Circuit : Rodney Byne, UK
Email :

Description
The purpose of this multimeter calibration aid does away with the need for an oscillator. It offers a mains-derived and fully variable 50 hz AC volts down to millivolts for comparison and to check accuracy, of two or more parallel connected multimeters.

Addition of dual power supply
This updated Test Gear item 23 of the previous 50hz AC calibrator prototype seemed incomplete, so the relative importance was changed to include a simple dual dc power supply in the box and the calibrator part is now an extention of the psu, as per the new circuit.

The requirement was thus met to have an alternative constant power source to supply my Item 31, 7-waveform generator test jig, as against its two 9V batteries which invariably lose volts over time and run down.

Extra parts fitted
The calibrator's original ex clock radio mains transformer was inappropriate for this new task, so was replaced by a Maplin store item 12-0-12 250mA having a centre tapped secondary winding, allowing construction of this additional -10v, 0v, +10v dc power supply.

Two new smoothing capacitors, small block style bridge recifier, two zener diodes with feed resistors, 4mm binding posts, green LED with series resistor and two extra back-to-back battery connectors as interface for this external psu.

Build Notes
As per the photo, simplicity and space restriction inside the box required two shunt zener diodes BZX85C10v max rated at 1.3w or 130mA for stabilisation, as preferred over series regulators.

The LED & resistor were wired across the two electrolytic smoothing caps for indicating power on, but mainly to cause the caps to discharge when the psu was unplugged without a load being connected.

To suppress incoming mains spikes and diode conduction spikes from within the bridge rectifier itself, a 10n ceramic capacitor was added across its AC input terminals.

PSU Testing
The completed unit was mains tested for AC & DC voltage measurements and zener currents as shown on the updated circuit drawing.
(NB. Transformer AC outputs and rectifier DC outputs may actually be higher than assumed)

Initially, it was calculated to set the two PSU zener currents to 25mA by the two equal value zener feed resistors. These values can be easily changed if required.

The 7-waveform generator test jig as the subject current load, had measured +10v at 11mA and -10v at 12mA, those being well within the zener capabilities.

WARNING; when using back-to-back battery connectors with this test jig, continuity-buzz the wiring for correct supply polarity to the jig BEFORE switching the jig volts on.

As shown in the photo, in this special case a black wire goes to the red positive supply and a red wire goes to the black negative supply.
Ignoring this safeguard will result in both TLO71 ICs destroyed.

The joined common red and black wires go to the 0v terminal.

After lengthy on-load soak tests, the mains transformer became warm as expected.

Waveforms viewed on the laptop software showed no sign of ripple or worsened shape distortion compared to batteries and no evidence of scope mis-triggering was seen.

Calibrator Circuit Notes
The raw AC 50hz neg 20dB sine waveform however, showed flattened tops and bottoms caused by the bridge rectifier action.

This effect was fixed to a pure sine wave per the screenshot, by wiring a 47mfd, available choice of 50v or 63v electrolytic across the AC output panel sockets.

Sometimes older instruments of this type are favoured to respond better at low frequencies than higher audio frequencies, which are better suited for measurement by wideband audio millivoltmeters.

It is to be called to attention that unless otherwise stated by the manufacturer, both mechanical and digital multimeters display all pure and irregular wave shape inputs including square, triangle, sawtooth and integrator, as readings in Mean or Average Form-factor.

Obviously the higher the mains transformer secondary voltage output delivered, the higher the AC switched ranges or auto-ranging which can be measured. The choice of transformer and resistor attenuator values can thus be adapted to suit individual voltage range requirements.

Caveat: To comply with Health and Safety, ideally a small adhesive warning sign should be affixed to the box lid to state "Danger Mains Inside". The prototype does not have one displayed as not being available. However in this case a screwdriver is required to open the lid, therefore that situation is in my control and at my risk.

Circuit description and calibration procedure
The centre-tapped mains transformer, in this case delivers about maximum 15v AC across a variable resistor network.

AC feeds a multiturn pot VR1 which sets the reduced base voltage and then in series to two tandem-wired 1k lin pots labelled coarse and fine. For convenience of adjustment nearer to the hole in the box lid, VR1 is recommended to be supported on a tag strip.

The raw secondary 50 hz output waveform at my home is rather ragged as viewed on my laptop pc scope.
So to improve the waveform to a more acceptable sinusoidal shape, this can be further low-pass filtered by a single 2.2 mfd or two 1mfd polyester capacitors wired in parallel across the two pots combination as shown.

Tip: Operationally, the sine shape on a scope improves if the coarse pot is kept at lower levels, while using the fine pot for small range adjustments.

The calibrated Avo 8 MkIII test meter I possess, expectedly loads the commoned monitoring line when in the lowest AC range position of 2.5v. However this is immaterial as the other multimeter on test also responds to the same volt drop.

If after building the cal aid unit correctly as shown in the prototype example photos, this must next be established to be working correctly.

Don't plug into the mains just yet.

Connect any good preferably calibrated multimeter to hand, via 2mm probes to the 2mm meter monitor sockets and a second meter if applicable to the main output socket, in this case a BNC type via adaptors.

Turn both coarse and fine pots to maximum clockwise for full output.

Set the multimeters to a suitable AC volts range, in this case 0 to 10v if manual and the other may auto-range if applicable.

For initial meter overload safety, using a trimming tool turn VR1 anticlockwise at maximum resistance for very low output.

Plug the unit into the mains.

Then turn preset VR1 quickly clockwise to minimum resistance.

This test starts at high volts to compare the maximum end of each respective 10v meter display range for accuracy. (The Avo is reputedly more accurate at the high end of its DC and AC scales.)

Then continue, slowly turning VR1 anticlockwise down to check their respective 2.5v ranges for accuracy. Stop at 0.5v.

Though absolute accuracy of the two or more meters under test isn't guaranteed at this stage, then turn the main coarse & fine pots slowly down to track the display or scales points together and compare their very low AC volts and millivolts performances. It will be seen on the Avo that useful scale resolution ceases at 100mV, but the Beckman is capable of lower resolution to millivolts.

For any further sensitive electronic work, it is suggested to leave VR1 AC base reference at about 0.5v to 1v AC across the coarse/fine pots.

Conclusions

The above tests should show the meters at least conform to "95% accuracy satisfaction" as demanded by test equipment manufacturers and calibration houses.

The photo above of two meters on test with the same 5v AC readings, seems to show they have acceptable 95% satisfaction, but it is also intriguing to ponder on this thought-provoking question to finish the article:

What is the difference between accuracy and precision?
Answer: From what can be researched, accuracy is a measure of fine detail or small differences. Precision is a measure of closeness to the true value.

It appears in life there is still much striving required to achieve perfection, though common sense compromises sometimes have to be allowed!

It was concluded that a worthwhile objective of two useful bits of test gear in one box, to power up other project items and check AC multimeter calibration had been achieved.