Low-cost RF Detector Circuit

of

As long as there's only a single dominant signal, it sounds like you want a 3 or 4-digit frequency counter that goes up to 120MHz.

If you take your signal and just feed it into a bench frequency counter, does it read what you want?

Tim.

Hi Ken,

Thanks, that's a pretty interesting idea. One question is, how stable would the frequency oscillator be be without the PLL and how accurately would the correlation between the PIC PWM signal and the actual frequency be?

Also, could you elaborate a bit on the second part of your reply about the PIC being run from the VCO and the sampling from the ADC? Thanks in advance!

Jeff

Ken Smith wrote:

of

amplitude.

to

FM

that

or

have

can be

then

go

it. If

find

run

can

Let me be the first to say "Use a PIC".

A PIC has an ADC built in. The ADC can't sample at a high enoungh frequency to actually digitize the signal you have but, you may not have to.

The PWM output of the PIC can control a VCO. The output of the VCO can be mixed with your signal and the result low pass filtered. The PIC can then sweep the VCO up and down while watching the low pass filter. As you go past the signal, the low pass filter will have a beat frequency on it. If the PIC watches for the beat being large an low frequencied, it can find the carrier.

You may even be able to skip the mixer and low pass. If the PIC is run from the VCO, the sampling of the ADC would do the mixing. If you can figure some way to reject the various aliases, you could have a very simple circuit.

--
--
kensmith@rahul.net   forging knowledge

Let me throw out there some ideas

you have a mixer? great. Now get an oscillator that goes your range of 90 to 120M cycles

you start the oscillator and send it into the mixer. coming out of the mixer maybe you add an integrator or just a capacitor to ground, then sample that output with your a/d

you increase the oscillator to the next point, give it time to stabilize and integrate and sample that

The output of your oscillator probably needs to be constant in amplitude

your code would look something like:

define low_range 90M freq value define high_range 120M freq value define step 50khz frequency step value define stabilize 10us delay time define read_buffer 10Mb linear buffer in memory define samples_to_take 1000 I just guessed 1000, could be more/less

import ADC_routine.h

variable osc = current frequency used variable adc = sample from adc

reset everything set osc = low_range set read_buffer pointer = 0

For count = 0; count is less than samples_to_take; count +=1 wait stabilize adc = Get_ADC_Sample read_buffer = " freq is @" + osc + " data is *" + adc read_buffer += CRLF set osc = osc + step

Next For

end

When this runs, you get the data in a text file thats human readable, and you have unique characters to help sort the values for a graph

The idea isn't that the VCO is stable. What I intended to explain was sort of like this:

Imagine that you are the PIC and you have your hand on a knob that controls the VCO. As you turn the knob you hear:

ssssssssssEEEEEEOOOOOOUUU.UUUOOOOOOEEEEEEssssssssss ^hiss ^high freq ^low freq ^high freq

as you sweep past the frequency of the signal you are looking for. The sound stops breifly as you actually pass the frequency of the carrier. You could rate the amplitude based on how loud the "UUU" part is and get the frequency by looking at the frequency where you got, lets say 3KHz as the beat and averaging the two values.

In the PIC, the idea is that once it sweeps past the carrier once, it can stop sweeping the full length and just run back and forth over the carrier to get the information it needs.

If the sample and hold of the PIC is quick enough, the idea may work. The ADC in the PIC is timed by the clock input. you may not even need the VCO to do this.

A sample an hold produces a voltage at its output that is ideally the voltage at the instant it was asked to sample. In real life, it is more like an average over a short time near the time it was asked to sample. If the sampling is happening at, lets say, 50KHz, the signal is effectively being mixed with a spiky waveform. This waveform has all of the harmonics of 50KHz in it.

Imagine that you sample at 50KHz and see a 3KHz beat signal. This means that the carrier must be at one of these frequencies:

80MHz - 3KHz 80MHz + 3KHz 80MHz + 50KHz - 3KHz 80MHz + 50KHz + 3KHz ... etc ...

If you than change the sample rate of the ADC, from 50KHz to 33.333KHz, the frequencies in the sampling will be different so the beat note will be at a different frequency. By selecting the right set of ADC sample rates, you can make is to that based on the beat frequencies, you can rul out all but one carrier frequency. If the characteristics of the sample and hold are well known, the amplitude of the carrier can be judged from the amplitude of the beat note.

If the sample and hold in the PIC doesn't have a short enough aparture time, you could perhaps use some timer output to drive an external circuit.

--
--
kensmith@rahul.net   forging knowledge

You're lucky. You will have to go to pretty remote places to have no or little signal is this range. Whatever signal you want to detect must be stronger than the usual FM and air trafic AM signals. I guess it is pretty hard to get a frequency assigned there. So in order to not lock onto an FM or air traffic transmitter, you're going to rebuild a spectrum analyzer. Only after masking the locally known frequencies you're going to display anything at all.

Rene

--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net