Wire worked by not wireless

[dhouston]

I've used the TWSxxx and RWSxxx (where xxx=MHz) extensively over the past 7-8 years. I have attached a couple of GIFs (captured at the receiver output) that illustrate the problem you have to deal with.

One GIF is of a marginally weak signal. The upper trace is the LINEAR output pin while the lower trace is the DIGITAL output pin. This happens to be an X-10 RF signal (which uses the NEC IR protocol). You would need to wait in a loop until you receive a valid preamble (e.g. using PulsIn) and then capture the next 32 bits.

The other GIF shows another approach with a preamble that is repeated a few times prior to the actual code. It takes 2-3 copies of the preamble to set the receiver AGC to a point where the preamble emerges from the noise. You would need to wait in a loop for the preamble and then capture the code which follows.

As I recall, Holtek merely repeats the entire code several times with the first copies used to set the AGC so you would have to listen all the time for a valid code.

One more tip: If you do not have access to a 'scope you can record these signals using LineIn of a soundcard. You just need a couple of resistors to form a voltage divider so that the voltage input to the card is in the 0-1V range. This has the added advantage that you can hear the codes, as well. The GIF with the single trace was recorded this way.

[mankan]

Hello,

I am making a Tx port pin High/Low for some duration of time and then on the receiver side, I am just using PULSIN to check the pulse.

I found that transmission and reception is very fine and this brings to the opinion that SERIN command has some problem.

Am I correct? Is there any other way of implementing SERIN for sending RF data?

Thanks for you support.
Mankan

[dhouston]

mankan,

Here is a tutorial that shows how to do manchester RS232 RF with a PIC using ASM. Note that he uses what he calls a header (equivalent to what I've been calling a preamble) to alert the receiver, set the AGC, etc.

• http://www.winpicprog.co.uk/pic_tutorial12.htm

The problem you will have on the receiving end is the same - you'll have to monitor full-time for the preamble/header to detect the beginning of a message.

How much data do you need to transmit?

[mankan]

Hello,

I have to transmit only following command pattern:

Header - 1 byte
Command - 2 byte + 2 byte
End of Packet - 1 byte

Hope this should not be a problem for RF transmission.

Thanks,
Mankan

[dhouston]

Panasonic uses a 48-bit protocol that is very similar to the NEC IR protocol. You could adapt the NEC protocol to your needs.

If your header and end-of-packet bytes serve no other purpose you can drop them, using those that are built-in to the NEC protocol. Here's a URL to documentation for the NEC protocol.

• http://www.cpcares.com/pdf/UPD6121G-001.PDF

It uses a 9.5mS pulse followed by a 4.5mS space as the preamble or Start-of-Packet and a 0.5mS pulse followed by a 40mS space as End-of-Packet. I use a 0.5mS pulse followed by 1.5mS space as digital 1 and 0.5mS pulse followed by a 0.5mS space as digital 0. I would send your 4 data bytes plus one checksum byte for error detection.

Experiment to find whether you can send just one copy or need to send multiple copies to guarantee clear reception.