Envelope Detector

[ThermalRunaway]

Hi everyone,

I'm thinking of using an envelope detector in order to detect the presense of a 200Khz waveform. When an envelope detector is used for the task which it was originally designed for, you feed in an AC signal which has been modulated with a much higher frequency waveform, and the envelope detector then strips away the high frequency part leaving just the original AC signal as in the attached diagram.

However, all I want to do is detect the presense of a 200Khz AC signal. If I were to feed a standard sine-wave 200Khz signal which hasn't been used as a carrier into an envelope detector, am I correct in thinking that I would get a DC-level output directly related to the amplitude of the original 200khz waveform?

Any thoughts on this?

Brian

[/img]

 

[Nigel Goodwin]

Hi everyone,

I'm thinking of using an envelope detector in order to detect the presense of a 200Khz waveform. When an envelope detector is used for the task which it was originally designed for, you feed in an AC signal which has been modulated with a much higher frequency waveform, and the envelope detector then strips away the high frequency part leaving just the original AC signal as in the attached diagram.

However, all I want to do is detect the presense of a 200Khz AC signal. If I were to feed a standard sine-wave 200Khz signal which hasn't been used as a carrier into an envelope detector, am I correct in thinking that I would get a DC-level output directly related to the amplitude of the original 200khz waveform?

Yes, it's simply an AM detector - or more simply, a rectifier. You input AC to a rectifier and you get DC out - this is no different.

Personally I think you are making it sound FAR more complicted than it is by calling it an 'envelope detector' :lol:

 

[ThermalRunaway]

Well I just reffered to it as the diagram is labelled, but I'd agree it is a simple little circuit and it's just as accurate to call it a demod circuit or whatever.

Thanks for the advice.

Brian

 

[red4925]

Rectifier or Envelope detector - all depends on how you use it, but yes, it's the same circuit.

If you're trying to detect a certain frequency, you could run your signal through a high-Q bandpass filter before rectifying it, and then use some sort of comparator to see if there's a decent voltage level at your rectifier output. That seems like it would detect a certain frequency for you. I don't know what other kinds of signals you must differentiate from though.

Another idea would be a simple mixer in line with a low-pass filter, but you would need a 200kHz local oscillator. And the mixer would probably attenuate your signal a lot, not sure though.

Hope that helps you out.

 

[ThermalRunaway]

Part of my design does incorporate filters before this rectifying stage, but that's mainly because the 200Khz signal has to be extracted from a 50HZ mains AC waveform before it can be processed. By the time it reaches this rectification stage, all I really need to be able to do is detect is the presense or absense of the waveform and provide this information, in TTL form, to a microcontroller. I don't need to maintain accurate reproduction of the 200Khz signal, I only need to be able to detect the presense of it.

With that in mind, I'll need a buffer stage perhaps of the schmitt trigger type after the rectification stage, in order that the correct DC levels for logic 1 and logic 0 are provided to the microcontroller.

Thanks for the reply red.

Brian

 

[Optikon]

Part of my design does incorporate filters before this rectifying stage, but that's mainly because the 200Khz signal has to be extracted from a 50HZ mains AC waveform before it can be processed. By the time it reaches this rectification stage, all I really need to be able to do is detect is the presense or absense of the waveform and provide this information, in TTL form, to a microcontroller. I don't need to maintain accurate reproduction of the 200Khz signal, I only need to be able to detect the presense of it.

With that in mind, I'll need a buffer stage perhaps of the schmitt trigger type after the rectification stage, in order that the correct DC levels for logic 1 and logic 0 are provided to the microcontroller.

Thanks for the reply red.

Brian

Trying to do digital communication over the power lines?
Have you considered frequency modulation? Apart from more involved transmit & detectors, it offers some advantages like potentially being more immune to trash that you will get on the mains lines. Think hair dryers etc..

 

[_3iMaJ]

Yeah but FM offers something terrible as well! Its bandwidth is infinite! As well as being terribly inefficient.

If you look at some of the more popular digital schemes such as QPSK, GSK, QAM, you'll see they are far more efficient than any analog scheme could ever hope to be.

Mainly the appeal to analog schemes is their ease, they don't offer anything else that could hope to compete with the best of the digital schemes.

 

[Nigel Goodwin]

Yeah but FM offers something terrible as well! Its bandwidth is infinite! As well as being terribly inefficient.

FM isn't terribly inefficient, FM radio is more 'usable' than AM of the same power - which probably means it's more efficient than AM?.

Your bandwidth worry isn't true either, although theoretically FM has infinite bandwidth, it's also at infinitely low levels. However, an FM signal does take up more than AM, for the same bandwidth.

If you look at some of the more popular digital schemes such as QPSK, GSK, QAM, you'll see they are far more efficient than any analog scheme could ever hope to be.

Mainly the appeal to analog schemes is their ease, they don't offer anything else that could hope to compete with the best of the digital schemes.

The crude AM detector scheme isn't likely to perform terribly well, and the FM suggested probably wasn't what was really intended?. The usual method is to use FSK (Frequency Shift Keying), where you transmit one frequency for HIGH and another for LOW, and you use an NE567 PLL tone decoder (or similar) to restore the digital signal at the receiver. This is simple, efficient, and reliable - the transmitter can be nothing more than a 555, and the receiver an NE567.

 

[ThermalRunaway]

I understand that using a rectifier isn't likely to perform very well in terms of reproducing the 200khz signal, but as I'm sure you've already understood the reproduction quality, for my application, isn't that important.

Optikon: Thanks for your comments. I hadn't actually considered using FM modulation, mainly because my current method is simpler and hopefully should do the job. However, should it prove difficult to get working properly then I will have to consider other options. I'm not sure how FM modulation of the data would work out in my particular case as I'm only transmitting data on a 1-bit at a time basis. I understand that this makes for incredibly slow communication, but we're only talking about transmitting and receiving 8-bit codes (which make up one single command) and communication only takes place when a new command needs to be issued (which isn't very often), so we're not talking about constant exchange of data between devices. Therefore, speed is not an issue with my design.

Thanks again for the comments everyone

Brian

 

[_3iMaJ]

How are you transmitting your data through the channel? Is it done using some sort of square pulse? If so, you may want to reconsider that (depending on what the frequency response of the channel looks like). Recall that the fourier transform of a square pulse is a sinc function (that pesky infinite bandwidth thing again). So when you pass your pulse through a channel you're probably not going to get a nice looking received wave form.

For this case a quick and dirty OOK (on - off keying) would work wonderfully. Instead of transmitting a square pulse simply transmit (for example) a few cycles of a 24khz sine wave for your 1s. And for your 0s you can simply do as you did before.

Now to receive this you can use that rectifier circuit you were thinking of (or envelope detector to us comm people). Simply rectify it and use a LM 311 comparator to clean up your waveform and convert it back to TTL levels.

 

[ThermalRunaway]

_3iMaJ:

No I'm not transmitting square waves over the mains, I'm transmitting pulses of 200khz which, via my filter stage which will later extract the 200khz pulses and a detector which will convert them to a DC level, I hope to "create" a square-wave type signal which I'll buffer so that the proper TTL logic levels are presented to the Microcontroller. So the transmission starts out as pulses of 200khz and finishes with logic 1s and logic 0s which depend on the presense or absense of the 200khz pulses.

If you're familiar with X10, this idea is extremely similar.

Brian

 

[_3iMaJ]

Oh then yes you're doing OOK just as I stated above. You're envelope detector will work nicely. All you need to do is choose your RC time constant significantly longer than that of the 1/200khz period. I'd say like the time constant should be 5-10 times longer than the period. I could give you a better estimate but I'm not certain if your signal is unipolar or bipolar.

Once you've done that use an LM 311 comparator (set up in TTL mode) to give you a very nice clean TTL signal that you're looking for.

Good luck.

 

[lord loh.]

Part of my design does incorporate filters before this rectifying stage, but that's mainly because the 200Khz signal has to be extracted from a 50HZ mains AC waveform before it can be processed.

I cannot think of 200KHz over a 50Hz carrier.

This would certainly make your diagramme (imput s{t}) look false.

If a rectifier is to detect a wave, the modulating wave ought to be higher than the carrier.

And the capacitor shall cause an envelope voltage shall overshadow all the 200KHz signal.

 

[Nigel Goodwin]

Part of my design does incorporate filters before this rectifying stage, but that's mainly because the 200Khz signal has to be extracted from a 50HZ mains AC waveform before it can be processed.

I cannot think of 200KHz over a 50Hz carrier.

This would certainly make your diagramme (imput s{t}) look false.

If a rectifier is to detect a wave, the modulating wave ought to be higher than the carrier.

And the capacitor shall cause an envelope voltage shall overshadow all the 200KHz signal.

The 200KHz doesn't 'modulate' the 50Hz, it's just carried along the same wires - although there's no reason you can't modulate a low frequency with a higher one, but as you say a simple AM detector wouldn't work (unless you use a mixer to up the frequency to a higher IF).