Historical Notes: The Water Cooling Systems



If we accept the edict of Sadi Carnot regarding the efficiency of the prime mover in a motor car, we cannot go far in power production without being brought face to face, as it were, with “heat rejected.”

In a previous article we saw that a lot of this went down the exhaust pipe, but an approximately equal amount would “heat up the ironmongery” to a dangerous degree if something were not done to keep the temperature down, and this necessitates what we know as a cooling system.

This may range from the “bare nothing” of Pennington to the modern pressurised liquid system. Of course, in the early days you were expected to descend from your motor car so often during the course of a 50-mile journey that once more to put in a gallon or so of cooling water made no appreciable difference, especially as steam cars had made the practice popular. In consequence, the first expedient resorted to was a water tank, pure and simple, in conjunction with a water pump, and very surprising quantities of water some of the early cars carried in this simple container. Anything up to 20 gallons, in fact. Even as late as 1904, light cars were averaging six or seven gallons of water capacity. This was, of course, a natural legacy from the internal combustion engine before the motor-car people got hold of it, nearly all stationary versions at that date relying upon a large static tank, topped up, if need be, by pump from the factory plumbing or a nearby pond or river. One wonders if some of the early versions did not rely on this for their successful operation, but that is another story. Indeed, the same basic idea is frequently used today for test-bed work, as it gives a good and simple method of controlling the water temperature and “upping the output” if need be! Necessarily all this threw a great responsibility on the water pump, usually driven by friction from the flywheel, very conveniently exposed in those days. It is astonishing how difficult it proved to be to make a trouble-free pump, too, especially as the principle was, even then, very old, and stuffing boxes and suitable greases were both readily available. Probably it was the unskilled driver who caused much of the actual trouble, but the real root of the matter appears so absurdly simple that, even at this date, one still marvels that no one realised that the poor wretched pump was grossly overworked, although this did not become quite so apparent until radiators took their great step forward in 1900. Before going on to radiators, however, it is perhaps illuminating to see what the water consumption of some of the earlier cars was.

The Automobile Club ran some trials over 50 miles in 1899, an average horse-power of the entrants, equipped with water tanks and pumps, being about 4 h.p. (declared), and the average water evaporated about 1.5 gallons, or 15 lb. per hour. Not allowing for what heat was lost by radiation, this appears to represent some 4 1/2 h.p. and probably meant about 6 h.p. in actual fact. Thus we have an interesting idea of the thermal efficiency of the engines of those days (in fact, about 18 per cent. in 1906).

Among those fitted with primitive radiators during the same trials (of a mean declared horse-power of 5), the average water evaporated was about 4 lb. per hour, adequate testimony to the added heat-carrying capacity that even the earliest radiators allowed. The coming of the “gilled tube” radiator was therefore the first real step forward in the development of the cooling system, but, astonishingly, they were always arranged “in series,” and up to 1900, never in parallel. Considering the amount of general knowledge that existed at the time concerning industrial heat exchange problems, to say nothing of steam-engine boilers, one may correctly describe this fact as astonishing. But the only remedy that seemed to suggest itself to the early technical folk was to ask the poor old water pump to supply a greater and greater flow at higher and higher speeds, to cope with the inevitably increasing heat rejection as the horse-power at the shaft went up. An early parallel radiator system was made by Mr. Escourt, rather complicated, but interesting as one of the earliest examples of the “thermo syphon” system that dispensed with the water pump. Very little water was lost from this system in 160 miles of motoring and it is to be hoped that Mr. Escourt reaped the due reward of his pioneering effort. Of the many different types of tubing used in these early radiators it is not possible to write fully, as everyone seemed to have his own pet idea. Some had “crinkled” gills, some round, some square, and some with little bits of early wire sprouting out of the main tube, rather like those frilly bits of paper on the ends of chicken legs in “high class” restaurants! Pathetic to think that even plain tubes, properly applied, would have solved the problem and halved the number required! The appearance of the first Mercédès car has already been adequately documented, the honeycomb radiator being, of course, the most prominent feature, and at once the gilled tube and the water tank started to die a natural death. Nevertheless, there were still a number of cars so equipped in 1906, and it is odd to read the following description of the 1903 Napier of 15 nominal horse-power: “The flywheel drives a rotary water pump, the water tank holding a supply for about 200 miles …”

There were some motor-car trials held at Hereford in August, 1905, and the following details are quoted from Beaumont’s second volume. The area of tube surface in gilled-tube types ranged from 1.84 sq. ft. for the 8-h.p. Rover up to 5.38 sq. ft. for the 14-h.p. Belsize, whereas similar figures for the honeycomb-radiatored types went from the 54.5 sq. ft. of the 8-h.p. Rochet Schneider up to the 119 sq. ft. of the 30-h.p. Humber, adequate testimony to the improvement wrought by the Mercédès example, it not being difficult to see why the water tank rapidly became redundant, and shrank to the now common proportions of the “header tank.”

The tubes were not small at first, the average figure at the Hereford trials being about 0.38 in. diameter, the smallest being the Humber with 0.25 in., these figures comparing with the 0.5 to 0.75 in. of the older tube-type radiators. It is interesting to note that some motorcar people were early attracted to the idea of flat tubes, but the 1911 Prince Henry Cup was the event which proved the fallacy of that one, which was that if the water boiled, the steam pressure distended the tubes slightly, thus reducing the “air-passing capacity” as if by magic, the mountains being needed to bring the mysterious trouble out.

The next step in the story of the radiator, once experience had dictated the sizes and flows necessary for the varying sizes of motor car, was that of overcoming the problem of making them in quantities. “… the radiator is a complex affair, and soldering it used to be a matter of skill. There are 95 tubes in a radiator. Fitting and soldering these tubes in place by hand is a long operation, requiring both skill and patience. Now it is all done by a machine which will make 1,200 radiator cores in eight hours . . .” writes Mr. Ford; just like that! — but his mathematics were somewhat astray. But this sort of production and the “crinkled strip” method do remind us all of how much we owe to the production people. One wonders what Mercédès paid for a radiator in 1902? Passing on to the “vintage” era, we find the tube and corrugated-strip tyres competing for the honours, but to describe in detail what was, after all, the first thing you saw in a motor car in those days, would be merely boring. The writer has purposely avoided reference to this component being the trade mark of the manufacturer concerned, but, of course, this practice reached its zenith in the ‘twenties, but we approach the time when pride of possession in motor cars waned, and they started to become mere commonplace things and had to be counted on to serve commercial travellers for day to day journeys, or ladies on shopping trips, and this meant that in an increasing number of cases the engine stood no chance of warming up properly, so that thermostats and radiator shutters began to make their appearance. They are certainly better now, but there was a tendency for the earlier thermostats to cause trouble, gradually reducing themselves to rusty bits of tin that some people were only too pleased to throw away. But perhaps that is a bit unfair on a mass-produced article that must have saved a few worn cylinder bores in its time. It is hardly necessary to write that the thermostat’s main job was to isolate the engine from the rest of the cooling system during the warming-up period, and the first proprietary models had a nice knurled knob that you could turn to cut the device out if you so desired. Later, presumably, even knurled knobs became too expensive and the whole thing ceased to become optional and disappeared out of sight right inside the water system. In essence the idea comprised a metal bellows, charged with some suitable volatile liquid which, as it boiled, expanded the bellows and opened a valve which permitted the full circulation to take place. In many cases, especially where expense was not so important, a similar device was used to operate “venetian blind” shutters ahead of the radiator. The writer has purposely avoided reference to the artistic aspects of the radiator, but it is impossible to recall the beautiful appearance of some of the shuttered radiators of the past without a sigh of regret for their passing, the 8-litre Bentley, for example: but perhaps we should rejoice that the Rolls-Royce remains to cheer us up. In the cheaper versions of the idea, the usual “post-vintage” faults soon manifested themselves. The pivots of the shutters were often mere tongues of thin metal dropped into holes pressed in even thinner metal and thus they jammed and rattled abominably after a short period of unlubricated existence, so that a little quiet work with the welding torch, a coat of paint and a shrewd puncture of the bellows with a sharp screwdriver was sometimes the best treatment.

But the day of the specialist had dawned, and most manufacturers began to buy their radiators from the specialist. It was a good thing too, as much research was needed into the materials used in construction and the problems of heat transfer generally, and today, as we see a radiator specialist get out his curves and tell us how big it should be, and how much water circulation there should he and how fast the air should pass, let us realise that those curves represent the essence of years of expensive laboratory work, which we get for about 6d. a radiator. It’s not bad value.

Before leaving the subject of the radiator, we must notice the modern pressurised affair, used so that the boiling point of the water may be raised and the engine thus allowed to run at a slightly higher mean temperature; all to the good from the thermal efficiency point of view. The writer will not readily forget the horror of the moment when he unscrewed the Silver Lady from a post-war Rolls-Royce in the sacred precincts, only to discover that the resulting aperture did not communicate with the header tank. It was as though the stars had halted in their orbits, but, of course, the slight tell-tale brown stain of the pressurised cap obviously could not be allowed to show itself in public, the real cap being quite nice, and at least there is a good technical reason for the change, which is more than can be said of some changes these days! There remains the fan, a component that the author admits he imagined propelled the car until someone lent him Bramley’s “Motors in a Nutshell” at the tender age of nine. As with most of the other components, the fan does seem to have “degenerated” from the technical puritan’s viewpoint. Consider the beautiful thing that graced the front of the 38/250 Mercédès , and compare it with the bit of bent tin that is nowadays used; and yet honesty forces the conclusion that the latter is just as efficient as the former, and many times cheaper. The early ducted fans were seemingly based on a misapprehension, that is that they increase the efficiency of the fan. They certainly do, but only over a very restricted range, the fan actually becoming a hindrance above a certain speed. The ducted fan has therefore disappeared, except in those cases where the engine has to work hard at low road speeds, notably on commercial vehicles.

As to water pumps, these have not changed greatly down the years. They removed themselves from alongside the flywheel quite early, and usually found themselves driven in tandem with the -dynamo and/or magneto in the vintage era, but the coming of the endless vee belt made it usual to combine the fan drive with that of the pump, and the two are now invariably placed immediately behind the radiator, and there they seem likely to stay. Needless to say, in spite of the seeming perfection of the modern cooling system, it is designed to cope with a given heat flow, proper to the engine concerned, but there is still, to this day, the same old misunderstanding about how the engine operates thermally. Take a very simple example. A sports-car equipped with a 40-b.h.p. engine. Let us say that we step up the output by 25 per cent, with a low-pressure supercharger, if we allow that the blower requires about 5 h.p. for itself, the insides of our modified engine are now called upon to produce some 55 h.p., and the cooling system and the other heat rejection paths can be very nearly 40 per cent. overloaded. These figures mount alarmingly if greater increases in output are sought. Why don’t we notice these things? Again, simply because we cannot normally keep the load factor high in this country, and a genuine “soak” temperature is never reached, but just try the straight at Le Mans once or twice if you want visual proof of how right Carnot was. What is the thermodynamic phrase? “The thermal distribution of the world tends towards a maximum” -and how! -” A. B. C.”