| View previous topic :: View next topic |
| Author |
Message |
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
| 18F66J90 UART2 problem [topic extended] |
 Posted: Sun Nov 22, 2009 11:04 am |
 |
|
Hi there,
I'm using fairly new 18F66J90 PIC for our new project. Previous version was using 18F66J10 and was working fine with UART2. Since both PIC's are pin compatible, I thought that I can easily replace it.
And it stayed as a thought, not as a reality.
Right now I'm fine with UART1 and CCS' Software UART. But UART2 is working at 1.250.000 baud which is weird. I just have done anything to make it work at 9600 baud, but no luck.
Note: The given baud rate is measured via scope.
Here is my related code.
| Code: |
#FUSES HS /// Using 20 MHz Crystal to get away from IOSC related frame errors
#use delay(clock=20000000, restat_wdt)
#use rs232(stream = AUX, baud=9600, parity=N, bits=8, stop=1, restart_wdt, UART2, ERRORS)
#byte SPBRG2 = 0xF64
#bit CREN = 0xF60.5
#bit SPEN = 0xF60.7
SPBRG2 = 32; /// See 18F66J90 datasheet page 272. (SPBRG2 Baud rate generator formula)
CREN = 0;
delay_us(10); /// See the ERRATA document of the silicon. enabling and disabling the UART needs 2 Tcy. I just added 10 us to be on the safe side.
CREN = 1;
SPEN=0;
delay_us(10); /// Same story as above
SPEN=1;
while(1)
{
fputc('U', AUX); /// A nice character to see the baud rate via scope. (0b01010101)
delay_us(10);
}
|
So... Any ideas, I've run out of them 
Configuration:
- CCS 4.093
- 18F66J90-I/PT Rev. A1
- MPLab 8.40
- PICKit2
_________________
/// KMT
/// www.muratursavas.com
Last edited by KaraMuraT on Mon Dec 07, 2009 6:33 am; edited 1 time in total |
|
 |
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
| Flash news |
| Posted: Sun Nov 22, 2009 11:23 am |
|
|
I checked the SPBRG2's actual value after changing it to 32, but I see that it's remaining as 0. Changing the sequence of the procedures didn't help at all.
I don't know whether this is a particular error of my PIC or an errata of the silicon. It does not sound sane to release a silicon without capability to change of a well known SPR.
_________________
/// KMT
/// www.muratursavas.com |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Sun Nov 22, 2009 12:41 pm |
|
|
Just switched back to 18F66J10 and it works just fine.
I hope Microchip has some sort of explanation about that. Else, the 120 pieces of 18F66J90 is just trash for me.
_________________
/// KMT
/// www.muratursavas.com |
|
|
MikeP
Joined: 07 Sep 2003
Posts: 49
|
| Re: 18F66J90 UART2 problem |
| Posted: Mon Nov 23, 2009 1:53 pm |
|
|
| Code: |
#FUSES HS /// Using 20 MHz Crystal to get away from IOSC related frame errors
#use delay(clock=20000000, restat_wdt)
|
Do not know if this error was a paste error or if it will fix the problem.
#use delay(clock=20000000, restart_wdt) |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Tue Nov 24, 2009 4:19 am |
|
|
Thanks Mike, for your attention. It was not a paste error, but also not related to this UART2 error.
Fortunately you sure have saved me from a lot of headaches in the future I haven't activated the Watchdog part of the software yet.
_________________
/// KMT
/// www.muratursavas.com |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Mon Dec 07, 2009 6:52 am |
|
|
I got analyzed a new PCB and 18F66J90, and seems like the problem was individual.
But this topic also revealed another CCS Bug. If you setup your code like this:
| Code: | #use delay(clock=16000000, restart_wdt)
#use rs232(stream = AUX, baud=9600, parity=N, bits=8, stop=1, restart_wdt, UART2, ERRORS) |
you'll expect that UART2 will work 9600 baud. Nope, you have to add this line (thanks to PCM Programmer)
| Code: | | setup_oscillator(OSC_4MHZ | OSC_PLL_ON); |
This will make your PIC really run on 16 MHz. But does it also solve the UART2 problem? Unfortunately no.
Here is the additional code to the lines above:
| Code: | #bit BRGH = 0xF61.2
#byte SPBRG2 = 0xF64
BRGH = 0;
SPBRG2 = 25; /// Check the formula, 18F87J90 datasheet page 272 |
Here is what causes the problem:
With standard #use RS232 line, somehow CCS compiler activates the BRGH bit. That causes the baud rate 4 times faster than normal.
Then you have to set the SPBRG2 register as 25. Because the CCS compiler sets it as 64. This causes an unexpected baud rate and framing errors. (4000 baud)
Hope this helps to your future implementations.
And the extended part of the question.
I need to set the FUSES as "INTRCPLL_IO". But looks like this also does not work as expected (MPLab show that this line chooses HS, instead of INTRC). I always have to set the config bits manually after every compile from the "Configuration Bits" window. I looked to PICC project wizard ot the CCS's own IDE. The project wizard does not even show an option as INTRC for 18F66J90. I think it was an false implementation of the PIC, am I right?
Is there any way to set the config bits manually in the code? (without using CCS #fuses)
_________________
/// KMT
/// www.muratursavas.com |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Tue Dec 08, 2009 3:53 am |
|
|
Thanks PCM Programmer, you've saved the day, again
_________________
/// KMT
/// www.muratursavas.com |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Wed Dec 23, 2009 10:57 am |
|
|
Great 
It was working fine since my last post, but now somehow CCS adds another fuses line to the hex file, like that:
| Code: | :06FFF800A0F001F7F0F09B
:06FFF800A0F4DFFFF2F1AE |
Where is that second line coming from? It ruins the config bits and I have to set the bits again for every compile within MPLab.
P.S: Tried to reinstall CCS, but didn't help.
_________________
/// KMT
/// www.muratursavas.com |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Wed Dec 23, 2009 1:23 pm |
|
|
Post a very short test program that shows the problem.
It must be compilable.
Post your current compiler version. |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Thu Dec 24, 2009 2:05 am |
|
|
Here is the config:
- CCS 4.093
- 18F66J90-I/PT Rev. A1
- MPLab 8.43
- PICKit2
Test code:
| Code: | #include <18F66J90.h>
#device *=16 ADC=10
#ROM int8 0xFFF8 = {0xA0, 0xF0, 0x01, 0xF7, 0xF0, 0xF0 } /// Config Bits
#use delay(clock=16000000, restart_wdt)
#use rs232(stream = AUX, baud=9600, parity=N, bits=8, stop=1, restart_wdt, UART2, ERRORS)
#byte SPBRG2 = 0xF64
#bit BRGH = 0xF61.2
void main(void)
{
setup_oscillator(OSC_4MHZ | OSC_PLL_ON);
output_high(PIN_D2);
delay_ms(200);
output_low(PIN_D2);
delay_ms(200);
output_high(PIN_D2);
BRGH = 0;
SPBRG2 = 25;
while(1)
{
fputc('U',AUX);
delay_ms(50);
}
} |
Generated hex code:
| Code: | :100000001FEF00F0EA6A060EE96EEF5011E0050EF0
:10001000016EBF0E006E0400002EFDD7012EF9D731
:100020002E0E006E002EFED700000400EF2EEFD73C
:10003000000C0400A4A8FDD7626E4BEF00F0F86A34
:10004000D09EEA6AE96A0086A00E646E010E016E17
:10005000A60E616E900E606EA86AAA6AA96AC15067
:10006000C00BC16E070EB46E056A400E9B6E600E2B
:10007000D36ED35095948C84C80E066EC3DF9594CE
:100080008C94C80E066EBEDF95948C846194190E14
:10009000646E550ECED7320E066EB4DFFAD703006B
:06FFF800A0F001F7F0F09B
:06FFF800A0F4DFFFF2F1AE
:00000001FF
;PIC18F66J90
;CRC=E506 CREATED="24-Ara-09 09:45" |
Correct fuses line: A0 F0 01 F7 F0 F0
Incorrect added fuses line: A0 F4 DF FF F2 F1
As you can see, it messes up with the Oscillator Bits (underlined) and makes the program useless.
P.S: It messes up only with 18F66J90. I tried same program with different PIC and there were no problems.
_________________
/// KMT
/// www.muratursavas.com |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Thu Dec 24, 2009 4:50 am |
|
|
Here is another thing not related with HEX CONFIG BITS issue, but with UART2 settings.
In this post I've really struggled with UART2, and it seems CCS has more problems.
I've found that "setup_spi2();" function has a false SFR address and ruins the UART2 settings. Be careful with using it 18FXXJ90 series. I don't know whether CCS fixed or not this issue on the newer versions. [mine is CCS 4.093 as mentioned before]
_________________
/// KMT
/// www.muratursavas.com |
|
|
dyeatman
Joined: 06 Sep 2003
Posts: 1923
Location: Norman, OK
|
|
| Posted: Thu Dec 24, 2009 7:19 am |
|
|
FWIW, here are the fuses in the .LST file when your code is compiled in 4.103
| Code: |
Configuration Fuses:
Word 1: F4A0 NOXINST STVREN NOWDT NOPROTECT
Word 2: FFDF E4_SW WDT32768 FCMEN IESO PRIMARY HPT1OSC T1DIG
Word 3: F102 NODEBUG CCP2C1 RTCOSC_T1
Word 4: 0000
|
And the actual hex code:
| Code: |
:100000001FEF00F0EA6A050EE96EEF5011E0050EF1
:10001000016EBF0E006E0400002EFDD7012EF9D731
:100020002E0E006E002EFED700000400EF2EEFD73C
:10003000000C0400A4A8FDD7626E4AEF00F0F86A35
:10004000D09EEA6AE96A7C86A00E646E010E7D6E1F
:10005000A60E616E900E606EA86AAA6AC150C00BAF
:10006000C16E070EB46E046A400E9B6E600ED36EB6
:10007000D35095948C84C80E056EC4DF95948C94EF
:10008000C80E056EBFDF95948C846194190E646E62
:0E009000550ECFD7320E056EB5DFFAD703003E
:06FFF800A0F001F7F0F09B
:08FFF800A0F4DFFF02F100009C
:00000001FF
;PIC18F66J90
;CRC=4396 CREATED="24-Dec-09 07:01"
|
_________________
Google and Forum Search are some of your best tools!!!! |
|
|
KaraMuraT
Joined: 16 May 2006
Posts: 65
Location: Ankara/Turkey
|
|
| Posted: Fri Dec 25, 2009 7:17 am |
|
|
Then it means this bug still exists in the latest version.
Thanks dyeatman.
_________________
/// KMT
/// www.muratursavas.com |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
