CCS (PCW C compiler IDE) baud error

[xxjoelxx]

Hey, I'm currently in a situation which I'm trying to solve. Currently I am using the PCW compiler version 4. Everytime I compile my program, it pops a baud rate out of range error. Whereas if I use the older version, PCW compiler version 3, it does not have this error.

This is the code:

[rnielsen]

This depends on what frequency your crystal is running at. Your #use delay() tells the compiler what frequency to use. If this frequency is not close enough to bring the baud rate within range then you will get this error.

What frequency is your crystal and is your #use delay() statement correct?

Ronald

[xxjoelxx]

Thanks Ronald for the reply. My motherboard is using a M.Crystal 25 MHZ frequency.
In my program I did a "#use delay(clock=25000000)", which matches with the clock frequency I am using. Thus, the fault should not be there correct?

Joel

[xxjoelxx]

HEY people. Is there anyone who can help? Would be greatly appreciated.

[PCM programmer]

[Ttelmah]

As a general comment though:

Sync serial, requires the baud clocks at each end to be fairly close. Just how close, depends on a lot of factors (the sampling strategy used, the word length being used, etc. etc..). Historically, the older V3 compilers, allowed a fractionally bigger error in the clock actually being generated, from what was requested, than the V4 compilers allow. However the change is tiny, and if V4 complains, it is borderline on whether the clock will be reliable. The allowable error, is just under 4%, with a 3.1% error being accepted by both, but an early compiler accepting 4.2%, while the latter compilers complain by about 4%. The change was 'pre V4', with latter V3 compilers having similar limits to V4.
You can always 'cheat' to avoid the problem. If (for instance), you want to generate a 300Kbaud serial, witha 40MHz clock, you sit down and calculate the BRG yourself. So:

40MHz/300000 = 133.33333

With BRGH, the prescaler is 16, so you get 133.3333/16 = 8.33333

Nearest possible divider is 8 (7 in the BRG), giving division by 128 (8*16), and a baud rate of 312500. A 4.16% error, beyond the level that anyone would accept as 'sensible', and the compiler _will_ therefore reject the 300K rate. However if you were totally confident that the system will work OK with this error, all you need do, is set the baud rate in the #use RS232 statement, to the 312500 rate that the hadware can manage. The compiler wll then stop complaining, and accept the rate, and program this rate....

Best Wishes

[PCM programmer]

Download PicBaud. It will quickly show you the percent error in your
selected baud rates, for your given oscillator frequency.
http://www.ccsinfo.com/forum/viewtopic.php?t=37250&start=1