RF and PIC basic

[dhouston]

I found another datasheet that does describe the protocol used by this chip. While the specifics for a "1", "0", and "F" differ from Holtek, the overall approach is very similar to the HT640. A bit period depends on the encoder oscillator frequency (set by an external resistor) you still need to capture some codes to determine the bit period. And there are some encoder pins that can be used either in the address or in the data so you'll need to determine that, also. Then you should be able to adapt Bruce's HT640 code to work with the PT2260.

• http://www.spelektroniikka.fi/kuvat/PT2260.pdf
  http://www.picbasic.co.uk/forum/show...ghlight=holtek

[Megahertz]

Any ideas (Bruce's code for HT640) if this will work if HT640 is used with other RF modules, like the normal el chipo style?
I will give it a go to decode PT2262 soon.

[Ioannis]

I tried to decode a UM3750/MM53200/HT12 type encoder with this either as subroutine or as Interrupt Driven Routine:

GET_CODE:
    PULSIN DIN,0,NUMS
    IF NUMS > 8200 and NUMS<12000 then        'Detect Header
        goto get_pulses
    else
        RETURN
    endif
get_pulses:
      FOR X = 0 TO 11           ' read 12 pulses
        PULSIN DIN,1,NUMS        ' 
        if NUMS>220 then
            if NUMS<300 then
                code.0[x]=1
            endif
        endif
 
        if NUMS>450 then
            if NUMS<560 then
                code.0[x]=0
            endif
        endif
        if NUMS=0 then
            loops=0:code=0:flag=0
            RETURN    '   temp=0:
        endif
   next x
   if loops<3 then
    if prev_code=code then
        loops=loops+1
    else
        prev_code=code
        loops=0
    endif
   endif
return

But I got a delay in decoding. It just gets one every 4 transmissions. So for a 4 times check of equall transmitted codes, it takes more than 250ms to give a valid. With the Logic Digital Analyser it is confirmed that it takes 1st transmission, then 4th etc.

The numbers in NUM are PIC and OSC clock dependant.

If anyone can spot why, please reply.

Ioannis

[mackrackit]

What var size is NUM?

[Ioannis]

Word sized.

It does work but gets one every 4th. Very strange...

I have not here the Logic. In the afternoon I will post a shot of the pulses to make it more clear.

The posted routine was tested also as part of Darrels DT-INTs on a fast 16F1827 chip at 32MHz clock.

Ioannis

[Jerson]

Ioannis, not sure what you mean by gets one every 4th. The way I understand your statement is that you get a valid pulse once in 4 transmissions. If that is correct, the code which does that is the loops part. if loops < 3 then ..... When it finds a code 'repeated' 4 times, it transfers to the Prev_code. Is that what you're looking for??

[Ioannis]

Hi Jerson. No, it is not that the case. It just gets the first train of pulses but not the second one after the header of 11ms. It waits for a reason that I do not understand and gets the 4th train again. I try to get a shot from the Logic screen and post it next.

Here is the picture:



The first channel shows the data received from the Rx module.

The second shows every time the routine detects the header like this:

 PULSIN DIN,0,NUMS
    IF NUMS > 8200 and NUMS<12000 then
        toggle portb.4
        goto get_pulses
    else
        RETURN
        'goto get_p
    endif

That delay is constant and not random. So it must be either a shortage ofthe lanquage or my mistake (possibly).

Ioannis