Hal, replies inline. On Mon, Apr 4, 2011 at 10:21 PM, Hal Murray <[email protected]> wrote: > >> Here are the results I got using your suggestion. The numbers may be a bit >> different than last night as I'm not sure I'm using the same cable. > >> A+, B+ = 18.9 ns >> A-, B- = 19.4 ns >> Obviously there is some difference in delay between the A and B channels. >> Otherwise, the two numbers would have been identical - correct? > ... >> If I take the average of the two readings, I get 19.05 ns, which is more >> precise than the readings I'm taking. I'm rounding to the nearest 0.1 ns on >> the readings. > > There are several possibilities for asymmetry. > The input signal could have other than a 50/50 high/low ratio. > The trigger levels could be set at other than the half-way point.
See my reply to John about trigger levels. > Do you have a scope? Yes, I will dig out the scope and check. The last time I checked the Spectracom 8140 (long ago), it was ok. I'll check it again. > There are probably second order problems with different rise/fall times. > > >> A+, B- = 5014.6 ns >> A-, B+ = 5023.5 ns >> If I take the average of these two readings and subtract out the 5 us for >> 1/2 period of the 100 KHz square wave, if get 19.15 ns. This agrees >> very closely with the above average. > > I read that as everything mostly works. That was my assumption, but it's nice to have someone more knowledgeable confirm it. > The question is how well does it work and/or what can you do to get > more/better data? > > You can get another set of numbers by swapping the A and B inputs so the > signal goes from B to A rather than A to B. > > Some of the errors might be more obvious if you deliberately offset the > trigger levels. > > NB: That's for all 4 modes: > A+ to B+ will be slightly less than a whole cycle. > A+ to B- will be slightly less than a 1/2 cycle. I'll try swapping inputs and see what happens. I'm not sure if I'll get to this tonight. If not, tomorrow night. Joe Gray W5JG _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
