just not operate. It doesn't necessarily mean 'don't tell the opperator' just don't let the operator be affected....
-
gotcha thanks for the explanation.
-
Hi,
the TDR is possible with only one chip. Many of the new 16bit PIC's have a CTMU component *charge time measurement* This unit allows for TDR and capacitive touch. it can detect a change in time down to a nanosecond. May not be fast enough for smaller robots, but might be helpful in larger projects, like lets say your creating an automated system to check the like 6000 miles of cable between your room and the laundry room in your dorm basement.
For Time Domain Reflectometry useable in even the largest robot, you need to keep your risetime and resolution within say 10ps.
That apart, it would be quite a bit of overkill, unless you have a very special app. (like a life support system able to cross rough terrain or similar). I doubt you will find any useable PIC for this. You need more resolution, faster risetime and much less jitter than any commonly available controller can deliver.
Why do you find a need for this in a generic board anyway? If your leads are prone to breaking, use physically stronger cable.
When I need to test for impedance changes along a line, I grab a scope and a pulse generator for it, add some common sense and old skool trickery (which in an automated system would need to be replaced by, at the least, a measurement of the S-parms [to get the "lead/lag" added to the basic propagation data for the correctly terminated raw but "perfect" cable], if the reflectometry results is supposed to be used for something).
Capacitive sensing, on the other hand, can be implemented in any available controller with a few discrete glue parts, as it is very tolerant to frequency deviation and can be designed to just about any frequency above the audible range.
But... Having a lot of possibilities in a single unit will most of the time mean that it is overloaded with something you don't need (MS Word springs to mind ;)) The more you fit, the more can go wrong - simplicity is the key to lean apps.
-
This has nothing do with PIC's but has anyone used the Wiring Board (Atmega128 processor I believe)? Seems to be a cheap alternative to the Axon(although I already purchased an Axon) that will keep me from running proto boards with my horrid soldering abilities. I'm trying to look for board to run about 18-24 servos (hexapod or quadroped w/arms ) and have several IR sensors/sonar/ possibly bumpers in the range of 50 to 70 bucks.
-
@Soeren
It was just a possible application of a module on the chip that i have in mind. It does not add or take away reliability, it's integrated into the IC, it's the same module that does capacitive touch. You're right it couldn't be implemented into a robot, but it might be a useful tool for a robot to use. like lets say your creating an automated system to check the like 6000 miles of cable between your room and the laundry room in your dorm basement.
either way, can we please drop the TDR issue and move along to more of your preferences, dreams, hopes and desires for your next board.
Penth,
you should have started a new thread for that one.
but no i haven't.
-
@trickynekro
I know some of them are supported on chips already, but the board would be an easy access to the PIC that i would be using, so chosing the appropriate PIC is important. Also what do you mean by 7seg?
(http://media.digikey.com/photos/Stanley%20Electric%20Photos/STANLEY%20ELECTRIC%20CO%20NAR131SB.jpg)
That's a 7 seg or 7 segment... LED array that can display numbers....
Although there are 7 seg that can display alphanumerical character they are not 7 seg cause they include more than 7 LEDs in the display...
And that is often called LED display not to mismatch with OLED displays...
Lefteris
-
ahhh, i see what you mean by 7 segment LED, gotcha
graphical OLED maybe the standard for this board.
thanks.