Electrokinetica The Electro-mechanical Museum

Nevelin Keepalite mercury-arc emergency lighting system

Background

Many buildings are required to have emergency lighting which remains alight for a minimum time in the event of the main electrical supply failing. Before small maintenance-free rechargeable batteries were available economically, emergency lighting installations were often supplied through a separate system of wiring from a central battery large enough to serve an entire building or wing. As nickel-cadmium and sealed-lead-acid batteries became more reliable and cost-effective, a trend started towards self-contained emergency light fixtures, but existing installations and high-specification designs continued to favour the more durable central-battery system. Although extra wiring is required to distribute the emergency supply, the system has a number of advantages including ease of maintenance and testing, increased backup time and simplicity of integrating backup lighting into any normal light fitting.

Many smaller central-battery systems used a battery voltage lower than that of the mains, typically 50-100 volts DC, to minimise the cost and size of the battery. For large buildings a mains-voltage battery (240V in the UK) could be justified by the lower cost of wiring (which could be of lighter gauge owing to the lower lamp current) and standardisation of lamp voltage. The batteries were commonly wet lead-acid accumulators built of individual 2V cells, which on account of their large size and ventilation requirements were often housed in dedicated rooms. Thus a large building might have one or more emergency lighting battery rooms adjunct to electrical switchrooms which would also house the chargers. The function of the charger or rectifier units would typically be threefold: To recharge the battery after discharge during a supply failure, to provide a top-up or 'trickle' charge to keep it in good condition, and to power the lighting installation without discharging the battery. Motor-generator sets, dry plate rectifiers, thermionic valves and mercury-arc rectifiers were all used for this purpose, with the mercury arc rectifiers being best suited to the largest installations.

'Nevelin / Chloride Keepalite' plant with mercury-arc rectifier

Keepalite rating plate

Keepalite rating plate

This emergency lighting control and charging set was originally installed in the Guildhall in London, where it supplied the lower floors of the building from a switchroom and battery room in the basement. Nevelin, a division of Lancashire Dynamo that specialised in mercury-arcs, was the manufacturer of the rectifier unit. The complete system including the battery was supplied by Chloride Batteries Ltd, whose Keepalite trademark was used on their emergency lighting equipments, appearing on this one alongside the Nevelin name. Our unit, which dates from the 1950's, was retired from service some time ago during a modernisation from wet lead-acid cells to sealed types which require a different charging routine.

Recovering the Keepalite

Keepalite plant in switchroom

Keepalite plant in switchroom

The Corporation of London Electrical Engineering Dept. sought a home for the unit rather than the scrapyard, and whilst the removal initially looked easy enough, careful measurement revealed that the cabinet was half an inch too wide to fit through the switchroom door and one inch too high to fit through the door to the loading bay. The Works Dept. therefore arranged to remove the switchroom door and frame before we arrived, enabling the unit to escape into the main corridor. There we deployed special lifting equipment to tilt the bottom-heavy cabinet onto its side before it reached the final door. The bulb was removed and transported upside-down, to avoid swirling mercury causing impact damage to the glass. There was also mercury in the battery cutout contact cups which had to be carefully extracted before moving the unit. Insulating materials containing asbestos were sealed up in case of damage in transit. Once back at the works, the cabinet and bulb were cleaned and re-united, a few wiring repairs were made and a supply and load rigged up. The power was switched on but at first the bulb failed to fire correctly. The trouble was traced to an open-circuit resistor and a bad connection in the regulator circuit. Then one anode refused to light due to a fault in the associated firing circuit. After some further attention to bad contacts all was well and the rectifier was put on load.


The Nevelin rectifier bulb.

Nevelin rectifier bulb

Nevelin rectifier bulb

In total, 12 electrodes are fitted to this bulb: 3 main anodes, 3 control grids, 2 auxiliary anodes, 2 excitation anodes, ignition dipper and cathode.The three main anodes are housed in angled side-arms; this protects against reverse current flow that would result from ionic bombardment if the anodes were in line-of-sight of the cathode pool. Also in the main arms are the control grids, connected via small terminals projecting from the side. This bulb is unusual in having auxiliary high voltage anodes designed for control circuit supply. These are mounted in small angled arms high up on the bulb wall. The excitation anodes are mounted conventionally in small arms near the base of the bulb to keep the forward voltage low. Nevelin usually arranged the cathodes of their dipper-ignition rectifiers to lead-in through a side-arm rather than through the base of the mercury pool as is the case on this bulb. This allowed greater flexibility in mounting and minimised the overall height of the cradle. The serial number on this bulb does not match that shown on the cradle, so this is almost certainly a replacement bulb.

The rectifier circuit and equipment.

The unit contains all the equipment needed to operate the battery and lighting installation. The mercury-arc rectifier is employed to supply the heavy current required for recharging the battery when depleted and for powering the lighting installation, whilst a small selenium rectifier is incorporated to provide a continuous low current maintenance charge to the battery. The output is fully regulated between 230 and 316 volts at up to 54 amperes, to allow for the different requirements of charging the battery and powering the lights. To permit this adjustment and regulation, a grid-controlled three-phase bulb is used, which operates with phase-angle control not unlike a modern lighting dimmer or speed regulator. 415V AC mains is supplied through the main switch to a 15kVA air-cooled transformer, wound with an interstar-connected (zig-zag) secondary to feed the rectifier main anodes. Tertiary windings on the transformer supply the control, excitation and auxiliary circuits. A conventional dipping-electrode ignitor and magshunt transformer-fed excitation circuit is provided, along with fan cooling and Varistor surge-protection for the bulb. The DC output is taken from the secondary winding neutral point and the bulb cathode, by way of a heavy smoothing choke and a current sensing relay. A master function-switch routes the interconnections between rectifiers, battery and load according to the mode selected, at the same time putting into circuit the correct voltage feedback network. Panel meters are provided to indicate rectifier and battery currents and system voltage, whilst pilot lamps signal rectifier operation and battery discharge. A manual voltage control allows fine adjustment of load voltage or charge current, according to the mode in use. The trickle charge rectifier is equipped with an independent supply transformer and current control rheostat.

View through cabinet door

View through cabinet door

Side cover removed

Side cover removed

3-phase transformer

3-phase transformer

Principle of operation of the voltage regulator

The main output voltage is controlled by variation in phase, relative to the anode supply, of the firing impulses applied to the rectifier bulb control grids. The latter are biased negatively to prevent conduction to their respective anodes, but are driven with positive impulses to trigger conduction that continues until the next current-zero without further influence by the grid. The point during the anode voltage waveform at which firing occurs determines the mean voltage impressed upon the output circuit during the conduction period. The impulses are generated by three identical firing circuits, one per phase, sharing a common negative-bias and feedback circuit. For each phase, an inductor with a core of silicon iron having progressive saturation characteristics is used to generate a triangular current from a sinusoidal supply voltage. This current passes through the primary winding of a 'peaking' transformer having a core of nickel iron, which exhibits a very sharp saturation transition. The winding is arranged such that the core is unsaturated for only a small fraction of the triangular waveform period, during which time the firing voltage is induced in the secondary winding and applied to the rectifier grid. A control winding on the transformer carries a direct current proportional to an adjustable fraction of the output voltage, opposing its DC ampere-turns against those of the primary and therefore increasing the minimum primary current required to prevent saturation. Hence the control voltage adjusts the point on the primary waveform at which the unsaturated period occurs, and thus the ignition delay. An interesting feature of this circuit is the use of specialised magnetic devices in place of valves or other active components. Custom magnetic devices are expensive but offer high reliability and stability.

Grid control circuit

Grid control circuit

Internal plate

Internal plate

Specifications

Chloride serial No. K6378
Input 400/415V 3 Phase 50 c/s
Maximum output 230/240V 54A
Charging output 230/316V 28/14A
Bulb No. DD2050GF F0111

-----

All content © copyright Electrokinetica 2007-2017 except where otherwise stated • Valid XHTML 1.0Valid CSS