Evershed & Vignoles meter restoration

 This old 7.5" diameter ammeter had been virtually immersed in seawater late in its life and the result was very heavy corrosion inside and out. The interior was in poor shape because the glass had long since disappeared. The movement was seized solid.

 The original movement and calibration were something like 30mA and 7.5amps respectively. Across the terminals, inside the case, was a small shunt to give the original full scale deflection of 7.5amps. At some point in its life this internal shunt had been cut and a rough, home-made brass shunt (centre-right of the picture) had been bolted externally across the terminals to give a new full-scale reading of 75amps. The new calibration marks had been written, not too carefully, in blue biro.

After much discussion the owner, who is refurbishing the boat from which the meter came, asked for a new full-scale deflection of 150amps and, to make matters quite tricky, wanted to fit the meter in such a way as to have it show charge as well as discharge. This meant resetting the movement to give a centre-zero reading. At least it started off that way, but later had to be biased to the right to give the customer's new requirement of a full scale "charge" reading of 150amps and a "discharge" reading, to the left, of 200Amps.

Some of the work was straightforward, some not. The tricky bits were removing lots of rusty flakes from within the magnet-coil space. This was not easy as the stuff, being iron oxide, clung tenaciously to the magnet.

Once the pointer was free to move; or at least once the coil was free to move, because the needle was cut off before work started as it had been damaged in the past and was a trifle wobbly; I was able to carry out basic tests to ensure that the project was feasible.

The movement appeared to be sound and measurements showed 30mA fsd.

Initially I planned to apply a de-ruster after scraping off the worst of the corrosion but as work progressed, incompatibilities in the original finish, traces of which were still present, and the various paints I planned to use became apparent and I finally settled for a smooth black Hammerite spray over bare metal. This proved to be not straightforward because when I smoothed away some blemishes prior to applying a final coat a weakness in Hammerite showed up. One just cannot spray onto a rubbed down finish without it ending up like the surface of a prune and having discovered this, I removed the wrinkly Hammerite and all traces of the original paint by disolving some sodium hydroxide in water and applying this to the finish. After an overnight soak the meter front panel had been stripped down to bare metal.

Several fresh coats of Hammerite, applied over several days, then produced a very good durable finish. The outer case also ended up looking in good shape, with all the over-painted heavy pitting ensuring that it looked the part of a re-painted old meter case rather than a shiny new one.

The old scale numerals and biro marking came off by gently rubbing with isopropyl alcohol. Retaining as much of the original artwork; scale markings and manufacturer's information; ensured that the dial still looked it age. Refitting the old needle was akin to brain surgery as it is made from extremely thin aluminium tubing, a little over 0.5mm in diameter. It was broken off at the point where it met the coil former and due to the dog-leg to raise it to the level of the dial meant some creative thinking. I used a fine nylon brush bristle inserted into the tube ends. This allowed the flexibility to position the pointer correctly, and once held in position, a tiny quantity of superglue was applied to the joint. This flowed into the tube ends by capillary action and, after allowing time to set between applications, further glue was applied until the joint was secure and had negligible extra weight.

I checked the calibration and confirmed that the new marking requirements were feasible by using a simple test set, then applied new numerals with rub-on stencils.

The trickiest part of the project was to design a suitable high current shunt, without the benefit of a DC power supply capable of delivering more than 20amps.

I calculated the resistance of the new shunt such that it would give the correct full scale readings, then fashioned it from galvanised sheet steel (I won't go into the maths here). Using a simple 20amp test set, I was able to determine the amount of drilling and cutting to complete the shunt. To do this I used a digital multi-meter set to milli-volts, a dummy load made from four high power resistors, and a 15 volt 20amp PSU. Once the shunt had been finished I was able to get the correct number of milli-volts across the meter terminals when a current of some 20amps flowed through the circuit. Because the scale markings are tailored to the original meter movement, which was untouched, it should be just as accurate as it was previously when used with the smaller shunts. True there will be a little inaccuracy because the new shunt will heat up slightly, and therefore change its resistance when current passes through it, but hopefully this will be small.

Below is the meter, repainted, and with its newly marked non-standard dial



 The original meter had been calibrated to 7.5amps and later modified to 75amps but I was unhappy about merely making a beefier shunt without changing the method of fitting it. The entire ship's 12-volt supply would be passing through the meter terminals and these were not specially large so I designed the new shunt so that the main power feed could be connected across the shunt rather than the meter terminals. This would ensure that the meter would never see more than its rated 30mA, even in the event of loose connections at its terminals.

Above: The new shunt is shown here before being fitted to the meter, by the smaller diameter bolts.

 Final testing will take place once assembly is complete. This will involve the use of a car battery capable of supplying enough amps into a dummy load to give full scale deflection. To confirm the number of amps flowing a "standard" shunt will be used. I have a suitable component shown here, marked "250amps" and 75milli-volts.

 These figures indicate the resistance of the shunt is 300micro-ohms; therefore, dividing the number of milli-volts across the shunt by 300 micro-ohms, when the meter reads full scale, will determine the current indicated by the meter.

If the new meter shunt is good the reading on the milli-voltmeter will be 45, corresponding to a forward deflection of 150amps and 60milli-volts in the reverse direction, indicating 200amps. If the dial readings are low I can file away some of the shunt material to increase its resistance, diverting more current through the meter.

Return to current projects page