How To Measure The Amplitude of a Sine Wave

oscilloscope, analyzer. Some hand held meters will do this, too.

I need to build my own circuit.. I have two options so far, Analog to Digital Converter (ADC) or a microcontroller..

You probably want to search for 'True RMS Converter'. They are usually integrated with other circuitry (ie. sample and hold) and provide a few outputs to describe the waveform being measured.

For measuring peak amplitude voltages try an internet search for "peak voltage detector", that will allow you to capture and convert the peak voltage into an analog signal that can then be fed into an analog (moving coil) or digital display for direct representation (although the digital display will need further conversion depending on the display technology).

A microcontroller and some custom code can do it easily, but you will need more programming hardware, software and a computer (which you may not have access to or the budget for).

<sigh>

You have so many unknowns here and provided us so little information to help us help you. You also ask such a basic question that I'm a little stumped about where to begin.

First, it would be very helpful to identify the anticipated voltage amplitude level and the anticipated frequency. There will be completely different concerns in measuring power distribution voltages and the the voltages produced by a Fender Telecaster pickup. Depending on your voltage source, you may not have only a sine wave to measure. You may have added interference or multiple (Telecaster six note chord) sine waves to measure simultaneously. To paraphrase this in economic terms, are you trying to measure the GDP of Lichtenstein or the daily cash register results of your favorite restaurant?

Second, how accurate, stable and precise does your measurement have to be? These are three different but related parameters that should be known to make an appropriate suggestion.

Lastly, why don't you just purchase a test instrument instead of making one your self with the A/D converter or microcontroller you mentioned earlier. There is a lot of engineering already dedicated in these instruments.

5 Vpp between 0.5K-Hz to 5M-Hz and at least +/-(1/4) dBm accuracy .. what do you think

5MHz is higher than most DVM's rolloff, which usually start rolling off between 100-200KHz. You would have to check datasheet's for True RMS Converter's capabilities. ADC's would need a good amount of signal conditioning to handle that frequency as well. You may need to look for an R.F. ACVM and matching probe to handle 5MHz.

Let's not forget Nyquist's sampling theorem here. To measure a 5 MHz signal digitally one must sample at least twice as fast as the maximum signal to be measured. So the A/D converter must be capable of sampling at a non-trivial rate of 10 MHz. Technically one can apply some conditioning tricks in front end filtering and software data analysis so that sampling must be at least double the bandwidth of the signal. In this case the bandwidth (4.9995 MHz) is nearly identical to the highest frequency so it doesn't matter.

Well now, you mixed a few parameters together that confuse me but I believe I can explain the confusion.

First, the good points. You do have a relatively narrow band-with of acceptable frequencies to be measured. Oops. My mistake. I thought at first you only had one decade of measured frequency amplitudes. I now see that you have four decades of acceptable frequencies.

Second, the confusing parameters and how it baffles me.

You indicate a single fixed voltage level of 5 Vpp (peak to peak) and a 1/4 dbm accuracy. The term dbm is a measurement of power and not voltage. The db scale is a very powerful measurement technique to measure any two identical units of signals, typically "output" over "input" or "after" over "before" but the magnitude of any two values can be used. 20 db is 10 times bigger that originally and 40 db is 100 times bigger that originally. By convention dbm is a single point measurement of power where the denominator is fixed at 1 milliwatt of power. So 20 dbm is 10mW. 40 dbm is 100 mW. 60 dbm is 1W. Similarly -20 dbm is 0.1mW. So 5 Vpp→ 5/√2 Vrms≈3.5 Vrms into a 600 ohm load as (5^2)/2 Vrms^2/(600 ohm) → about 20.8 milliwatts or 26.3 dbm if the load impedance is 600 ohms. If the load impedance is different then the dbm value is different for the identical voltage peak to peak.

Now to show my compounded confusion. The frequencies you've chosen are just above the audio frequency range where 600 ohms would be standard load and at the lowest range where RF standards of a 50 ohm load would apply. So not only have you mixed units of power and voltage, your frequency selection precludes me from making a plausible assumption how to convert.

To return to an economist example (I hope you don't mind me picking on your handle like this. ) this is like asking somebody how many Italian lira a $100 US bill is worth without giving them a date in time for this exchange or asking if you could survive on $100 US a day without identifying when or where one must survive in the about the 220 year history of the US dollar.

By convention dbm is a single point measurement of power where the denominator is fixed at 1 milliwatt of power. So 20 dbm is 10mW. 40 dbm is 100 mW. 60 dbm is 1W.

When talking about 'POWER' in the log scale, the multiplier is 10 (not 20), so 20dBm is 100mW (2 factors of ten), 40dBm is 10,000mW, or 10W (4 factors of ten) and so on.

When staying just in the voltage domain of the log scale the multiplier would be 20, but all reference would just be dB (not dBm, which is specific to power, actually referenced to 1mW).

10 x LOG( P1/P2 ) vs. 20 x LOG( V1/V2 )

You are absolutely correct. I was confused and it was late.

We still have no idea what load impedance to use for any calculation. So my confusion remains.

Thanks for the feedback, I was confused with LOG scale calculations until I used a slide rule in my Junior year of high school. It became a 'natural' way of looking at numbers for me after seeing that relationship on the scales.

Since the impedance is not known, and the frequency range is so broad, it is very difficult to recommend a decent or proper measuring system. I think a cheap o'scope would be the best bet, but if he is being asked to come up with a circuit, a high speed ADC with proper signal conditioning sounds best. That just leads down another path of interfacing to either a microcontroller, processor, or external PC.

This is turning out to be more difficult than I thought. I would have to do some reading on sampling theorems first,

In addition to several good posts already given:

* Do you really only need amplitude? As you develop whatever you are working on will you need quality also (total harmonic distortion, amplitude of specific harmonics, etc.)

* Don't forget that real world signals have noise to corrupt measurements and sometimes surge energy or ESD to damage components.

* Any chance you might also need phase with respect to some other signal or clock?

Without knowing more about what you are doing it is hard for others to give better suggestions.

The amplitude of a sine wave is very easy to measure. You measure its peak and divide by SQRT of 2. If you are dealing with a non sine wave, things start to get messy, and depending on the scope of your measurement, you may need P2P, peak, average or RMS. S.M.

RMS = 1/√2 = 0.70711 of peak sinusoid wave, both Voltage and Current.

AVG = 2/π = 0.63662 of fullwave peak sinusoid wave, Voltage only, function of exponential RC-time decay for 120Hz.

AVG = 1/π = 0.31831 of halfwave peak sinusoid wave, Voltage only, function of exponential RC-time decay for 60Hz.

RMS is used to 'equate' peak Voltage and Current to their effective "heating" equivalents for computing "average" power, ie: P(avg) = V(rms)×I(rms).

Notice the relationship: 90% ≈ AVG/RMS = 0.63662/0.70711 = 0.90032

There's really no need to reinvent/re-engineer such a circuit, and I'm willing to bet there's an IC out there that will do exactly what you want. Try looking at the LTC1968 as an example, there are many others.

If you handed me that problem the first thing I would do is get out my digital oscilloscope and clip the leads to the circuit so that I could see the signal to be sure that it's truly a sine wave. I would then use the internal processing power of the 'scope to digitize the parameters of interest from the circuit and output them from the 'scope to my PC for further analysis.

Good luck in your research.