The first steps



Watt, c. 1793

The first of Watt's pressure-indicating gauges had been tried by December 1793. It consisted of a small cylinder, no more than an inch in diameter with a bore six inches long, which contained a solid-head piston made with the greatest accuracy that could be obtained. The cylinder terminated in a small cone-tipped pipe, which could be inserted in a hole bored in the cylinder or condenser covers. Communication between gauge and receptacle was controlled by a system of cocks.

A long spiral spring connected the piston rod with the supporting frame, and a pointer attached to the rod-tail lay against a graduated scale allowing the pressure within the cylinder (or, alternatively, the vacuum in the condenser) to be determined. Calibration was undertaken by referring to a mercury barometer.

Watt's primitive gauge worked well enough to enable the operating characteristics of individual engines to be determined, but had an important drawback: pressures could be observed only by watching the movements of the pointer during the piston stroke and then simply writing the results down. This process was open to error, even though the ponderous movements of the early beam engine were slow enough to facilitate observation.

Indicating devices of this type soon proved so useful that refinements were made. The Science Museum in London still has a machine in the form of a small beam engine, with the piston on one side of the supporting column and the spiral spring on the other, taking the place of the connecting rod or pump rodding on the full-size engines.

The minuscule beam, with two small arch heads, supports an elongated metal rod running upward to serve as a pointer. A graduated scale could be attached to a board attached to the vertical arm of the frame, which in its entirety resembled a large slender round-headed 'T' on a low four-leg stool. Elongating the pointer rod magnified the movement of the pointer against the scale, facilitating observation, but the process still demanded great care if the fluctuating pressures within the cylinder were to be recorded accurately. Exactly when this instrument was made remains in dispute, though the Boulton & Watt Papers, now in the care of the Birmingham Museum of Science & Industry, contain a variety of references dating back to 1794.

Watt & Southern, c. 1796

The indicator was soon adapted to provide a written record of each individual application instead of merely a transient observation. This was a tremendous analytical breakthrough, allowing, as it did, an accurate picture to be formed of the pressure of steam at any time during the movement of the piston. The inspiration was due to John Southern (1758-1815), Watt's draughtsman, who recorded in a letter dated 14th March 1796 that he had 'contrived an instrument that shall tell accurately what power any engine exerts'.

By August 1796, Southern was expressing doubts that diagrams that had been supplied from the Salford Cotton Mill engine were accurate. He went on to note that 'It would be better if instead of drawing the board uniformly forward, a pair of wheels was applied so as to make one revolution for a double stroke of the engine and crank fixt [sic] upon one end of such a length as to give the stroke you wish for the board to move. The exactness of the beginning and ending might be ascertained very nicely, and as the pencil would go over and over again the same track or nearly, the mean might be taken with some precision'. The 'closed loop' was obtained by fitting a tablet that reciprocated in phase with the piston.


Genuine Watt indicators are exceptionally rare. This replica built by Bruce Babcock of Amanda, Ohio, on the basis of information supplied by the Science Museum in London, has been shown at many steam fairs in the USA. The drawing, taken from Terrell Croft's Steam Engine Principles and Practices (1922), shows a typical Watt-type indicator of the 1820s.

Watt moving-tablet indicators were made only in small numbers, though construction and design often differed greatly in detail; for example, a long spiral spring was often substituted for the cord and weight. Many of them were still being used in the 1850s. They were invaluable to the engineers of their day, even though excessive friction in the moving parts promoted inefficiency. Consequently, the moving-tablet indicators were eclipsed first by the McNaught instruments and then by the many 'high speed' designs deriving from the Richards pattern.

However, there were many who mistrusted the ability of the spring-driven reciprocating cylinder to provide accurate diagrams, pointing to the dependence on inherently elastic driving cords and on the assumption that the performance of the spring in the recording cylinder would be consistent throughout the entire range of movement.

One of the earliest attempts to resurrect moving-tablet systems was reported by John E. Sweet to a meeting of the American Institute of Mining Engineers held in Chattanooga, Tennessee, in the summer of 1879. Even at this early stage in the development of high-speed engines, a need had been identified, said Sweet, for efficient indicators that dispensed entirely with parallel motion. The Thompson indicator (then regarded as the most modern design) had dramatically reduced the weight of the parts in the pointer linkage, but engineers were already predicting that running speeds of 1000rpm would be achieved. No indicator available in 1879 could provide legible diagrams at this speed.

Sweet claimed that the indicator made by Fred Halsey, one of his students, had worked very well, providing diagrams as good at 330rpm as they had been with the Thompson instrument at 270rpm and the original Richards machine at only 220rpm — though the rigid connection between the reducing gear and the sliding tablet undoubtedly explained part of the improved performance.

The most sophisticated moving-tablet indicator to be introduced prior to 1914 was patented in 1893 by an Englishman, Moses Wayne, who acknowledged the prior existence of instruments working on similar principles. A lightweight moving tablet was combined with a rotary piston controlled by an externally-mounted helical spring.

Steam from the engine cylinder entered by two channels through diametrically opposed admission ports, twisting the piston by its vanes against the pressure of the control spring. The steam or water that leaked past the piston simply dissipated through two escape ports. The pointer was a radial arm attached directly to an extension of the piston spindle, drawing its trace on a piece of paper attached to a concave mounting plate on the tablet. This was moved forward and then back by the connection with the crosshead (or suitable reducing gear), producing a conventional indicator diagram.

However, the Wayne instrument could also be fitted with a detachable limiting mechanism (later known as a 'liner'), enabling it to be used in circumstances — e.g., on railway locomotives — where vibration hindered the creation of a single diagram. The limiting mechanism allowed the operator to produce a continuous line-by-line summary of performance as the tablet reciprocated. The movement applied to the piston by the worm gear ensured that each line was drawn at slightly greater pressure than its predecessor.

The piston could be moved from its 'off' position (when the cylinder pressure was less than that the opposing spring) to 'on' when the pressure finally overcame the spring. Though the absolute pressure was not shown, the point at which it rose above the controlling spring was clearly marked on each line and allowed the points to be joined to provide continuity. Experience soon taught the operator how fast to turn the limiter crank-handle, until the individual lines were no more than a twentieth of an inch apart. At this spacing, the pressure line became all but continuous.

Elliott Brothers made Wayne moving-tablet indicators until 1900 or later, though the quantities involved were probably small. The simplicity of rival rotating-drum patterns relegated the Wayne pattern largely to experimental or laboratory use, even though the ease with which legible 'lined' traces could be obtained from engines running at 600rpm (or more) was greatly in its favour. Elliott turned instead to the equally short-lived Simplex drum instrument, with its detachable 'tong' springs.

McNaught, c. 1827

The first major advance in design was made by replacing the reciprocating tablet with a revolving drum, which was much more compact, easier to manage, and offered less frictional resistance to the recording stroke. The instigator of this system is generally believed to have been John McNaught, who began trading on his own account in Glasgow in the 1820s having previously made Watt-type indicators for (possibly among others) the engineer John Farey.

McNaught relied on the piston stroke to make half a rotation of the drum and a spring within the drum to enforce a return. The date of this advance has not been satisfactorily determined, though some evidence was laid by McNaught before the Society of Arts for Scotland in 1829, including a pamphlet entitled Description and Use of Macnaughts Improved Indicator for Steam Engines, published anonymously in Glasgow in 1828 but almost certainly McNaught's own work. There were also several testimonials, including one from 'Mr Alexander', who claimed to have been using a McNaught indicator for 'more than two years'. This suggests that the development had been completed by, at the latest, the winter of 1827.

The original McNaught indicators were of the 'co-axial' or 'in-line' design, with the paper drum around the cylinder. They were suitable only for low-power engines, the scales usually ranging a vacuum of 15lb/ to a pressure of 15lb/ above atmospheric level. Soon, perhaps inspired by an ever-increasing enthusiasm for compounding (and undoubtedly also the introduction of railway locomotives), a high-pressure 'parallel axis' variation had appeared. The recording drum had been moved to a bracket projecting at right angles from the cylinder.

A catalogue published in Glasgow in 1831 shows both patterns. By 1842, however, McNaught had abandoned the co-axial indicator, and the distinction between pressure ranges was being addressed with differing pistons. The high-pressure and low-pressure patterns had areas of 1/ and 1/ respectively. Pressures ranged as high as 130lb/, which had advanced only to 140lb/ by the mid 1850s. Some, if not all high-pressure McNaught instruments were also apparently capable of indicating a vacuum; however, it is suspected that most would have been used in conjunction with non-condensing locomotive engines. The pressures were far too high for the stationary engines of the day.

By 1856, and the final version of McNaught's promotional leaflet (which had become a book), the separate-cylinder indicator was being offered for pressures of 60, 100 and 140lb/ In addition, a specially enlarged version was being offered, capable of giving a six-inch diagram instead of the customary 3.75in version.

McNaught indicators—and the copies that they inspired—came from a variety of sources. Some may have been made for McNaught by Joseph Chadburn and Chadburn Bros. of Sheffield; many were made by the Novelty Iron Works of New York (apparently from 1847); and others came from Joseph Hopkinson of Britannia Works, Huddersfield. One example in the Science Museum collection bears the mark of John Hannan of Glasgow.

A McNaught parallel-axis indicator, probably dating from the 1840s, with a replacement spring. Photograph by courtesy of Bruce Babcock, Amanda, Ohio, USA. Drawings taken from Indicator Diagrams and Engine and Boiler Testing by Charles Day, published in 1898. A later indicator, made by J. McNaught & Son of Glasgow. This large example, with replaceable pressure-graduation plates, may date from about 1858 and is believed to have been used in a military factory in India. Museum of Making collection.

McNaught indicators were successful, particularly for use with slow-running engines, and set a trend that lasted for more than thirty years. They were still regarded as standard in the Royal Navy (where boiler pressures had remained exceptionally conservative) as late as 1882; the advent of compounding then had an effect, and McNaughts were soon being preferred only 'for general use on ordinary service'—the Richards design was used 'for the records of steam trials and other special services'.

The indicator promoted from the early 1850s onward by Joseph Hopkinson of 'J. Hopkinson & Co., Engineers of Huddersfield and London', was the most interesting variant of the McNaught system. Hopkinson's indicators returned to the co-axial design, which he considered to be more resistant to vibration. The operating cord ran around the base of the drum, and around a pulley attached to an arm held to the base of the tube by a collar-and-thumbscrew assembly. The pencil pointer lay on a spring-steel arm, attached directly to the piston tail rod to work directly on the trace-paper. Additional springs could be supplied, each suited to differing pressure ranges. Unlike virtually every other design, however, these springs were added to the piston rod above the cylindrical casing, where they were retained by a locking nut.

Hopkinson's design was simple and compact, remaining in vogue even after the first Elliott-made Richards instruments had been distributed in Britain in the 1860s. However, it soon lost favour once high-speed engines became common, as the inertia of the heavy spring/piston unit contributed to excessive vibration and irregular trace lines. The original axial or 'in-line' design was discontinued in the mid 1870s.

Hopkinson had attempted to make a parallel-axis indicator, protected by British Patents granted in 1869-70, but this fragile-looking design offered little improvement on the axial pattern. The inventor was still championing the direct-reading system, and a flimsy curved arm, with a slender cylindrical tail rod, was simply slipped on to the piston-rod extension and clamped in place with a small threaded nut. The tail rod was supposed to steady the assembly by passing down through a small horizontal plate protruding above the cylinder cap, which allowed the whole tracer unit to turn until the pointer was brought to bear on the paper.


The new-pattern Hopkinson indicator, clearly inspired by the Richards design but lacking an amplifying trace mechanism, was made in small numbers in the early 1870s. This example is numbered '209'.
Museum of Making collection

Play in the tracer mechanism and the use of springs that were unnecessarily large, owing to the absence of amplification, were too much of a handicap to allow accurate readings to be taken. Consequently, the parallel-axis Hopkinson indicator was in vogue only for a very few years. For a long time, none had been identified; then, in the space of as many weeks, two examples were found.

Hopkinson indicators retained their popularity in northern England into the 1880s, losing ground there only as the mill and factory engines increased in size, speed and power, and there is circumstantial evidence to show that they were also popular in Cornwall.