is there a complete analog circuit that gives you an indication of the signaltyp

[RAPo]

Is there a complete (i.e. no Arduino/microprocessor etc or visual inspection) circuit that, given a signal in 10Hz-1MHz, minimal 2Vpp) gives you an indication of the signal type (sine, square, triangle/ramp, other)?

[shapirus]

Is there a complete (i.e. no Arduino/microprocessor etc or visual inspection) circuit that, given a signal in 10Hz-1MHz, minimal 2Vpp) gives you an indication of the signal type (sine, square, triangle/ramp, other)?

Peak detection, RMS calculation, then C = Peak / RMS, and then find C in a table of known crest factor values for different waveforms. Sounds like it should be doable, but not too easy.

[RAPo]

Thanks for your swift&nice reply. One and the same crest factor can indicate different signal types (https://en.wikipedia.org/wiki/Crest_factor). Do you have a reference to such a crest factor table?

Peak detection, RMS calculation, then C = Peak / RMS, and then find C in a table of known crest factor values for different waveforms. Sounds like it should be doable, but not too easy.

[Ian.M]

I suspect that crest factor fingerprinting will confuse too many distinct waveforms.  However, if you want to pursue this, its not difficult to generate your own table of crest factors using any programming language you can generate the waveforms in and implement peak detection and RMS calculation algorithms.  If you are familiar with Matlab or GNU Octave, I'd start there, though my personal preferences for manipulating and creating engineering data tables tend towards writing console apps in 'C' or scripting Excel.

In theory, yes it can be done, as its possible to characterise a signal by spectrum analysis, comparing the relative power and phase of the fundamental and its harmonics, (which for a fixed frequency periodic signal is sufficient information to reconstruct the original signal subject to a bandwidth limitation of the highest harmonic used) and pure analog swept spectrum analysers can be built, as can analog circuits to sample the various harmonics phase and quadrature components and window compare their levels, but its impractical to do it that way, (and has been impractical for over 30 years) due to the physical complexity (I estimate a 19" rack mount cabinet full of eurocard size PCBs)  vs sampling the fundamental and harmonic levels with a MCU (which became viable approx. 40 years ago), or even sampling the signal and using DSP techniques which has been viable for over 25 years.

[shapirus]

One and the same crest factor can indicate different signal types

Well, then my idea won't work . At least, without checking additional conditions, or beyond a few basic waveforms. What other formal parameters can we use to identify a given waveform? Probably integration and differentiation could be useful. At this point it gets more interesting theoretically, at the same time becoming less feasible practically.

[reboots]

Run the signal through a cathode-ray oscilloscope. Detect peak amplitude and period to automatically adjust gain and sweep for a consistent single-cycle display. Shroud the display along with a phototransistor, photomultiplier or other sensitive device. Characterize typical overall brightness for each waveform; e.g. square excites more screen area than sine, then triangle, etc. Use those signature levels to trigger indicators for each type.

My friends have a saying that in the early age of electronics, when they wanted something done really badly, they did it, really badly. I believe I have honored that spirit of invention here. Perhaps there is a better way to characterize the information density of a cycle in the analog domain.

[shapirus]

the information density of a cycle

it's basically the same characteristic as crest factor, I guess. It works as long as you don't need to discern more than one different waveforms that have the same crest factor. If you do, you need to analyze something else in addition to that.

[TimFox]

Of the "elementary" waveforms listed in the Wikipedia article, the pairs that have identical crest factors have different DC (mean) values, which could differentiate between the two candidates with that crest factor.
Of course, one can generate lots of "non-standard" waveforms to add to the textbook examples in that list.
Specifically:  DC and zero-mean square wave have crest factor = 1.0, and sine wave (zero mean) and full-wave rectified sine wave have crest factor = 2^1/2.

[jonpaul]

scope and eyes, brain

[Ian.M]

Even if there was a simple(ish) circuit to characterise waveforms in the analog domain, the complexities of switching time constants to handle a five decade fundamental frequency range would make a MCU/DSP based solution far far more attractive! 

[RAPo]

Thanks, I can program in the languages/applications you mentioned, but before doing, I would check the knowledge of this forum if there wasn't an existing solution

[RAPo]

Yea the main gist is: this will require a lot of tinkering.

[TimFox]

Even if there was a simple(ish) circuit to characterise waveforms in the analog domain, the complexities of switching time constants to handle a five decade fundamental frequency range would make a MCU/DSP based solution far far more attractive! 

Analog circuits for measuring crest factor (peak over RMS) and mean value (DC level) have been made to cover a wide frequency range, and have been used in successful analog voltmeters since the -hp- 400 series.

edit:  clarification--"since the time of the -hp- 400 series"

[RAPo]

That goes beyond the question. Just like we have testers for if a transistor is npn or pnp, I would like to have a small (think the seize of a DMMcheck) all-contained box with less indicating the most probable type.

[RAPo]

And what if I lower the specs to two decades?

[shapirus]

I would like to have a small (think the seize of a DMMcheck) all-contained box with less indicating the most probable type.

If you retract the requirement for it to be analog, you can rather easily implement it with an ADC and a microcontroller or a single-board computer to perform the analysis. It will easily fit in a small box.

[TimFox]

That goes beyond the question. Just like we have testers for if a transistor is npn or pnp, I would like to have a small (think the seize of a DMMcheck) all-contained box with less indicating the most probable type.

You asked for a "complete" circuit, not a miniature circuit.

[RAPo]

true

[TimFox]

Anyway, there are analog ICs that can measure RMS, peak, and mean values of a continuous waveform that could build such a complete circuit in a reasonable size box, as well as ratios.
(Not necessarily in a single package.)

[2N3055]

No.

[artag]

sine, square and triangle all have different harmonics. An analogue circuit to do this for the general case might be difficult, but it wouldn't be too bad for a specific frequency.