# How to measure the frequency response of a waveform generator?

I was looking at the data sheet of an arbitrary waveform generator and found a plot of the frequency response of the instrument, then wondered how it was measured. I know how to measure the frequency response of a simple electrical circuit. Basically I have an oscilloscope that can perform frequency response analysis, so I connect channel 1 to the input of the circuit and channel 2 to the output, then use a sweep sin wave over a range of frequencies, then the oscilloscope handles everything else and I get a bode plot of gain and phase. However, doing the same for the waveform generator will require taking apart the casing and some internal parts to get the circuitry, which may damage the device, let alone knowing where to connect the probes.

So does anyone have a suggestion how to measure it?

• What do you even mean by "frequency response of a waveform generator"? That doesn't sound like a well-defined concept to me.... Oct 17, 2021 at 17:49
• depends if it has an external frequency control. If that happens to be a knob, for example, you set the knob to a selection of frequencies in turn, and note the output voltage for each in your lab book. Later, you can plot these on graph paper, and join them with a smooth curve. Hey presto - a frequency response graph! If it has an input allowing voltage to control the frequency, and your scope has an X axis output (ramp voltage) it gets even easier. Oct 17, 2021 at 18:06
• Why do you need to take it apart to measure a sine wave? This is not making sense to me. Oct 17, 2021 at 18:18

'Frequency Response' has to be understood in context.

For a circuit with an input and an output, it's as you say, the complex ratio of the output to the input signal versus frequency.

For a generator with only an output, it's ratio of the output power to the requested output power versus signal frequency.

To measure it, you'd connect it to an ideal receiver, and sweep its output frequency over its advertised range. The problem with this method is that you're measuring the response of both the generator and the receiver together.

If the receiver is an oscilloscope, you might hope that it's reasonably flat up to 50% or so of its advertised maximum frequency, as long as you are taking care to terminate any cables you use for connection.

For precision measurement of a source over frequency, we usually use a power meter, which has a far simpler construction than an oscilloscope or spectrum analyser, and is easier to make wideband.

• If I understand correctly then for a generator, it is the ratio of power of the signal I want to generate, to the power of the generated signal versus frequency. But, the generator will ouput a signal identical to the one requested for the same frequencies, or not? So the power will not be different, or am I missing something? I also never heard of power used in frequency response analysis, do you have any reading material on that? Oct 18, 2021 at 11:12
• @Valdi In an ideal generator, the output power will be what you asked for. In a real generator, it will be slightly different. Even in a proper;y calibrated generator, you can expect the usable maximum power to decrease as the frequency increases. Oct 18, 2021 at 11:26
• Isn't the power an integration of voltage times current over a fixed period? If power decreases as frequency increases, then voltage or current must have also been affected. Then we should be able to use voltage to measure frequency response, so why are you refering to power? I don't understand this! Oct 18, 2021 at 22:25