Changing ble_app_hids_mouse for trackpad application

[Matthew K]

Hello guys. I'm helping a customer to work with a trackpad application using the nRF52840 and the nRF5_SDK_17.0.2_d674dde.

For this customer, she wants to use Microchip's touchpad solution (like https://www.microchip.com/en-us/solutions/machine-learning/smart-human-machine-interface)

and add nRF for BLE HID. Since the rep_map_data[] already has the mouse buttons, scroll/pan, mouse motion, advanced buttons (play pause volume down up etc),

I thought I simply need to call mouse_movement_send() - ble_hids_inp_rep_send(INPUT_REP_MOVEMENT_INDEX, INPUT_REP_MOVEMENT_LEN), using the trackpad as the parameter input.

First, for trackpads, do I need to change the rep_map_data[] or can I use it as it is?

We want to create a simple trackpad, like a mouse with volume control gestures, no keys/keyboard.

I was wondering whether a trackpad like this needs additional report map data during the initialization in hids_init().

Next, is it true that DIS - PnP characteristic is mandatory for wireless trackpad/mouse/keyboard applications?

Also, regarding DIS PnP,

PNP_ID_VENDOR_ID_SOURCE = 0x02 /**< Vendor ID Source. */
PNP_ID_VENDOR_ID = 0x1915 /**< Vendor ID. */

am I allowed to use these DIS PNP vendor ID values for this nRF product?

I'm slightly confused with this Nordic thread here  nRF52840 - USB Vendor ID and USB Product ID 

so I was wondering whether I have to register my own vendor ID.

 

Lastly, I want to check how to send a volume key. Is it ble_hids_inp_rep_send(INPUT_REP_MPLAYER_INDEX, INPUT_REP_MEDIA_PLAYER_LEN) ?

I read that I have to send a "mask" value to stop increasing or decreasing the volume so I need your help how to do this properly.

Thank you so much for your help!

REFERENCEs

https://www.bluetooth.com/specifications/specs/hid-over-gatt-profile-1-0/

https://www.usb.org/sites/default/files/hid1_11.pdf

[Edvin]

Hello,

This sounds like a solution that would require a driver from Microchip. I think you should reach out to them to ask what usb report maps they intended to be used with the trackpad that you are using.

If they don't have anything to share, this is a more generic USB question than a Nordic Specific. Trust me, if I knew exactly how you needed to set up this report map, I would tell you, but I think you need to look at how report maps are built up and how they are used. 

Best regards,

Edvin

[Matthew K]

Thank you for replying Edvin. I think the USB HID spec is the document that describes how report maps are built.

Let me search for the Microchip driver.

Come to think of it, if I use Microchip’s USB HID report map for this BLE trackpad application, do I have to use Microchip’s vendor ID for the DIS PnP characteristic?

[Edvin]

Matthew K said:

Is it true that DIS - PnP characteristic is mandatory for wireless trackpad/mouse/keyboard applications?

I believe it is, yes. But please verify with USB.

Matthew K said:

I read that I have to send a "mask" value to stop increasing or decreasing the volume so I need your help how to do this properly.

Where did you read that?

BR,
Edvin

[Matthew K]

Sorry I accidentally pressed VERIFY ANSWER. Here's the Nordic Devzone Thread 8 years ago

 RE: HID keyboard example - What does the buffer do and why it needs to be dequeued? 

There's a phrase

"After that I send 0 mask at BLE_EVT_TX_COMPLETE event and I did not call buffer_dequeue."

May I ask what this is about?

[Edvin]

Hello,

I am not sure what the customer from the phrase meant by 0 mask. Perhaps it is a HID message saying that no keys are pressed. If you don't do it, it will look like the user is holding the last keypress forever.

BR,
Edvin

[Matthew K]

Hi Edvin. Yes, I think that's what it meant by 0 mask as you have discribed.

I can't find the code doing this 0 masking from the ble_app_hids_mouse (SDK v17) example.

May I ask for your help how to send that "no keys are pressed" HID message?

[Edvin]

Hello,

In the ble_app_hids_keyboard sample, you can see how they do it. They use the function hid_kbd_state_key_clear() to set the internal state of the key to 0 (not pressed), and then it actually sends the keyboard state using key_report_con_send() -> bt_hids_in_rep_send().

BR,

Edvin

[Matthew K]

Thank you Edvin. Then do you mean the mouse example is not sending 0 (not pressed)?

Shouldn't there be a code for sending 0 in the mouse example?

[Edvin]

Oh, sorry. Yes. Well, this sample only shows moving the mouse. So each packet consists of a move in x,y directions. Mouses only send changes in position, and not speed, so there is no need to send a message saying it is no longer moving. If you do mouse clicking (left mouse button and right mouse button, for instance), then you need to send a new message with the button no longer being pressed. If not, it will think that the mouse button is being held down forever.

I don't remember the values for mouse button clicks, but you can use bt_hids_inp_rep_send, with rep_index = INPUT_REP_BUTTONS_INDEX (= 0)

[Matthew K]

Thank you Edvin. Then back to my original questions,

I want to check how to send a volume key in the mouse example.

Is it

uint8_t media_buffer[INPUT_REP_MEDIA_PLAYER_LEN] = {0xEA} // from rep_map_data[] Volume Down ?

ble_hids_inp_rep_send(&m_hids, INPUT_REP_MPLAYER_INDEX, INPUT_REP_MEDIA_PLAYER_LEN,

media_buffer, m_conn_handle) ?

I wasn't sure what value should I send for Volume Down.

Back to your Microchip touchpad driver part,

https://github.com/MikroElektronika/mikrosdk_click_v2/blob/master/clicks/touchpad/lib/src/touchpad.c

just like this code, the drivers I found were SPI drivers.

The input on the touchpad was sent to a microcontroller using SPI.

In that case, I can simply use that SPI data to map the mouse data (x_delta and y_delta) and

send it using

ble_hids_inp_rep_send(INPUT_REP_MOVEMENT_INDEX, INPUT_REP_MOVEMENT_LEN)

right?

Or do I still need to modify the rep_map_data[] in this case?

Thank you so much for your help!

[Edvin]

So what do you see when you apply the SPI driver? I assume it will be able to provide you with x and y coordinates, but is it supposed to handle double touch? How about clicks? Does it handle that at all, or is that something that you need to implement on your own?

BR,

Edvin

[Matthew K]

Hi Edvin. Yes, the SPI driver handles:

x y coordinates - touchpad_get_touch_coordinate()

read single touch / click - touchpad_get_event_state(), I need to implement double touch on my own

gesture (zoom in/out, two finger touch for mouse right click) - touchpad_get_gesture_type()

Is a double touch something similar to a mouse double click where the HID rep_map_data[] has to change?

Thanks for your help