Hi all,

Sorry if this sounds vague, I have not posted any schematics.
But I'm just looking for suggestion in the implementation aspects.
I have uploaded my actual and expected waveforms

In matlab/simulink I have a relational operator where I compare the relative amplitude of the signal during each symbol duration to check if it is "1" or "0". Having said that, you can see the image "Original_Demodulated_Signal" which consists of original signal and the other one "Demodulated Signal_with_Expected_results" which consists of expected waveform [I have marked in red, which tells what I am looking to regenerated form this waveform].

Here in the waveforms my bit or symbol duration is 0.5ms.

I have the idea: where I will check for certain threshold value at the starting of each symbol duration, and based on the threshold value the high(some voltage) or low (0 voltage in case of unipolar and -voltage in case of bipolar) is assigned.

I am looking for implementation suggestions. please let me know if someone knows it.
Regards

 

If I were going to try this I would start by amplifying the signal so that layout issues and offsets don't have much affect on the operation of the circuit. After a little amplification the signal can be passed through a comparator with a little bit of hysteresis to make a logic level signal, then a non-regriggerable monostable at about 1/2 the smallest interval between pulses to take out any the ringing that might get through. After that it is a matter of decoding the pulse interval.

If the amplitudes are not consistent then you can either use AGC or make the comparator threshold dependent upon the peak signal amplitude.

 

Dear DickCappels

Sorry I did not mention in my question about what kind of modulation I am using.

I am doing frequency modulation of a binary data(pulses in case of LTspice) and I am passing it over the inductive coupled channel, and the waveform what you see in my question is what I get at the output.

Now I passed the output through PLL circuit, and I get the amplitude variable signal (for high one value of amplitude and for low one more value of output), so now as you said, I am trying to come up with the circuit which defferenciate this amplitude variations and makes the decision between a high pulse or low pulse.

If you know details about this kind of circuit, please share your knowledge.

Thanks and Regards

If I were going to try this I would start by amplifying the signal so that layout issues and offsets don't have much affect on the operation of the circuit. After a little amplification the signal can be passed through a comparator with a little bit of hysteresis to make a logic level signal, then a non-regriggerable monostable at about 1/2 the smallest interval between pulses to take out any the ringing that might get through. After that it is a matter of decoding the pulse interval.

If the amplitudes are not consistent then you can either use AGC or make the comparator threshold dependent upon the peak signal amplitude.

 

What frequency is your original signal?
The XR2211 will do the whole job - signal in, decoded logic level signal out - in one chip, up to 300kHz.

 

What frequency is your original signal?
The XR2211 will do the whole job - signal in, decoded logic level signal out - in one chip, up to 300kHz.

Dear Albert

My input signal is of 2 to 3Mhz, so I cant use XR2211 if its limit is upto 300KHz.

Considering if I get a output as like the one attached with this reply, can you please guide me with what kind of circuit I can differentiate the amplitude levels and get back the original pulse ?

Regards