[neonlamp]

For those of you with Avo 8 multimeters, I was puzzled by the circuit diagram for the A.C. voltage measurements. This shows (at least on the diagram I have in a Mk 6 booklet) that the resistor in series with the meter when set to A.C. volts is <=5k.

When the meter is set to A.C. volts, the transformer is switched in. For 3V and 10V ranges, the input is such that 10mA is needed for 3V and 1mA for the 10V range. The transformer then gives a step-up ratio of 20 and 2 respectively. This is, of course, because it is being used as a current transformer. It means that the output current will be about 0.5mA for these inputs, and once higher voltages are selected, this current is supplied directly through the multiplier resistors.

The A.C. output is rectified by what at first sight seems a dubious circuit with resistors in one arm of a bridge rectifier, while the other arm uses OA95 diodes. In calculating the voltage which appears across these resistors, it seemed that a series resistance of 5k, plus the meter, which on A.C. volts is 2667 ohms as the current shunt is disconnected, would seem to be too low to limit the current to 37.5uA. The resistors are quoted as being 889 ohms, so allowing for 37.5uA in the meter, the voltage drop across one carrying most of the current would be 411mV. This needs a 10.9k resistance in total.

But there is a form factor of 1.11 to allow the average voltage to be read as RMS, meaning that the resistance should be 1.11 times lower, and there could also be another factor due to the transformer not being exactly 100% efficient. These bring the resistance down to about 9500 ohms.

Now, looking at the circuit board attached to the transformer there are some printed resistors and the one in series with the meter is marked green, black, red indicating 5000 ohms. Since the resistance is adjusted by scratching some material away, it must be greater than 5000 ohms. Indeed, another meter read this as 6200. Totalling, with the meter coil, 8870 ohms approx., which is much closer to the expected value!

So the original mystery is now solved. The Avo circuit should have been written R>=5000: the sign was the wrong way round.

In investigating this measurement, the use of resistors in one arm of the "bridge rectifier" is a neat trick by the designers. This means that the voltage applied to the meter is dependent more on the resistor than the diode. Therefore, if anyone has to replace their OA95's, it is possible that ordinary silicon diodes like the 1N4148 will work. This is also aided by the potentially high voltage output from a current transformer, which means that non-linearities will be quite small. I haven't checked this for non-linearity, but may do if I have time.