Dear Matthew and Pishan,

I looked at the signal from channel 14 in your archive M1236949301. There was no signal from No14. You are not receiving bad messages at any significant rate in your lab. Your interference does not have the correct modulation frequency to appear like one of our transmitter messages.

For more about bad messages see here:

http://www.opensourceinstruments.com/Electronics/A3017/SCT.html#Bad%20Messages

Instead, the interference in your lab is corrupting messages from the transmitters. Each transmitter appears to get one corruption every two seconds on average. Compare this to my basement lab, were I get less than one corrupted message every hundred seconds. It is possible for our transmitters to interfere with one another, but unlikely, because we designed the receiver to be proof against such intereference. The cause of your corrupted messages is some other source of 900 MHz power. It could be a 900 MHz cordless phone, some 900 MHz wireless data acquisition system, or 868 MHz and 940 MHz cell phones.

For more about corrupted messages see here:

http://www.opensourceinstruments.com/Electronics/A3017/SCT.html#Corrupted%20Messages

To corrupt a message, the interference must corrupt the bit sequence in the message so that our error-checking is defeated. My guess is that we are losing a hundred messages for every message that is corrupted but not rejected.

The system appears to work okay with the antenna centered upon two cages, but it no longer works with the antenna centered upon the four cages.

I will plan to boost the transmitter power in the next design. A factor of 10 boost will give a factor of 3 increase in range. If necessary, I'll design a new receiver that has four independent antenna inputs, and which receives messages through all antennas, rejecting duplicates and storing the unique messages in memory.

For now, I sent a new antenna, cable, and a T-junction to Matthew for Pishan to use as a second antenna. It may be that two antennas each centered upon two cages will give better performance, even though all we're doing is combining the signals at the receiver. When I try the same arrangement in my lab, I don't get much improvement, but I don't see any loss of performance either, so it does no harm.

There is also the faraday cage option, rejecting interference. Perhaps we should try that, although it seems a lot of work. We would not have to reject all interference power, only 90% of it.


Yours, Kevan