Internal Osc question.

[muddy0409]

Hi everyone.
I am using an 'F819 for a little thing I am doing.
I am doing some serial stuff using DEBUG at 600baud.
This works perfectly when using XT osc with a crystal, but as soon as I go over to INTOSC and define osc 4, the sent stuff turns into junk. I am viewing these items on the PBP terminal view.
This begs the question, just how accurate is the internal oscillator on these (& other) PICs. It's just going to be a pain to have to use a crystal, since I have already done the PCB for the project, or at least pinched one from another project that doesn't use serial coms.

[Acetronics2]

Hi, Peter

I did some measurements on the intOscs ( not '819 ...)

16F84 ... +/- 20% is to consider for the EXT RC osc ... ( includes components tolerance )

16F628(A) ... +/- 10 % of nominal freq is to consider

16F88 ... +/- 2 to 3 % of nominal freq to consider.


But you have the "OscTune" register to play with ... The use of MPLAB + ICD2 could be helpful here !!!

Alain

[Pic2008]

Can you try to use 2400 for DEBUG_BAUD. I'm using other PIC, and it is working fine with internal OSC.

[mackrackit]

For what it is worth, I have never been able to use the internal reliably to serial. It may work on the test bench some, but move it someplace else (different temperature ?) and it will start sending "junk".

At least if you use a crystal with built in caps (resonator) it can be pretty much placed anywhere on the PCB with jumpers and super glue

[muddy0409]

Thanks folks.
No, the speed must be 600 'cos that's what the receiving device wants. I will probably end up using a resonator, just 'cos I can!