Homebrew Calibration For Test Equipment

If you work for a large company, you probably have test equipment that is routinely calibrated. Some companies have their own metrology labs and others send out to an external lab. In a garage lab, you are less likely to do calibrations and — in our experience — that isn’t usually a problem. Still, it is nice to be able to do at least a sanity check on your gear. Also, if you buy old test gear and repair it, it would be nice to be able to check it, as well. [IMSAI guy] built his own little calibration setup, adding to it over the years, and he shares the details in a recent video, which you can see below.

The board started out simply as one voltage regulator and some 0.01% resistors. Over time, though, he added a few more bells and whistles. The setup isn’t going to rival a NIST-traceable lab setup, but for your garage it is perfectly fine.

The regulators are really precision voltage reference ICs that are readily available, so they should be better than your old bench supply. However, we didn’t think you’d really want to slavishly copy this design, but the idea of having a single calibration kit on a board is something you can grow organically out of your junk box and hamfest finds.

Among the additions to the board are a precision oscillator module from an old GPS and a second voltage reference. The initial reference was a 10V device that is specified to have a maximum error at room temperature of +/- 0.05%. He might have wanted to put some diode protection on the device, though, since reverse wiring it destroyed it. On the plus side, it caused him to look to see if any new better devices.

So when he replaced the reference he also added an AD587 as a second 10V reference which is just as accurate and has an ability to trim the output (although he doesn’t use that capability).

Of course, if you get obsessed with calibration, you might want to get a rubidium standard — in fact, one shows up in the video. There are also a variety of precision resistors we’ve looked inside of in the past.

8 thoughts on “Homebrew Calibration For Test Equipment

  1. I did something similar with a 5V reference from voltagestandard dot com and a bunch of 0.1% resistors from China. The V ref came pre-built, calibrated to 5.0000V and cost less than $30. The resistors cost pennies but if you buy a bag of 100 or more you can simply test them with your best meter freshly calibrated to the above V ref and choose the ones that are spot on. I built a simple voltage divider circuit out of some of the best 1K ohm resistors and a piece of perf board and put in taps for 2.50V, 3.00V, 3.50V, 4.00V and 4.50V using the 5.0000V reference as source. Going between the taps gives me any value from 0 to 5.0000V in 0.5V increments. The circuit also gives me exactly 1.000 ma end to end which is the V ref’s optimum current output. This may not completely be lab quality but it’s better than consumer quality and more than enough to calibrate my 50 year old VTVM and keep my motley collection of meters and scopes in tune. I also have a 1 MHz OCXO oscillator on order to help me calibrate the frequency functions of my equipment.

  2. Can make into a high quality, temperature monitored (and I guess pressure and humidity also and maybe RFI/EMI) decade LCR and more standard box.

    I’ve developed and validated standard references and can’t see why this can’t be done DIY homebrew style in the electric, magnetic and electromagnetic world outside of spectroscopy. Kinda of newbie over ambitious… though would be interesting to get more critiques of improvements and others methods. 2 comments only so far?

    Electrochemical standard cells comes to mind for starters for voltage.

  3. A major consideration with voltage standards is not their initial reading but how they drift. The initial reading can always be “calibrated out”.

    I have a few saturated Weston Standard Cells (the oldest being from 1949!) that can be used as room-temperature thermometers with a 40uV/C sensitivity.

    Roughly speaking, the current crop of ic based voltage references come in two varieties…

    1) poor initial accuracy, but they don’t drift as much over time
    2) impressive laser-trimmed initial accuracy, but they drift over time. My sample-of-one has drifted 0.5mV in 5V, initial accuracy ~800uV.

    If you want to delve into voltnuttery and timenuttery, there are dedicated usenet groups with highly experienced practitioners and discussions going back decades. A more, ahem, accessible starting point for youngsters would be http://www.eevblog.com/forum/metrology/

    Be warned: getting that extra digit is expensive, elusive, and fascinating.

  4. Pffft 0.1%. I have one of those Aneng multimeters recommended by That Australian Electronics Youtube Idiot… and it changes its gain by about 15% when it auto-ranges. 15% is better than 0.1%, right?

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.