FSK optimal Deviation

Hi its been a while since i have done this much RF work..

Here is my problem,

i have a FSk transmitter on the 433 ISM band transmitting at +17 dbm with a datarate of 2.4kps. Using a 1.5ppm Tcxo for frequency acuracy.

My Reciver is a narrow band with an AF Bandwidth(-3db) of 4 KHZ, and a sensitivity of -120db SINBAD .

I was using a deviation of +- 3 Khz and while this works ok , i am trying to optimise the range of this link. I have been told that this shouldnt work as my theoretical bandwidth is 8.4khz yet the reciver is only 4khz.

Though the way i understood the theoretical calaculations ( i had to look that one up - quite confussing) was 2(bitrate + Deviation) = 10.8 khz..

so here is the question what is an optimal deviation ? what is the problem if the deviation becomes to small? I can live with 1.2kbs datarate but should i then have a deviation of 1 khz i.e a theoretical bandwidth of 4.4khz or am i missing something all together?

Narrower bandwidth of the receiver after the demodulator (on audio side) doesn't affect the SNR of the receiver, so doesn't improve the link budget.
If you want to improve the link, have to narrow the bandwidth before the demodulator.

If the deviation becomes too small you will see a very small change in the output of the phase detector.

I see your from Australia. Not WA by chance?