>Just curious here, do you have problems with BRGH = 1 ?
>Microchip data says -
>"Note: At the time of this printing, the asynchronous high speed mode
>(BRGH is set) may experience a high rate of receive errors. It is
>recommended to have the BRGH bit cleared. If you desire a higher baud
>rate than BRGH=0 can support, refer to the device errata for
>additional information."
>

I posted my findings on this to the list about 12-18 months ago
and I concluded that there was NO problem using BRGH=1.

I detected error rates of less than 1 packet in 10E7. Using
HDX RS485 short cabling and exactly matched baud rates.

My conclusion at that time was that the reported increased error
rate was due to increased sensitivity to mismatched baud rates due
to changing from x64 sampling to x16 sampling.

Most reports of the problem revolved around folks who had
applications that connected to PC's for various purposes and
found inexplicable errors on Brand_X but not on Brand_Y etc..

The fix was almost always to drop back to x64 sampling (ie BRGH=0)
sometimes tweaking the fosc frequency to reduce the error as well.

So if you have COMPLETE control over the device that the PIC is
talking to then BRGH=1 works fine.

All of this is precisely summarized in the Micrchip Errata which
categorically states that.

        "may experience a high rate of receive errors"
         ^^^
Exactly!

For most applications I would recommend stick with BRGH=0
and only use BRGH=1 when you have complete control over both
ends of the link, (This mostly rules OUT talking to PC's)



Ray Gardiner Techinical Director DSP Systems ray@dsp-systems.com
private email to:- ray@netspace.net.au  http://www.dsp-systems.com