FSK transmssion with moving objects

In the attached drawing I show a simple, modern, S5 FSK transceiver chip, say at 2.45 GHz, that is attached to a wheel that is rotating. It is transmitting data while the wheel moves to a fixed receiver, with a similar chip. I know that these chips use digital frequency demodulators that look at the leading edge of the received carrier and determine if the frequency is above (a "1") or below (a "0") the reference.

So, what happens to the frequency demodulator as the wheel rotates? In position 1, the two antennas are in the same plane and in the same phase. Later on, in position 2, the two antennas are suddenly 180 degrees out of phase (ignoring the path length change), so the carrier must make a sudden 180 degree jump. Then a moment later, it is back to position 1, and the carrier again is in-phase.

Does all this momentary carrier phase shifting cause the frequency demodulator to lose its mind? Or does it just coast over this sort of stuff as background noise.

~ 1 mbps data rate at 2 FSK, wheel rotating at maybe 200 revolutions per second.

Anyone know for sure?

Isn't this just a (harsher) variation on what you would see if you put the 2 Tx on a vehicle and drove towards the Rx? i.e. you would get Doppler shift.

In your case the shift would vary (back and forth) differently for each Tx due to the angular nature of the movement and opposite fixing points so the peak shift could be quite high. I guess the amount of peak Doppler shift would depend on how big the (200Hz spin rate) wheel was.
Assuming it isn't very big then I would have thought that the doppler shift would not cause an issue at the 1 mbps detector in the receiver.

Think of it as regular FSK modulation signal summed with sinewave Doppler shift induced signal summed with modulation.

It all depends on rotation rate and diameter of wheel (or just ground speed of wheel), regular FSK modulation deviation, and rate of data.

I would be surprised if logic is sampled on leading edge. That would be rather poor design. Should be integrated over bit width.

For FSK, doesn't matter much the fact that antenna is turned upside down.

Hi, biff44,

FSK is freq shift key modulation, that should be two different freq to denote 1 or 0. In your case, baud rate is 1Mb/s, such as say 1 denote 30MHz carrier, and 0 denote 10MHz carrier. So you should check the chip two frequencies.

Then calculate the doppler freq shift: 200 around/s, T=5ms, say wheel diameter=10m, maximum V about is (10/(5/2))*1000=4000m/s. doppler freq=(say 30*1e6)*4000/3*e8=40Hz. That can be neglected compared to 10M or 30M.

Cheer!

I am probably not being clear in what I am asking. I have already calculated the doppler shift and I would have to be spinning 10X faster before it was a real problem.

I was specifically wondering what the sudden 180 degree phase shift, back and forth, as the wheel rotates, would do to a modern FSK receiver. Something like a nordic nRF24L01
transceiver chip.

Where is the sudden 180 phase shift?

As the antenna orientation changes, there is an abrupt change between one phase and the other. In the attached view, if the TX and RX antennas are spaced the same physical distance, changine the orientation one of the monopoloes will get a carrier phase change of 180 degrees.

I have measured this effect extensively for position locating systems, and it is a sudden dramatic phase reversal when it happens.

Maybe a better depiction would make the sudden reversal clear. Somewhere near the horizong the signals wink out amplitude-wise. But before that point the carrier phase was X degrees, after that point it is X+180 degrees. (this is ignoring the gain patterns (which also change) of the antennas of course, just talking about phase).

Hi, biff44,

Don't worry about the phase hop for GFSK modulation. For FSK modulation, 0 & 1 detones two defferent freq. The receiver don't care phase. Just think about FM radio, although the signal received phase changed differently such as multipath, but FM radio can work properly. Normally the GFSK allow freq tolerance about 160KHz. So you can ignore doppler effect or phase hop.
you can ref Agilent application note: bluetooth measurement fundmentals.

I am not fully convinced. If FM demodulators were completely immune to a sudden 180 degree phase hit, then why do FSK systems have phase noise requirements (typically -70 dBc/Hz at 10 KHz offset for DS1)?

I guess I am missing one piece of knowledge--the exact type of FM demodulator that the Nordic chip has inside of it.

From a purely theoretical point of view, I do not see a lot of difference between FM and PM modulation. They are both angle modulations. One is just he integral of the other.

I had designed microwave point to point telecomm system for several years. I should say FSK system (PDH) have much better performance to prevent phase hit than other modulation, such as QPSK or QAM.
We use 4FSK system, and the maximum rate can reach 24E1, i.e. 48Mb/s. In the system, there are many component are sensitive to the phase hit, but more than 90% our PDH products can pass three temperature cycles, from -35deg to 60deg, no any error bit for about 24 hours. I can't say that the other 10% was caused by phase hit. There are so many factors can cause error, such as vibration (microphonic?). Normally we neglect phase hit for 4FSK system.
But SDH system (QAM modulation) is very sensitive to the phase hit. We have struggled several years to prevent phase hit.
So I should say for 2FSK and system rate is only 1Mb/s, you can have much better performance than our product in preventing phase hit. For your system, maybe vibration is the main source for bit error.