Spark tracers

The meteoric rise in the popularity of the internal-combustion engine was due partly to enthusiasm for motoring, and partly to the intervention of the First World War. Conflict often accelerates technological progress, and the years between 1914 and 1918 were no exception: the aeroplane, in particular, made a spectacular advance from the rickety craft of the first days of fighting to the gigantic bombers that were being built when the hostilities ended.

Though methods of testing engines on the ground had been provided, maximum-pressure indicators were too unsophisticated, pencil-arm autographic indicators were much too delicate, and optical indicators were too cumbersome to enable experiments to be undertaken in flight.

Development of the 'RAE Indicator' — named after the British Royal Aircraft Establishment, South Farnborough, Hampshire — began in 1919. The instigator was Squadron-Leader Geoffrey Norman, who sought, in the words of W. Sydney Smith, Superintendent of the research establishment, to provide 'an indicator which would record diagrams from the engine with entire satisfaction at high speeds, either on the bench or in the air'.

The basis of the indicator was a small disc valve with which air pressure on the top side, supplied from a separate bottle, could balance the pressure generated by combustion in the engine cylinder. When the latter just overcame the former, the valve lifted — a rapid and almost imperceptible movement that broke the primary electrical circuit and allowed a high-tension induction coil to induce a high-tension voltage in the secondary coil. This was forced to jump an air-gap on its way to earth in the form of a spark, scorching a mark on the diagram paper. The pressure continued to rise to a peak, then fell to make another mark on the opposite side of the pressure curve. The air pressure was then increased, giving another pair of marks, and the process was continued until the entire diagram had been constructed. The unusually large diagrams (7½ inches tall; 14½ inches wide) were easy to read, assuming that the spark points could be clearly seen.

In the earliest design, each diagram was a composite of 125 revolutions obtained as the spark-point moved from one end of the drum to the other. The drum was driven from the engine by either a shaft or a short chain, by way of a special clutch. The spring cradle on the casing accepted two 'main tension springs', attached to special anchors or 'spring horns', and a pair of compression springs on the central strut. The outer springs could be replaced, depending on the pressure-range required, or detached to enable light spring diagrams to be obtained with the assistance of the central springs alone.

A preparatory Preliminary Report on Electrical Indicator for High Speed Internal Combustion Engines was submitted on 16th January 1920 and the first prototype was completed in the summer of 1921, to be put immediately to the test. Tried for six months with an experimental aeroplane engine, taking more than 1500 diagrams, it proved to be very successful; six examples of an improved design, with the batteries, high-tension coil and distribution box built into the indicator body (or 'casing') followed in 1922. An application for a British Patent was made in April 1922 on behalf of Harry Wood of the RAE and Jessie Norman, widow and 'legal representative' of Geoffrey Norman; it was accepted in July 1923 as British Patent no. 200595. Later, in January 1924, Wood submitted an improved contact-breaker design. This was patented in April 1925.

The experimental indicators were initially confined to bench testing. On 3rd February 1923, however, one was successfully installed in the rear cockpit of a two-seater a DH 9 biplane and used to test the 400hp twelve-cylinder Napier Lion engine at altitudes of 500, 5000 and 10000 feet. This enabled the major goal of the project to be achieved: the investigation of engine performance, particularly at high altitude.

ABOVE
Left: the first successful in-flight test of the 'RAE' or Farnboro indicator was made on 3rd February 1923. This picture shows the instrument (one of the six improved examples made in the Royal Aircraft Establishment) installed in the rear cockpit of a DH 9 biplane. Right: the original Dobbie McInnes-type Farnboro, from the handbook published c. 1925.

Series production of the RAE Indicator was licensed to Dobbie McInnes & Clyde Ltd of Glasgow, and work on what was known commercially as the 'Farnboro' (after the site of the Royal Aircraft Establishment) began almost as soon as the first patent had been granted. A review by A.W. Judge appeared in The Automobile Engineer in January 1925 and the first handbook seems to date from the same period: one of the illustrations inside the back cover shows the improved paper-winding system patented by Walter Clyde and Dobbie McInnes & Clyde Ltd in August 1923.

The earliest manual reveals that the indicator was supplied in a partitioned wooden case, accompanied by four pairs of tension springs, a pair of light compression springs fitted in position, three two-volt accumulators to be connected in series provide six-volt electrical power, a disc-valve unit with an 18mm thread, and a hundred black diagram sheets. The box was 24 inches long, 16½ inches high and about 11 inches wide. Tension springs were identified either by colour or, in written material, by individual letters: blue ('B') for 1in of diagram height per 40lb/sq.in., a maximum pressure of 300lb/sq.in; white ('W') for 1:80 or 600lb/sq.in.; yellow ('Y') for 1:100 or 750lb/sq.in; and red ('R') for 1:150 or 1125lb/sq.in. The compression springs were suitable for pressures ranging from 12lb/sq.in. below to 30lb/sq.in. above atmospheric pressure.

The Farnboro indicator was an instant success, but there were those who questioned its accuracy. The most damning report was made in the USA by John H. Collins of the Langley Memorial Aeronautical Laboratory: Alterations and Tests of the "Farnboro" Engine Indicator (Technical Notes, National Advisory Committee for Aeronautics, no. 348) subjected the standards of manufacture and the disc valve to trenchant criticism, as '…modifications were made to the instrument to improve its operation. The original design of disk valve was altered so as to reduce the mass, travel and seat area. Changes were made in the recording mechanism, which included a new method of locating the top center position on the record. The effect of friction on the motion of the pointer…was eliminated by providing a means of putting pressure lines on the record…'

There is little doubt that the modified NACA Farnboro performed better than the Dobbie McInnes & Clyde instrument in static testing; the new valve weighed only about a fifth, had a seat-width about one-seventh and a vertical travel of about one third of the original dimensions. This made it much more responsive, though it must be questioned whether the theoretical superiority claimed by Kalman DeJuhasz in the 1930s (1:156 in favour of the NACA pattern) has much validity.

Many experimenters subsequently used Farnboro indicators, though changes were sometimes made to suit specific requirements: Harry Ricardo used one in 1929–30 to investigate fuel-line pressures, and W.J.R. Roach & J.G.G. Hempson (Ricardo employees) used another in 1951–2 to investigate the performance of large marine diesel engines. Though Dobbie McInnes capitalised on Roach & Hempson's research, published in the British Shipbuilding Research Association Report no. 92 and also in The Engineer, the authors had been critical: "With the original valve, trouble was experienced due to gas leakage, over-heating, contact-face corrosion and internal fouling…' A new balancing unit had been developed to overcome the perceived problems, with a large flexible diaphragm. This was subsequently made commercially by Dobbie McInnes as the 'Pick-Up Type P' (for normal pressure ranges) and the 'Pick-Up Type L' (low pressure), replacing the original disc-valve system.

ABOVE
Left: a drawing of the Farnboro indicator from the Roach & Hempson report, 1952. Right: a standard Farnboro indicator of the type being made by Dobbie McInnes in the late 1950s. This example, no. GM 764, lacks the detachable Pick-Up unit.

Another type of drum-type balanced diaphragm spark-trace indicator was promoted in the early 1930s by Massachusetts Institute of Technology, relying on two circuits (one for the rise and one for the subsequent drop in pressure) and sparks generated in the secondary winding of the spark coil to mark the card. This was designed to minimise the tendency of the Farnboro indicator to miss points when the valve opened too slowly to cope with high pressures of increasingly shorter duration.

Spark-trace indicators were something of an evolutionary dead-end, as they suffered from several practical drawbacks — e.g., the leakage of air into the engine side of the valve or conversely, of combustion products into the air-pressure side. Consequently, the balanced-valve indicator developed in the early 1930s by Brandt, Viehmann and Urbach for the Deutsche Versuchsanstalt fur Luftfahrt (DVL) relied on a trace that was recorded photographically.

The rise of non-autographic systems based on cathode-ray tubes made the spark systems obsolescent, though the British Railways Board was still using the Farnboro in the 1960s and new instruments were being offered by EMPI Ltd, successors to Dobbie McInnes Ltd, as late as April 1972. By this time, a wide variety of accessories and two additional types of tension spring had been introduced: the springs were coloured green for 1:200 or a maximum pressure of 1500lb/sq.in., and orange for 1:250 and 1875lb/sq.in.