|
|
View previous topic :: View next topic |
Author |
Message |
SuperDave
Joined: 22 May 2008 Posts: 63 Location: Madison, TN
|
|
Posted: Wed Oct 29, 2014 9:41 am |
|
|
"5v operation and huge 25ma sink/source per pin!!! "
Quick caution. Since the Pics read/modify/write I/O pins, connecting a led directly to a pin to get the full 25mA (and maybe more) will cause all kinds of trouble since the led clamps the pin voltage in a no man's land. ALWAYS use a resistor. Even if you set the current to 25mA, the resistor will allow the voltage at the pin to rise to the power supply (assuming you use the correct resistor.) |
|
|
ELCouz
Joined: 18 Jul 2007 Posts: 427 Location: Montreal,Quebec
|
|
Posted: Wed Oct 29, 2014 3:23 pm |
|
|
SuperDave wrote: | ALWAYS use a resistor. |
Of course! I think this is the basics of electronics !
Quote: | Really, when it comes to questions regarding what's the best or most powerful processor, it all comes down to another question: what's your application? |
I wasn't asking for a specific application...
It's just that I'm hitting the limit of PIC18. I need a truck load of I/O pins.
My PDIP 40 package there's no pins left and I need to add more things (sensors) to the PIC.
I was think adding a port expander but that will just make the project harder to implement since you add another chip in the debugging.
Most of dsPIC and PIC24 come in a package with minimum 64 pins or more.
I don't have any experience in the PCD land. Tried twice with 4.060 few years back and it was awful!
IMO CCS only truly support a handful of dspic and pic24 which are bug free. The rest, I feel is non-tested chips that are buggy and partial support from the compiler.
That's why I've asked, what's the most powerful pic the PCD support fully!
Even if I do a parametric search (which I did) like PCM Programmer suggested, I feel I will end up with a chip barely supported/tested with CCS. (Because it's new)
Clarification: I think PIC16F877 or PIC18F2550 has been tested to death from CCS (both device/drivers/library/examples) but in the 24 bit land ...??? minefield?? _________________ Regards,
Laurent
-----------
Here's my first visual theme for the CCS C Compiler. Enjoy! |
|
|
temtronic
Joined: 01 Jul 2010 Posts: 9236 Location: Greensville,Ontario
|
|
Posted: Wed Oct 29, 2014 4:49 pm |
|
|
hmm...
perhaps you could use a second PIC for the additional I/O ?
Without knowing what all your pins do, there may be a logical 'grouping' of I/O functions that could be 'offloaded' to a smaller PIC. Since the PIC is programmable , you can utilize onboard peripherals easier than a dedicated I/O expander.
You've already invested a lot of time and resources into getting PIC18s 'up and running' and the learning curve for another processor series might cost you a lot of R&D time and the loss of a lot of hair!
While it might not seem cost effective, based on PIC prices, they are cheap especially compared to what usually happens when migrating to another micro and compiler.
It's the route I've done in the past as overall it saves time and money,gets the product out the door to happy clients.
hth
Jay |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19535
|
|
Posted: Thu Oct 30, 2014 1:41 am |
|
|
Have to agree wholeheartedly with Temtronic here.
I don't think any but the smallest little interface board I have built in the last 20 years, has ended up with just one processor.
Looking at current project, have three little PIC12's, doing odd I/O functions (and one using it's capacitive sensor), and 'reporting', to a small PIC24. This then uses these and it's ADC's, and a number of I/O pins to control the whole 'front-end', and reports to the 'user interface' (another PIC24), which handles RS232, RS485, USB, SD card, LCD, RTC and keyboard. There is also another PIC18 here that controls relays, and some general I/O. Six PIC's, which also helps reduce the board complexity, since the control chips are in each case 'close' to what they control. Especially recently, with cheap PIC18's, with internal clocks that are sufficiently accurate for most operations, trying to connect a maze of signals 'back' to one processor, costs more, and is more liable to lead to interference problems, than using 'distributed I/O'.
Also debugging is far easier. I'll often 'stress test' the processing chips, with massive amounts of I/O from a PC, cycling through rates of change and combinations they should never meet, while at the same time working on the front-end chips. On the current unit, the linearisation maths was solved and checked on the PC, before it ever went near a PIC, which also allowed working out where lower maths types can be used. |
|
|
ELCouz
Joined: 18 Jul 2007 Posts: 427 Location: Montreal,Quebec
|
|
Posted: Thu Oct 30, 2014 2:27 pm |
|
|
Thanks guys...
Maybe i'll go the way Ttelmah says...split up single huge pic into tiny slaves ones!
You are right for wire complexity, closer the pic is to the source better it is!
For me it was more for challenge to fit everything in a single package than the lazy way to put a lot of pics for simple tasks.
Lots of boards I've seen from production (consumer products) using PICs use the cheapest one for the project so they tend to cram a lot of task into a single unit.
[...]
While we are around the subject...
From all the boards I've seen while tearing apart things, 70% was discrete stuff, 10% ARM , 10% proprietary IC, 5% Atmel and rarely Microchip, I've seen more FPGA than PIC in a consumer / prosumer products.
Market penetration seems low for Microchip products (Cost per unit too high?).
Is there even demand for 8-bit anymore (consultant side) Ttelmah? _________________ Regards,
Laurent
-----------
Here's my first visual theme for the CCS C Compiler. Enjoy! |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19535
|
|
Posted: Thu Oct 30, 2014 3:49 pm |
|
|
Funnily enough quite a few of my products use PIC's implemented in FPGA's....
However the big market penetration for PIC's is small stuff. A huge number of the processors in things like RFID tags, and smart cards, are PIC's.
PIC's ran over 8% of the total MCU market share last year, but you just don't see them as the 'big' visible processor. Not really their part of the market.... |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|