Post by thread:16c74 using BRGH=1

I am using the 16c74 with BRGH=1, and after extensive network tests have concluded that for my application the error rates are vanishingly small. Alexej Vladimirov has posted some test code which I have rewritten and run, with the intention of locating once and for all exactly what the BRGH=1 problem really is. The weird thing is, I can't reproduce the errors that he and others have reported. The complete code is attached below. No errors after 10,000,000+ bytes My "real" error rate test runs the networking software and detects errors on a transaction by transaction basis. The code presented here is just a simple loop back test, consequently does not expose any sensitivity of the part to mismatched baud rates, or sensitivity to line noise. Nonetheless, this is the type of test that others have used when reporting the BRGH=1 problem. If anyone cares to run the code below and can produce errors. I would like to know. Until then I will continue using BRGH=1, and suggest that others wanting to use the 16c74 with BRGH=1 evaluate according to the application. ; ; Loop Back Test program for pic 16c74 comms ; Ray Gardiner 1996 ; ; Notes: ; 1. Assemble with MPASM 1.4 ; 2. Link RXD and TXD (RC,6 and RC7) ; list p = pic16c74 include "c:\mpasm\p16c74.inc" errorlevel -302 cblock 0x20 ErrorCount:4 ; 32 bit counter for errors GoodCount:4 ; 32 bit counter for chars TxChar ; what we sent RxChar ; what we got back endc ; stuff for 32 bit counters ; Inc macro a local x incf a,F bnz x incf a+1,F bnz x incf a+2,F bnz x incf a+3,F x endm Zero macro a clrf a clrf a+1 clrf a+2 clrf a+3 endm ; main program starts around here ; org 0 goto Start org 4 goto Start ; no interrupts Start clrwdt bsf STATUS,RP0 movlw 0xFF movwf OPTION_REG clrf PIE1 clrf PIE2 clrf INTCON ; initialize comms 153600bps with SPBRG=1,BRGH=1 ; I am using Fosc = 4.9152 Mhz movlw b'10100100' ; | ; +-------------- BRGH ; movwf TXSTA movlw .1 movwf SPBRG bcf STATUS,RP0 movlw b'10010000' ; | ; +----------------- SREN ; movwf RCSTA bcf RCSTA,CREN ; clear any errors at start bsf RCSTA,CREN ; ; Zero ErrorCount ; clear counters Zero GoodCount ; clrf TxChar main incf TxChar,F clrwdt call Tx call Rx movf TxChar,w subwf RxChar,w bnz NotGood Inc GoodCount goto main NotGood Inc ErrorCount nop ; <<< insert break here >>> goto main ; ;-------; ; send TxChar Tx btfss PIR1,TXIF goto Tx movf TxChar,W movwf TXREG return ; ;-------; ; receive RxChar Rx btfss PIR1,RCIF goto Rx movf RCREG,w movwf RxChar btfss RCSTA,OERR return bcf RCSTA,CREN bsf RCSTA,CREN return end Ray Gardiner, Shepparton, Victoria, Australia spam_OUTrayTakeThisOuTnetspace.net.au

Hello Ray! 06 Sep 96, .....rayKILLspam@spam@netspace.net.au writes to all: P> I am using the 16c74 with BRGH=1, and after extensive network tests P> have concluded that for my application the error rates are vanishingly P> small. P> Alexej Vladimirov has posted some test code which I have rewritten P> and run, with the intention of locating once and for all exactly P> what the BRGH=1 problem really is. The weird thing is, I can't P> reproduce the errors that he and others have reported. P> The complete code is attached below. No errors after 10,000,000+ bytes P> The code presented here is just a simple loop back test, consequently P> does not expose any sensitivity of the part to mismatched baud rates, P> or sensitivity to line noise. Nonetheless, this is the type of test P> that others have used when reporting the BRGH=1 problem. P> If anyone cares to run the code below and can produce errors. I would P> like to know. Until then I will continue using BRGH=1, and suggest P> that others wanting to use the 16c74 with BRGH=1 evaluate according to P> the application. The main difference with your type of loopback test is, that in your case transmitter and receiver powered from the same clock generator and fully sinhronized. In my application PIC was used for RS232 communication with PC, so clock sources for transmitter and receiver was different. Another interesting point is, that I check this code with different silicon revisions and found some working without this problem. Below is list of silicons, checked by me. I will be glad, if some PICLIST users check their PIC16C65//65A/73/73A/74/74A/74ME micros with external loopback and BRGH=1 (send to PIC from PC, loopback in PIC, receive by PC and compare, as in my test) to clear up situation with current silicon revisions. Work without errors: - PIC16C74/JW 9508 CAT Work with errors: - PIC16C74 ENG SMP 9430 CAT Rev A2 - PIC16C74 ENG SMP 9502 CAT Rev A2 - PIC16C74/JW 9524 SAT - PIC16C74ME A2 BOND 9443 CAA - Emulator pod Alexej Vladimirov avladKILLspammail.ormix.riga.lv http://www.ormix.riga.lv/eng/mchip/mchip.htm ...more than 170 Microchip-related links now... --- GoldED/2 2.50+

Alexej Vladimirov wrote: > > P> I am using the 16c74 with BRGH=1, and after extensive network tests > P> have concluded that for my application the error rates are vanishingly > P> small. > > Below is list of silicons, checked by me. I will be glad, if some PICLIST > users check their PIC16C65//65A/73/73A/74/74A/74ME micros with external loopback > and BRGH=1 (send to PIC from PC, loopback in PIC, receive by PC and compare, as > in my test) to clear up situation with current silicon revisions. > > Work without errors: > - PIC16C74/JW 9508 CAT > > Work with errors: > - PIC16C74 ENG SMP 9430 CAT Rev A2 > - PIC16C74 ENG SMP 9502 CAT Rev A2 > - PIC16C74/JW 9524 SAT > - PIC16C74ME A2 BOND 9443 CAA - Emulator pod > Has Microchip a comment to these investigations? Could they explain to us, what the nature of the BRGH=1 problem is. Maybe, some of us can decide, to use it or not. regards Wolfram -- +------------------------------------------------+ ! Wolfram Liebchen, Forschungsinstitut fŸr Optik ! ! .....liebchenKILLspam.....ffo.fgan.de ! +------------------------------------------------+