nRF Sniffer wireshark bad interval connection

RE: nRF Sniffer wireshark bad interval connection

MPS — Thu, 07 Dec 2017 13:45:01 GMT

Yes, but I don't understand why the event counter is increased. I'm going to search this point. Thanks for your help ... I think I'll have another question to ask ... I'm starting.

RE: nRF Sniffer wireshark bad interval connection

David Edwin — Thu, 07 Dec 2017 13:36:22 GMT

See answer below, the sniffer is capturing exactly as the connection exists, i.e. there is no master transmit for the event counters that are not seen in the sniffer as the scanner is running at that time.

RE: nRF Sniffer wireshark bad interval connection

David Edwin — Thu, 07 Dec 2017 13:09:50 GMT

ok, since multiple roles are in the central I think the connection events are being skipped so that it can scan for other devices and occassionally the scanner need to use the radio. This is confirmed by the fact that there is no packet loss (as the SN, NESN are in sync). I think the sniffer is capturing exactly as it happens.

Edit: The event counter is used to see the time line, so even if a event occured and a tx did not happen on the master side, the counter has to be incremented. This is generated by the sniffer and is not on the air.

RE: nRF Sniffer wireshark bad interval connection

MPS — Thu, 07 Dec 2017 12:53:56 GMT

I use ble_app_multilink_central (Central) and ble_app_blinky (Peripheral) with S132. I don't know if they use timeslot API. I think it's time used by scanning ... because if I reduce at only 1 peripheral to find, now I have always ~20ms ...

RE: nRF Sniffer wireshark bad interval connection

David Edwin — Thu, 07 Dec 2017 12:11:34 GMT

What are you running in the 2 nRF52-DK, is it an example from the SDK or other custom software, Are you using the timeslot API in the code ?

RE: nRF Sniffer wireshark bad interval connection

MPS — Wed, 06 Dec 2017 17:08:59 GMT

Yes, sniffer is between the 2 nRF52 DK (5-10cm on each side). Another idea for this problem ? I use "JLink 6.16c" with "sniffer_pca10031_51296aa.hex" and wireshark 2.4.2.0.

Thanks for the time explanation. I know what time to use now.

Many thanks for your support.

RE: nRF Sniffer wireshark bad interval connection

Sigurd — Wed, 06 Dec 2017 16:36:26 GMT

Are you placing the sniffer between the devices you are sniffing, as shown at page 3 in the sniffer user guide?

Regarding your question on the "Time column", this is explained in this post, quote:

The delta time ("end to start") is the one that is reliable. The "start to start" is not correct one. The time stamp on the TIME column also not reliable since it's the time stamp on UART on PC.

RE: nRF Sniffer wireshark bad interval connection

MPS — Wed, 06 Dec 2017 13:34:12 GMT

Thanks for your answer. Do you know why sniffer misses some packets. It's strange because it's always the same cycle : 3 received, 2 lost ... There is no correlation between the column "Time" with the reality ? Sorry but I started with nRF and this sniffer.

RE: nRF Sniffer wireshark bad interval connection

Sigurd — Wed, 06 Dec 2017 12:57:23 GMT

Hi,

I don't really see anything very wrong here.

You can see from the delta time in the log, that the master send a new packet every ~20 ms, and that the slaves respond after 150 µs. Looking at the event counter, you can see that the sniffer sometimes misses some packets, and in that case the delta times will increase.

RE: nRF Sniffer wireshark bad interval connection

MPS — Wed, 06 Dec 2017 12:38:13 GMT

Hi,

You can find in attached file the sniffer trace. Thanks in advance.

Best regards

sniffer.pcapng

RE: nRF Sniffer wireshark bad interval connection

Sigurd — Wed, 06 Dec 2017 12:09:28 GMT

Hi,

Could you save the sniffer trace(.pcapng file) and upload it here ?