Chirp or Pure Tone? 1

[kjiang]

Hello, I was wondering whether to use a linear chirp or a single frequency in measuring the doppler shift. We are using a monostatic setup and are emitting a frequency of 21k. Standard radar practices say to use a linear chirp instead of a pure tone. Why is that?

Thanks

 

[IRstuff]

see:

http://www.rfcafe.com/references/el...-technique-january-1965-electronics-world.htm

for one explanation

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[VEBill]

Determining Doppler given chirp (combined with variable or unknown range), or optimising chirp when there's Doppler, are both non-trivial.

"21k" = 21 kHz? Audio? Is this a school project?

 

[kjiang]

I should have made my question more clear. Sorry. I am using a sound generator that emits a frequency of 21 kHz. By measuring the change in frequency of the object (whether it be moving toward or away in a linear path directed along the path of the source) one can determine the speed of that object solely. (Using the Doppler Equation) However, is there a way in which we can measure the distance as well? An obvious answer would be to measure the time delay from the emitted signal to the received signal; however, how can I get that information by looking at the Fourier Transform? Do I need to plot it in the frequency vs time domain instead? I was also wondering whether there is any other possible method? This is when I started looking at the benefits of using a chirp.

Thanks for the website IRstuff. However, I don't believe that range nor resolution is an issue in the setup I have.

 

[IRstuff]

Hypothetically, yes, but it's only the phase that should be different for a pure delay.

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[kjiang]

If the object weren't moving then the frequency vs time graph would simply be a horizontal line followed by a time delay then another identical horizontal line. (Assume I'm using a pure tone) The phase difference in the frequency time domain gives me the difference in time (time delay) and thereby I can calculate the distance to the object (knowing the speed of sound of course).

The other case, where the object is moving (let's say towards you), will look like a horizontal line followed by a time delay (this time delay will be different in duration, shorter in fact, then the previous case) and then another horizontal line (higher then the first due to the doppler effect). Given the time delay and the speed of sound, is it possible to find the distance to the person?

The person / object is moving; so the object moving towards you will have a speed v1 and the speed of sound will be c. The time it takes to receive the signal is t. So the equation to find the distance from the source (which emits the pure tone) to the object would have to involve calculus because the distance is constantly changing. Correct?

Edit: Calculus is not involved. I just though about it and the only thing that matters is the distance and instant the sound wave and the moving object collide. This can be treated like the first case, at that instant, the object is stationary and the sound wave reflects off. The only thing to account for then is the velocity of the object after the sound wave hits it.

 

[kjiang]

Furthermore, once the sound wave hits the object, (it will reflect of course) and returns to the receiver, the hypothetical equation from m previous post (let's hope it exists) will only be a measurement of the distance of the object at the instance the sound wave hits the object. Not the final position (when the sound wave returns to the object).

 

[IRstuff]

If you are that concerned about position, then you should be using a laser, not sound

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[VEBill]

Decades ago, one would have had to make a decision one way or the other. These days, when almost everything is programmable and DSP'ed, one could emit, receive and process a different waveform with each pulse.

If one was really ambitious, one might use two waveform concepts at the same time (within a pulse). Either at two adjacent frequencies, or perhaps even (conceptually) stacked hierarchically.

 

[GregLocock]

Given that ultrasonic tape measures exist obviously distance measuring is feasible. As to the advantages of chirp over pure tone, I'd have thought 10 minutes writing a simulation would give you an idea of the relative merits.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?