Continuous recorders

Watt and McNaught indicators gave a picture of the internal workings of an engine, but, obviously, only of one particular stroke. When academic research into the theory that lay behind steam-engine practice began in the second quarter of the nineteenth century, experimenters soon realised that analysis based on an isolated cycle was doomed to fail: there was nothing to show that the individual record was typical of the performance throughout a working day or if, for example, the working load on the engine had been changed.

The need for indicators that could record continuously was answered in the early 1840s, and some surprisingly sophisticated designs had been tried by 1860. These instruments came in several classes. Some recorded progress on conventional diagrams, either as a series of individual single-cycle traces that had to be analysed separately or as a single trace made from components of as many as a hundred cycles. The former type customarily relied on paper rolls, driven mechanically from one spool to another; the latter emboded stops and springs to take only a tiny horizontal 'slice' from the full-size trace that would otherwise have been produced during each cycle. A third class of 'totalisers' or 'totalisators' embodied the gears and wheels necessary to analyse (by mathematical integration) the performance of an engine over a period of time.

Multiple-trace indicators

The continuously-recording machine proposed by Daniel Gooch (1816–89) was the first indicator to be used successfully on a railway locomotive at speed. It owed its origin to the 'Battle of the Gauges', a controversy that raged in Britain in the 1840s over the ideal dimensions for railway track. One powerful faction, led by the Stephensons, claimed that the 'standard' gauge (then 4ft 8in) was superior to the Broad Gauge (7ft 0.1/4in) championed by the maverick engineer Isambard Kingdom Brunel (1806–59). Among the sharpest differences of opinion concerned the resistance afforded by air pressure, friction and other factors to a moving train, and it was this that Gooch sought to investigate

The Gooch indicator comprised a horizontal-cylinder apparatus incorporating two half-elliptical springs and a slide valve instead of a turn cock. It was also the first instrument of its type to incorporate an amplifying recorder: a pencil bar, attached at one end to the piston rod was mounted so that the movement of the piston was magnified approximately fourfold. Unfortunately, the diagrams produced by the Gooch indicator took an unusual form, partly because of the continuous-recording feature but also because the recording pointer moved radially.

LEFT A French MacNaught-type indicator fitted with a continuous-recording system. It was probably made by Duvergier of Lyon in the 1860s.

A few attempts were made to fit McNaught-pattern indicators with continuous-roll apparatus, most notably by the French clockmaker Paul Garnier in 1857-65. However, the success of the Richards design held immediate attention. Some continuous-record instruments relied on three drums — one holding the paper roll, one recording, and the other receiving the used roll — though others had only two, and a few attempts were made to contain the entire assembly in the recording drum. Elliott Brothers of London made Darke and 'Improved Richards' indicators of this type, but the universal acceptance of pillar-type torsion springs at the expense of flat spiral springs in the drum base subsequently favoured external paper feed systems.

ABOVE
Two Dobbie McInnes continuous explosion recorders, intended for use with gas and oil engines — the Mathot pattern (left) and the Cippolina type (right).

T.S. McInnes & Co. Ltd of Glasgow (and their successors, Dobbie McInnes Ltd) made a few Mathot-type indicators, combining a conventional vulcanite-clad external-spring body with a clockwork-driven paper drum. The design was patented in Britain in 1900 by Rodolphe Mathot, an engineer living in Brussels, Belgium, and was intended specifically for use with 'explosion engines' (gas, oil or petrol). The Mathot indicator was superseded by the Cippolina Continuous Double Diagram Recorder, production beginning in Glasgow in 1904 or 1905. Protected by a British Patent granted in 1902 to an Italian naval engineer, Giuseppe Cipollina of La Spezia, the instrument was basically two Dobbie McInnes No. 1 indicators sharing a common platform. Complexity and high price kept production to a minimum, even though production continued into Dobbie McInnes days.

Cumulative-trace and integrating/planimetringindicators

Perhaps the first of the limiting or 'slice' indicators was the work of an Alsatian engineer, Gustave Hirn (1815-90). Alterations had been made in the 1850s to a co-axial McNaught-type indicator. Hirn preferred the direct-reading instrument where absolute accuracy was concerned, but acknowledged that the oscillations that affected the pencil posed a real problem when the engine-speed rose above 20-25rpm. An answer was found by limiting the recording to a fraction of the cycle, then repeating it section-by-section until an overall picture was created. Hirn had first applied the method crudely, simply by pressing down on the pointer with his finger, but then progressed to mechanical stops. The slice diagrams or tranches minces were subsequently employed in France by Deprez, in the USA by Burkitt Webb, and in Britain both by Farnum & Bodley and by Moses Wayne.

Among the first attempts to develop a 'work meter' was made by a Briton, Professor Moseley, who published 'Result of a Trial of a Constant Indicator upon the Cornish Engine at the East London Water-works', undertaken in 1842, in the Minutes of the Proceedings of the Institution of Civil Engineers in 1844. The advent of the Amsler polar planimeter (see Accessories) inspired work to begin again. Though calculating indicators were developed — the Ashton & Storey design of 1869 was made in small numbers — the integrating indicator had greater potential.

The first of this type may have been the work of Henry Lea, patented in Britain in 1877. Built on the basis of a Richards indicator, Lea's design had a trace arm with a tiny wheel where the pencil-point would normally lie, and an abnormally long rearward overhang to accommodate the drive to the gear train of the recorder; a separate gear-train in the top of the drum recorded the number of strokes. The Lea planimetring indicator was theoretically capable of recording a single stroke, though it gave no clue to the steam distribution that would have been obvious from a pencil trace. Its true role was to record and calculate the mean performance of an engine over a period of time.

ABOVE
The Lea planimetring indicator, patented in 1877. From Jacques Buchetti, Guide pour l'essai des Machines à Vapeur... (Paris, c. 1887).

The planimetring indicator designed by the Briton Charles Boys, patented in June 1880, achieved much the same goals as the Lea pattern. However, it was also capable of simultaneously providing the trace of a single stroke or an accumulation of strokes produced at the same engine-setting. It had a cylindrical brass piston housing, flaring vertically into three standards to support the platform that carried not only the integrating arm but also a drum mounted on a spring-loaded reciprocating carriage. The carriage was driven from the engine cross-head or reducing gear, a drive rod from the drum to a gear train allowing readings to be taken from a large dial at the front of the instrument.

Patented in 1894-7, the integrating indicator developed by W.G. and C.W.G. Little of Bexley, Kent, was intended to provide a continuous record of the output of an engine working under varying load, keeping an accurate tally as work proceeded. Though the Little indicator could not analyse individual strokes in the manner of other instruments, it proved to be surprisingly accurate. Tests suggested that, over the course of 150 revolutions, the error amounted to only about a half of one per cent. However, Little indicators were not robust enough for run-of-the-mill engineering work and were generally confined to laboratories, schools, colleges and educational establishments. Production, never large, came to an end with the rise of the optical indicator in the early twentieth century. It is believed that the instruments were made by Elliott Brothers, though no survivor could be traced for inspection.

The Lea, Boys and Little indicators were made only in small numbers, and then, perhaps, only used for experimentation. Interest in instruments of this type seems to have lapsed for several years, only to reappear in earnest in Germany. Among the earliest of this group was designed by Heinrich Eicke, but more popular was the indicator patented by Anton Böttcher in 1908-9. This had an integrating wheel set in a housing hinged to the top of the recording drum, a spiral spring connecting the housing with the platform, and a prominent pillar (rising from the platform between the drum and the amplifying gear) that anchored a rocking link between the piston-rod extension and the planimeter head.

A few of these instruments were made by Maihak of Hamburg prior to 1914. During the 1920s, however, an improved version appeared with a sheet-steel carriage and bell-cranks between the piston-rod extension and the integrator. Finally, in the mid 1930s, Maihak produced a much-modified integrating indicator driven by a combination of gearing at the base of the drum and a rocking rod engaging a collar on the piston rod.

ABOVE
The pre-1914 (left) and 1920s (right) versions of the Böttcher planimetring indicator, made by Maihak of Hamburg. From Stephan, Verbrandingsmotoren (1917), and de Juhasz, The Engine Indicator. Its Design, Theory and Special Applications (1934).

All three of the Maihak/Böttcher planimetring indicators could also produce conventional diagrams. However, they were too complicated for most applications and were usually confined to research or teaching facilities. It is difficult to judge the true extent of production on the basis of a few survivors.