Smart Bluetooth Forum Discussions

I have a requirement to change the GATT data base on the fly. Has service changed indication mechanism tested in bcm20737? If so, which platform and can you share the code snippet to achieve that?

Hi,

One weird finding on using GPIO pin as LED control.

The symptom is found at using P27/PWM1 or P14/PWM2 as GPIO function for driving the LED.

The GPIO pin state won't stay at low (to turn on the LED) statically while we programmed it to be low from API:

#define GPIO_PIN_LED1 14

#define GPIO_SETTINGS_PIN_LED1 (GPIO_OUTPUT | GPIO_INIT_HIGH | GPIO_LED)

....

gpio_configurePin (GPIO_PIN_LED1 / 16, GPIO_PIN_LED1 % 16, GPIO_OUTPUT_ENABLE, 0);

  It behaves as a short-period 'low' then back to high. So the LED will be just a short and dim blinking in every call of API gpio_configurePin (we configure it in every second).

  But we don't see this issue while changing the PIN to P13/PWM3.

HW setting:

VDDIO: 1.8v (also the same voltage level for LED)

Resistor for LED: 100 ohm

  Any comment?

how can i use to i2c between bmi160 sensor and BCM20732(EV Board)?

Development Environment: WICED-Smart-SDK 1.1

BMI 160 connection
- Power : 1.8V power line connected to VDD, VDDIO of bmi160 sensor
- I2C :  R5, R4 connection
- Interrupt pin : P25, P24 connection

I was refer to i2c_temperature_sensor simple source and "WICED-Smart-Hardware-Interfaces.pdf"  for bmi160 sensor I2C implementation.
I ues cfa_bsc_OpExtended function for I2C.
I saw the message(No slave device with this address exists on the I2C bus.).
bmi160 sensor i2c address is 0x68.
Our hw team say that sensor connection is no problem.
but i can not use cfa_bsc_OpEtended function for i2c.
cfa_bsc_OpEtended function normally operated when i use I2C address of EEPROM(0xA1,0xA0).
I need  additional work for I2C when I added device(bmi160 sensor)?

We could not find API for current SPIFFY 2 pin configuration. Please help below

Application uses SPIFFY 2 only (not UART) – in slave mode. Could we use pin configuration (CS/SCLK/MOSI/MISO – P32/P3/P4/P25) as below:

Hello. Thank you always for your help.

I was sent to 236byte the length of characteristic. BCM20732(SDK-1.1.0) transmission performance.

When I check in Android, it takes about 10 seconds, the transfer is complete. 236byte / 10 => 23byte => 184bit .

It's too late to transfer.

According to BCM20732(SDK-1.1.0) transmission performance.The maximum throughput is 300kbits/s around.

How can make it around the 300kbits/s?

If possible the other way, let me know how possible to 300kbits/s.

Hi,

  Is there any way for scanner to know the channel that one advertising packet is coming from leAdv callback?

According to speed_test example, it is possible that it is transferring data to fine_timer of 12ms.

I saw in the discussion related to the previous one. Re: Maximum BLE throughput for reliable data transfer

"So at the app layer, if you send 20 bytes / packet, assume an average of 4 connection events / connection interval and 133 intervals (7.5mS) / second, you will get ~ 85Kbps."

In this article, what the meaning of "average of 4 connection events"?

If the calculations, 20 * 4 * 133 * 8 = 85120bits -> about 85kbits / s.

Is the means by 20byte simultaneously transmit four times?

If this is correct, how do this?

If not, let me know how.

In speed_test is confirmed by sending once every 12ms.

According to "connection interval and 133 intervals (7.5mS) / second", that the transfer once every 7.5ms. This is possible?

I do not know too much about the ble.

Thanks you for your cooperation.

If there is an error that is returned from the write callback where does that actually end up?

From client persepctive is this just seen as an abrupt error?

Case in point if it is a requirement to start indications on a handle and someone attempts to write to the handle but client side has not yet started indications how can that error be propagated back to the client? We cant send an indication to the client for this handle because the very mere fact that indications have not yet been setup ...

So, is there something in the stack itself that handles these types of errors that client can trap?

Hi Sir,

Is there any limitation of the service quantity and characteristics quantity?

Besides, can I create 2 or more notify characteristics in same service?

Thanks,

When using the bleprofile_WriteNVRAM write command, I can only write to 28 NVRAM Ids

Example:

If I start writing at ID 0x10, my last valid write is at location 0x2b.  I can't write out past that NVRAM ID. ID 0x2c Fails.

If I start writing at ID 0x30, my last valid write is at location 0x4b.  I can't write out past that NVRAM ID. ID 0x4c Fails.

Is this a limitation?

Is there a setting that allows me to write all IDs between 0x00 and 0x6F?

Thanks,

Mike