Both the 3456A and 8505A have been on for approx 18 hours and I have some interesting results.
Test set-up. One of the AD584-M's powered by USB and powered continuously during the test. It was set to the indicated 9.99691V. That voltage is achieved at 25 C. I noticed during the Winter months that both my AD584-M's read lower due to ambient 21 C and take a while to warm up.
Results: Within approx 2 hours the 3456A stabilized and didn't vary all night. The 8505A did not fair well. It started out agreeing with the 3456A within one or two digits but as the test progressed it's measurement kept continuing to drift lower and lower and it wasn't until maybe 12 hours into the test that it stopped. The 3465A stabilized at 9.99693V to 95. The 8505A finally stopped drifting at 9.99676V to 78. In real world applications that doesn't mean a hill of beans. But to a volt-nut that's tragic. And if you compare it to the other three AD584-M voltages it begins to tell a story....
AD584-M 3456A 8505A
2.50066V 2.50061V 2.50062V
4.99794V 4.99792V 4.99790V
7.50011V 7.50012V 7.50000V
You can clearly see a pattern here. As the voltage increases the 8505A becomes more and more inaccurate.
And as further proof that something is amiss with the 8505A. Here is the other AD584-M
AD584-M 3456A 8505A
2.49937V 2.49938V 2.49938V
5.00030V 5.00034V 5.00030V
7.50042V 7.50050V 7.50038V
10.00096V 10.00108V 10.00094V
I don't have the ability to perform a full calibration on the 8505A and I don't think it's worth the effort and expense to send it to a cal lab. And unfortunately there are no cal labs within reasonable driving distance. But I suspect there's more going on here than just a calibration. The non-linear delta as the voltage increases is very suspicious. And in order to troubleshoot this DMM you need an extender board which is almost impossible to find. So my choices are sell this 8505A as a parts only unit or just quietly retire it and hope that one day an extender board shows up. I haven't decided what to do yet.
And as a side note. Speaking of the Siglent SDM 3055. I've decided that when the budget allows I'm going to sent it in for a full calibration rather than attempt to “tweak” it myself. And here it is measuring the same 9.99691V after it's been powered up for over 12 hours. I would expect at least a reading of 9.9968V or 69. And it's off at least that much on all the other references.