I noticed that the wiced sense stops sending advertising sometimes after one or two seconds and do always complete the advertising durations of 10 seconds + 10 seconds.
Note that there is no connection happening, as the connection stops the gpio interrupt and stops the led from blinking and make a buzz.
The led keeps blinking, without packets sent, once it stops after 10+10 seconds, the wake up wakes it and another test of counting advertised packets start.
This is a reminder and also to chekc if I have something wrong.
On wake up, the wiced sense tag will be sending advertisement at high rate for one duration, then at a lower rate for another duration.
in the "wiced_sense_cfg" table:
- high_undirect_adv_interval : 32 => 32 x 625 us => one every 20 ms
- high_undirect_adv_duration : 10 => 10 seconds, so we would expect 500 advertisements during this period
- low_undirect_adv_interval : 1024 => one every 640 ms
- low_undirect_adv_duration : 10 seconds, so we would expect about 15
Issue description and logs:
1) High at 32 slots
with 20 ms delay, the advertisements are very unsteady, and stop sometimes much sooner than expected. So for 500 packets expected, I made multiple retries and here is the list of how many I get every time :
|Number of Adv Packets received on every test|
Note that the advertisements were interrupted even after the run following the build without battery, so totally powered from usb, so I don't think this is a battery issue.
2) High at 160 slots
with 160 slots which take us up to 100 ms, the advertising is much more stable, but sometimes (once every two to three tests), I still get interruptions and no more advertising.
The attached file adv_160_Good.txt, shows the good case, where no problems happen, just to confirm the numbers
The attached file adv_160_Bad-Stop.txt, simply shows one example of when the advertising was stopped
3) High at 320 slots
everything is working fine
I would totally agree with any one criticizing this test as not using a billion euro official testing equipment, but a simple tool. However, statistically speaking, I made multiple crosss validations, used different devices, the tool might skip some packets, some times, but it is very unlikely that it get 100% of packets till time t, then 0% afterwards. And that is why I am asking for other opinions maybe some others have a better explanation.
These deep tests were lead with a modified source, but a similar behavior was observed with all binaries, I hereby would ask if any binary can be used for reference. If I use an out of the box tag (I do have one), maybe someone will ask, no do the update, which update ? My latest tests included wiced_sense.c.From_WICED_Sense_SDK 2.1_1.3Final and wiced_sense.c.From_WICED-Smart-SDK-2.2.1
Thank you already for reading so far, your comments are appreciated.