2 antennas, destructive interference

[othla]

I'm experimenting with two radios (IEEE 802.15.4 .. cf = 2.475GHz, bw = 5MHz), transmitting the same data on the same channel. Both radios have lambda/4 antennas, vertically polarized, 2dBi gain. I'm looking at the received power and decoding on the receiver end (through a mixed domain oscilloscope).

When the antennas are placed far off (~10-20 cm), I see power in the channel increases (~6dBm) and can decode well. Understandable.

But when they are placed one touching the other, then power in the channel decreases, and received symbols are corrupted! Can somebody explain the phenomenon here, please?

 

[othla]

To add ..

I'm experimenting with two radios (IEEE 802.15.4 .. cf = 2.475GHz, bw = 5MHz), transmitting the same data on the same channel. Both radios have lambda/4 antennas, vertically polarized, 2dBi gain. I'm looking at the received power and decoding on the receiver end (through a mixed domain oscilloscope).

When the antennas are placed far off (~10-20 cm), I see power in the channel increases (~6dBm) and can decode well. Understandable.

But when they are placed one touching the other, then power in the channel decreases, and received symbols are corrupted! Can somebody explain the phenomenon here, please?

I've ensured that the two nodes are synchronized at nanosecond level by external power source. So the waves on the channel must almost be the same, except for a little phase difference.

 

[zz0468]

Can somebody explain the phenomenon here, please?

Not with the question as stated.

Do the radios have two antennas each that you're changing the separation, or are you changing the separation between the two radios?

At 2475 MHz, one wavelength is approximately 12 cm. Dipoles (or monopoles with ground plane) spaced one wavelength apart and fed in proper phase will exhibit some gain and directivity. Placed at the wrong spacing, they could exhibit almost complete field cancellation.

 

[othla]

We did try changing the separation for various distances, and we got the results almost as you guessed it at lambda/2. It does not totally cancel out since the two nodes will not transmit with the same phase (in theory, they should and at lambda/2 the phases are 180 degree out of phase). But there were lot of bit errors.

The interesting question was to see what would happen if all antennas were placed together. In this case, we found power on the channel decreased and there were lot of bit errors, and also many packets were lost. I don't know why ... I suspect it is something to do with the antenna and some physics behind it ...

 

[SCPD]

Have you thought about mounting one antenna directly above another? We used to do this with dirty cheap repeaters, and had good success. I'm gonna make a lame attempt at making a picture:

I_I
I
I_I
I
I

Not too bad. Basically one vertical directly under another. Like I said, I've seen setups like this, without a duplexer, and the output not getting into the input on a repeater.