aero and automotive engines and engineers


Some Notes and thoughts on the ill-fated 20hp Armstrong Siddeley of September 1938
by Richard Hodgson

This article first appeared as an article and follow-up article in about early 1990 and 1991 respectively in "Sphinx", the magazine of the UK Armstrong Siddeley Owners' Club.  The version below is a re-edited and expanded version formed from an amalgamation of the two original articles and some notes privately distributed in 1997.  Please use your Web Browser's "Back" button if you want to return to this article after visiting any hyperlink which are shown in bold dark blue.  Please click here to go to the list of articles on this site's home page.


In September 1938, two completely new Armstrong Siddeley models were announced, namely the 16 and 20hp.  The 20hp was of especial interest as it was the first ASM car to be announced featuring independent front wheel suspension.

Some of the contemporaneous overviews and articles of the 20hp are to be found in:

        AUTOCAR,  Sept 16th 1938,   pp 513-515.   Overview of 1939 models, including new 20hp model - illustration
        shows "AS 18" registration (see below).

        MOTOR,  Sept 20th 1938,  pp 285-8.  Overview of 1939 models, and road test of 16hp saloon.  Again
        illustration on 20hp shows "AS-18" on the plate.

        MOTOR TRADER,  Sept 21st 1938,  pp 304 & 306.   Overview of 1939 models. Illustration of 20hp model
        shows "AS-20" on plate.

        AUTOCAR,  Sept 23rd 1938, pp 545-6.   Road test of 16hp saloon.

        PRACTICAL MOTORIST,  Sept 24th 1938,  p.762.   Overview of 1939 models.  Illustration of 20hp model
        shows "AS-20" on plate.

        AUTOCAR,  Oct 7th 1938,  p.642.   Technical details in table form.  No output nor turning circle quoted for
        the 20hp, though given for 16hp.

        MOTOR TRADER,  Oct 12th 1938,  p.51.   Exhibits on Armstrong Siddeley's stand, including 20hp Ensign
        and Atalanta models.

        AUTOCAR,  Oct 14th 1938,  pp 716-7.   Details of Armstrong Siddeley's stand.

        PRACTICAL MOTORIST,  Oct 15th 1938,  p.865.     Details of the stand.

        THE AUTOMOBILE ENGINEER,  Nov 3rd 1938 - various pages.    Detailed account of most makes' new
        models - diagram of 20hp's ventilation system at p.420.

        AUTOCAR,  Mar 10th 1939,  pp 408-9.   Listing of new cars, includes 20hp.

        MOTOR TRADER,  May 17th 1939,  p.349.   Current new car prices, no 20hp listed.


Mention is made above to the illustrated registration numbers on the ill-fated 1939 20hp models (some shown as 18hp).  There may well have been a last minute change in size.   The 20hp car was only some 3 cwt heavier than the 16hp, so not too much extra power would have been required to give it the same performance as the 16hp, which as the Autocar stated in its enthusiastic road test (Sept 23rd 1938):

"No Armstrong Siddeley previously has performed in relation to its engine size as this new 16hp model does".


Comparing various original materials indicates that the specifications given below are not too controversial and are taken from:  the above articles;  Glass's 1939 Register of new models;  and two Armstrong Siddeley brochures (the front page of one is green at the top and grey at the bottom;  the other green at the top and yellowish/fawn at the bottom.  The latter gave no prices, and referred to the models by both their RAC and bhp outputs - 14/45; 16/65; 20/85; 17/60; and 25/85.  Both at the National Motor Museum, Beaulieu, are dated September 1938).  The prices for the 20hp models were to have been:

Chassis 400
Ensign Coach saloon 535
Town & Country saloon 555
Atalanta touring saloon 585


The engine and general method of construction were along similar lines to the 16hp model, but it was to have been a somewhat larger and roomier car.  The Automobile Engineer of 3rd November 1938 stated that the main constructional difference between the engines was that whilst the riser beneath the downdraught economiser carburettor on the 16hp was short, in the 20hp the riser "T" junction was cast integral with the exhaust pipe, the inlet gallery tracts being separate and of aluminium.  It added that the arrangement of separate aluminium galleries had long been employed by Armstrong Siddeley on the six port induction systems.  In the 16hp engine, described as having a compound inlet manifold, there was a two port external aluminium pipe, which in turn fed a long single gallery cast within the cylinder head, there being a restriction in the centre of the main gallery.  It must be noted that The Motor at p.287, 20th September 1938, indicated that the external branch and internal manifold of both engines were the same.  If there were an error, the evidence of The Automobile Engineer is usually to be preferred, or it could be an indication that the design of the 20hp engine was not 'finalised' at its launch, or that the original engine for the 20hp was to have been the new 16hp unit bored out to 18hp.

Play was made of the fact that the 20hp models had a proper inbuilt heating and ventilating system using a small fan with its own radiator (though The Automobile Engineer stated that there was no fan).  Fresh air was drawn from the front of the car.  The system provided heating and ventilation into the front and rear either separately or simultaneously.  The radiator and outlets of the system were underneath the front seats.  A double dashboard of special construction was to give complete insulation from under bonnet fumes.  All heat and noise were to be extracted beneath the chassis by a special kind of fume extractor.  According to The Automobile Engineer, the double dash had its forward face concealed against the bonnet, and extracted fumes and heat-laden air were carried through a duct in the double dash to an orifice beneath the car, extraction being assisted by a venturi.  There was also a deflector plate at the top of the forward dash to deflect the heated air, this action being assisted by the direct air stream created by the [engine's] fan.  Thus a cool compartment was to be obtained, and the engine was stated to be working under better conditions as well.

The heating and ventilating system was almost identical to that disclosed by patent 504,093 (in the names of Armstrong Siddeley and Christopher Sibley Oliver, the then General Manager) whereby air was to be drawn in from the front part of the car and released under the front seats through a special radiator.  The patented system was silent as there was no fan and released air at a low velocity to ensure correct ventilation.  A fan was only required if the system had to work when stationary when it was to be introduced outside the car body.

The application for this patent was only made on 4th August 1938 - only about 6 weeks before the car was publicly launched, and an even shorter time before details would have been released to the press.  ["Motor" were able to include a full test report of the 16hp at the time that car was first announced].  If details of the patented system had been disclosed or "leaked out" before the 4th August, this would have been most strong grounds for holding the patent invalid.  There would have had to have been secrecy amongst those necessarily few [and trusted] employees who would have been involved in its research and development.  The inclusion of this system on the 20hp as announced was probably a last minute addition.  Given this, the differences in induction between the 16hp and 20hp, and the question of a last minute increase in engine capacity, it is possible that the 20hp was a project which was brought forward and "marketed" before it was really ready.

The Luvax piston type shock absorbers were stated by the Motor to have had finger-tip control.   The chassis was automatically lubricated by operation of the "clutch" pedal.  The brochures stated that a feature of the chassis was "Automatic lubrication using engine oil" - an ambiguous statement.

The biggest difference between the 20hp and the 16hp was the (Andre) Girling independent front suspension on the 20hp.  The inventor was Albert Henry Godfrey Girling, who applied for a patent on October 24th 1934, the patent (432,157) being granted on July 22nd 1935, and an improvement (492,771) was patented by James Henry Pratt, Gilbert Edgar Manley and Girling.  Application was on April 28th 1937, grant on September 27th 1938 - roughly speaking there would be provisional protection at any time after the application was made, so that the invention could be commercially exploited after the application date.  On grant, the provisional protection becomes 'absolute'.

This suspension was not the best idea to have sprung from the fertile mind of Girling, due to conflicting motions of the links and many rubber joints (and rapid wear if rubber joints were not employed), which lead to imprecise steering.  Depending on the way the patent was then worked the impreciseness in steering could occur when new - see. for instance, the criticism in "The Motor Vehicle", 11th edition, 1989, by Newton, Steeds & Garrett, published by Butterworths, a standard automobile engineering reference work.  As a matter of record, Rover adopted a Girling type front suspension in its P3 range of 1948, the Autocar review of the Rover P3 (Feb 13th, 1948) stating that:

"The Rover engineers have been somewhat slow to move over to independent front suspension for reasons which can well be appreciated.  I.f.s has been, and still is, by no means an easy nut to crack, for with the easier riding and slow motion springing it is apt to introduce imperfect steering and a less certain stability".


The Automobile Engineer stated (November 3rd 1938) that this sort of suspension demanded an exceptionally rigid frame, that all the joints of the linkage were fitted with Silentbloc bushes to allow small deflections, and that in the frames employing it provision was made accordingly.  At page 406, the problem of the various motions in the components of the suspension and steering were mentioned, The Automobile Engineer then commenting:

"All of these points have apparently been considered, and a most ingenious arrangement is provided .... "


Whilst Armstrong Siddeley used a similar engine to the pre-war 16 for its first range of post-war cars, it adopted a very different independent front torsion bar suspension system.

In all the overviews and newspaper articles I have seen so far, the 20hp was relegated to the status of an "also-ran" in comparison with the 16hp from the moment it was introduced in September 1938.  I have yet to find any reference to any reviewer or journalist having even driven the car, yet alone put it through a road test.  Note that that the 16hp was immediately graced with detailed cut-away drawings and full-page photogravure plates.  There has to be more than a possibility that quite apart from considerations of the factory reducing output of cars due to rearmament, the 20hp was a 'lemon' from the outset from the steering and/or engine point of view.  In the years before the war (and not only then!), motoring journalists would not run down a model from a respected British manufacturer, but simply would give little coverage and then quietly forget about its existence.

It is interesting that the usually accurate Glass's 1939 Register of New Models that appeared in early 1939 lists the 20hp as having semi-elliptic front [sic] suspension.   Similarly, Culshaw & Horrobin's "The Complete Guide of British Cars", Macmillan, 1974, shows the 20hp as having semi-elliptic suspension.  Whilst they do not quote any specific source of reference, the entries for weight and wheelbase in the latter work are blank, whilst the weight is given in Glass's 1939 Register.  There is thus the possibility that there was a plan to change the front suspension to the traditional layout, but as with most matters concerning the 20hp, this remains speculation.

The 20hp models on show at the Earls Court show in October 1938, stand 130, appear to have been the Ensign and the Atalanta.  Note that there is no separate external illustration of the 20hp Town and Country Saloon in the brochures mentioned above.


Other specifications were:

Engine:   6 cylinder,  75 x 105mm,  2783cc,  20.9hp.   4 bearing crankshaft.   No output quoted in contemporaneous articles, unlike 16hp.    Brochures state 85bhp.

Carburettor:   Zenith, large air cleaner and silencer (much larger than 16hp).

Wheelbase:   10'.  Track:  4' 9" front,   4' 10" rear.

Road clearance:  6 3/8"

Length:   15' 6" }
Width:    5' 11" } not necessarily for same body, as was some variation.
Height:   5' 9"   }

Weight:   30 cwt.   The chassis did not have the steel floor pan of the 16hp.

Turning circle:  39' - according to Glass's Register of new models that seems to have appeared in or about late February 1939.

Petrol tank:  14 gallon - 2 gallons from the reserve tap.

Brakes:   Bendix Cowdrey, pistol-grip type handbrake.

Gear Ratios:  4.7;  6.4;  9.36;  15.95

Rear Axle:  Hypoid bevel - unit in the brochures looks just a little odd.  The 16hp started with a spiral bevel unit, but changed to hypoid unit at about chassis U3611.

Suspension:    Andre-Girling (front),  long semi-elliptic (rear).    Silentbloc bearings.  Double piston-type shock absorbers (see above).

Tyres:   Dunlop 6.5 x 17.  Rim: 3.62 x 17

Silencer:   Burgess dual multi silencer mounted on rubber.

Jackall hydraulic jacks.  Illuminated panel with satin-silver instruments.

Biflex headlamps - two pass lamps.

Colours:    Black, Blue, Green, Grey,  Fawn and Red.

One of the bodies won a coachwork prize at the October 1938 Motor Show.  As already stated, the 1939 Glass's Register of New Models of about February mysteriously lists the car as having a traditional non-ifs front suspension.  However, the car disappeared from public mention after about mid-March 1939.


So what might have happened to the 20hp?

The ASM Company Minutes of 13th September 1938 show that the car was not ready, but that it was due to be completed during that week.  The 20hp was announced in the 16th September edition of Autocar, which would have gone to press a day or two earlier!   C. S. Oliver, the then General Manager, told the Board that he "could not undertake to produce the 20hp car in quantities until February 1939.  Output would be satisfactory but not rapid.  The coach work for the 16 and 20hp family saloon only would be manufactured by Burlington and other types of bodies would be the subject of subcontractors".  Frank Spriggs (later Sir Frank), the managing director, pointed out that "on the 16hp the [profit] margin was insufficient and that in the 20hp car a loss was inevitable.  He wished to express to all concerned that this was the last occasion on which sanction would be given for such a step and he hoped that every effort would be made to eliminate such [loss] margin".  On 13th October 1938, the minutes reveal that the first 20hp demonstrator car was on the road.   Specifically referred to (page 118 of the Minutes) is the fact that on the same day the company entered into a licence agreement with Andre Components Ltd to use the Andre-Girling suspension.  Whether or not, as has been suggested in the past, there were any plans to take a licence from Mercedes for the independent front suspension, the 20hp was to have had the much criticised Andre-Girling system.  The licence fee for this suspension was probably low, probably very low.  More importantly, part of the Andre sales pitch had been the claim that existing chassis designs could be used with little alteration compared to the extensive stiffening (often aided by substantial x-frames) usually required when independent front suspension was used.  By late 1938, this claim was already being treated with caution, as witnessed by the above comments of the Automobile Engineer.  As an interesting aside, ASM had had experience of negotiating the grant of licences from German companies in the 1930's, including, in the mid-30's, a licence for a German steam-cooled gas-turbine design (which was never built by ASM).

There is no further reference in the minutes to the 20hp, save that on the 22nd March 1939, the Board decided that the 20hp be withdrawn "as a matter of policy" and that a short 25hp be made available for the next Motor Show.  No further details are given.  This brief entry begs more questions than it answers!

At least three 20hp's were initially made - An Ensign and Atalanta (plus a chassis) for the 1938 Motor Show.   Shortly afterwards, there was demonstrator.  The exact number produced remains unclear, as it now [1997] seems that, inter alia, not all car experimental cars from late 1938-1941 were entered into the official books .... [The books matter is a very different story that may be written about one day and almost certainly relates to the books of the entire business that it was considered might have to be produced - and which indeed had to be produced in May 1941 - to various official bodies and government inspectors].  The final number is probably between 5 and 7.  Lucas has reference numbers for some of the 1938 20hp's electrical equipment.  A Floataire wall chart for replacement shock absorbers indicates a replacement unit for the car.  Interestingly, the original unit is shown as a Girling - not the Luvax unit referred to in "The Motor".  The Jig book for the period lists fewer jigs than might have been expected had the car been put into limited production.  However, it is probable that much of the manufacture of the car was to have been subcontracted.  Note the reference, already in the September 1938 minutes, to some of the bodies being due to be built by subcontractors.  Quite apart from the then worsening international climate, economies in the cost of production were already being demanded of the car division - again see the September 1938 comments of the managing director as to margins on cars.  After the war, such demands were to reach a crescendo during the early work on the replacement of the 16/18hp range.  It is clear from the comments of Spriggs in 1938 that certain models had made a loss in the near past.  As an example, a large (and quite total) loss had finally to be accepted by the company about 8 months before Sprigg's warning to the Board.

If the 20hp did suffer from technical problems, it is possible that front suspension rather than the engine was the main area of concern and that this coupled with the loss on each car resulted in it being dropped.  A reason for favouring the suspension as prime technical culprit is that the 1943 16hp prototypes [of the Lancaster], which were discussed during the Board meeting of 5th January 1943, were first designed with beam front axles.  Even then, this project was called the "Post-war motor car".  On 2nd February 1943 (page 194, Minute Book), Sopwith asked the Board that:  "the question of providing an independent front wheel suspension be given serious consideration".  No doubt as a result of this, only the first car ("16/1") was fitted with a beam front axle.  On 15th March 1943, the Minutes record that a second body was not available and that the first body might have to be used as a sort of "utility" body for both vehicles.  The gearbox is recorded as being "well forward".  By 7th October 1943, it is stated that: "The I.F.S. recommended by Mr. Sopwith was promising in the second car, but not quite right yet".  The second car 16/2 had by then done 475 miles with the "utility" body, whilst 16/1 had done 5000 miles.  16/3 was due to be ready in about 6 weeks with a production gearbox.  The "first body was ready to mount, the 2nd body was ready for painting and trimming, a 2nd utility body should be ready in 2/3 weeks".  From this, it seems that not only did Sopwith want a prototype with ifs, but also had specified the type of ifs to be tried.

However, the real reasons behind the Board's "matter of policy" in withdrawing the "proposed 20hp" remains unclear.  If the matter of policy related only to the problem of the loss on each car, the simplest solution would have been a slight change in specification justifying an increase in price.  However, when the company's Minutes over a number of years are studied, the phrase "withdrawn as a matter of policy" does seem euphemistic.  The company secretary did not pull too many punches in what he set out in the Minutes when it came to financial matters (unlike personnel and certain, usually adverse, technical matters).  For the record, I believe that had the sales potential of the 20hp been considered not great enough to justify putting it into production by reason of its already highish price, then this fact would have been recorded.  After all, what in effect amounted to the reading of the financial riot act by Spriggs to the Board in September 1938 was recorded.  For anyone interested, the minutes also record that the financial riot act was read again at the Board meeting already referred to on 5th January 1943, and the strongest of readings took place on 13th June 1944, when it was pointed out that certain adverse matters at Armstrong Siddeley could have a serious repercussions on the Group [which I take to refer to Hawker Siddeley given the seriousness of the matter and subsequent firing].

There must be the distinct possibility that the car was a "lemon".  The fact that the 1943 16hp prototype was designed with a beam axle is, I suggest, a strong indication that the 20hp had had suspension problems.  The further fact that design of 16/2 was able to be modified in a very short time to incorporate independent front suspension is a strong indication that there was no physical or financial reason (such as war-time restrictions) why ifs could not have been designed for 16/1.  The Lancaster and Hurricane, publicly announced by mid 1945, had an ifs system that required a [patent] licence licence from Citroen (see Automobile Engineer, 27th November 1952, p.482).

There are three other reasons (all arising since the follow-up article was first published in 1991) which also strongly point to the car being a lemon.

1.    During the spring (or possibly early summer) of 1938, Dr. H. S. Rowell was retained as an employed consultant.  Though a chemist, he had first entered the car and lorry industry in the 1920's, and made a name for himself, including his work on various industry wide committees.  Directors of ASM also sat in on these committees and, no doubt as a result of that, he was asked to come to Parkside.  The timing seems to be more than coincidental.  Whilst the other candidate at Parkside for a consultant was the Deerhound MK 11, there has never been any suggestion from the engineers, some senior, who worked on that project that Dr. Rowell was involved with it at that time.  I suggest that he was initially brought in to sort out the mess on the 18/20hp model.   Unfortunately, his experience covered both suspension and engines, so that appointment does not clear up what might have been the main cause of the problems.  Dr. Rowell was later to become General Manager of all of ASM when C. S. Oliver was required to leave.  Rowell, in turn, was required to leave in 1944.

2.    Rover had enormous problems with the Andre-Girling suspension system that they had started to experiment with - probably rather earlier than Siddeleys.  As [the late] Gordon Bashford, one of Rover's then designers, told me:  "The geometry was simply wrong - it did not work properly.  It did not line up in three dimensions when used in a real chassis.  You had to use enormous rubber bushes - this lead to poor handling.  The stress set up by the suspension in real use, even with the big bushes, flexed even the firmest chassis.  It took a long time and a lot of calculations - including complex projective geometry ones - to get it right.  It was worth pursuing though, because once the system did line-up, there would be less stress on the chassis than with one or two other ifs systems of the time".  When the solution was found, Rover sought a patent for the solution.  The engineer who calculated the solution is a now totally forgotten genius by the name of Paul Scott-Iversen, a Dane who came to England in about 1936.  He sadly died of cancer in 1943 aged just 37.  His work on gas-turbines, including wide range of flow atomising burners, was quite crucial, though no credit has ever been given to him - though another then Rover gas-turbine engineer never felt the need to redirect credit after Scott-Iversen's death.  I have never seen his name appear anywhere.  An article will, I hope, appear about him one day.   The suspension system was successfully used by Rover after the war.

3.  In about June 1939, in extraordinary circumstances, Stewart Tresilian became chief engineer of Armstrong Siddeley at the age of 35.  He was to be fired on 19th January 1942 for even more extraordinary reasons.  His overriding task was to sort out very severe problems with the Deerhound Mark 11 aero-engine, barely touched upon in the Minutes.   Sopwith, Sigrist and Spriggs also wanted him to oversee the car side.   Tresilian had had much experience on both the car and aero-engine side at Rolls-Royce, including the still-born Phantom III V12 ohc [sic] engine, and at Lagonda where he was the chief designer of the V12 ohc engine and car, including its suspension.   He was interested not only in piston engines but also in all aspects of car design, including suspension and steering.  He had raced his type 35 Bugatti in the late 20's and early 30's and won at least one cup at Donnington.  His next car was a type 55.   By the time he came to Parkside, he drove a Lagonda and a Derby Bentley.   Possibly as a result of having owned, raced and designed very high performance cars, he claimed to be reluctant to become involved with the car side of the company.   W. H. "Pat" Lindsey, a brilliant aero-engine engineer, later Chief Engineer and who was ASM's last Technical Director, told me that Tresilian had told jokingly him shortly after he (Tresilian) joined the company that he had asked the Board to put it in writing that he would not have anything to do with the car side!  Not that this stopped Tresilian from causing a car to be constructed to his own design (which apparently did not go through the books).

In any event, there was a clear dichotomy between the car and aero design offices at Armstrong Siddeley at this time.  When Tresilian was first invited, as chief engineer of Templewood Engineering, to Parkside in about March 1939 (at the firm request of Sigrist), he was asked to test drive what was much later recalled as a "new model".  In December 1990, Mrs. Tresilian clearly recalled that her husband had been "most rude" when asked to give his view about the car and what should be done to improve it.  The author is confident that the car was the 20hp and that the comments would not have been entered verbatim in the company's Minute book.  In fact, there are no references at all to Tresilian or even to his appointment anywhere in the Minutes even though he was the chief engineer, though he never had a contract of employment ....  But, as they say, that is another story!


These "most rude" comments, together with the projected financial loss, almost certainly caused the directors to decide that as a matter of policy the car be withdrawn.  However, even as of 2000, whether the engine or suspension - or both - was or were the main technical culprit or culprits is not completely clear (at least not to the author).

Until recently, the author believed the suspension to be the prime technical suspect, especially due to the problems faced by Rover.  But on re-reading and re-editing for inclusion on this web-site the articles concerning Allen's investigation on how to increase - urgently - the output of the post war 16hp engine and the "resulting" Thornett design of an all-new aluminium engine of nearly 24hp (RAC) whose actual output was only marginally greater than that of the pre-war 20hp engine, I am now less sure.  Looking at the Thornett aluminium engine project, it would seem to have made more commercial sense to make use of and develop the 1938 20hp engine had it been in any way half successful rather than design an all-new engine from scratch which was almost bound to take much longer and cost more (though it would have been lighter).  According to the late John Densham, there was never any mention of using the 20hp engine when the car division was considering, as a matter of urgency, how to increase the 16hp's output.  When I asked him in 1993 about the then likely position of 20hp drawings, he said that sufficient engine and chassis drawings - but not production drawings, tooling and the like - were likely to have survived the war.  He recalled that car engine and chassis drawings said to have been lost/destroyed during the war had a habit of turning up on the quiet if absolutely required.  The fact that the 20hp does not appear to have been considered in any serious way after the war may be a pointer to the "final" version of the 20hp engine being less than satisfactory.

Given from the review/overview articles set out above that the pre-war 16hp and 20hp engines seem to have been essentially the same design, save, importantly, for cylinder head and external manifold design, it is interesting that the stroke as well as the bore of the 20hp were larger than that of the 16hp.  The 16hp engine was similar to the post-war 16hp, and when the time came to increase the latter's output, only the bore was increased, taking the RAC rating to a nominal 18hp.  18hp is the rating that the first two 1930's articles (above at the beginning of this article) very much tend to indicate the 20hp as being, as all the cars illustrated had registration plates bearing their RAC h.p. - i.e. the 16hp is shown with 'AS 16', 25 with 'AS 25', and 17 with 'AS 17'.  Further, there can be very little doubt that the "20" of the "AS 20" on registration plates of the 20hp models illustrated in the two sales brochures mentioned above have been retouched quite crudely - unlike any of the other registration plates illustrated therein (see pages 14 & 18).

An increase of the 16hp engine from RAC 16hp (actually about 15.7) to RAC 20hp (actually about 20.9) would of necessity have required an increase in the bore - stroke playing "no part" as such in the RAC formula.  Such an increase in bore (assuming it was physically possible), might then have caused a certain weakening in the engine structure if the bore increase were simply accomplished by boring out.  The bore of the '16' was 65 mm, that of the post-war '18' (actually about 18.2) - and no doubt any possible pre-war '18' - was 70 mm.  The bore of the 1938 "20" was 75 mm.  It is interesting to note that the stroke of the post-war 18hp was the same as that of the 16hp - the engine had been "merely" bored out.

Further, and strangely, the 16th September 1938 Autocar overview stated that the "modern trend as evidenced in this new Armstrong Siddeley engine" was towards a shorter stroke.  This may be a further indication that the 20hp initially had a 16hp engine bored out to 18hp.  Further, it might be quite possible that the 100 mm stroke of the 16hp engine bored out to 18hp [or yet further to 20hp if possible] was then increased from 100 to 105 mm at the last moment to increase power yet further (but such an increase in stroke is not reflected in the RAC hp figure).  However, a last minute increase in stroke is usually very unlikely as it nearly always entails a quite major - to very major - change of design if the original smaller stroke engine has been well designed in the first place from the point of view of minimising overall exterior dimensions and weight.  Such design change was and still is highly expensive and usually would include the crank, crankcase and cylinder block (especially if the upper half of the crankcase and block are cast as one).   Such a last minute increase in stroke (assuming the increase happened at this time) could well point to the suggested original 18hp [or to an "intermediate" 20hp 100mm stroke design if that were possible] version of the engine being underpowered and/or a bit of a lemon.  An increase in bore brings greater mechanical losses, which is probably the very last thing the 16hp engine needed.  The post-war engine 16hp engine, similar to the pre-war 16hp unit, was found by Allen in his urgent Report for the Board to be suffering from mechanical losses to an unusual degree.  An increase in stroke may thus have been required to obtain the desired output when the bore was increased to 75mm.

Importantly, the apparent last-minute increase in stroke and engine RAC rating from 18 to 20hp might also go some way to explaining why, initially at least, a loss was going to be inevitable on the 20hp.

I suspect that it is quite likely the engine had to be redesigned in a rush not only to allow a longer stroke, but also to allow the use of the desired larger bore (75 mm c.f. the assumed 70 mm for the 18hp) with some degree of confidence.

Possible reasons for a last minute increase from 18hp to 20hp (actually 20.9) with a longer stroke include:  (1) The 18hp did not give sufficient extra output compared to the 16hp and/or was rough in part of its running range.  The 20hp was rated as 85hp and had a different induction arrangements compared to the 16hp (the induction system on the generally similar post-war 16hp engine was the subject of adverse comment by Allen);  and (2) at least one other manufacturer of a competing range to the new Armstrong Siddeleys, Rover, had a new high output "14" engine (1901cc) which had been publicly announced by mid-August 1938, and from which it was clear, I would suggest, that the Rover "16" and "20" outputs were to be increased shortly thereafter, as indeed was the case by about mid-1939.   Rover enjoyed great success in sales in the mid to late 30's, their models having been dubbed the poor man's Rolls-Royce and the "Rolls-Royce of light cars".

From all of the above, I would submit that it is not implausible to suggest that Armstrong Siddeley were originally going to release the 1938 20hp model as an 18hp with a bored out 16hp engine, but decided at the last minute that yet further power was going to be required and decided on [i.e. had thrust upon them the need for] a radical engine redesign.  Unfortunately, neither the late Mervyn Cutler nor the late John Densham, both of whom were only the aero-engine side of the business before the war, could through any light on this point.

Finally, it should be recalled that compared to the very well received 16hp cars, the cheapest 20hp model was over 40% more expensive (and still due to sell at a loss).  A yet further increase in performance and/or refinement might well have been expected by customers.



At the moment [2000], and for what little it's worth, my view tends towards both engine and suspension being troublesome.  This is, of course, without prejudice to the fact that the car was going to sell at a loss.


1989, 1990, 1991, 1997, 2000  Richard Hodgson - may not be reproduced in whole or in part without permission


With thanks to Leland Crosthwait, editor of "Sphinx-NZ" - the journal of the New Zealand Armstrong Siddeley Car Club, for pointing out some typographical errors in the previously posted version of this article - RAH 2nd February 2001


Back to List of Articles on Main Page of www.wolfhound.org.uk