With 12 or 16 channels to cope with, it is not a simple or easy task to test the accuracy, the stability and the voltage calibration of the entire DA unit. This simple machine does just that for all channels in a matter of minutes. Furthermore it allows quick selection of the most accurate and/or most stable channel when a critical measurement is required, as well as providing data for close calibration of every input.
The calibrator is built on a small perforated prototype board. The external supply can be a lab PSU or a standard 9 V battery, if portability is needed, as the unit only draws 3.5 mA. To allow two point calibration/check, unit provides two outputs: at 2200 mV and at 480 mV. The outputs are fed in parallel to all 12 channels of the 1012 or all 16 channels of the 1216. Outputs are properly buffered by Op Amp.
With reference to the schematic, 9 V supply voltage is connected to M1 and scaled down to 4400 mV by virtual zener IC D1, featuring high temperature and aging stability, as well as adjustable breakdown voltage. Zener voltage is adjusted by multi-turn trimmer R3. Total trim network resistance has been calculated for an adjustment range of 4280 to 4470 mV. Diodes D2 & D3 provide correction of temperature drift induced by the trimming network. Resistors R5-R6 scale zener output down by a factor of 2, so that final output range is 2140 to 2235 mV. All asterisked resistors should be high stability film and selected / matched to achieve values required.
Input impedance of each 1012/1216 is nominally 1MOhm (the writer checked and actually found exactly 1 MOhm spot on!) so the parallel impedance of 16 channels drops down to 62.5 kOhm. The drive of this impedance is really a very easy task for the Operational Amplifiers IC1A and IC1B. R8-R9-R10 scale the 2200 mV to 480 mV. The output voltage is selected by connecting J1 either to M2 or M3.
For accurate performance the calibrator board should be kept enclosed in a plastic or glass enclosure to avoid short term effects which could be caused by sudden local air drafts.
The writer was pleasantly gratified to find that the 1012 under test (after 3 years of use) was well within specification on all channels, as shown in the enclosed PC shots. The spec calls for 1% overall accuracy. Without the ± symbol, I take it to be ± 0.5%. At 2200 mV the spec translates into ± 12 mV. No channel was found off by more than 6 mV.
The calibrator has been working in a 16 to 22 °C ambient temperature range: 48 hour stability is better than ± 500 μV, after it has been turned on for at least 6 hours. The PC screens shown are for 10’ runs with samples taken every 60”. I also enclose the corresponding .PLWfiles.The tests are done in the range 1998 to 2002 mV.