Hi,
In my application I use 5 devices and 1 host to communicate. I have only a few packets to transfer so the data rate can be low (22Byte/transmission). The devices wake up between 1x/day and 2x/seconds (that's possible, normally it is about every 10 seconds). So I decide to use the gazell protocol with following setting:
init_ok &= nrf_gzll_set_max_tx_attempts(7);
init_ok &= nrf_gzll_set_xosc_ctl(NRF_GZLL_XOSC_CTL_AUTO);
init_ok &= nrf_gzll_set_tx_power(NRF_GZLL_TX_POWER_4_DBM);
init_ok &=nrf_gzll_set_device_channel_selection_policy(NRF_GZLL_DEVICE_CHANNEL_SELECTION_POLICY_USE_CURRENT);
init_ok &= nrf_gzll_set_datarate(NRF_GZLL_DATARATE_250KBIT);
init_ok &= nrf_gzll_set_timeslot_period(2700);
init_ok &= nrf_gzll_set_sync_lifetime(0);
in the best case we got this (no retransmission; R_shunt=20Ohm):
and in the worst case this (7 retransmission; R_shunt=20Ohm; peak->15mA):
-
So my first question is: Are they parameters from the gazell protocol optimal, or is there a way to save some power? Especially the timeslot period, retransmission rate, xosc, tx_power..... I know I can change the datarate and the timeslot to 2MBit and change to TX_Power to a lower level, but then the range is not so good. So is there an optimisation from nordic (or someone else) which parameters give the best compromise between range and power?
-
I must reduce the burst current, because of a large voltage drop in the power supply. So I want to stabilize the power supply with a capacitor. But if I want to reduce the voltage drop to 100mV with a average worst case peak current from 15mA for 20ms (worst case) I have to add a 3mF capacitor which is to large for my application. So I try to get a short idle-time between the retransmission peaks to recharge the capacitor. If we do so we only got 3ms to stabilize the supply voltage and then we only need a 450nF capacitor which is much smaller. So is there a way to reduce the current between the retransmissions? Or if you have a better solution I would also be happy:)
Thanks for your help.
ps: there is the try to increase the timeslot (no good idea,..:) R=18Ohm)
