Simulating Signals for Electrical Engineering Projects

[shekoofy]

Hi every body

I'm an electerical engeneer with a fair backgraund in telecommunication field.I'm working on a project in which I'm supposed to simulate the whole signal a receiver gets on Earth (from suppose a satellite)with all the specifications like the doppler shift of the transmitted signal and the acceleration of the reciever as well.(in fact I'm going to fool a stationary receiver as if it is receiving true signals whit doppler shift and has a certain acceleration itself)
untill now I've managed to simulate the baseband signal plus the atmosphere errors in matlab.now I'm wondering how can I simulate the doppler shift and the acceleration?should I do it in baseband?how?or I should wait for a vector generator and show the doppler in the phase of the signal(just as we see in equations)

I've read tens of papers but since I'm working individually It't very hard to make a final decision.
ANY suggestion is welcome
thanks

 

[sophiecentaur]

If you mix up and then mix down to a slightly shifted frequency, I think this should be a sufficient approximation to a Doppler shift as long as your fractional bandwidth isn't too much.
Doppler will shift frequencies by a given ratio, whereas up/down mixing will shift all signal frequencies by a constant frequency. But, unless your transmitter and receiver are moving extremely fast wrt each other, I don't think you could tell the difference.

What effects of acceleration are you to investigate? Varying the frequency shift (above) would give this effect - with the same caveat.

 

[vk6kro]

If you measure the frequency of a transmitter in a satellite coming towards you, it starts off at a frequency that is higher than its real frequency and steadily drops in frequency until it is directly above you or until it is passing at right angles to you.

At this time, the satellite transmitter is heard at the right frequency.

As it moves away, it is heard at an increasingly lower frequency.

This website shows some nice graphs of this with some actual frequency shifts for different frequencies:
(click on the next line)
http://www.qsl.net/vk3jed/doppler.html

Note that Doppler shift only depends on the relative velocity of the receiver and transmitter for a given frequency, so accelleration doesn't matter.

The maximum height of passes is given in degrees. These are passes where the satellite does not go directly overhead but instead rises from the horizon, reaches a maximum height at right angles to the observer and then drops towards the horizon again. This may happen over a period varying from zero for a distant pass that barely rises above the horizon, to about 20 minutes for an overhead pass.

 

[shekoofy]

Thanks vk6kro for the nice web site,it seems I should have real scenarios of both satellite and receiver trajectories and then dynamically update the doppler shift in the generated signal.Am I missing something?

"Note that Doppler shift only depends on the relative velocity of the receiver and transmitter for a given frequency, so accelleration doesn't matter."

really?!?sounds odd!It means if we only check the doppler shift diagrams we can not realize if the receiver had an accelaration or not?it doesn't show any effects on doppler shift?

 

[shekoofy]

thank you very much sophiecentaur for the precise reply,now I have came to this question:science there is a massage carried by the carrier with different(much lower) chiprate(read frequency)do I have to show the doppler shift on it as well?if I mix up and then mix down to a slightly shifted frequency it will do a doppler shift only on the carrier.
in fact I want to know up to what frequensies the doppler shift doesn't have any significant effect on the signal? is it just not necessary to simulate the doppler shift in lower frequencies or it is a complete waste of time?what sort of things should I take into consideration?(the satellites in my project supposed to be in MEO orbit and the carrier is in L1 band)

 

[sophiecentaur]

Mixing up and down will shift Everything. How could it not?
The only problem is that the ∆f is the same value for all the sidebands. Doppler produces a fractional change in frequency. It just depends whether there is a significant difference in the final effect at the frequencies of your satellite. In the case of Doppler, I think the effect will be to alter the baseband (demodulated) frequencies whereas, in the frequency shift simulation, there will be a change in the carrier frequency but not in the relationship between carrier and sidebands. To examine that, you would also need to change the modulating signal frequency- by changing the clock? perhaps. For a complete simulation, you'd need to do both, I think. Either the downconverter and demodulator could be crucial, depending on the bandwidth of the modulation system.

You can calculate the doppler shift if you know the orbit diameter and, hence the relative speed. This will tell you the maximum ∆f in each case. (And the rate of change - as the velocity changes)
You may even be able to look it up somewhere using Google, in any case. But you can do that calculation for yourself, surely. That's the first thing to tackle.

 

[vk6kro]

really?!?sounds odd!It means if we only check the doppler shift diagrams we can not realize if the receiver had an accelaration or not?it doesn't show any effects on Doppler shift?

Acceleration has obvious effects on velocity, but the Doppler effect only depends on velocity and the frequency involved.
This formula applies for satellite applications:
[PLAIN]http://dl.dropbox.com/u/4222062/Doppler.PNG

For example, if the relative speed of transmitter and receiver is 2 Km / second and the signal frequency was 5000 MHz,
Doppler shift = 5 E 9 * 2 / 300000 = 33333 Hz

up to what frequencies the Doppler shift doesn't have any significant effect on the signal?

Doppler shift becomes a problem if the received signal moves outside the passband of the receiver or if it moves enough in frequency to upset the demodulation process.

You may be able to compensate for this by manually tuning a receiver to keep the signal within the passband, but this would mean the Doppler shift was significant.

A narrow band receiving technique like single sideband requires the receiver to be receiving within about 50 Hz of the correct frequency to resolve speech adequately. Quite small Doppler shifts mean the receiver has to be retuned.

Incidentally, all frequencies in the transmitted signal are shifted almost equally for a narrow band signal. The higher frequency ones will be shifted very slightly more as you can see from tha above formula.

 

[sophiecentaur]

@vk6kro
Of course the Doppler shift only depends on velocity. But, if you need to lock to a very narrow band signal, then the rate of change of velocity / frequency (acceleration) could, in principle, be relevant. Hardly likely to be a relevant factor with a satellite, though.

Your calculations look right but it's not likely that the OP is referring to SSB Voice transmissions. Also, manual tuning is not likely for a satellite receiver.
My point still stands that it is easy to obtain an fixed overall frequency shift with a simulator but that it is not the only effect of Doppler.
One would need to know the actual figures for the modulation system and the data rates involved to know what phase / frequency distortion would be relevant.