BLE advertising and cellphone scanning

[UniWindsor]

Hi, I have a general question for BLE advertising, if several BLE peripherals broadcast simultaneousl(the advertising packets contain information the peripherals want to send out), and the cellphones scan for these advertising packets to get the information. Let's say these peripherals just broadcast for a while. My question is: is the above method a stable way of transmitting information from peripherals to cellphone ?( my concern is several BLE peripherals broadcasting simultaneously which may cause interference)

[John]

I believe the general answer to your question is yes, it is a stable way of transmitting information. It would be highly unlikely for the advertisers to be so perfectly synchronized in their transmissions to consistently interfere with each other. Also, I believe the BLE standards have them jitter their timing by a random value from 0 to 10mS each time they want to transmit to prevent something like this from occurring.

I know that we have hundreds of devices within range of my desk here at work and both our mobile app and my BLE sniffer both have no problem seeing the advertising packets even with them all using the same basic transmission pattern.

[UniWindsor]

How about these peripherals just broadcasting for a short period of time(say less than 1 s)? I tested with 4 beacons with my smartphone, it turned out to be that sometimes the smartphone couldn't receive all the advertising packets.

[butch]

Yes, I have the same question: what is the shortest time you can advertise (and stay awake) and with high probability a scanning central (say cell phone) find you and make a connection to transfer information to you? My application is a little different, since I can assume a quiet environment where there are many peripherals, but they don't all advertise at the same time (the central stays awake, the peripherals periodically wake, advertise, and receive writes to a characteristic.) (If I am hijacking your thread, I will start another.)

[UniWindsor]

I think broadcast in time division is a solution. I tested with TI CC1350STK, a broadcast in about 0.1 s has the high probability of being found by the central device(cellphone)，that's to say, I let 4 devices broadcast one by one with each broadcasting for 100ms, I think 400ms is OK for me

[butch]

That is essentially what I am doing. I have many peripherals, they sleep most of the time, they randomly choose whether to wake, advertise, and accept a write to a characteristic. If they get a write, they spread the value among themselves using another protocol.