Radio Timing and PPI

RE: Radio Timing and PPI

Torsten Robitzki — Tue, 11 Jul 2017 08:03:00 GMT

@Petter Yes, sorry, the uploaded hex didn't work. I've attached a new hex file (and tested it before).

RE: Radio Timing and PPI

Torsten Robitzki — Mon, 10 Jul 2017 14:55:31 GMT

That would be awesome if you could ask one of the developers. Currently, this method is the only method available to me to check timings. That method served me well in the past and I would like to understand, where this additional 10us comes from to be able to use the method verify my code. TIA Torsten

RE: Radio Timing and PPI

Petter Myhre — Mon, 10 Jul 2017 14:21:00 GMT

My logic trace is very similar to yours. Around 200 us. I'm not sure why it's not 190 us. I can try to check with a developer if you want me to, but vacation has started. Let me know.

I tried your hex, couldn't get any pin changes on 23 or 25...? :( The Ellisys sniffer shows 190 us though.

RE: Radio Timing and PPI

Torsten Robitzki — Fri, 07 Jul 2017 13:19:16 GMT

@Petter Thanks a lot!!! I've attached a hex file to the original message. The radio activity is routed to pin 23. Pin 25 gives a short pulse, when a scan request is received to be used as trigger. I will also add the code, that I've used to configure PPI and PGIOTE.

RE: Radio Timing and PPI

Petter Myhre — Fri, 07 Jul 2017 13:01:22 GMT

I'm not sure what the mistake is, your reasoning is makes sense to me. I did some testing on our stack with an Ellisys sniffer. For the T_IFS + preamble and access address I got between 190 us and 192 us, never 200 us. If you have a hex file I can try to test your stack as well. I will try to hook up into the ADDRESS and END events and see if I can get a logic trace soon.

RE: Radio Timing and PPI

Torsten Robitzki — Wed, 05 Jul 2017 15:12:56 GMT

Yep: github.com/.../bluetoe. Interesting is, that for the advertisment (the first rectangle) and for the scan request (the second rectangle), this measurement method delivers the correct values. But not for the low pulse (T_IFS) between both rectangles. But I'm quite sure, that T_IFS is correct because I always measure the same values irrespectively of the used GATT Client.

Thanks a lot for having a look into this.

RE: Radio Timing and PPI

Petter Myhre — Wed, 05 Jul 2017 15:04:38 GMT

I need to look into this. You have developed this Bluetooth stack?