Hi Group: I've got a project that I'm contemplating that needs:
To read a keypad with 20 switches. Control 7 digits of 7 seven segment displays. Generate pulses for a small stepper motor. (full step and micro step) Possible control of a brushless servo motor with encoder feedback. Monitor a few simple sensors. A few hundred bytes of RAM is enough. Program size with lookup tables ect, should be under 4-6K. Flash would be nice.
I've never worked with the PIC and looking at their product line up there's just too many to chose from. I have worked with the Motorola stuff (HC05) but they are not making anymore HC05 parts. My question is, What PIC would you use to do the above. Or maybe use several PIC's to get the job done? Or would you just go for one of the bigger HC08's?
Flash is essential these days! Anything else is a pain and not worth touching.
One PIC is plenty, unless you have a compelling reason to separate tasks. A basic decent size PIC is the 16F877, so I'd look at that to start with.
Is this a small or high volume thing?, does cost or board size matter?, how many pins do you need? (is the keypad and LED display MUXed?)
The PIC is no less powerful than the HC08, or any other 8bit micro for that matter really.
C is the go, forget assembler unless you really really want to. I've heard the Microchip 18series C compiler is essentially free now but with very minor limitations, so maybe use a newer 18series PIC?
wrote in message news:zfGdnb4JbbkxFrrZnZ2dnUVZ snipped-for-privacy@comcast.com...
MicroStepping, is quite a complex thing in itself. You need to generate sinusoidal drives, to two different power drive branches. Now there are several PICs that can do this, but you would end up using both PWM outputs (most PICs have 2*PWM), which then leaves you without PWM's for the servo control application. I suspect I'd probably look at the Allegro A3977 to drive the stepper, which removes the need for seperate power FETs, clamp diodes etc.. This then leaves the PWM outputs available for servo control. These days, flash is cheaper, an simpler for any limited size production run, and is 'standard' on most parts. Now there is the question of 'what interface' for the 'sensors'. Mst PICs will offer ADC's, I2C. SPI, and sync/asynch serial. Depending on the number of 'encoder counts' for the servo motor, and the rotational speed, there are some PICs with encoder inputs, which would allow handling at very high speeds, while most PICs could handle lower rates, using the 'interrupt on change' ability. The LED drives, and keypad scanning, are just down to number of pins. You need 7 bits to control the segments (possibly 8, if you have a decimal point?), and three bits (using a 3-8 line decoder) to address the digits (or 7 bits for direct drive). On the keypad, 4*5 lines. So assuming you use the Allegro chip, say 11 bits for the LEDs, 9 bits for the keypad, 2 bits for the stepper (step/dirn), 3 bits for an H-bridge for the servo (a PWM output, and two control bits), 2 bits for the quadrature inputs, and leave perhaps 6 bits for the 'sensors'. This puts you as needing a minimum of 33 I/O lines. I'd probably look at the PIC18F4331, or 4431. The 'F' means 'flash'. These chips are 40 pin DIP packages, with 36 I/O lines, including up to 9 ADC inputs (or eight, if you use an external reference voltage). They have a quadrature encoder built in, and 2 PWM outputs. 768 bytes of RAM,
4096/8192 'words' of program memory respectively, 256 bytes of EEPROM, UART, SPI, & I2C.
I've been using the OTP eprom PICs for a long time, but now I have to upgrade to a FLASH version, propbably the 16F877. I've read conflicting reports about the relative slowness of programming FLASH devices. A 2K OTP part such as the 16C57C takes only 3 or 4 seconds using the parallel programming algorithm.
Question: Using in circuit serial programming, how long does it take to program a FLASH device such as the 16F877A?
schreef in bericht news:zfGdnb4JbbkxFrrZnZ2dnUVZ snipped-for-privacy@comcast.com...
Ron,
Most important constraint seems to become pincount. Your keypad and display will require 23 pins. As a common 40 pins (8 bits) micro has only 36 IO-pins you may not have enough pins left to meet your requirements. IMHO you hve several options:
- Bigger micro (more IO-pins, 16 bit type)
- Extra electronics to reduce the number of IO-pins required (for instance external registers/buffers and decoders to handle the displays)
- Extra micro. Same as above. A 20/24 pins micro can handle the display and communicate with its master using only 2 IO-pins.
The HC08 is the flash-based replacement for the HC05 in Freescale's lineup, though they have recently announced a stripped version HC08/5/4 for really cost-sensitive consumer applications. I forget what they call it, but it's probably not of interest to you anyway. The 08s are nice parts.
The core micro is not so important except how it might fit with your experience and tools. If you're used to the 05, you'll probably find the PIC rather irritating, especially the lower-end ones. Unless you're very cost-constrained start with the 18f, 24f or dspic series.
Is the potential BLDC motor/tachometer input (quadrature input?) concurrent with the stepper or 'instead of'?
You've got a number of fine-time resolution things going on at once that can only tolerate a limited amount of jitter or delay in reading them. That will determine the hardware you need. I see no reason to go to multiple processors or expensive controller chips, it can all be done with a single processor. Sounds like you need 4 or 5 timer-counter capture compare modules, which should simplify your search. If you need quadrature input, that will further limit things or you might choose to use some external logic there.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
A 4K (words) 16F88 takes about 8 seconds to program and verify the entire flash area. A 2K 16F628 takes about 6 seconds total using my PICALL programmer.
In my experience my code typically gets programmed in under 10 seconds. Yes, some consider that to be slow but it is still much faster than Windows booting up or "burning" your firmware onto floppy disks (believe it or not our PC104 embedded controllers still use them to install/upgrade software).
Consider also the AVR series from Atmel, and any one of the bazillion
8051 derivatives out there.
To wade through the mess of different chip styles you can do what I do:
- list your requirements in terms of I/O pins, PMW outputs, ADC inputs (if you need it) and whatever else you need
- sort through your list of microprocessors to get viable candidates
- choose one based on price, your attitude toward the vendor, and any other 'soft' considerations you may have. I always look for a pin-compatible upgrade, unless I know that what I've chosen is severe overkill (in case I look for a lower-price pin-compatible 'downgrade').
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Posting from Google? See http://cfaj.freeshell.org/google/
Did you consider TI's MSP430 series? This has flash, is programmable from the serial port by connecting DTR and RTS and it is loaded with I/O. One of the advantages is that is has a flat memory model and it is fast. This makes it very easy to program in C. A free compiler is also available (mspgcc).
--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
Nah. Shared lines mean 18 pins max, with no external logic. Might be possible to do it with 11, depending, still with no external logic.
There are plenty enough 8-bit micros (as well as 16 and 32-bit) in 64,
80 and 100 pin packages.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
You might want to look at the PIC18F6410 which I'm designing into a PWM design I'm doing at the moment. Amongst other things it's got 3 PWM output ports and 64pins. It's SMT, and I plan to program it in circuit using an ICD2 clone from Sparkfun:
formatting link
The device is supported by the Microchip development tools MPLAB and C compiler C18, you can download both FOC. MPLAB apparently recognises this programmer. Some optimisation limitations come in at 60 days, allegedly may not matter much.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.