My previous articles described the basic operation
of the two-stroke engine and pointed out that with modern knowledge its
operation is not as simple as first thought. For this reason early attempts to
tune two-strokes were not that successful and for the between the war years,
four-strokes dominated motorcycle racing. The main exception being the supercharged DKWs in the 1930’s that
were developed under a “no expenses spared” racing budget. They were famous for
the noise they made and their fuel consumption and it is no wonder that they
became the best known pre-war racing two-strokes.
There were other companies who tried their hand with
two-strokes and one of them was Levis, an English motorcycle company. The company
was owned by the Butterfield brothers and the engineer responsible for
development was Bob Newey. The name Levis came from the Latin Levis et Celer
(light and fast). In the early 1920’s the company had considerable success in
the 250cc (lightweight) class with a simple deflector piston two-stroke powered
bike ridden by Geoff Davison.
Through a mutual friend who knows Geoff’s son,
Martin Davison, I have been able to loan copies of his late father’s books
“Racing Through the Century” and “Racing Reminiscences”. These describe the
trials and tribulations of the Levis racing team during their brief successful
period. I am sure that these will be of interest to BTSC members and a synopsis
follows for your delectation.
The power output of any given engine is given by the
simple formula
Power
= PLAN (divided by 33,000)
where P
= bmep (brake mean effective pressure)
L =
Stroke
A =
Bore
N =
Revs Per Min
You do not have to be a mathematical genius to realise
that increasing any or all of the above terms will increase power. Some are
obvious, if you increase L and/or A you will have a bigger engine and hence
more power. Increasing the revs (without blowing up the engine) will produce
more power and this approach was successfully adopted by Honda with their
four-strokes. This leaves P which in simple terms equates to getting the most
mixture into the cylinder, compressing it and igniting it efficiently which as
you can gather is a very important factor.
Put yourself in Bob Newey’s shoes and ask the
question how are you going to make your 250cc deflector piston two-stroke
produce more power based on the knowledge of the 1920’s? You should bear in
mind that Bob was aware of the fact that if you raised the compression ratio
above a certain value you got pre-ignition and that experiments with an
aluminium piston had been disastrous. They had melted!. So cast iron pistons
would have to be used. Bob was under no illusions on the task ahead of him in
preparing a winning race engine for the 1922 Lightweight TT. As he told Geoff
“I cannot give you the speed of the New Imps or the OK Supremes but I can give
you better acceleration and greater reliability”. Geoff for his part
concentrated on lightening the cycle parts. By picking the lightest of
everything, 10lbs was saved and by using narrow bars, dispensing with footrest
rubbers, drilling holes here and there, fitting a bicycle saddle etc, another
10lbs was saved. When finished his bike was 20lbs lighter than the other works
bikes. He gave every attention to his attire for the race and he started the
race in a cricket sweater, drill breeches, stockings and dance shoes. To stop
the breeches flapping at speed he fitted his girl friend’s garters round his
thighs. The rules on race wear have changed since then!!
A special cast iron piston was machined with 3 thou
clearance at the top reducing to 1.5 thou at the bottom. In order to maximise
the “P” term in the power equation, the
C.R. was set to be as high as possible without causing pre-ignition. To
keep the cylinder temperature down, the top fins of the cast iron barrel were
machined off and an alloy finned replacement was shrunk on. Crankcase
compression was increased by alloy stuffers which extended right up under the
piston so at bdc (bottom dead centre) they almost reached the gudgeon pin. The
1922 racing engines ran on roller bearings and had a bore and stroke of
62.5 x
82.5mm giving a capacity of 248cc. The engine was not a “high revver”
having a limit of around 4000rpm but with the attention paid in it’s assembly
together with its light weight it was reasoned (quite correctly as it was to
turn out) that gains would result in reliability, braking and acceleration. The
port timing of this engine would appear to modern eyes to be on the short side
with 115, 140 and 128 degrees for the transfer, exhaust and inlet periods
respectively. A Brown and Barlow racing carburettor of around 1” bore was
employed with a variable jet and the inlet pipe was straight and short with the
bell mouth projecting sideways into the air stream. the centrally disposed
single exhaust port fed a short pipe of around 1.5 inch diameter which in turn
fed a cylindrical box at the front of the frame that had a short exit pipe. The
engine was noisy....! Ignition was by a
chain driven magneto. Lubrication was by a worm pump with adjustable sight feed
which fed the flywheel side cylinder wall and the inlet pipe. Castrol R was the
oil used. The crankcase, as it now used roller bearings instead of plain
bushes, now needed crankcase seals and this was acheived by long glands
machined in the shaft beyond the main bearings.
Though the new TT bike employed nothing in the way
of an innovative breakthrough in the art of two-stroke tuning, it did accept
the known limitations of the two-stroke and its design and preparation took
these into account to end up with a competitive machine.
Geoff did his early TT practice on “a gutless bike”
but when the new engine arrived “in a brown paper parcel under Bob’s arm” four
days before the end of practice, it was immediately run in with an over rich
mixture which caused a lot of four-stroking during practice. When the engine
was stripped down they were surprised
to find that the plug and piston were covered in aluminium. It was clear
that the con rod had been scraping the crankcase stuffer blocks. The con rod
had been cut down too much - “not
enough H section” was the conclusion drawn. The continual four-stroking during
practice had caused it to flex. For the race the mixture was weakened to stop
it four-stroking and this cured the con
rod problem.
On race day the Levis ran faultlessly to win by over
14 minutes from a four-stroke Rex Acme and in doing so set a record speed of
just under 50mph. Max speed of the little bike was about 60mph. With wins in
the French and Belgian Grands Prix as well, the Levis-Davison combination
cleaned up the Classic events for 1922.
For 1923 Bob Newey decided that the 1922 engine
represented very nearly the limit in efficiency for a straight forward
two-stroke design. So he thought (like many over the next three decades) that
the only way to increase the power output was to force more mixture into the
cylinder prior to ignition and the way to do this was by some form of
supercharging. He chose to add another piston to the crankcase such that more
mixture was sucked into the crankcase and it was compressed higher when
compared to a normal two-stroke’s crankcase compression. This extra crankcase
piston had a bore of 55mm and a stroke of 50mm (giving 119cc) and was operated
by a crank on the offside of the engine shaft. The crankcase had a small
aluminium cylinder cast in its side to accommodate the supercharger. Similar
crankcase pumps were later developed with considerable success by DKW as their
“Lade pumpe”.
The new engine gave an extra 4mph plus considerably
improved acceleration. “Little Roger” was the name given to this new
development. However it suffered from one major snag, namely vibration. test
sessions at Brooklands showed that frames lasted 5 miles before they broke. The
extra piston should have been balanced but this factor had been overlooked and
as the season had already started, the project was abandoned and the engine
ended up under the bench in “The Holy of Holies, the Levis Comp Shop”.
So for 1923, the “standard” engine was re-worked and
turned out to be even better than the previous year’s model. Unfortunately it
was not ready for the TT but it did win the French GP and was second in the
Belgian. At Brooklands it averaged over 65mph before seizing whilst in the
lead.
For 1924 it was agreed that the four-strokes were
not only faster but were proving more
reliable in the 250cc class, so it was unlikely that a balanced “Little Roger”
would be competitive. the decision was made to concentrate on a new racer for
the new 175cc Ultra Lightweight class. Its tuning “secret” was to employ a much
higher compression ratio than ever used before. To cope with the necessity to
get rid of the resulting extra heat, copper would be used instead of aluminium
for the head and fins. Bore and stroke for this new racing engine were 56
x 70.5mm and it was nick name “Copper
Nob”.
In TT practice, the bike went well during the early
morning sessions but during the race, after taking an early lead, the motor
began to “pink”, it lost its edge and finished a disappointing fourth. The
trouble was diagnosed as being due to the copper fins not radiating enough
heat. they finished third in the French GP with the same problem since there
had been no time to do anything to try to solve the issue. The solution was an
application of elementary physics. When the burnished copper fins were painted
black it made all the difference to getting rid of the excess heat. Geoff went
on to win the Belgian GP by over half an hour and finished third in the 250cc
class.
This success at the end of the 1924 season was to be
the last for Levis. Two-strokes were losing their appeal for the sporting
motorcyclist and in 1926 they launched a 350cc four-stroke. So ended their
two-stroke development and for the UK at least, two-strokes became solely
associated with cheap, economical utility transport. Foe those of you wishing
to know more about the Levis marque I would recommend getting a copy of The
Classic Motorcycle of September 1993. the late Brian Woolley’s article features
his own racing Levis and historical details of the company.
James
Barnett Jnr