sooooo.... i have four 5.5 digit DMMs, including a Fluke 8840A, on my bench. given the limits of my volt nuttery, i am happy that they all measure the voltage of a DMMCheck within 0.004 volts but...
they are kinda, i dunno, drifty, fluctuating, each and every one of them, in the least significant digits. and the changes are fast (every update when on internal trigger) and seemingly random (doesn't seem to be consistent in amount or direction). they don't settle down if I let them warm up. i guess it is a noise issue (WAG). i haven't done a lot of interwebbing to see if there is a reason or a fix and figured i'd ask the collective brain trust here for some guidance before i crack the cases open.
edit: to be clear, i see these fluctuations with the leads shorted, too, though the delta is a lot smaller. it is worse on the fluke than the HP meters. so it seems to be more than just the random nano current across a large resistance...
Could be external noise, e.g. spikes.
Could be internal noise, e.g. caps or the voltage reference or dirt/deposits.
My Solartron 7081 recommends that after being left in storage it is wrapped in a blanket, turned on, and left for a day. The internal temperature rises to ~40C. Mine improved significantly after that process.
Like wise both my 5.5 digit DMM's, Siglent SDM 3055 and Fluke 8810A, need minimum 12 hours power on after a cold start before achieving best accuracy and stability.
I've always accepted that the more digits you have on a meter, the more the least significant ones will change and this is to my mind, to be expected as external circumstances will begin play a part. Lets examine if you will a set of scales, as the scales or transducers get more sensitive to measure minute weights, the more external events like breeze, vibration temperature etc will cause them to fluctuate their readings so shouldn't similar forces play a part when it comes DMM's?
You can get some very cheap 3.5 digit meters that will agree 100% with say a Fluke 3.5 digit meter but as you get into the world of 4.5 and higher digit meters, the cheaper brands drop off rather quickly because they are far more complex to design and make to be reliable and this can demonstrated with my own set of meters, I can put a shorting plug between common and the V and
and the last digit on all but the 3.5 meters will change over time and I only go upto 6.5 digits so if your talking about 8.5 digit meters then this level of accuracy is a magnitude more difficult to obtain and to be honest, I'd be perfectly happy if I had such a meter for the last 2 digits to change.