I'm not so worried about the power consumption but more interested in the instruments stable state. Something as sensitive as the 3458 seems to need some real time to "settle" into its most stable state if it's running correctly. On one hand I feel like I might be using up my 3458's life but on the other I know this is it's stable state after running for a few months non stop and I can characterize it's base running capabilities without wondering if it's still going through a recovery from being turned off for an extended period. In many ways it's similar to the way an OCXO needs time to become frequency stable (as much as it ever will be) after being turned off and it's oven cold... maybe a simple analogy but that's how I like to think of it.
Bill
Bill,
the 3458A generally reaches its stable state (i.e. its specification limits) after a 2..4h warm up phase, with constant room temperature and an ACAL. There is no further settling time like an OCXO always shows.
Only if you make observations beyond (below) its specification limits, then you might observe subtle relaxation / hysteresis effects.
Practically speaking, when I let my 3458A settle longer, maybe 12h over night before making measurements, one can observe the last tenths of ppm drift, mainly caused by better stabilization of the temperature of the instrument or the room.
The question arises, what is really relevant for you, and which quantitative limits you are able to observe.
These limits you should define for your use case initially.
Also keep in mind, that running 24/7, the old version really consumes the lifetime of these EL2018 comparators, because they were getting quite hot, which is not the case any more with the Black Edition.
The inner temperature of the B.E. rises about 5°C less than the old version, due to use of modern components.
Let me illustrate this problem a bit deeper with several examples from practical experiences.
Remember, that no DMM is accepted as a reference standard (for Volt, Ohm) by any Standards Institute, even if Fluke or HP have once promoted the 8508A or the 3458A as metrology grade instruments.
Reason maybe that their internal references drift too much in comparison to pure DCV and Ohm standards.
The 3458A is used for metrology purposes at Standards Institutes as a Transfer Standard mostly, using its ultralinear A/D of < 0.02ppm INL, and they let it also run 24/7 to avoid the warm up and stabilization time of 2..4h.
Of course these 0.02ppm INL are not officially specified, but they were able to characterize each of their instruments, to achieve much lower specification limits, maybe including timely and temperature stability of the LTZ1000A reference and of the A/D.
Like the Standards Institutes, us volt-nuts may also be able to characterize and use the 3458 beyond (below) its specification limits, see TiNs efforts and measurements on his "Golden" DMMs, with near zero ppm drifts, and proven 0.02ppm INL.
My 3458A, with an oven set to 65..70°C and intermittently used, shows a DCV drift of only 1.8ppm over 7 years.
I'm comparing my continuously running LTZ references (and my VHP202 resistors) once a month, and the achieved repeatability, provided by the 3458A, is on the order of 0.2ppm only. If it would be powered 24/7, I would not expect better numbers, which are already below the limits of what the specification delivers.
In contrast, crucial was the recent use of low thermal PTFE cables with reversed measurements, which obviously improved this repeatability.
I observed greater shifts up to 0.5ppm, reversible and non-reversible ones, when either the 3458A, or the references were physically transported.
Otherwise, judging from these limits, I could not observe any further instability disadvantages, or hysteresis effects by intermittent operation, which you might have suspected.
You should define for yourself, which metrological drift / stability parameters are relevant for your, by defining quantitative limits.
Then you should perform a reality check, if these limits are really affected by intermittent operation, I doubt that, and what your own capabilities are concerning stability figures.
I mean, if you expect e.g. 0.1ppm stability for a specific parameter, and you were not able to check that, or your other equipment is much less stable, what would be the benefit?
The 3458As (sub ppm) stability, on 24/7 or intermittent operation, in the end will always play a minor role only, I assume.
Frank