Thoughts on instrument calibration -- professional vs. DIY, gear needed, etc

[nsubulysses]

So every time I see a Triplett or Simpson or Tektronix something or other that doesn't work, kind of works, or is unknown in functionality for very cheap it is hard for me to NOT buy it. You know, like when you come across a Triplett 630 untested for $20 and it doesn't look beat up. Ok, sure I'll give it a try.

I am enjoying exploring and attempting to repair and calibrate this kind of gear. Wondering if anyone has comments on what could be a good route to try to cal this old gear to the best of my ability.

My thoughts:

1. Pay a pro for a thing or two and use it as cal reference for my other gear. There is a Tektronix site that is not far away. They will cal VOMs and DMMs for about $65 each, although it's $77 if you want cal documentation with it. As a scope $$$ reference I asked about a Tektronix 465 cal, which was about $155. A higher level of NIST cal is also offered. I don't know how much it costs but I don't think I even want to.

2. Use my own references which are probably pretty good, but not "spot on."

I have a
- Tektronix TDS210 digital oscilloscope. This is might be my best option because it is self calibrating and passes all its own calibration checks. I hope it is telling the truth and is pretty accurate. It's still from the early/mid 90s though I think.
- Tektronix 465 which was last calibrated in 2002. Seems basically agreeable when compared to my Tek TDS210
- Fluke 27/FM true RMS DMM. These are totally old but I have two that were army calibrated semi-recently. 2008 and 2012.
- BK Precision 2709B DMM. This the only piece of new gear I own so I would hope it would come from production pretty accurate. All other test equipment I have is like 15 years old minimum. This meter is is 18 months old and reads slightly off with the Flukes!!

3. Purchase semi-espensive gear that I'm not that familiar with as a true reference and go DIY. Maybe stuff like this.

- Extech Voltage/Current Calibrator -- Does 0-24mA and 0-20V
- Fluke 382A Voltage/Current Calibrator -- Does 0-2A and 0-50V.
- Martel BC 2000 Voltage/Current calibrator Does 0-200mA and 0-20V.
- Buy precision resistors or calibration resistors for ohms settings

[Chuck H]

Repairing and calibrating scopes is beyond my pay grade, but I do have a logical consideration on the matter. Why not pay to have one scope calibrated, the one with the tightest specs, and then use THAT as a reference for calibrating all your other repaired units.

[Mick Bailey]

Something that springs to mind about calibration - not test gear, but crimp tools;

I bought a job lot of crimp tools from a nearby Air Force base. I asked the guy why they were being sold and he says they'd failed calibration. They looked brand-new to me and I bought the lot for next to nothing. When I looked up these tools, most of them were £400 each at least - some much more. Not one of them give a poor crimp or had anything remotely wrong from an automotive/amp repair perspective. I kept what I needed and moved the rest on. They'll outlast me and have done countless motorcycle and car re-wires and amp terminal installs. Then again, nothing I do is expected to perform whilst under enemy engagement or at high-altitude.

So, I'm think about the accuracy of an instrument relative its intended use. For my work on amps and guitars I can't say I've ever needed a calibrated instrument - just one that actually works to a reasonable degree of accuracy.

But if you're looking to calibrate your own gear as an interest in its own right, then you may want to establish some reference standards such as voltages, resistances, frequencies and a means of standardizing temperature at a reference level 20c or whatever. I used to have some physics lab standards I got from the local University that would have been useful in calibration - a 1v cell, resistance and capacitance standards etc.

[km6xz]

Most lab style test instruments have extensive calibration procedures in the manuals, particularly if older when shops/labs were regularly calibrating their gear. One or two reference standards are enough to calibrate everything, say a time base and a precision resistor. But unless you are working on precision instruments, having high degrees of accuracy, precision and resolution is a waste of time. Three significant figures is all that are valid anyway in electronic circuit....except for time, where high res and precision time bases are more common. 10 digits and 1 year stability of 1x10 to the neg 9th is possible with TCXOs and more with aged crystals in ovens. Whatever you design, using 0.05% parts will be lucky to come out to a couple digits of prediction, no matter how well optimized a model is used. One good, well calibrated meter in the shop is enough to calibrate the rest of the general meters and scopes. Analog scopes are intended as precision measurement devices, meters do much better if you need accuracy.
The worst error inducing element of a typical shop is probes and lead resistance. As probes wear, corrosion, oxidation and wire strands breaking all build up gradually but cheap modern Chinese probes start life worse than those 30-40 year old Fluke lab series(rack or counter top style, not the portable style)or H-P will still be better. About once a year I take a day to check general cal consistency between meters and the lab type seldom need adjustment. The biggest variation are true rms readings and resistance(only test accuracy using Kelvin connections/ 4 wire connections so your probes do not throw everything off.).
If you want to calibrate time bases the NIST maintains time and frequency broadcasts that are more than accurate enough using short wve frequencies 5, 10, and 15 MHz at high power and 20 Mhz and a new test 25Mhz carrier frequency. For more accuracy, WWVB broadcasts at 60Khz. WWVH located in Hawaii broadcasts on the same frequencies as WWV located in Colorado. Those signals carry a lot of time information such as absolute carrier frequency, time interval(seconds) and tone, used to calibration audio generators. Any instrument that measures frequency or time should be calibrated against one of these periodically. Cheap digital scopes use internal time bases to get the reference for the digital display of frequency or time and they drift. Check against these stations and put them back on time. That time base impacts every measurement it makes so calibrate the time base, clock oscillator first.
Most of the newer cheap instruments do not have range calibration so if it is off on one range, you are either stuck with it or need to change some divider resistors.
In my old large shop I collected hundreds of HP, GR, and Tektronix test instruments for example 54 HP 400FL AC volt and DB meters, dozens of 465b scopes, and 73xx and 74xx series lab scopes and RF gear for up to 18ghz so that cal exercise was more time consuming but also easier because I also collected lots of reference standards for volts, resistance, capacitance, inductance, time etc. For repair it really was not needed but some people collect guitars, I collected test gear and played with it.

[Leo_Gnardo]

For low voltage DC scales, the old field test was to measure a brand new D-cell, though I'm sure a C or AA or even AAA would work - but it had to be the zinc-oxide type, standard back in the day. 1.56 volts, maybe a hundredth or two more or less depending on temperature. Dynaco made 1.56V DC their preferred bias test point voltage for Mark III, Stereo 70 and some other amps, no coincidence there.

[nsubulysses]

Thanks for everyones replies to these questions. Reading through has definitely altered my perspective of how to go about this, which is exactly what I needed.

At this point I am leaning towards paying Tektronix to calibrate my DMM and then using that as a reference to cal my other DMMs and VOMs. Would come in handy with checking the old scopes too. I will probably use my newly calibrated DMM to check against the self-calibrating Tek digital scope and take it from there. If it's good, I will use it as a reference for the older analog scopes. This seems like the cheapest, best way to start going about this.

What is a good brand or place to get the Kelvin connectors from? Also, noob question, how would I use the 4 connector Kelvin plug with my DMM?

[Tom Phillips]

... how would I use the 4 connector Kelvin plug with my DMM?

First verify that your meter has the capability to use a 4 connector Kelvin plug. If the meter only has the two conductor input then you can't use the Kelvin leads. There is no 2 to 4 adaptor. (Unless you intend to use the Kelvin leads as a standard 2 conductor probe which has no purpose.)

Google "Kelvin Clips" All the finished assemblies and the clips to make them are surprisingly expensive