This forum has moved to forum.jeelabs.net - this read-only archive is for reference only!
Question:
In the RFM12.cpp file the bitrate is set in the 0xC6 field: rf12xfer(0xC606); // 57.6Kbps (38.4: 8, 19.2: 11, 9.6: 23, 4.8: 47) Are you sure that 06 is the command for 57.6 kbps ? I think it is 05 So the speed now is 49 Kbps the formula is 10Mhz/29/(r+1)/1+cs7) see Hope RF12b programming guide.
regards Jeroen
found a nice tool to configure the rfm12b:
(
RFM12 (old label, producing by HOPE MICROELECTRONICS CO) = SI4420 (new, produced by Silicon Labs)
(edited by jcw: added <>'s to make the URLs live)
Let's see. If we assume these calculations work in integer truncation mode:
10000 / 29 = 344 344 / (1 + 6) = 49.142 KHz
Oops. You're right!
So that's around 20 µS per bit, i.e. 160 µS per byte transmission time - which is also the interrupt rate, btw.
I'm going to have to leave this as is - too many implications if we change now, I'm afraid :(
Note that there's no relation with the 57600 baud of the serial port. I just wanted to attain a reasonably high transfer rate.
Don't know the implications. Could be substantial, I think update the //comment field with the right value is perhaps sufficient.
Just noticed this change in the source, but notice that it no longer lists other baud rates in the comments. I found this really useful rather than having to refer to the datasheet to select another baud rate.
Oh, and I perviously found that at 115,700 baud, the 's' command to send a full message would fail. I'm wondering if the previous '115' setting was also miss-calculated, causing the RFM12 to go out of range?
At 115,700 baud, there would be an interrupt every 10 µs or so. Might be too much for our poor little ATmega (at least with the current C-code interrupt overhead). I've added it to my somewhat gigantic TODO list as something to look into - maybe there's an easily-fixed bottleneck.