Got a bunch of test gear? How often do you get it calibrated?

[Lightages]

As I start accumulating a pile of new gear I start to wonder what I need to do to keep on top of things. It seems to me that if you are serious about your test gear, you should also be serious about keeping it within spec.

I have a number of voltage standards and a DMM Check, so I can check my multimeters, but I have no frequency standards, capacitance, nor current standards.

What do you do and what solutions have you found for your situation. Do you have a good way of dealing with certain calibration needs? DIY?

A true multi-instrument calibrator cost thousands of dollars. I am not likely to buy something like this because I cannot recover the costs by hiring my services out for calibration where I live. There just isn't any business for that here.

I was thinking of getting a good bench meter with much higher accuracy and stability than all the rest of my gear and then maybe send that out for calibration when needed. The problem is having all the different stable sources at the values needed to use to calibrate the other devices whilst using the accurate and calibrated meter for a reference.

So let's hear what you do.

[zapta]

If the display comes on then it's perfectly fine.

[VK5RC]

Never, but a bit of gear,  some old - some new,  so I  cross check especially after new purchase of older gear,  so far nothing that bad to bother re-calibrating (if not broken).

[pickle9000]

In my experience most issues (measurement related) occur from not checking gear. If you plan on using a scope connect it up to your sig gen using the probe you plan to and see what you get. Do it  before you start the days work, simple and worth the effort.

I have some precision resistors, a rubidium source and a few other things in a box with marked values so I can do rough equipment checks. Just to make sure I'm not crazy.

[Negative-Bias]

I  guess if you run a company where  precision matters it's maybe worth the cost,some you tubes  channels are a bit anal about calibration,personally :I'm not bothered if my test gear are a couple of counts out,sometimes I press the self/cal button on the scope. What I'm trying to say is life is far to short ,In my situation  TV's come in broken and leave working = food on the table,bills paid.

[mikeselectricstuff]

I have never calibrated any of my gear, but rarely have any need for accuracy, and have enough stuff that there is plenty of scope for cross-checking between different makes/models.

Scopes are rarely used for anything that requires serious accuracy - the timebase on a digital scope comes from a crystal, which is never going to drift enough to care about, and if one channel gain drifts significantly it will be obvious by comparing with other channels. It's pretty unlikely to get a fault that makes all channels drift the same way.  This is not as true for analogue scopes, but you're even less likely to use these for anything that needs accuracy
 

[HighVoltage]

Some instruments I am using for certification work and I must proof to my customers, that this gear is calibrated.
Other gear I cross check and if needed, I will do the calibration.

I built myself a set of
- resistor references
- got a good 10.0000 V voltage reference
- some precision inductors
- some precision capacitors

I took these items to a calibration lab and got all values and then compared them with my gear.
It can get very expensive to calibrate all gear and in some cases it is totally stupid to do so.
But if the customer requires it, then you have to do it.

[German_EE]

Time and frequency get calibrated on a continuous basis as I have a 10 MHz source locked to DCF77

For resistance I do the same as Dave, three 0.01% devices wired to big chunky terminals, these get checked once a year.

The voltage standard gets checked once a year as well, once this is done I can check current.

RF power can be checked once I have verified my dummy load really is 50 ohms and that my scope is calibrated correctly. 100W is 70.7 volts RMS across 50 ohms.

No standards for inductance or capacitance.

[Electro Fan]

Observations from a rookie with test equipment:

I don't have nearly the knowledge, skill, or experience of many of the veterans here so take this with a BIG grain of salt.  (And this applies to my thinking as an enthusiast, not as a pro who is designing/building/testing/repairing gear for a fee.)

My experience is that some (generally newer and more expensive) equipment will be somewhat more accurate than other (generally older and not as expensive) equipment; however, sometimes older and/or less expensive can be better than newer and more expensive.  - But if it's worth keeping it should all be in the ballpark.

How do you know what is accurate?  You could spend a fair amount on collecting all the gear needed to calibrate all your other gear or you could send it out periodically to have it calibrated.  Or you can spend some $ on products like a DMM check, etc. and use these to check your other equipment (in this case for voltage).  In the process you will probably find that some of your DMMs and other equipment disagree with each, or they might agree at some voltages but not others.  What you will probably come to conclude is that one of your pieces of equipment gives you the most confidence and it will be a "ruler" or reference point.  You might be able to adjust some of your other equipment to come into alignment to some degree with your "ruler" but it's possible that this won't be feasible (or even desirable) with your other gear.

So what you wind up with is a bunch of gear that varies to some extent (maybe hundredths of a volt or thousandths of an amp).  IME, voltage measurements (at least on some equipment) are somewhat harder to get dialed-in with precision (some number of zeros) than frequency measurements.  As mentioned elsewhere in this thread oscilloscopes don't seem to be real good at measuring voltage (I trust my DMMs more than my scopes on voltage).  On the other hand, my scopes and frequency counters and frequency generators seem to consistently line up with one another to a decent level of precision on frequency.  (This could just be a coincidence among my gear.)  Someone here will point out that accuracy and precision are not the same - which is true - but when several pieces of my equipment (most of which have not been tweaked to agree with one another) line up on a reading I begin to have confidence in the accuracy and as the precision goes farther out in unison I get still more confident.  (Even so this is faith and not science since not much of my stuff is freshly calibrated, but hey I'm not launching any mission critical projects on most days.)

I'd really like everything to line up precisely all the time but I've come to the conclusion that I have to work with reasonably accurate values and often just relative or comparative values and/or trends - and therefore I have to settle for concepts or principles being manifested with less accuracy than I would like. To me one of the beauties of test equipment is that it provides tremendous feedback for learning. And no doubt if you had complete confidence in all measured values it would be easier to ferret out the principles at work.  If the budget wasn't a consideration it would be cool to having everything measureable to micros, nanos, picos, and femtos.

I'm pretty new to test equipment (a couple years) but this is what I've found - so far.    YMMV

PS, having said all that, the couple posts just above mine look like pretty good approaches....

[jpb]

It is a bit early for me to answer this question as I bought a calibrated Keithley 2000 and used that to compare my other DMMs. I also got a bunch of (expensive) high tolerance resistors and put them in boxes.

I may send the Keithley 2000 off for calibration again in a year or so.

Timewise GPS, even a simple module, provides high precision especially when combined with a good but cheap OCXO off ebay.

I've just bought a Time calibrator (out of calibration) off ebay - it's no good directly as a calibrator as it is not recently calibrated and anyway is low spec but it provides a very steady voltage/current to compare my DMMs with each other.

I bought a couple of boards of Joe Geller (he's stopped trading but was happy to sell me boards) and will build myself up a 10V reference and may get one of Doug's 10V ones to get a calibration point.

I plan to be more systematic about recording measurements for future comparison.

For a home lab, the key things to have are a DMM that you trust and a frequency reference. You can then check the dc Volts on a scope and its time base, the AC side is more problematic but if the dc bit is ok on a good scope then the ac part probably hasn't drifted much.

[KJDS]

I've seen more gear damaged in transit too and from cal labs, or cals messed up by junior techs adjusting things badly than I've ever seen equipment drift.

Sometimes equipment goes wrong, and if a part is replaced in a critical area then I'd consider getting it cal'd, otherwise I'll keep it safe on the bench unless a customer needs a cal'd measurement.

[G0HZU]

Agreed

At best, it's the equivalent of paying someone to throw your test gear parcel(s) in a transit van and drive (as in shock/vibrate)  it halfway round the country for a few days and then bring it back with a little calibration sticker on it.  At worst, it means that a spotty junior technician may have tinkered inside it to make it 'pass' their calibration limits. Often the limits are so wide that the calibration sticker is of little technical value anyway.

Obviously, a company needs to show/prove to their customers that they send their gear off for calibration every year or so calibration is essential in this case but I wouldn't let a typical UK calibration house near my RF test gear even if they offered me a totally free door to door service including the courier costs.

However, when it comes to home/hobby calibration then I think that many hobbyists get their priorities wrong. Many seem fixated on wanting to calibrate  megadigit DMMs to unrealistic requirements and also wanting to calibrate frequency references to unrealistic requirement levels. I would suggest that this adds little or no value and pretty much everybody could get by with a few basic checks using a cheap voltage reference chip and a few precision resistors.

In my case, I am more concerned about the accuracy of my RF/AF test gear than I am about trying to get silly levels of accuracy from a bench DMM. So I adopt the cross checking approach using tools like precision attenuators, signal sources and some basic physics when checking my test gear for accuracy. But even then, the really important thing to to be able to quantify and manage things like overall measurement uncertainty when trying to make good measurements with test equipment.

This is a key skill that easily outweighs the value of a current 'calibration' sticker on the front panel

[Fsck]

I don't have any gear that can't be calibrated in town so I usually check gear out to make sure they're more of less within spec. if not in spec, I send them out to cal.

[TunerSandwich]

It can never really hurt to have a couple of transfer standards in your lab.  I check before every critical measurement, against one reference or another.  Usually send the transfer standard out when more than one measurement (from it) is off.  Example DVM 1 reads accurate, DVM 2 reads accurate....everything is probably good.  If BOTH are off, then transfer standard goes out.  If one is on the money and the other is off....that second unit gets check against a third and then goes for cal. 

I like to verify ALL T&M gear I plan on using in a project, before any measurement gets started (in that project).  I have found it just saves time, to not have to figure out something is wrong, while you are neck deep in a prototype, or verify production units (if you do that in house at all). 

For time, you can get a nice programmable rubidium reference on ebay for under $300....check it against a known counter that is stable....if it holds there is your timebase (although I have never had any piece of gear actually drift to a point where I thought it needed to go out).....

You can get a few nice thin film resistors (vishay precision are nice) and throw them in an isolated enclosure, with some nice banana jacks.....good enough for most purposes.....or get a nice used decade box you can trust....

Same deal on capacitance, get a midrange value for each order of magnitude you think you might need to measure......stick them in the appropriate enclosure, with the appropriate leads....

AND...the BIG one.....a null detector.....

There is a point where it's prudent to do this stuff, and a point where it's obsessive and wasteful.   

[Richard Crowley]

The response to your poll likely depends a great deal on whether the person answering is a home hobbyist vs. someone operating a business or dealing with customers who need traceable (or demonstrably accurate) measurement capability.  So without knowing who is responding, dunno how any results will be indicative of anything?

[Lightages]

Valid point Richard.

I am going to assume that anyone who sends their gear out every time on the right date is a professional operation and those who have their own full calibration setup will probably be too.

The rest? Well I would assume hobbyists would occupy the other part of the spectrum although I am sure there are some professional operations that don't follow the letter of the law but also won't admit it

[TSL]

Likewise I have a GPSDO that gives me time and frequency and a couple of Rubidium standards for good measure.

As for voltage and resistance, I bought a transfer standard from these guys....

http://www.voltagestandard.com/

Fro a hobbyists lab this is more than enough accuracy

I check everything at least once a year.

regards

Tim

[SeanB]

I use calibrated masspieces at work to verify massmeters ( they have a separate calibration check cycle, this is for QC ) and they get taken in the van by me every 2 years to be compared with a reference derived from the standard kilogram. As I take care with them they do not drift appreciably, though there was a step one year when I took them a week before ( knowing they would not be used that week) and cleaned them then repainted them to get them looking better than the scruffy look they had since I got them. I do have a 1g masspiece which is 1.0000g with an uncertainty of 5mg. By accident the last cal cycle they also did the 5mg masspieces, so I know that they are accurate to the microgram level. To calibrate the one massmeter that calls for a 50.00g reference mass I have to use the 50g ( actually 49.995g) masspiece along with a 5mg masspiece. This does allow me to see the single count difference with removing the foil mass afterwards.

[radiomog]

work equipment is recal'd 1/yr, costs for which are built into the projects in which they're used.

home equipment is on an as-needed basis.  periodically checked for accuracy.
I recently acquired a GPSDO, so that should help things out a bit

[Smith]

Most of my work stuff gets calibrated once a year. Not all, because some is just used for basic debugging not for calibration or adjustment purpose.

Most of my stuff at home is calibrated at work Thankfully I work a lot with the people who do the calibration. I dont recalibrate unless it has been repaired.  I do check my stuff once in a while by comparrison because I have quite a selection of hardware.