I'm using the function generator on my 5444B to make square-wave output at 16 MHz. I find that the actual waveform jumps between two states, each with different on-state duration. I see this both when using the PicoScope to measure the signal, and when using a separate stand-alone scope.
I've included a screengrab below  of persistence mode to show the two state problem. Notice (e.g.) the second pulse has a single rising edge, but two falling edges (one rarer than the other), so the on-state duration is different in the two cases. There's a periodicity of ~1 second to the cycling between the two states. I can record and post a screencast showing this if that would be useful.
Also, at a lower frequency (e.g. 10 MHz), I do not see this problem .
Thanks for your help.
 Persistence mode at 16 MHz showing the two-state problem
 Persistence mode at 10 MHz showing normal behaviour.