Hi, we are designing a factory acceptance test (FAT) jig for a custom board containing an NRF9160 SiP.
We are testing more than one thousand boards, and probably more than ten thousand.
The main problem we face is the amount of time it takes to program the modem firmware of a signle NRF9160.
It takes around 30 seconds which is severly eating into our time budget. We are hoping that it is possible to
cut this time in half.
We are currently programming the device using a Segger jlink BASE with nrfjprog using the following command:
nrfjprog --snr <id> --program firmware/mfw_nrf9160_1.3.6.zip --verify
Will upgrading to a Segger jlink PLUS improve the programming speed?
I have seen these forum posts
Can we merge nRF91 app and modem hex to speed up production flasing.
and
Speed up modem firmware and application software programming
and they are both linking to a dead web page. However, I found the following application note on production programming of the NRF9160:
https://infocenter.nordicsemi.com/pdf/nan_041.pdf
Under "4.4.1 Recommended method" it mentiones "This method uses a double-buffered approach to increase the speed of the operation." Can you confirm or disprove that this method is the one that is actually implemented in nrfjprog ? Could we achieve a speed improvement by implementing this method ourselves?
We tried increasing the programming clock with the optional nrfjprog flag -c, but it does not affect the total time. When scoping the clock of the jlink programmer it seems that there is a fixed 10ms delay between some of the packets which dominates the total time. What is the cause of this delay? Can it be avoided?
Best regards,
Christian
