Hello TI experts,
We've previously highlighted this issue to local support but wanted to post here as it is now a project show-stopper.
The CC1120 data sheet and other postings in this forum highlight the device's ability to lock the AGC on very short preambles (for example, 4 bits).
We are working on an ad-hoc networking system - meaning that received power varies very significantly - and have so far been unable to get a stable working setup for two CC1120EMK-868-915 using the TrxEB and SmartRF studio, or on our own (broadly similar) hardware.
In fact, we've only so far been able to get good results with the default 32bit preamble, 32bit syncword. At 62ksym/s (we are targeting about 125kbit/s 4GFSK, but preamble and sync are sent with 2GFSK) this takes 1ms, which is hugely wasteful and means that we don't get the required capacity. Unfortunately increasing the channel bandwidth is not a solution due to regulatory limits on channel width.
We've spent a couple of weeks trying to shorten the preamble and sync (making necessary changes to relevant thresholds) but we get severe packet loss at higher receive powers and non-zero packet loss even in clean conditions. We're not helped by the documentation being thin about register meanings at the zero IF, higher data rate modes.
Please can you provide CC1120 settings to give something like:
- 4 bits preamble, 16 or 24 bits syncword
- 125kbit/s modulated (payload) data rate
- 125kHz filter
- 868 or 915MHz
- in the absence of interference, zero packet loss between at least -95dBm and -30dBm received signal strength (note that the RSSI measure saturates at high levels)
- also zero packet loss when two TRXEBs are located approximately 1m apart with the stock antennas and line of sight in a typical building with nearby walls (this seems a pathological case)