Match rileylink settings

[oskarpearson]

(This is based on my previous PR - #3 )

This patch updates the modem settings to match the Rileylink project https://github.com/ps2/rileylink/blob/master/firmware/cc1110/main.c#L71-L104

With these settings, I seemed to get fewer CRC errors (in the UK at least). I have to say that the detail of the differences is not something I'm 100% clear on. Please let me know if I've done something wrong :)

[oskarpearson]

I chatted to @ps2 - the author of RileyLink in Gitter. He commented as below.

I'd be interested in your thoughts on changing these, and whether there's a reasonable way to experiment with the optimum settings.

Some of the diffs, like FREND and AGCCTRL, are fine tuning params that I haven’t had the time to really nail down. I picked ones suggested by the docs, but they have some tuning setups you can run to get them tuned well. Not sure why JBerian picked the ones he did.
the PA_TABLE diff is odd to me. That table is mainly used for doing other types of RF modulation than OOK. If doing OOK, you’re supposed to use PA_TABLE0 and PA_TABLE1, but jberian is using PA_TABLE2 for some reason.

MDMCFG1 and MDMCFG2 are channel spacing. Looks like he uses a bit wider channel spacing than I do.

MDMCFG4 is RX filter bandwidth. You can make it wide to receive data across a wider range of frequencies, at the cost of not picking up frequencies close to ‘center’ as well. mmcommander has it set wider, whereas mine is narrower.

DEVIATN is not used for OOK.

[oskarpearson]

I'm closing this request, as we've found that with certain pumps the original mmcommander settings seem to work better.