Peer manager and chip/SD/SDK select

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Thu, 21 Dec 2017 13:33:03 GMT

ok,I will try tomorrow and go back here .Thanks a lot

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Thu, 21 Dec 2017 13:06:15 GMT

I don't know why your DFU procedure goes faster. It could make sense to call it once, but not every time you get BLE_GAP_EVT_CONN_PARAM_UPDATE the event.

You don't have to set the connection parameters in the peripheral, if you get BLE_GAP_EVT_CONN_PARAM_UPDATE they are already set.

Why would you call sd_ble_gap_ppcp_set()? That would just set Set GAP Peripheral Preferred Connection Parameters.

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Thu, 21 Dec 2017 13:01:37 GMT

Or maybe I should call sd_ble_gap_ppcp_set() instead of sd_ble_gap_conn_param_update() at the BLE_GAP_EVT_CONN_PARAM_UPDATE event?

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Thu, 21 Dec 2017 12:57:04 GMT

HI，Peter, we added sd_ble_gap_conn_param_update() there Since we found that if we call it there our DFU procedure is much faster with some android phone. And in our opinion, by BLE_GAP_EVT_CONN_PARAM_UPDATE event, we can get the connection parameter from smartphone and set them to 51822(it seems work for the DFU). If I were wrong, please correct me ,thanks a lot.

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Thu, 21 Dec 2017 11:27:53 GMT

I missed that. Why exactly are you calling it every time you get BLE_GAP_EVT_CONN_PARAM_UPDATE? That explains why you get so many BLE_GAP_EVT_CONN_PARAM_UPDATE events...

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Thu, 21 Dec 2017 00:48:23 GMT

Hi, Peter, I have searched through my project ,and only found sd_ble_gap_conn_param_update() called in the function which I provided above, is there any other way to initialize a parameter update from the peripheral? thanks a lot.

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Wed, 20 Dec 2017 16:13:36 GMT

I checked.

Are you calling sd_ble_gap_conn_param_update() anywhere? I suspect you do and the central rejects, and you get a BLE_GAP_EVT_CONN_PARAM_UPDATE with the same connection interval. See variant #2 here.

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Wed, 20 Dec 2017 02:01:53 GMT

HI, Peter, I have updated the question above ,please check .Thanks a lot.

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Tue, 19 Dec 2017 12:50:39 GMT

I still don't understand how you got to the conclusion that it is the nRF51 that rejects?

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Tue, 19 Dec 2017 08:39:07 GMT

HI, Peter, I use two different android smart phone,the parameter update requesting after connecting to my 51822 is different, I added debug output for gap_evt.params.conn_param_update.conn_params at the event "BLE_GAP_EVT_CONN_PARAM_UPDATE", the output is like that :

min int 6 max int 6 slave lat 0 timout 2000

min int 39 max int 39 slave lat 0 timout 2000

when another android phone connect ,the print is like that:

min int 6 max int 6 slave lat 0 timout 2000

min int 9 max int 9 slave lat 0 timout 2000

the unit is 1.25ms. if anything wrong, please correct me, thanks a lot

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Tue, 19 Dec 2017 08:27:08 GMT

What do you mean? How do you know the smart phone is requesting 7.5ms? And how is the nRF51 rejecting?

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Mon, 18 Dec 2017 03:21:53 GMT

HI,Peter, I know that smartphone will limit the connection interval for the resource share problem, but in my case,it is the 51822 limit the connection interval, since my smartphone is requesting 7.5ms ,but 51822 refuse the value,so smartphone have to change the value to a quite long one.

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Fri, 15 Dec 2017 10:23:37 GMT

I believe this is on purpose limitation introduced to save battery power. And also because most smart phones have one chip and one antenna that handles BLE, Bluetooth Classic, and Wi-Fi, so they have to share the total radio time.

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Fri, 15 Dec 2017 08:25:05 GMT

I just call sd_ble_gap_conn_param_update with the parameters the event provided to enable the new value, but it seems 7.5ms is not acceptable?

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Fri, 15 Dec 2017 08:22:09 GMT

Yes, I know it all depends on the central, but why 7.5ms can not be set as the final value? Do I have to do something special to enable 7.5ms connection interval?

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Fri, 15 Dec 2017 08:20:24 GMT

So if you get a connection interval outside the min or max, you have three options, send a connection parameter update request, accept the values even though they are outside what you prefer, or disconnect.

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Fri, 15 Dec 2017 08:18:12 GMT

It is the central device that dictates the connection interval. The initial connection interval is decided by the central and is included in the connection request. Then the central can do a connection parameter update to change the connection interval. The peripheral can only request the central to do an connection parameter update.

RE: Peer manager and chip/SD/SDK select

hunterliuyi — Fri, 15 Dec 2017 06:05:35 GMT

Hello, Peter, We have checked that our original connection interval is 20ms for the min and 40ms for the max ,which are the shortest value complying to the requirement of IOS .we found a problem that some android phone will try to update the value to 7.5ms lower than the min value directly, after several times of this trying, they change the value to 48.75ms,exceeding the max range of 40. Some other android phones will change the value to 11.25ms which is also below the min value,and we found that the DFU time of later android phone is much faster then the former phones .So I have questions, why the connection interval can not be set to 7.5ms and 11.25 is valid? and it seems that the original min and max is useless, because the final value can be either larger or smaller than the range,why?

RE: Peer manager and chip/SD/SDK select

Petter Myhre — Thu, 14 Dec 2017 12:47:16 GMT

You can try to reduce the connection interval. See this on how to check what it actually is.

I'm assuming that you are not bonding at all now? You can try to use Device Manager, which Peer Manager replaced. This is available in SDK 10. Check out the ble_app_hrs example. Try setting IS_SRVC_CHANGED_CHARACT_PRESENT to 0 to not include the service changed characteristic.

I'm not sure how easy it is to shrink the Peer Manager. If you just need simple bonding to one device you may be better of implementing your own bonding library, but this would take some work as well.