09-21-2022 10:59 AM
I want to use the BMA456 for anymotion detection. I'm using the config file from the bma456_an.c of the SensorApi on GitHub.
But I don't get any interrupts (neither at the INT1-pin nor in the INT_STATUS Reg). Here is what I do:
busy_wait(TBW_1us(450));//wait for 450us
bma456WriteReg(0x59, 0x00); //Adr. 0x59: INIT_CTRL
(void)i2cWrite(BMA456_SLAVEADR_WRITE, (uint8*)bma456an_config_file, sizeof(bma456an_config_file));
bma456WriteReg(0x59, 0x01); //Adr. 0x59: INIT_CTRL
busy_wait(TBW_1ms(140));//wait for 140ms
bma456WriteReg(0x41, 0x00);//Adr. 0x41: range
bma456WriteReg(0x53, 0x0a);//Adr. 0x53: INT1_IO_CTRL
bma456WriteReg(0x55, 0x01);//Adr. 0x55: INT_LATCH
bma456WriteReg(0x56, 0x10);//Adr. 0x56: INT1_MAP
bma456WriteReg(0x28, 0x38);//Adr. 0x28: FEAT_EN_1
bma456WriteReg(0x7D, 0x04);//Adr. 0x7D: acc_en
//read reg 0x2A, bit4 == 1? init_ok
adr = 0x2A;
(void)i2cWriteRead(BMA456_SLAVEADR, &adr, 1, buf, 1);
printf("bma456 init: 0x%x", buf & 0x1F);
//read anymotion config
adr = 0x5E;
(void)i2cWriteRead(BMA456_SLAVEADR, &adr, 1, buf, 4);
printf("bma456 config: 0x%x, 0x%x, 0x%x, 0x%x", buf, buf, buf, buf);
//write anymotion config
buf = 0x5E;
buf = 0x0A;
buf = 0x00;
buf = 0x05;
buf = 0xE0;//axes enable
(void)i2cWrite(BMA456_SLAVEADR, buf, 5);
reading Register INTERNAT_STATUS returns 0x01 (ASIC initialized)
Reading the anymotion config returns 0xc8, 0x2e, 0x00, 0x2e. This should be default values 0xcc, 0x00, 0x05, 0x00 according to Application note "Hearables feature set"
Solved! Go to Solution.
09-21-2022 02:55 PM - edited 09-21-2022 04:27 PM
ahh, what I read from Adr. 0x5e is what I've been written to it: it is the first 4 bytes of the configfile: 0xc8 0x2e 0x00 0x2e
How does it work? It is the same adress to write the configfile to and to read the feature configuration from. It is not the same memory, is it? Is the separation done by writing the 1 in reg. 0x59?
The configfiles begin with 0xc8 0x2e 0x00 0x2e 0xc8 0x2e 0x00 0x2e 0xc8 0x2e 0x00 0x2e 0xc8 0x2e 0x00 0x2e
the feature configuation differs from that...
09-27-2022 10:44 AM
If you can't write into 0x5e it's likely that the configuration phase went wrong. You should check the status from 0x2a (should be 1).
Burst writing the configuration data requires that the I2C buffer is large enough. You can write it in chunks following the method that's used in the API (see stream_transfer_write in bma4.c). The trick is to write an offset value into reg 0x5b and 0x5c before sending the chunk's data (set 0x5b to (offset>>1)&0xf and 0x5c to offset >> 5).
If the status register looks OK, you should be able to write into 0x5e. I also had to add a soft reset (write 0xb6 into reg 0x7e) in the beginning to make sure the chip gets reset when the MCU restarts.
09-27-2022 12:04 PM
09-27-2022 04:36 PM
thank You for your reply. The first problem was, that I didn't reset the BMA456 after downloading a new software to my target. I work with a debugger and make a target reset after downloading new code, but of course that didn't trigger a power up reset to the BMA456. After fixing this, the reading of the configuration worked.
But still no anymotion interrupt!
Then I enabled both the anymotion and the nomotion interrupt... and: there was an interrupt when moving the board!
Disabling nomotion again: no interrupt
Enabling nomotion and disabling anymotion: interrupt when moving the board.
Conclusion: I have to use the nomotion interrupt to get an interrupt when the bord is moved? Can anyone confirm this?