Jump to content

emte

EstablishedMember
  • Content Count

    377
  • Joined

  • Last visited

Everything posted by emte

  1. Could be used either way, but yes I was thinking on the master. Out of curiosity are you using the serial port on the slave? Since RC6 is both /SS and TX... And I also do not see you setting RC6 as an input...
  2. The SSPIF interrupt will trip for any SSPIF event, you need a second flag to detect which slave you want to deal with. Something like: if ((test_bit(pir1, SSPIF))&&_slaveOne) . . else /* for all other slaves */ { clear_bit(pir1, SSPIF); /* clear for second slave */ tmp = sspbuf; /* empty buffer to avoid overflow */ } Oh and also related, some chips have multiple slave type settings, would have to check the datasheet though.
  3. emte

    Sb 6.84 Registration Problem.

    It is flakey under wine, what does wines log tell you? It could simply be waiting for the correct window type, which sometimes means you need tell wine to use an older OS othertimes its a missing font etc.
  4. We would have to see your interrupt routine, it could be as simple as not detecting which slave or just a line outside your loop.
  5. Just to elaborate a bit clearer, First you control the SlaveSelect(SS) line from the master, then the master controls the clock line. After that, data is transfered in a normal fashion, but keep in mind only one device can control the clock line at a time and only the master controls SS. Just to question something in the fog of my brain, isn't there an issue with using the bit functions in an interrupt? I've been away from PIC programming for a few months so some things are a bit rusty.
  6. Just to question the obvious, but I've done it myself, are you sure your using the right hex file or that it was generated?
  7. emte

    Rs232 Docs

    I am curious why you need more documents... Serial/RS232 libraries are arguably to help new users only, you will find far more useful information in the datasheet. I suppose I should qualify that, to initialize serial you set a handfull of flags and that is the hardest part of setting up serial. Most serial libraries(not all) utilize functions that result in extra processing time for something that takes one cycle. The classic example of this is checking the transmit flag. Anyway here is a simple init and remap you can drop in a header and use: #ifndef _SERIAL_H #define _SERIAL_H #include <system.h> volatile bit _SER_TRISIN @TRISC.7; volatile bit _SER_TRISOUT @TRISC.6; volatile bit _SER_SPBRG @SPBRG; volatile unsigned char _SER_TXREG @TXREG; volatile unsigned char _SER_RXREG @RCREG; #define _SER_RCREG _SER_RXREG /* Naming Convention Macro */ volatile bit _SER_TX9D @TXSTA.0; /* Transmit 9 Bit */ volatile bit _SER_TRMT @TXSTA.1; /* */ volatile bit _SER_BRGH @TXSTA.2; /* */ volatile bit _SER_SYNC @TXSTA.4; /* */ volatile bit _SER_TXEN @TXSTA.5; /* */ volatile bit _SER_TX9 @TXSTA.6; /* */ volatile bit _SER_CSRC @TXSTA.7; /* */ volatile bit _SER_RX9D @RCSTA.0; /* */ volatile bit _SER_OERR @RCSTA.1; /* */ volatile bit _SER_FERR @RCSTA.2; /* */ volatile bit _SER_ADDEN @RCSTA.3; /* */ volatile bit _SER_CREN @RCSTA.4; /* */ volatile bit _SER_SREN @RCSTA.5; /* */ volatile bit _SER_RX9 @RCSTA.6; /* */ volatile bit _SER_SPEN @RCSTA.7; /* */ volatile bit _SER_TXIF @PIR1.4; /* */ volatile bit _SER_RXIF @PIR1.5; /* */ #define _SER_RCIF _SER_RXIF /* Naming Convention Macro */ void _initSerial(void) { /* Specific pin setup */ _SER_TRISOUT = 0; /* Output */ _SER_TRISIN = 1; /* Input */ /* TX setup */ _SER_TX9D = 0; /* 8bit mode */ _SER_TXEN = 1; /* Enable Transmit */ _SER_SYNC = 0; /* Asynchronous */ _SER_TRMT = 0; /* Clear TSR */ _SER_BRGH = 1; /* High Speed */ /* RX setup */ _SER_SPEN = 1; /* Enable Serial */ _SER_RX9D = 0; /* 8bit mode */ _SER_CREN = 1; /* Cont. Receive */ /* 115.2 baud @ 20Mhz * BaudRate = Fosc div (16(x+1)) * 113636.3636 = 20Mhz div (16(9+1)) */ _SER_SPBRG = 10; _SER_RXIF = 0; /* Clear RX Interrupt */ } #endif /* EOF */ Hope that helps.
  8. Let me see if I understand ... The problem you are having is that your toolchain and target are getting reset everyday when you restart or power up your development system? Are you using a Source Control System? If so, did you add your workspace file?
  9. emte

    Very Slow Compile

    I've tried SB now a few times on different systems and a pattern seems to be emerging, then again I could still be delusional. But, depending on which windows you have installed or more directly how much MS has altered the MS-DOS compatability for different versions, you get vastly different performance out of the compilers. Older seems better, with that said, I am only about to try and see what happens if i try to get it to run the compiler under cygwin. ...providing i can get SDCC to behave with this 8052 code first.
  10. Looking for a little bit of lecture, little bit of example, and a lot of "you can't do that fool!" about fuzzy logic implementations in PICs (specifically integer based ones). PID stuff is not wanted in this thread, just embedded fuzzy logic implementations. Prefereably without Fuzz-C being used to generate the code(This would be good in a parallel thread as would be NOVOS implementations) i am just venturing into this area and would appreciate any insight or examples of successful implementations.
  11. emte

    16 Bit Pic Micro Support ?

    Incorrect, what i said was that for most people and opensource, the limited compiler works just fine and that any license under $1k is reasonable for commercial (for profit) usage. As well, using a heavily mantained parent like GCC means there are less problems with long term support and migration. Why reinvent the wheel if you already have a round rim and the rubber?
  12. emte

    Build Or Buy That Ultimate Pic Programmer?

    It is quite trivial to do this, but the USB tranceivers can get quite expensive comparativly and they require a bit of extra firmware work.
  13. emte

    Are You Fuzzy Yet?

    I shall read as soon as if find time, which seems to be in very short supply for me. Not only do I work 70 hours a week, i am also researching how to fabricate a waveguide laser. emte Not truly on topic, not being about PICs, but you might be interested anyway. I've been around a long time in electronics, seen quite a few things come - and quickly go! The one-bit processor? The General Instruments PIC? - it's still here A favourite columnist over the years has been Bob Pease of National Semi. You might enjoy his take on fuzzy logic from back in the '90s, when it was the 'next big thing'. http://www.national.com/rap/Application/0,1570,25,00.html His stuff on audio accessories like speaker cables is fun too. Regards, John
  14. Aggressive optimization is better to be used on a function-by-function case, that way if it does something bad you know where the issue is. #pragma OPTIMIZE "a" int someFunc(short var) { blah } Use this before the function you want to optimize.
  15. Unless you remap your tags, all your pin tags etc need to be lower case, just so you know. Uppercase is very common with a lot of compilers. An easy compatability work around is: #define PORTB @portb #define TRISA @trisa
  16. Strange there is so much about this very common task ... Dave's method does work but it invites cyclic includes which ends up chasing multiple defines, the simple solution is: *** export.h *** extern int expose; *** export.c **** int expose = 123; **** main.c include expose.h void main(void) { expose =42; while(1); } This avoides the cyclic include issue and exposes variables that you choose to expose as global. As the variable is only defined in the C file and exposed in the H file there are no multiple defines. Hope that helps.
  17. First off, Hyperterminal is VERY unreliable at high datarates and is far from a standard terminal device. From the table in the datasheet: 20MHz Crystal 33.6k Baud 0.55% error SPBRG=36 BRGH=1 57.6k Baud 3.34% error SPBRG=20 BRGH=1 As for best/worst case, that depends. With a 20MHz crystal your max reliable is ~57.6K and any error rate under 5% is pretty good depending on how it is transmitted. If you were using RF, you would definiatly want it under 1%, but if it is short cable then 5% is decent. Redundancy cleans up most error issues anyway, with the most basic being a parity bit.
  18. emte

    Spi

    The same way you would if it were a direct program in the chip, only the interface handling changes. eg. for a read you send an SPI command that the pic recognises, the pic then processes it, then sends data back.
  19. emte

    Instant Posting

    Even with the setting at 10 a few get through every couple months. As it is, Dave and Pavel and etc are fairly quick to review posts.
  20. It was/is not a compiler issue it was a silicone issue and something to do with indirect addressing and interrupts if i recall correctly.
  21. Hi emte, Nice to see you back in the forum after a small gap. Well, i suppose you are not impressed with my project . Hey man.. atleast give me a word of motivation. I am not going for a competition nor am I trying to sell some thing, I am only sharing my experience in constructing this project. Let me see if you would change your mind once i have posted all the project stuff. I had gathered these information with great difficulty & I am sure its going to benifit many. note: The project has got nothing to do with pogo pins. Info on UTM:UTM Zwick UTMs Regards Raghunathan. Bad me, bit tired since i currently work two jobs. Nice project, bit disapointed you used a 16f877 but i am sure some will find it useful and a good reference. My comment is only based on personal experiance using high flow production testing equiptment where it is a waste of time to "connect" external devices so pad contact using pogo pins is used. You just mash the boards into the jig and 5 seconds or so later your done. It is fairly simular with the exception that the test CPU is usually an FPGA not a PIC.
  22. emte

    Are You Fuzzy Yet?

    So much for getting caught up ... i now have one job, and a production contract that are chewing up my time, i do have new notes but not clear enough to post here yet, hopefully soon.
  23. This is actually a bit incorrect, variables that you wish to retain values for and/or wish to access in your main program should be declaired globally. If only defined staticly in your interrupt, you may lose the values or they may be corrupted as interrupt variables are not handled the same way as regular program variables. There used to be an application note about this odd behaviour on Microchips site, i have not looked for it in some time and do not know if this has ever been fixed.
  24. i hate to tell you this but your overselling your self, or your project atleast. A universal testing machine is a bed of 60-500 pogo pins with specialized protection circuitry so all you do is program the board for a specific setup and press the PCB down onto the bed of pogo pins. For those who do not know what a pogo pin is: It is a gold spring loaded pin that compresses when you push on the tip. Pogo Pin As for your projects thread, it is is good especially for those who skip the basics and try to go headlong into a project.
×