Hi If you slew rate limit the square wave (which is reality) you get a sin(x)/x frequency response. It doesn't go to infinity, but the lobes keep going for quite a while.
Things like cables and connectors have upper frequency limits as well. A square wave will only be happy with a linear phase shift. Both cables and connectors depart from a linear phase vs frequency response if you go high enough in frequency. Some get quite nasty. The original Heliax with the constant spacing in the inner supports was really crazy when hit with a fast pulse. It pretty much rang forever and ever… A simple square wave implies a DC level for a switch point. With long runs that gets messy. Most practical systems go differential. You double the cable cost, but get back to a simple switch point. No free lunch. Bob On Nov 3, 2012, at 8:46 PM, Magnus Danielson <mag...@rubidium.dyndns.org> wrote: > On 11/04/2012 01:13 AM, Peter Gottlieb wrote: >> Of course you can't have a perfect square wave! That would imply zero >> transition time > > ... oh, THAT would be useful! :D No trigger point jitter! > > ... and it would be a hell to contain within the cables and connectors we > have, as they leak a lot as you get up into frequency. > >> and since frequency is inverse to time that implies >> infinitely high frequency bandwidth is required to achieve that perfect >> square wave. Getting a "square" wave with a "fast enough" slew rate >> between high and low levels is certainly achievable and better than that >> perfect square wave. Be careful what you ask for, because with a perfect >> square wave you would have such high frequency content that you would >> get induced noise everywhere. > > Indeed. Bob has shown this. > > Another aspect of it is that phase-shifts at higher frequencies may eat into > the phase shift of the zero transition, so it may for some systems give a > worse environmental/temperature dependence than a sine would. > > Then again, if you check your signal properly, a square-wave may be exactly > what you want and need. It's just that again, your milage may vary. > > Cheers, > Magnus > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
