Function Generator Frequency Range for Calibrating Oscilloscope

[TinkerGDS]

I have a very rudimentary question about Function Generators.  I do not have one so I can't test anything to know the answer.  If I want to calibrate a 400Mhz oscilloscope, do I need a function generator that has a range up to 400Mhz?

[tggzzz]

No you don't need a function generator. If you think about the duality between time and frequency, you can simply use a voltage step with a sufficiently short rise time. The rule of thumb is that BW = 0.35/t_r

[rstofer]

No you don't need a function generator. If you think about the duality between time and frequency, you can simply use a voltage step with a sufficiently short rise time. The rule of thumb is that BW = 0.35/t_r

Which presents another problem, where do you find a signal source with less than 1 nS rise time.  Not that they aren't available but what do you buy off the shelf?

[tggzzz]

No you don't need a function generator. If you think about the duality between time and frequency, you can simply use a voltage step with a sufficiently short rise time. The rule of thumb is that BW = 0.35/t_r

Which presents another problem, where do you find a signal source with less than 1 nS rise time.  Not that they aren't available but what do you buy off the shelf?

A 74LVC1G14 has a risetime of that order. Three of them in parallel each with a 130ohm series resistor will drive a 50ohm line with ~650ps (not 650pS, of course) transition time. Make sure they are well decoupled.

I use that when calibrating my Tek 485 (which also has a built-in 1ns/1MHz calibrator waveform).