Robert Baer said:
* ---^--- Decent idea; is there any experience as to the recovery time?
Saturation doesn't work that way -- takes flux (volt seconds), not
voltage. At best, you could clamp the later part of the spike, once the
ferrite's charged up (so it tightens to a narrower pulse, but the peak
value remains). You still need freewheeling to discharge it, which means
as soon as that pulse is gone, you're going to have at least as big of a
tail (at minimum, the waveform will look like dV/dt with a peak voltage
similar to the original; at best, high frequencies will be attenuated in
the ferrite and it will relax slowly).
A rectangular pulse, 600V * 4ns, is only 2.4 uVs, so you'd need quite a
small core (about 8 mm^2 cross sectional area, about what's on a large
ferrite bead-on-lead). I don't think any ferrites go fast enough, anyway;
even the highest are all but RFI shielding by those frequencies. It would
work (in the sense of doing what it will do, not necessarily what's
intended
) on pulses of, say, 10V and 100s of ns.
I have another question for the OP:
If these signals are so immensely weak... how is it even possible for the
end purpose to care? How does it know? Surely there will always be some
leakage between buckets, and the goal is simply minimizing the amount, and
mitigating the error resulting therefrom?
Since this is physics we're talking, wouldn't it be possible to add two
"kickers" to the line, timed to the big pulse, which neatly remove the
leading and following buckets? If that requires modifying or lengthening
the line (or, if the buckets are already zooming along near the speed of
light.. it may require a bend!), well... whoever designed this should've
thought of it in the first place..?
Tim
