How will using Manchester encoding help to prevent the false reception 
of data?  Wouldn't I need to first find the rising edge then ensure that 
it is still a one after say 35% of the pulse width, then again after 70% 
to ensure that it was not just noise I was seeing?  If it was noise then 
I need to keep looking at the input to detect the 'real' rising edge of 
the data.  Seems like that is going to take a large portion of the PIC, 
leaving little for other processes.

What am I missing?

Phil

Bob Axtell wrote:

> The sequence that seems to work best is a form of Manchester Code. It 
> is usually
> a bit time of 5-10Ms, where the transmitter is keyed ON at the 
> beginning of each bit.
> If it remains keyed for more than 50% of the bit time, its a '1', 
> otherwise it is a '0'.
> You will need to include a CRC method so that the receiver can verify 
> that the data
> was received correctly. Finally, just send it 4 times; when the msg is 
> received correctly
> for 2 or more passes within a tight time interval, the data is accepted.
>
> Done this way, I've had 100% reliability at 100meters on two different 
> projects.
>
> --Bob
>
> Phil Keller wrote:
>
>> I am working on a project that has a remote interface transmitting 
>> data to a central location.  Unidirectional at this time but you 
>> never know :). The data rates are low (2400 Baud is sufficient) and 
>> the distance is not excessive (50-100m).  I tried the TWS/RWS-434 
>> pair connected to the UART of a PIC but that does not function very 
>> well.  The problems appears to be that the receiver is always 
>> decoding data even when the transmitter is off line.  This, I assume 
>> is due to natural noise and is unavoidable.  Because of the noise, 
>> the RX UART has a very difficult time synchronizing to and receiving 
>> real data.
>>
>>  I could add in the Holtek encoder/decoder pair but that adds 
>> complexity and component count.  If that is the best solution, so be 
>> it, but......
>> - Is it feasible to write my own asynchronous routine to receive the 
>> serial data bits and to shift through the stream until an SFD is 
>> detected? (and still have some CPU left over for other things)
>> - Has anyone used the Radio Packet Controller from Radiometrix.  Cost 
>> may be an issue here? http://www.radiometrix.co.uk/products/frpc2.htm
>> - Other suggestions?????
>>
>> Phil
>>
>
>

-- 
===========================================
PK Consulting
Phil Keller
Fremont, Ca

Phil@PKConsulting.net



-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist