Problem connecting to peripheral

[sandeepmistry]

Problem connecting to peripheral

[freber]

Anyone successfully connected to a peripheral from MacOS 10.8 or 10.9? I've tried everything at this point. Maybe it's something about the connection window being to short? Can it be configured?

Best regards,
Fredrik

[sandeepmistry]

@freber I've got it working with 10.9 using LightBlue. Sometimes it does take a few attempts to connect, I'm not sure if it's a connection interval thing or just the way OS X behaves.

Connection interval could be configure if nRF Go Studio exposes it, we would need to figure out which bytes in the setup message control it.

[freber]

Interesting, I couldn't get it to connect in LightBlue, I can only see it advertise.

I will try to diff the service.h file with a change in the connection interval to see where that goes in the configuration.

Also, I couldn't strip down the working profile more than 11 config messages. I can see that you managed to get it down to 7.

Best regards,
Fredrik

[freber]

I manged to diff the service.h files from a few different configurations now.
It looks like the connection interval is set at 2 places.

baseSetupMsgs[8] and baseSetupMsgs[last]

diff services_min1000ms_max4000ms.h services_min10ms_max30ms.h

----- start -------------------------------------------------------------------
77,78c77,78
< #define GAP_PPCP_MAX_CONN_INT 0xc80 /**< Maximum connection interval as a multiple of 1.25 msec , 0xFFFF means no specific value requested */
< #define GAP_PPCP_MIN_CONN_INT  0x320 /**< Minimum connection interval as a multiple of 1.25 msec , 0xFFFF means no specific value requested */
---
> #define GAP_PPCP_MAX_CONN_INT 0x18 /**< Maximum connection interval as a multiple of 1.25 msec , 0xFFFF means no specific value requested */
> #define GAP_PPCP_MIN_CONN_INT  0x8 /**< Minimum connection interval as a multiple of 1.25 msec , 0xFFFF means no specific value requested */
133c133
<0x04,0x01,0x20,0x03,0x80,0x0c,0x00,0x00,0xff,0xff,0x04,0x04,\
---
>0x04,0x01,0x08,0x00,0x18,0x00,0x00,0x00,0xff,0xff,0x04,0x04,\
267c267
<0x06,0x06,0xf0,0x00,0x03,0xe5,0xf0,\
---
>0x06,0x06,0xf0,0x00,0x03,0xe7,0x42,\

----- end --------------------------------------------------------------------

I wonder why it's set in 2 different places. Also, I actually wonder what are reasonable values for these two configuration parameters together with the Connection supervision timeout ;)

Regards,
Fredrik

[freber]

I've updated the syntax markdown now..

[sandeepmistry]

Just curious which nRF8001 device and Arduino are you using?

[freber]

I'm using a Arduino Pro Mini (3.3v) and a Bluefruit LE (http://www.adafruit.com/products/1697)

[sandeepmistry]

One more thing to try out is to clear the CoreBluetooth/blued cache, I wonder if it's expecting the extra characteristics nRFgo studio uses (which I removed). See: http://stackoverflow.com/questions/23549859/clearing-corebluetooth-gatt-cache-without-removing-bond

[freber]

Sorry for late reply.

I've tried to clear the cache. Still no progress. I've pulled from the git repo today and is trying the starter.ino together with LightBlue. I've mailed Punch Through Design (makers of LightBlue) and asked what intervals they are using.

They use 20 milliseconds for connection interval and 5.9 seconds for supervision.

Best regards,
Fredrik

[sandeepmistry]

@freber do you have another Arduino like the Uno or Leonardo to test with? I don't have a Arduino Pro Mini (3.3v) to try out the lib with ...

You could also try manually modifying the setup messages (with the diffs from) earlier in nRF800.cpp

[sandeepmistry]

@freber debug output would also be useful, uncomment: https://github.com/sandeepmistry/arduino-BLEPeripheral/blob/master/nRF8001.cpp#L13

Also, Apple's Hardware IO tools has a BLE packet inspector, the trace from that would be useful as well ... http://iosdevblog.com/tag/hardware-io-tools-for-xcode/

[sandeepmistry]

@freber closing this for now, please re-open with the request debug info ...