I own a 1996 since 20 years and use it on a regular base. Mine will give a wrong measurement result from time to time. May run flawlessly for hours and then give one complete false reading. May be software, may be hardware. Remember the electrolytic capacitors in it may be 30 or more years old, having most probably lost some of their capacitance. I can well imagine that the device reacts today more sensitive to anything on the power line than when it was young.
Regarding the second part of your question: Imagine the 1996 basically as a time interval counter about which Racal Dana claimed that is has an single shot resolution of 1 ns. The 1 ns is a bit conservative. If one dives deeper into measurements with the 1996 and gets the COMPLETE results over the GPIB (which has more digits than you see in the display) then it becomes more likely that the true single shot reolution may be in the order of 300-400 ps but with an estimatiom of 1 ns you are on the safe side. Since the 1996 is intelligent, every measurement that it makes, regardless of being frequency, time intervall or phase can be reduced to an time interval measurement for wich the counter has a certain time availabale, i.e. app. the gate time you set it. If you apply the 1 ns absolute resolution to the measurement time availabale you come to a relative resolution of 1E-9 @ 1 s gate time, 1E-10 @ 10 s gate time and 1E-11 @ 100 s gate time. Up to that it is physiscs and electronics that increase the relative resolution. You may apply additional mathematics to your measurements but that will not increase the physical measurement resolution. Since the 1996 will always try to give you the full available accuracy and resolution of an result, there are chances that the 10 digit display is not adequate to display all significant digits. In this case the 1996 "overflows" which is by no means an indication for an error but simply for the fact that not all of the digits are displayed (while they are on the GPIB). In this case the "O/F" led will turn on and indicate that the number displayed has 1 or 2 digits left of the most significant displayed digit that are not to be seen. However the missing digit(s) are easily guessed: With 10 MHz applied and a display of 0.000000001 E 6 the missing digit is clearly a "1" while in the same situation an 99.99999999 E3 clearly shows that there must be two "9"s that we are missing. Best ragards Ulrich Bangert, DF6JB > -----Ursprüngliche Nachricht----- > Von: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Im Auftrag von tom jones > Gesendet: Samstag, 3. März 2007 02:39 > An: [email protected] > Betreff: [time-nuts] Dana Racal 1996 Questions > > > Anybody using either 1996 or 1995 racal counters. > I use my hp5061a for external reference to both counters > The racals are continously in mean average mode (10000 > seconds).with a different > 1pps signals to the racal counters.. > Occasionaly the counters present a reading like 10.0 -3 (10ms) > The expected readings should be 0.001 -9 (1ps) > > Anybody else using dana racal experience this. It seems > like a software bug? > > Question number two; > The readings from these racals seem off . > I get a reading like 0.001 -9 equals 1X10 -12 I belive > its decimal point mighy be off. > I'm thinking it should be .0001 -9 equaling 1X10-13 When > I interpet the readings with the extra digit of accuracy the > readings make more sense and correlate with my austorn loran > frequency & timing receivers and other counters and gps 1pps signals. > > To also support my theory the racal counters will resolve > 1X10 -11 using 99 (100s) second gate time. when I activate > the mean average mode this extends the gate time two more > decimal places (10000 second gate) wouldn't you assume a > resultt of 2 more digits in accuracy ? > > > > > > --------------------------------- > Don't pick lemons. > See all the new 2007 cars at Yahoo! Autos. > _______________________________________________ > time-nuts mailing list > [email protected] > https://www.febo.com/cgi-> bin/mailman/listinfo/time-nuts > _______________________________________________ time-nuts mailing list [email protected] https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
