Create test for specific frequency.

[bassguy1965]

Hello all,
I am primarily a software person, but I find myself in the hardware world as of late. Please forgive what maybe a newbie question.

I am trying to build a circuit that can detect a specific frequency from the following frequencies: 128kHZ, 129kHZ, 130kHZ, 131kHZ and 132khz.

If the frequency is detected, I would like the ability to pull up the voltage on a specific data pin.
only one of the frequencies will (should) be activating a ferrite core inductor at a time.

I have looked at band pass filters and have not sussed out exactly how I would implement this.

any help pointing me into the right direction would be greatly appreciated.

Thanks for the time,

Bassguy1965

 

[Sceadwian]

The specific frequencies you want absolute knowledge of and the entire spectrum you have to work with is pretty narrow. DSP's are a good solution because they can achieve high cut's (the narrow range of 1khz) where as with an analog circuit you'd need a very serious multipole bandpass filter, as the frequency in relation to the bandwidth and channels you're addressing is very high.

 

[ronsimpson]

Is there lots of noise on the input?
Are there more than one frequency present at the same time?

If there is only frequency at a time and there is little noise than it is easy.

 

[RadioRon]

One way that this can be done is using phase-locked loop detectors. These can be designed to have quite narrow bandwidths.

 

[Mr RB]

What makes the frequencies, what amplitude and are they continuous?

 

[alec_t]

Does your solution need to be able to detect all five frequencies (individually) and have some way of specifying which one? Or will it only ever have to detect a specific one?

 

[bassguy1965]

I cannot get too specific, but, my current system is an active RFID system that has adjacent a transmitter tuned to the frequencies mentioned above (130,131,132, etc.). The RFID tags respond on a different frequency whenever they are within range of these transmitters. (@15 feet) Currently the tags use the incoming frequency as a sync frequency encode some data on it and piggy back that onto the carrier wave returning to the main receiver. At the receiver the data is decoded with the same sync pulse. The tag is never aware of what frequency it received as it is just using the 130kHZ or 131kHz etc. as a sync pulse. The reason for this is that the tag can broadcast to a much larger area than the transmitter that activated it and we are very interested in knowing exactly which transmitter the tag was near when it was activated.

I am trying to have the tag be a bit smarter and be aware of the frequency it receives and map it to an address that I can encode into its outbound message.

Thanks again,

Bassguy1965

 

[alec_t]

So as I understand it you would like each tag to have the capability of detecting, with a resolution of better than 1%, the frequency of an incoming signal in the 128-132kHz range. I think that's going to involve extensive signal processing. Would it be possible, instead, to modify the transmitters so that each sends a unique code sequence which the tags can detect more easily?

 

[ronsimpson]

If the ID tag needs to know the difference between 130k and 131k then 1% error/resolution will not work. The difference is 0.7%, The transmitter will have some errors + the receiver will have some errors. I think a resolution of 1/2% is boarder line working.

If the receiver has a PIC in it, the oscillator needs to be crystal not ceramic or RC.

 

[bassguy1965]

I suppose it could send a code. Our current system uses the frequency as the code. I was just concerned about size and battery usage if there were to be data parsing involved. The beauty is as I am prototyping new, anything is still an option to me.

Bassguy1965

 

[alec_t]

I was just concerned about size and battery usage if there were to be data parsing involved

If you've got in-tag smarts IMO that wouldn't be much of a chore, whereas analogue (phase-locked loops etc) or digital (A/D, DFT etc) processing to identify a carrier frequency very accurately would use much more processing time and power.