nrf52832 scan channel switch time

[loquat]

I use nrf52832 + SoftDevices132_nrf52_3.1.0.

sd_ble_gap_scan_start() run with scan_interval=scan_window=5ms.

I look at scan-channel-switch time is about 1ms. (I analyze current consumption.) It's so many.

I want reduce switch time.

Add, I programed.

static const ble_gap_scan_params_t scan_params = { .active = 0, .use_whitelist = 0, .adv_dir_report = 0, .interval = 0x0008, /* 5ms / .window = 0x0008, / 5ms */ .timeout = 0, }; ... ... uint32_t err_code = sd_ble_gap_scan_start(&scan_params);

[endnode]

Are you sure that peeks in the power consumption curve are really related to switching of Rx channel?

[loquat]

yes.

I can not upload a analyzeing image. But It's cyclic a 5ms interval. and when change interval(10ms, 20ms, etc), interval is changed a setted interval, but switch time is about 1ms.

Also, I measure the recieve counts. interval:20ms 1500counts/minutes interval 5ms 450counts/minutes

Other case, I compare againest CSR BLE chip. CSR BLE chip is 1000countes/minutes on 5ms scan interval.

Is nrf52832 poor againest CSR BLE chip?

[endnode]

No, there are no signs that CSR chips would be superior to nRF52. I can only say that I've reached resolution of <2ms with nRF51 (meaning it registered >500 adv. packets per second) and I'm pretty sure that S132 stack on nRF52 should be in range of 500-1500 adv. events per second (at least). Note that during DTM chip is able to see one packet per 0.625ms without single failure. I'm afraid it's in your measurement set-up not in the chip (nor in the stack).

[loquat]

Thanks for comment.

I know that nRF52 recieve over 1000-adv per sec. But, I measured it on 100ms scan-interval. If change scan-interval to 5ms, how many receive advs?

I seem it to poor receiver.

[loquat]

I tryed by TimeSlot API for scan.

result On 5ms channel switch interval, switch time is very fast(<200us)!!

I think, problem is SoftDevice's sd_ble_gap_scan_start().