My Fluke 731B came from the Washington Naval Yard, Washington, D.C., Eastern Standards Laboratory, vintage 1976. As received, the battery pack had been removed, the unit operated without batteries for some years as I had not opened the unit. A few years ago I opened the unit, with the exception of the batteries, it was in excellent condition. I replaced the battery pack with NiMh, 3.3AH batteries, made a slight change in the series resistor charge circuit and also made some minor changes to the op amp circuit to reduce the noise a little bit at the output. The noise specification in the manual appears to be quite conservative.
It is important to note that the battery circuit aids in reducing noise in the power supply and voltage outputs. Without batteries, significantly more noise is passed on to the voltage reference circuits. If you don’t require battery operation (output noise is slightly better on battery power), a zener diode of the correct voltage paralleled with an electrolytic of moderate size will produce similar results to the batteries on line operation.
My unit has undergone several calibrations in the past few years to determine long term stability as my unit came without any cal data unfortunately. Initial calibrations were done at 6 month intervals and then at one year intervals. There has been very little drift over these measurements, the 1V output was adjusted once the last time, it was 100µV low, it appears that R19 may have developed a weak spot where the wiper was sitting, it was noted there was some slight variability at that setting (probably worn spot). It will be rechecked at the next calibration and replaced if needed.
These are the last set of readings for last year’s calibration; note that the 1.018V and 1.019V outputs were not readjusted from the previous calibration.
20C ==> 9.9999940 / 1.0000005 / 1.01806359 / 1.01910223 VDC
23C ==> 10.0000065 / 1.0000027 / 1.01807099 / 1.01910386 VDC
You will note that there is a small variation in the 10V output with temperature of 1.25PPM equaling a TCV of 0.41667PPM/°C, not too shabby for an unregulated temperature device. Al voltages were stable over time while at calibration. It was also noted that noise in the readouts were limited to 2-4 counts.
The 731B is also pretty much immune to being cycled on and off (for hours), it has repeatedly returned to the same value after being on for an hour or so, the cal lab did this at my request and those are the findings, any changes were too small to be significant, hidden in the noise.