Hi Dave, I think you need to review how to calibrate the 720A and Kelvin Varley dividers in general. The loading you need to worry about is the load you put on the output terminals, irrespective of the particular decade. I do not know how you are measuring the output, but it would appear, from your list of values, that your method of measuring the output is loading the divider, particularly as the value of error increases as you step up the "A" range. The divider was designed to look into an infinite impedance. I.E. a null detector driven from an equally stable and known DC source. Like, for example a 332 or 335 Fluke DC voltage standard. Any loading is going to drag down the output value and it will get worse as you go from the bottom of the range to the top. The calibration process assures that the ratio from one step to the next is equal. Your values indicate that there may be other issues you did not properly account for. Good luck with getting it straightened out. Bill....WB6BNQ David C. Partridge wrote: >I've added resistors in series with R1008 and R1044 to bring the resistance >for those parts of the A divider to a value that it would calibrate OK (two >of the resistor sets in the oil bath very low in value). I also had to >remove five turns (about 1.3 Ohms) from R1051 as the combined value of R313 >and R314 was too high to allow me to calibrate S2. I also replaced R1041 >as it was worn out (almost impossible to get a null). > >I've successfully calibrated the C decade as per the maintenance part of the >manual. > >Sadly the beast isn't giving me a nice linear voltage sequence when using >just the A decade. > >Input voltage 100.001,4V fed to 0/1.0 Input terminals All Decades set to >zero except Decade A. > >Decade A Output V > 0.1 9.999,88 > 0.2 19.967,8 > 0.3 29.936,7 > 0.4 39.903,8 > 0.5 49.874,7 > 0.6 59.855,6 > 0.7 69.852,4 > 0.8 79.871,3 > 0.9 89.919,6 > >See also attached graph. > >Any suggestions on where to look for trouble will be most welcome ... > >All sections of the A decade attenuator now measure about 9.999,2 ohms. > >It was suggested on eevblog that the next decade (or a subsequent one) was >loading it, but with all the other decades set to 0 I don't quite "get" how. > >Thanks >Dave > > > > ------------------------------------------------------------------------ > >------------------------------------------------------------------------ > >_______________________________________________ >volt-nuts mailing list -- volt-nuts@febo.com >To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >and follow the instructions there. >

Drats! That was most of the problem! Moving down to 10V where the 3458A has an input impedance of >10GOhm, and I see much more sensible results. Thank you Dave -----Original Message----- From: volt-nuts [mailto:volt-nuts-bounces@febo.com] On Behalf Of wb6bnq Sent: 07 August 2017 11:19 To: Discussion of precise voltage measurement Subject: Re: [volt-nuts] Update on 720A Hi Dave, I think you need to review how to calibrate the 720A and Kelvin Varley dividers in general. The loading you need to worry about is the load you put on the output terminals, irrespective of the particular decade. I do not know how you are measuring the output, but it would appear, from your list of values, that your method of measuring the output is loading the divider, particularly as the value of error increases as you step up the "A" range. The divider was designed to look into an infinite impedance. I.E. a null detector driven from an equally stable and known DC source. Like, for example a 332 or 335 Fluke DC voltage standard. Any loading is going to drag down the output value and it will get worse as you go from the bottom of the range to the top. The calibration process assures that the ratio from one step to the next is equal. Your values indicate that there may be other issues you did not properly account for. Good luck with getting it straightened out. Bill....WB6BNQ David C. Partridge wrote: >I've added resistors in series with R1008 and R1044 to bring the >resistance for those parts of the A divider to a value that it would >calibrate OK (two of the resistor sets in the oil bath very low in >value). I also had to remove five turns (about 1.3 Ohms) from R1051 as the combined value of R313 >and R314 was too high to allow me to calibrate S2. I also replaced R1041 >as it was worn out (almost impossible to get a null). > >I've successfully calibrated the C decade as per the maintenance part >of the manual. > >Sadly the beast isn't giving me a nice linear voltage sequence when >using just the A decade. > >Input voltage 100.001,4V fed to 0/1.0 Input terminals All Decades set >to zero except Decade A. > >Decade A Output V > 0.1 9.999,88 > 0.2 19.967,8 > 0.3 29.936,7 > 0.4 39.903,8 > 0.5 49.874,7 > 0.6 59.855,6 > 0.7 69.852,4 > 0.8 79.871,3 > 0.9 89.919,6 > >See also attached graph. > >Any suggestions on where to look for trouble will be most welcome ... > >All sections of the A decade attenuator now measure about 9.999,2 ohms. > >It was suggested on eevblog that the next decade (or a subsequent one) >was loading it, but with all the other decades set to 0 I don't quite "get" how. > >Thanks >Dave > > > > ---------------------------------------------------------------------- > -- > >----------------------------------------------------------------------- >- > >_______________________________________________ >volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to >https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >and follow the instructions there. > _______________________________________________ volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts and follow the instructions there.