The output filter probably has a sharp roll-off.
Nico, what would make you think so?
With a bandwidth of 100MHz a square wave of 30MHz won't look much like a square wave since only the 60MHzand 90MHz harmonics get through the filter.
True, IF it had only a 100 MHz BW. However, if you look at the specs you'll see it claims it can do 40 MHz square waves, and 30 MHz pulses, and that's simply not possible with a 100 MHz BW.
Hmm, those are the new limits, according to the OP, but Hanteks specs are quite different, even for the 100 MHz size models... basically
an anemic 4 or 5 MHz for
everything except sine and arb.
But that doesn't make any sense, because arb is rated to 30 MHz, with a 6ns rise-time. And the duty cycle of square waves is rated at 40-60% at "10-40 MHz", and 50% beyond 40 MHz? And Pulse periods are rated down to 33.3 ns (i.e., 30 MHz). So I'd say the "4-5 MHz" part of the spec chart is incorrect, and the OP has got it right.
To achieve those results would mean the output filter would be out somewhere around 400 MHz, not 100 MHz.