Running equipment 24/7? Particularly (right) after calibration...

[Rax]

As I'm waiting to get into the fray of "Round 3," I am starting to think on my current equipment, its state of calibration, best methodologies I should employ, if I should get a case of single malt or vino spumante for the celebrations. A lot of things to think about, so many decisions to make.

One thing I could use some input on is the best utilization for newly calibrated equipment. My reference meter is currently a freshly adjusted, and confirmed/verified calibrated Fluke 8502A. This was stickered (Z540, so with measurements, which have been invaluable for "objective" reference points in my environment) a few week ago, and I've been running it almost constantly since then. I have not really been concerned with its possible drift from the "characterized" values by calibration, as it's such an old piece of equipment I assume most drift has already occurred. Besides, most these were used in systems applications, running under automation and GPIB-controlled, and so designed for 24/7 operation.

Also, I've been busy checking and transfer-adjusting a couple of Digital Precision 8200s, a Fluke 8505A having been repaired and now adjusted, etc. So there wasn't very much downtime through this.

So what are experienced metrology hobbyists (and professionals for that matter) doing? Are instruments going off for extended periods at a time? Overnight? Staying on all the time?... I'd hate to run this too hard and have it drift a bunch right before getting the US CAL CLUB reference to examine here.

[alm]

If you run equipment only sporadically, then the drift will be low, but there is a small chance that a voltage reference will give a slightly different voltage after being powered off and on again. That's why commercial voltage references like the Fluke 732 series have a battery backup and an "in cal" light that turns off if the unit ever loses power. But I'd expected this hysteresis to be in the order of 1 ppm, and it appears to be somewhat rare based on anecdotal reports. So I'm not sure how much a concern this is for a DMM.

If you run equipment 24/7, then drift will be higher, but it will be more predictable. What I do is run primary references, in my case voltage references, 24/7. But run bigger things like DMMs and calibrators only when I'm doing experiments. I'll let them warm up for at least a few hours, if not over night, and will run them continuously for the duration of the experiments.

[Conrad Hoffman]

I have three Fluke 731s that are powered 24/7 and have been for a decade or so. They've needed various service over the years, caps and such. I don't use batteries in them, just large filter caps with zener voltage limits. The truth of the matter is I've never seen any significant change when they get powered down, or after a power failure, so I don't worry about it. I have one 3200 (8200) that only gets powered when I need it. For anything serious I'd power it several hours in advance, longer if calibrating. My ancient HP3455A, which is really excellent, is the same. Not relevant here, but my HP frequency counter with an ovenized crystal oven also stays on standby power 24/7. Alas, every single thing, plus my 845, needs a full checkout before round 3 of the traveling standard. What I dread most is the switches on the 3200.

[RandallMcRee]

I am keeping the Keithley 2002 powered on 24/7. Checking its drift using a bank of LTZ1000s (always on) and two 731Bs (powered on/off) and using resistance standards as well. Other meters get turned on and off.

As alm says, these other meters are on during and way before experiments.

The cal club is a great thing but you will definitely want to create or buy your own voltage standards. You (rax) already have resistance standards You are way down the rabbit hole. 

Thanks,
Randall

[Rax]

I am keeping the Keithley 2002 powered on 24/7. Checking its drift using a bank of LTZ1000s (always on) and two 731Bs (powered on/off) and using resistance standards as well. Other meters get turned on and off.

As alm says, these other meters are on during and way before experiments.

The cal club is a great thing but you will definitely want to create or buy your own voltage standards. You (rax) already have resistance standards You are way down the rabbit hole. 

Thanks,
Randall

Randall... I think you're right! It hasn't been that long, and yet I can't even recall when last I saw sunshine... 

At this time, the best I have for voltage is the 8200 calibrator. I feel darn lucky with it, but to serve my ESI RV722 right, I think I absolutely need to consider at least a 731... Which is a bit of a problem, as all decent voltage standards are pretty pricey.

That said, I just received my ADR1399KEZ, so maybe the DIY path is in my future (it tends to dominate, by far...).

Thanks a bunch, Randall!

[vk6zgo]

Running things with SMPS 24/7, especially older stuff, (1980s/90s) can be a trap.
Some power supplies of that vintage will run continuously with no problems, even if they have developed a fault in their start up circuitry.

It only takes one power break for the SMPS to refuse to start.
At my old work, we had four Barco Picture monitors in the Master Control area, which were powered up 24/7.
Following a Mains break, all four refused to start.

This was the occasion for much scurrying around finding temporary replacements.

All of them had exactly the same fault, so after that, I made it a habit to stroll through MCR flicking each one off at odd intervals.
Fixing one, was a lot less problem, than bulk failure.

[GigaJoe]

running 24/7 expensive equipment, would be wise to use at least AC mains spike protection,   or better noise filter , or combination of good sinewave UPS and spike\noise filter.

a while ago a fuse blown in my freq. counter, when lighting was,   so i scratch my head, and retrofit all power cord extender with caps, gas discharge tubes, and TVS MOV. 
for my lab i accidently found a couple heavy industrial AC filters with large toroidal filters inside, caps,  i add some MOV, TVS, GDT as well   filter even has LCD counter of events.

expensive UPS has some reasonable TWS circuit, and switching load when over\under voltage.  used one cost around 100-300 depend on type...

or build your own, with isolated transformer, taken from 2 cheapest broken UPS , adding filer and stuff listed above ...