| View previous topic :: View next topic |
| Author |
Message |
newby
Joined: 26 Sep 2004
Posts: 32
|
| strange RS232 behaviour ? |
 Posted: Tue Mar 15, 2005 2:31 pm |
 |
|
I have connected a GPS receiver with serial output (TTL) directly to a PIC 18LF2620 on RC7 (RX). The GPS output has high levels of 3 V and low Levels of 0 V as long as the PIC is not connected. The PIC is supplied by 3.3 V. If i connect the PIC, the low level of the GPS signal raises up to about 1.1 V. What�s that ?!? Why does the GPS signal changes its voltage if i connect the PIC ? The manual of the GPS receiver tells me that it is no problem to connect the GPS directly (without MAX drivers...) to a e.g. PIC.
That�s my 1st problem and probably also the reason for my 2nd problem.
The 2nd is, that the receive of the GPS data on the PIC is malfunctioning from time to time. That means sometime "nothing" comes into the RX buffer, sometimes only garbage and sometimes it works like it should...
I use a ceramic oscillator at 8 MHz (ICD says 8.05) and want to establish a 4800 baud connection to the GPS - that should not be a big problem... but it seems to be. Could it also be that the frequency is not stable enough for using RS232 ? Do i have to use oscillators instead of these smart ceramic oscillators ?
I also tried to use the "auto baud rate detection" module, but i could not find any clear information how to use it with CCS and it didn�t work just by setting it to "auto mode without waiting for 055 char".
Does anybody have an advice for me ? |
|
 |
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Tue Mar 15, 2005 2:43 pm |
|
|
| Quote: | The PIC is supplied by 3.3 V. If i connect the PIC, the low level of the
GPS signal raises up to about 1.1 V. What�s that ?!? |
To me, this implies that the PIC pin is configured as an output.
So you have two output pins fighting each other.
I suspect that you are setting the TRIS for port C, and it's not correct.
Post a small test program that shows the problem, and post your version
of the compiler. Show all statements such as #fuses, etc. |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Tue Mar 15, 2005 11:50 pm |
|
|
..no that�s not the problem, i have checked this already.
TrisC.7 ist set to 1 (input). But according to the data sheet of the PIC18F2620 it doesn�t matter, as soon as you enable the EUSART, the PIN is configured correctly - i tried this without setting TrisC.7 with standard and fixed IO, it�s all the same.
I�m on a business travel today and will post the fuses and code as soon as i�m home again. I will also have to reduce the code to the "main problem", it�s too much code...
Thanks a lot ! |
|
|
Douglas Kennedy
Joined: 07 Sep 2003
Posts: 755
Location: Florida
|
|
| Posted: Wed Mar 16, 2005 2:38 am |
|
|
| Did you try pulling the GPS low ( Ex 20k resistor to ground)..and are the PIC and the GPS grounds tied together? |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Thu Mar 17, 2005 2:21 pm |
|
|
I have reduced the code to the following (i use PCWH 3.221):
| Code: | #include <18F2620.h>
#device ICD=TRUE
#fuses HS,NOIESO,NOBROWNOUT,NOWDT,NOPUT,STVREN,NOLVP,NOFCMEN,NOXINST,NOPBADEN,MCLR,DEBUG
#use delay (clock=8000000)
#use rs232(baud=4800, parity=N, xmit=PIN_C6, rcv=PIN_C7, bits=8, stream = GPS_channel, ERRORS, BRGH1OK)
//setup_UART(UART_AUTODETECT_NOWAIT);
#use rs232(DEBUGGER, stream = debb)
#ZERO_RAM
#BYTE RCREG = 0x0FAE
#BYTE RCSTA = 0x0FAB
char char_dummy, rs232_buffer[50];
int8 rs232_i, i;
#int_RDA
RDA_isr() {
rs232_i++; if (rs232_i == 50) { rs232_i = 0; }
rs232_buffer[rs232_i] = RCREG;
}
void main() {
// setup_oscillator(OSC_8MHZ, 0);
port_b_pullups(TRUE);
setup_wdt(WDT_OFF);
setup_comparator(NC_NC_NC_NC);
setup_ADC_ports(NO_ANALOGS);
enable_interrupts(INT_RDA);
enable_interrupts(GLOBAL);
while(TRUE) {
i++;
char_dummy = fgetc(GPS_channel);
if (i == 50) {
i = 0;
fputc(char_dummy, debb);
}
}
}
|
Both the isr INT_RDA as the while loop in main recieves data, but garbage (from time to time, but quite often)... The data provided by the GPS is OK, ich checked it using the SIOW tool.
I have also tried to pull down the GPS Signal to GND by several resistor values, ranging from 1 kOhm up to 100 kOhm. But all what happens is, that the high & low level of the signal is shifted a bit (about 0.1 - 0-2 V)towards GND. |
|
|
ckielstra
Joined: 18 Mar 2004
Posts: 3680
Location: The Netherlands
|
|
| Posted: Thu Mar 17, 2005 4:59 pm |
|
|
| PCM programmer wrote: | | Quote: | The PIC is supplied by 3.3 V. If i connect the PIC, the low level of the
GPS signal raises up to about 1.1 V. What�s that ?!? |
To me, this implies that the PIC pin is configured as an output.
So you have two output pins fighting each other.
|
I do agree this sounds like a hardware problem with two outputs connected. Are you sure you have connected the GPS Tx to the PIC Rx pins? I mean a cross connection, also called null-modem.......???
Thinking more about it, this is not very likely the case as you said you sometimes do receive data.
A problem in your test program is that you have both the ISR and the main routine read the USART receive register simultaneously. Now it all depends on luck which routine gets the character and sometimes both routines get the character. Totally unpredictable. Change this so that only the ISR reads the data, the main routine can read from the buffer filled by the ISR. |
|
|
Douglas Kennedy
Joined: 07 Sep 2003
Posts: 755
Location: Florida
|
|
| Posted: Thu Mar 17, 2005 7:33 pm |
|
|
As well as the changes ckielstra suggested
I'd change
| Code: | if (rs232_i == 50) { rs232_i = 0; }
|
to | Code: | if (rs232_i >49 ) { rs232_i = 0; }
|
Also initialize rs232_i to zero
You probably have done this already if not
I'd look to test if the PIC rx pin that is now tied to your GPS Tx pin is configured as an input. Try disconnecting the GPS which you know to work and see if the PIC rx pin will drive a voltage across the 20k resistor you have tied to GND. If it's 3v its an output if its zero its probably an input but could be an output putting out zero. In case it could be pulling low tie the 20k to 3v and see if it actually does pull low. |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Thu Mar 17, 2005 11:49 pm |
|
|
the above code is only "quick and dirty" to show the problem. If garbage is recieved than not only 1 of 100 chars, it�s more than up a certain point it comes only garbage... and then it functions again (that can change in minutes or i have to "wait for an hour"...)
I use #zero ram, therefore all variables are set to 0 at startup.
How could the RX PIN be set to output when using the EUART ? I have tried already playing around with the TrisC register but it doesn�t change anything. In the datasheet of the 18F2620 is also mentioned that enabling the EUART will auomatically configure the RS232 pins... But nevertheless i will try your suggestion ! Thank you ! |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Fri Mar 18, 2005 12:02 am |
|
|
@ckielstra:
the above code is only "quick and dirty" to show the problem, it�s clear that it does "not function" in a way to receive a complete GPS data set.
In my program i use the isr to fill the buffer. I found out that sometimes the INT_RDA is not started even there are some chars which can be got using fgetc(). Which is indeed also very strange (to me)?
The Pic receives somewhat which can be polled using the getc functions, but it doesn�t start an INT_RDA... That is why i implemented both ways to look if this strange isr problem occurs again.
I will check my wires again very carefully. At the moment i can only say that the RX from PIC is directly (just on wire) connected to the TX from GPS, i don�t use the TX from PIC and the RX from the GPS. But that shouldn�t be a problem (it has also worked before...).
I guess if there would be a mistake, i would not be possible to receive (sometimes, for e.g. a couple of minutes) several complete and correct GPS data sets.... |
|
|
Ttelmah
Guest
|
|
| Posted: Fri Mar 18, 2005 3:01 am |
|
|
You need _two_ wires to connect even one way RS232. The signal itself and the ground (another poster suggested looking at this). I suspect what is happening, is that you are perhaps relying on a ground 'through' a power supply, or other connection, and it is either not reliable, and/or the quality of connection changes with different hardware layouts, and is generating your problem.
Best Wishes |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Fri Mar 18, 2005 4:52 am |
|
|
| Quote: | | You need _two_ wires to connect even one way RS232. The signal itself and the ground |
...i don�t understand this. The GPS uses the _same_ GND as the PIC, and that should not be enough to ensure that te GND potential is the same ?
How does it work anywhere else ?
Or do you mean because the GPS seems to have some kind of MAXxyz driver implemented, that the GND from the GPS after MAXxyz is floating ?!?
I will try this out and connect RX AND TX, thought that is not necessary (it hasn�t been till this project) !
Thank�s a lot, hope that�s it ! |
|
|
Douglas Kennedy
Joined: 07 Sep 2003
Posts: 755
Location: Florida
|
|
| Posted: Fri Mar 18, 2005 8:10 am |
|
|
The description of this issue seems confusing. The poster says the GPS has TTL output built in and if I read correctly the poster says it can also talk to the PC RS232 port and send data correctly according to the poster ( using SIOW) without a second level converter ( max 232 eg) Presumably the poster is selecting invert bytes in the SIOW menu. Now it maybe the posters voltage measurements are made with a VOM while the GPS is transmitting in which case they would represent an average voltage.
Now the UART port has a hysterisis so it might be that the UART which waits for a start bit before kicking into action is not seeing most start bits.
The PC hysterisis will be markedly different since it is expecting true rs232 and will see zero or less as a mark.
The poster says the PIC is has a 3v supply but I see NOLVP in the fuses setting.
The GPS supply is probably also at 3v but if its 5V it maybe the cause of the problem. |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Fri Mar 18, 2005 10:53 am |
|
|
@Ttelmah
I have connected RX and TX, it didn�t change anything.
@Douglas Kennedy
The GPS modul is powered with 5V, the PIC with 3.3 V. The RS232 Signal from the GPS has (without PIC) a high level of 3 and a low level of 0 V. I don�t know what a "VOM" is, i use a oscilloscope. Yes the GPS can provide TTL as "RS 232" levels - says the description and it "works/ed". |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Fri Mar 18, 2005 11:06 am |
|
|
Can you post the manufacturer and part number of the GPS unit,
and a link to the data sheet if possible ? |
|
|
newby
Joined: 26 Sep 2004
Posts: 32
|
|
| Posted: Fri Mar 18, 2005 12:00 pm |
|
|
Unfortunately i couldn�t find a datasheet, but here is the description http://www.fortuna.com.tw/U2.htm.
I�m "redesigning" my test bed at the moment to test it with PIC and GPS both running at 5 V. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
