NEC IR Protocol remote decoder, help me before I kill myself!

[dhouston]

I would allow for ±20% deviation in the start pulse. The NEC code has built-in error checking so grabbing the wrong code is not much of a problem.

Where are you doing your testing? Fluorescent lights can output a lot of noise in the 38kHz band that NEC used. Direct sunlight can also be a complication. Otherwise, IR tends to be mostly noise free so catching the start code is usually straightforward.

I would not use PulsIn. Here's some hastily modified and untested code that should work. Blame the forum software for the screwed up spacing/tabs, etc.

'=======================| GENERIC IR/RF RECEIVER |=========================
'GEN-RF-683.BAS
'PIC12F683 @4MHz                          uses MPASM

'==========================================================================

@ __config _INTRC_OSC_NOCLKOUT & _WDT_OFF & _PWRTE_ON & _MCLRE_OFF & _CP_OFF & _IESO_OFF & _FCMEN_OFF

DEFINE OSC 4                    
DEFINE DEBUG_REG GPIO
DEFINE DEBUG_BIT 2     
DEFINE DEBUGIN_REG GPIO
DEFINE DEBUG_IN 3            
DEFINE DEBUG_MODE 1                 
DEFINE DEBUG_BAUD 9600

@Timer1=TMR1L
Timer1  VAR      WORD EXT
TMR1ON     VAR    T1CON.0          'Timer1 ON/OFF control bit    
IR      VAR     byte[4]
leadin    var    byte[2]
period    VAR     byte
i       VAR     byte   
bits    VAR    byte
           
    OSCCON        =%01100001    '4MHz _INTRC_OSC_NOCLKOUT
        CMCON0         =%00000111      'disable comparators
        ANSEL        =%00000001    'AN0=analog
    ADCON0.7    =1        'right justify ADC
        TRISIO        =%00101011    'input=1,output=0
        GPIO            =%00101000    'high=1,low=0
        INTCON        =%00000000    'disable interrupts
        T1CON        =%00110000    'TMR1 prescaler=8,tick=8µS
                     
init:    RF[0]=0:RF[1]=0:RF[2]=0:RF[3]=0:
        While GPIO.1=1            'wait falling edge    
        Wend                            'falling edge        
        TMR0=0:OPTION_REG=%10000101    'TMR0 prescaler=64
        While GPIO.1=0                'wait rising edge
        Wend                'rising edge
    leadin[0]=TMR0:TMR0=0
        If (leadin[0]<154) Then init    '8500µS (154)
        If (leadin[0]>190) Then init    '10500µS (190)
        While GPIO.1=1            'wait falling edge      
        Wend                'falling edge
        leadin[1]=TMR0:TMR0=0           'clear TMR0                
        OPTION_REG=%10000011        'TMR0 prescaler=16
        i=0
        Repeat
          While GPIO.1=0                 'wait rising edge
        If TMR0>186 Then break      '3000µS      
          Wend                'rising edge
          While GPIO.1=1                'wait falling edge
        If TMR0>186 Then break      '3000µS      
          Wend                'edge  
          period=TMR0:TMR0=0        'reset TMR0
          If (period<49) Then init        
          If (period>93) Then        '1500µS        
            IR.0(i)=1                
          EndIf
          i=i+1
        Until (i>31)
break:  If (i<32) Then init 
        Debug hex2 i
        For i = 0 to 3
        Debug hex2 IR[i] REV 8
        Next
        TMR1ON=0:Timer1=0:TMR1ON=1
        Debug 13,10 
        GoTo init
    
    End

[Bonxy]

Hi Guys

Ok, I am totaly stumpped with this, I have tried to use a timer as suggested by dhouston & I have been trying the pulsin method as suggested by bogdan but I get the same problem, the start pulse is all over the place, it ranges from 0 to 255 etc, it seems random.
I have changed the ir reciver to eliminate a fault there and I checked the output on my scope and it seems to show a correct pattern.

God knows what I am doing wrong, I just wonder if it is down to the pic's i use 18f4550 & 18f14k50 ?, I have even changed the oscillator down to 4mhz to try and copy bogdan's pulsin code (shown below), also I dont know why the example uses a word var for the pulsin, am I missing something there ?.
Maths is not my strong point so maybe I'm just reading things worong I dunno but I've been at this a week now and cant get the same start pulse 2 times in a row.

Could someone check the code below and tell me why I dont get values in the range of 400 - 490 ?.


'Defines for EasyPic3
DEFINE LCD_DREG PORTB ' Define LCD registers and bits
DEFINE LCD_DBIT 4
DEFINE LCD_RSREG PORTB
DEFINE LCD_RSBIT 2
DEFINE LCD_EREG PORTB
DEFINE LCD_EBIT 3
INCLUDE "AllDigital.pbp" 'Set ALL ports to Digital i/o
DEFINE OSC 4 ;=4MHz
DEFINE PULSIN_MAX 490

Header VAR WORD '<-- why is this a 'word' instead of a byte ?
index VAR BYTE
IR_pulse VAR BYTE[31]
IR_data VAR BYTE[31]
PORTC=0
TRISC=%000001 ;PORTC.0 input for the IR



'Note, device is a 18f4550

IRIN:
PULSIN PORTC.0,0,Header

IF (Header < 400) OR (Header > 490) THEN GOTO IRIN '<-- problem, values I get are not between 400 to 490


goto irin

[Jumper]

Hi,

Please post your HW schematic with component numbers and values. It will be much easier for all of us if we can see what and how you connect to the PIC. Do you have the pic on 5V and the reciver also on 5V.. or are you running a 5V receiver on 3V ?

Step one must be to rule out HW and also make it possible for others to duplicate your setup for testing purposes.

/me

[Ioannis]

Your Header variable is a word because the values you expect are beyond 0-255 range that a byte can accept.

Keep in mind that Pulsin is clock dependant. The real clock your PIC is using, NOT the one you declare with DEFINE OSC line. At 4MHz it increment by a 10usec step. At 20MHz by a 2usec step.

How do you know that values are not within the range?

Do you have some kind of output check?

Ioannis

[Bonxy]

Thanks for your replies

Jumper, I dont have a schematic to post but Im using an easypic3 board with a 18f4550 at 5v with the tsop 2438 reciver on the same 5v with the output pin connected to portc.0 (pdf datasheet attached).

Ioannis, I realised as soon as I posted why it was a word var lol, I assume the valuse are not in the range as I am using the same xtal and code as bogdan's working code ?, as I said I am usless at math so I am taking bogdan's word that the value I should get at 4mhz is between 400 to 490.

[bogdan]

the code was for a samsung tv .... close, but not the same like the nec protocol

SAMSUNG IR protocol:
http://www.samsung.com/global/busine...000_090108.pdf (check "Figure 1. IR Signal")



NEC IR protocol:
http://www.mcselec.com/index.php?opt...=223&Itemid=57

... the leader/header for nec should be 9msec

change the
DEFINE PULSIN_MAX 990 (try with 10% deviation... 900*1.1)
- - - - - - - - - - - - - - - - - - - - -
IF (Header < 810) OR (Header > 990) THEN GOTO IRIN (again the -/+10% deviation)

[Bonxy]

I've looked at the nec ir protocol but i'm non the wiser :-(

Just if anyone is interested I have attached the actual code that came with my remote and reciver IC kit, it was written for the arduino and commented in japanese, I have translated the comments to english but I cant translate the code, if anyone wants to look at it it might reveal to someone more experienced what I need to do to get it to work maybe ?.

[Jumper]

Plan B:

Use an other pic and set up the HPWM manually for 38 kHz. Then you connect an IR diode to the HPWM pin using a resistor to get aprox 20 mA in the led.

Then in the code you make a simple main loop. Turn on the HWPM, delay 9 ms, and then turn off the HPWM, wait a while (i.e 5 ms) and then goto main again. Point this led towards the ir-receiver from a close distance and now you have a signal that you really know how it looks like. Then you put a shoebox upside down over both the receiver and the transmitter led so they get a nice dark space to be in.

Have you ever thought about catching the IR signal using the CCP module or by using interupts and a timer?

This way you can even build a text signal for the NEC protocol ... [startpulse] delay [bit0] delay etc..

Which way to decode the signal is the best? Well it all depends what other things your pic will do. If it will only decode the IR signal you can just read the pin in a fast loop and increment a counter when it is LOW.

Keep in mind that the receiver you are using is ACTIVE LOW.. so the pin will be high when there is no modulated IR light present.

[dhouston]

NEC IR protocol:
http://www.mcselec.com/index.php?opt...=223&Itemid=57

Here's a more authoritative source for the NEC protocol...

• http://davehouston.org/Nec6121.pdf

[Bonxy]

Thanks dhouston, but looking at more protocol descriptions is just giving me headache lol, I've been at it all week, I think I've actually looked at the description before when I was first googling the nec protocol.

What I need is a simple working code example that I can then develop...

If you or anyone can give me a code example for a 18f4550 at 4 or 20 mhz i would be extreemly gratefull..

Thanks guys

[cncmachineguy]

the code was for a samsung tv .... close, but not the same like the nec protocol

SAMSUNG IR protocol:
http://www.samsung.com/global/busine...000_090108.pdf (check "Figure 1. IR Signal")



NEC IR protocol:
http://www.mcselec.com/index.php?opt...=223&Itemid=57

... the leader/header for nec should be 9msec

change the
DEFINE PULSIN_MAX 990 (try with 10% deviation... 900*1.1)
- - - - - - - - - - - - - - - - - - - - -
IF (Header < 810) OR (Header > 990) THEN GOTO IRIN (again the -/+10% deviation)

Did you try this?

[Bonxy]

cncmachineguy

Yes I did, i still did not get valuses to match, I got values in the range of 552 to 614 and everything in between.

[bogdan]

please let us now the model # of the remote

(i will try to upload the codes into a universal remote and i will check the protocol later today)

[Bruce]

There are much better/faster ways to do this, but this should get you started with something that's fairly easy to understand.

Assuming you have an NEC transmitter that outputs something like the top portion of the graphic below;

This should return similar results to what's shown in the serial terminal window just above.

#CONFIG
  __config _XT_OSC & _WDT_ON & _MCLRE_OFF & _LVP_OFF & _CP_OFF
#ENDCONFIG
 
DEFINE OSC 4
 
Leader VAR WORD     ' will be up to 900 for a 9mS leader pulse
BtnVal VAR BYTE[32] ' holds 32 pulse results
DByte1 VAR BYTE     ' address byte
DByte2 VAR BYTE     ' inverse of address byte
DByte3 VAR BYTE     ' command byte
DByte4 VAR BYTE     ' inverse of command byte
X      VAR BYTE     ' loop count
 
Main: 
  PULSIN PORTB.0,0,Leader  ' leader pulse is ~9mS low-going
  IF Leader < 850 THEN Main
 
  FOR X = 0 TO 31          ' grab 32 incoming pulses
      PULSIN PORTB.0,1,BtnVal(X) ' now measuring high-going pulse widths
  NEXT X
 
  ' now we'll decode 4 bytes from 32 pulses
  FOR X = 0 TO 7               ' sort 1st 8 pulses
      IF BtnVal[X] > 150 THEN  ' > 150 x 10uS = > 1.5mS pulse period
         DByte1.0[X]=1
      ELSE
         DByte1.0[X]=0
      ENDIF
  NEXT X
 
  FOR X = 8 TO 15              ' sort 2nd 8 pulses, etc....
      IF BtnVal[X] > 150 THEN
         DByte2.0[X-8]=1
      ELSE
         DByte2.0[X-8]=0
      ENDIF
  NEXT X
 
  FOR X = 16 TO 23
      IF BtnVal[X] > 150 THEN
         DByte3.0[X-16]=1
      ELSE
         DByte3.0[X-16]=0
      ENDIF
  NEXT X
 
  FOR X = 24 TO 31
      IF BtnVal[X] > 150 THEN
         DByte4.0[X-24]=1
      ELSE
         DByte4.0[X-24]=0
      ENDIF
  NEXT X
 
  HSEROUT [BIN8 DByte1,13,10,BIN8 DByte2," Address",13,10,13,10]
  HSEROUT [BIN8 DByte3,13,10,BIN8 DByte4," Command",13,10,13,10]
  PAUSE 1000
  GOTO Main