HID keyboard and central BLE missing keyboard presses

RE: HID keyboard and central BLE missing keyboard presses

Håkon Alseth — Wed, 30 Oct 2019 08:24:27 GMT

Glad to hear that you found a way around the issue!

Kind regards,

Håkon

RE: HID keyboard and central BLE missing keyboard presses

canaengintern — Tue, 29 Oct 2019 23:30:24 GMT

After some trouble shooting and testing I was able to use two single shot timers to space out the time between when "app_usbd_hid_kbd_key_control" gets called in my repeated timer handler. After changing the timing of my single shot timers I was able to use 1ms as the perfect response and one of the fastest times it could execute. I also tried it with 0.5ms but it was to fast and "app_usbd_hid_kbd_key_control" was not ready to receive a keyboard press.

I should also say that the main.c code I posted above has changed a lot as I am no longer using the scheduler.

Thank you Håkon for the help and pointing me in the right direction. I would consider this case closed.

RE: HID keyboard and central BLE missing keyboard presses

canaengintern — Tue, 29 Oct 2019 18:23:51 GMT

Yeah when I was first trying to diagnose this issue it was I thought it might have something to do with nRF_LOG lagging.

Anyways, I'm not exactly sure how to check for NRF_SUCCESS but using VERIFY_SUCCESS(err_code) no errors were returned.

It would make sense that "app_usbd_hid_kbd_key_control" would not be ready to receive a key press if it was busy. So if 4 keyboard presses were to be recorded quickly and since I'm using a for loop to do this i could see it missing a press or too due to "app_usbd_hid_kbd_key_control" not being ready to receive press and the for loop iterating to fast.

I am going to try to work with single shot timers to take some relief off of "app_usbd_hid_kbd_key_control" and only send a keyboard press when it is ready.

RE: HID keyboard and central BLE missing keyboard presses

Håkon Alseth — Tue, 29 Oct 2019 13:01:41 GMT

The throughput over usb is fairly low compared to what it technically can handle (which is why I thought the LOG and CLI took too much time), but it would be interesting to see if your calls to "app_usbd_hid_kbd_key_control" return NRF_SUCCESS, or if they return anything else. Could you check this?

Kind regards,

Håkon

RE: HID keyboard and central BLE missing keyboard presses

canaengintern — Mon, 28 Oct 2019 17:20:21 GMT

Hi Håkon,

Thanks for getting back to me. It does seem like the USBD is being saturated.

I have disabled nrf_cli and nrf_log in my sdk_config file and commented out any reference to them in main. It still seems as though the USBD is still being saturated. I also have abandon the scheduler.

I set a repeated timer up previously to check to see if a keyboard press would need to be sent in hopes the USBD wouldn't be saturated and I could control how much data was being sent. But, no success.

If you have any other suggestions that would be great.

Thanks

RE: HID keyboard and central BLE missing keyboard presses

Håkon Alseth — Mon, 28 Oct 2019 14:41:17 GMT

Hi,

If I understand the scenario correct, it is that if you receive data quickly over BTLE, it will saturate the USBD.

I see that your main loop consists of nrf_cli and nrf_log handling as well as the scheduler. Are you using NRF_LOG, with serial as the backend? If yes, try to disable it and see if this helps the processing.

Same goes for nrf_cli; if you're not using it, try disabling it to see if this has an impact on your overall performance.

Kind regards,

Håkon