The other side of the coin . . . REDUCTION DRIVES, ARE THEY WORTH THE TROUBLE? By W. H. Ekin (EAA
65396)
Crumlin Co. Antrim, Northern Ireland
B,
> UT FOR THE fact that the Wright Brothers used a reduction drive in their first successful powered aircraft, the first flight by a heavier-than-air machine
would have been considerably delayed. But for the fact that the Spitfire and Hurricane Merlin engines employed a reduction drive, it's conceivable that The Battle of Britain would have been lost. (Incidentally the Wright's engine weighed 200 Ibs. — a little heavier than a VW — and developed 12.05 hp. Their reduction ratio was 3.3:1 while the Merlin's was 2.1 and 2.38:1.) No helicopter would ever fly but for an extremely large "Prop", albeit its axis is turned through 90 degrees being driven via a reduction gear box by its engine. If the reduction drive is so wonderful, why doesn't every aircraft engine use it? Anyway, how much more efficient is it, and why? I cannot, yet, give a complete answer, but as I have had experience with reduction drives for the last 4-5 years, I can perhaps shed some light on the subject. The latter part of the article gives practical details of my latest reduction drive. What we want from an engine/propeller combination is simply "thrust", which is the name we aeronauts give to that force produced by the prop, in the direction we want to go. Now, Newton formally stated that in order to get a "force", momentum (that is, mass x velocity) had to be increased. From the definition of momentum, we see that we can get a force, by either increasing the mass of something, or increasing its velocity, (or both). The "propeller" creates its force by increasing the momentum of the air passing through it by increasing its "velocity". That is, the prop approaches a particle of air at one speed and "ejects" or "propels" it away at a faster speed. Hence, a propeller can only create thrust if the air leaving it is faster than the air approaching it. As the momentum of the air is the product of two things, namely its velocity and its mass, we could create thrust in two ways. 1. Take a large mass of air and increase its velocity by a small amount (i.e., a helicopter rotor) or 2. Take a small mass of air and increase its velocity by a large amount (i.e., direct driven props on "high" speed engines or jet engines). Just as "love and marriage go together like a horse and carriage, etc.", so does the creation of thrust, which we fervently want, go together with the creation of kinetic energy which is dissipated in swirls and eddies of the prop — slipstream — which we fervently don't want. People cleverer than I have proved that Kinetic Energy equals '/•> mass x (velocity) 2 , and that means if you double the mass you double your loss, but if you double your velocity you clearly quadruple your loss which is obviously a less good proposition. Hence the reason WHY alternative 1. is better than 2.
(Photo Courtesy of the Author)
W. H. Ekin's VW reduction unit. This unit received a 50 hour static test at 85-100% power and came through with flying colors.
In order to try and find out how much better specific
reduction ratios would be, I ploughed through various N.A.C.A. and A.R.C. reports which purported to give an answer, but became despondent when calculations based on their charts and tables did not tally with the thrusts that I have frequently measured in my own backyard. Then my heart leapt with joy when I espyed a very simple empirical formula in Popular Rotorcraft Flying Magazine (July-August 71). Although, there was a misprint in it, it indicated that thrust was directly proportional to diameter, assuming the propeller efficiency was 84%. Now this is far easier to manipulate than the methods
suggested in my various technical papers and having done the calculations per the technical papers, I wouldn't think that its accuracy is any the worse in spite of its simplicity.
Let's consider what this pregnant and sensational statement means to us aero and gyronauts. It simply means that if we can use a 60" diameter prop instead of a 50" one, we can expect about 20% more thrust, or if a VW gives 300 Ibs. static thrust with the smaller prop, it will give 360 Ibs. with the bigger prop
according to the formula which assumes constant efficiency, but since the point of using a reduction is to increase efficiency, then on this count alone the 20%
figure will be too conservative and because of a combination of factors as explained below, we shall see that we'll get a much greater "dividend" than the conservative 20% above. Figure One is general for either aero or gyroplanes, and shows two lines. One is THRUST AVAILABLE and the other is THRUST REQUIRED FOR LEVEL FLIGHT. Now the interesting thing is, that it is the DIFFERENCE between the curves that is the POWER AVAILABLE FOR CLIMB/ACCELERATION. — That is, at 50 mph, 250 Ibs. thrust is required for level flight, but we have a total of 300 available. Therefore, we have a surplus of 50 Ibs. which can be used for acceleration/climb. Hence, it is scarcely surprising that 50 mph is the "best" climbing speed. Now if the thrust is increased by 20% (i.e. to 360 Ibs.), the excess power is 110 Ibs. (all other things being equal), namely, a whopping increase of 120% in the thrust available for climb/
acceleration. In actual fact, the picture is just a little less rosy because to get the extra 60 Ib. of thrust, there will be the additional weight of the reduction drive. However, (Continued on Next Page) SPORT AVIATION 37
REDUCTION DRIVES . . .
(Continued from Preceding Page)
this weight is considerably more than offset by the increase in performance. SUMMING UP: The POWER AVAILABLE FOR CLIMB/ACCELERATION is much greater than proportional to increase in thrust. So much for the theory, what figures have been achieved in practice? Just 2 words of caution: 1) Naturally, I didn't measure other people's figures, so they are not, strictly speaking, comparable. 2) A prop designed for maximum efficiency at 0 mph will give better static thrust figures than an equally efficient prop at a high speed. However, since most amateur built aircraft have a fairly small range between minimum and maximum speeds, static thrusts which have the supreme merit of being capable of measurement on the ground are of some value. ENGINE
H.P. ft; R.P.M.
McCulloch McCulloch VW 1600cc
75 (11 4200 90 (11 4000 66 ft; 4200
Barker VW 1834cc85 ft; 4000 Barker VW 2180cc 103 ft; 3900 "WM" VW 1600cc65 ft; 4000
"WM" VW 1700cc70 ft; 4000
STATIC WT. THRUST (LBS.) (IBS.)
SOURCES OF FIGURES
75 75 154 140 155
280 315 290 300 340
154 154 154
290 300 310
Bensen Manual Bensen Manual Bensen Manual Northwest Flyer Northwest Flyer and letters Catalogue of WM Aircraft Engines Ltd.
"WM" VW 1800cc75 (« 3900 WHE VW 1600cc 70 (Nom) ft; 4200 168
415
Own
measurements on best engine with 57" prop, (prop rpm 2800)
Now, assuming that you are convinced of the superiority of absolute thrust figures achieved with reduction drives, why doesn't every aircraft engine have them and what are the snags? Discounting the fact that all turboprops and helicopters do, my limited research also seems to indicate that MOST other high performance internal combustion engines, other than the smallest sizes did/do in fact use reduction drives. In spite of the fact that the vast majority of car engines almost from the time of their inception some 7080 years ago were fitted with not only 3 or 4 reduction drives, but also one "backwards reduction", the layman might wonder what the big problem is in fitting a reduction drive to an aero engine. In the design of aircraft, one wants the greatest possible lightness — which leads to increased flexibility and proneness to vibration. The power output of reciprocating engines especially if they contain only a small number of cylinders and the power absorbed by propellers is anything but smooth and this can set off vibrations which if they coincide with natural frequencies of propellers, crankshafts, gear-wheels, etc., can wreak severe mechanical damage in a very short space of time. Therefore, it is very often necessary to fit torsional vibration dampers which leads to increased complexity.
In particular, I was interested in the merits of Fenner "Concorde" Belts (a number of Vee belts with their backs joined together) versus ordinary Vee belts. However, Concorde belts were ruled out because their greater weight per unit length meant that the centrifugal force tending to throw them out of the pulley grooves would be greater, thus allowing less power transmission. Fenners manufacture 2 main groups of what you and I would call Vee belts, but in spite of the sections being similar, Fenners call them SpacesaVer Wedge Belts (4 sizes) and V-belts (5 sizes). In addition, the V-belt can be gotten in ordinary or premium quality, the latter having better power transmission/life qualities. In addition, the V-belts can have steel, fiber-glass or terylene tension members. Now to use Fenner's method of belt drive calculation requires a fair number of steps based on: 1) The smoothness of the power output (1 cylinder diesel engine being the worst; big electric motors being the best). 2) The smoothness of the power required by the driven equipment (quarry machinery, bad; centrifugal pumps and centrifugal compressors, good). 3) Driver pulley diameter (the bigger the better). 4) Driver pulley speed (power can be transmitted proportional to speed up to a certain point). 5) Speed ratio (reduction drives increase power transmitting capacity of the belts). 6) Length of belt (there is an optimum length of belt based on pulley diameters). 7) Arc of contact (the less arc of contact the less the power transmitted). 8) Amount and periodicity of use. So, I hope that one can now see that achieving a good V-belt drive design is not just a matter of going to the local hardware merchant and picking up the 3 or 4 belts that happen to meet one's fancy. Furthermore, I think few people realize the effect of under-belting a multiple V-belt drive and I quote Fenner's table below. It is based on 10 belts being the normal number required to drive the load. It will be seen that one V-belt less decreases the life of the whole set by as much as 30%. No. of V-Belts
% Left
10 .......................................... 9............................................. 8............................................. 7 .. ...................................... 6.............................................
The beauty of a gang of Vee belts is that, when properly
designed, the torsional vibrations are absorbed. Conciliation. Bigger diameter propellers rotated via a reduction leads to a significant increase in thrust for little increase in weight of engine/prop combination and this usually leads to a greater than proportional increase in the aircraft performance. REDUCTION UNIT DEVELOPMENT
As I had known that some people who tried Vee belt reduction drives in the past had considerable trouble, usually through excessive vibration and rapid wear, I decided to get the technical help of perhaps the best known British Belt Manufacturers, namely, J. H. Fenner, Ltd. 38 MAY 1973
(Photo Courtesy of the Author)
Another view of the aircraft version of the Ekin VW. The reduction unit weighs 29.4 Ibs.
100% 70% 45% 28%
was wanting a life of 2000 hrs. but didn't have enough
THRUST RETIRED fOO. LEVEUHJlSHT
excesc, -TWR.UST AVAI
3.250
JO
5o
/CO
FORWARD SPEED
F16URE.
3.
To cut a long story short, the best drive for my requirements turned out to be 10 SpacesaVer Wedge Belts, which can transmit up to 3 times as much power as ordinary Vee belts. In other words, if I had used ordinary V-belts, I might have had to go up to 30 for a reliable drive! The next problem were journal loads, i.e. the loads
imposed by the belts on the pulleys. With a gang of 10 belts I expected the loads to be high, but I couldn't decide whether they would be higher when static or when
running (Answer: when static). Also, I had to extract
from VW Ltd. figures for the max-journal load that could be put on the bigger end of the VW crankshaft, and I give this below with 2 other figures of interest. a) Max-journal load at flywheel end of crankshaft equals 176 Ibs.. b) Max torque at small end equals 35 ft. Ibs. c) Max thrust (for 10 minutes in correct direction equals 330 Ibs. Figure a) was much smaller than I required. The
fact that most VW aircraft engine modifications utilizing a direct driven prop on the smaller end of the crankshaft regularly and successfully exceed Figure b and c, indicated VW's figures to be very conservative. Anyway, conservative or not, I decided to design a dural casting with spigot and bearing to relieve the crankshaft of the journal load. I had wanted to get the first casting made in Belfast before taking the plunge and ordering up a worthwhile quantity from an A.R.B. Ap-
proved Foundry, but as the foundries that I tried mainly used scrap motorcycle gear-cases and pistons, etc., I declined! In the event, after one had paid for the pattern and core boxes, the cost of the casting itself is reasonable. For an extra £8, I got the castings X-rayed
to be additionally sure, but now after having done my proof-load tests, I would not now consider X-raying essential. BEARINGS. Choice of bearings and bore and shaft fits was an interesting exercise, but I wouldn't recommend the bearings handbook as light bed time reading. The main problem was the bearing for the lower or engine driven pulley. The VW crankshaft normally has 2-4 thou. of an inch approx. axial clearance and I wanted my bearing to allow this movement. Since the roller bear-
ing is the only type allowing small axial motion, in addition, of course, to the main rotary motion, it seemed to be the one. As it was essential that any bearing by coincident with and perpendicular to the crankshaft center-line, a self aligning bearing seemed desirable, but, unfortunately, it wouldn't accommodate any axial movement. In the end, I got over the problem by devising a method of machining the casting so that the spigot could
not help but be coincident with and parallel to the crankshaft center-line and then choose the roller bearing. I
room for a big enough bearing. How about using 2 identical bearings beside each other? After all, the shaft and bore diameters will be machined to fine limits and will be the same for both bearings and the bearings themselves are in any case manufactured to fine tolerances. However, the bearing manufacturers told me that in spite of the above, invariably there is unequal load sharing thus eliminating the advantage of the double bearing. Fortunately, there was plenty of room in the upper or bigger pulley and thus the life of its bearings are 3000 hrs. while that of the engine pulley is 500 hours. I got SKEFCO's Technical Department to check the choice of bearings and fits and they also confirmed that grease lubrication (as opposed to oil) was satisfactory. STRESSING. Since my days of stress analysis at University are now in the dim distant past, I thought the easiest way to get the stressing checked was to approach the head of the Aero Stress Department of our local University who was most helpful and produced 12 foolscap pages of calculations. Since then I have completed proof load tests of up to over 4 times the design belt tension. Also I reckon that my belt tensioning jacking screw will strip its threads before imposing 4 times the design tension on the engine. My latest series of proofload tests removed the haunting fear that I had of some ham-fisted individual jacking the tension up, especially if he didn't have a reliable means of measuring the tension until he cracked the fabricated mount or engine crankcase with potentially disastrous results. The Fenner method of checking my best tension is to tension them until V/2 kg. will deflect the belts by one eighth of an inch. I designed a tensionmeter to check meter against a gang of belts loaded to specified figures in a special rig to check the sensitivity of the meter because I was concerned about the possibility of excessive loads being imposed if the measurer didn't use the tensionmeter carefully. However, I reckon that there is little to fear in this area now. RUNNING TEST. My first series of tests lasted in short periods a total of 3 hours and was merely designed to demonstrate the principle that a VW could drive a propeller by means of V-belts without excessive vibration, belt-whip, belt turn-over and even belt escape from their grooves, as had been experienced by others. To effect this, I put up a three eighth of an inch thick steel plate over a window with a peep-hole in it, and positioned myself on the one side and the engine on the other. Fortunately, nothing flew off and the belts appeared to be happy. My next test was a 50 hour endurance run at about 857c-l007( max. power. In order to do this at all on the ground, it was necessary to put a 2nd engine in tandem with the first to blow cooling air over the test engine. In order to prevent the test engine "cooking", I used
heat sensitive paints, a cylinder head temperature gauge, oil temp and pressure gauges and tachometers. The throttle levers and stop switches were operated by means of cables inside my workshop, and while I used 450 gallons of fuel (!) I thought, my goodness, there must be easier and much less noisy ways of "burning" so much money!! RESULTS. Apart from one official complaint about
the noise, a temporary dulling of the hearing in spite of
my "ear-defenders" and a lightening of the old pocketbook, the only troubles that arose were: a) A slight seepage of grease from upper pulley which
has been completely cured by inserting an additional rubber "O" ring.
b) Continual heat transmission to wooden prop started to cause shrinkage of the wood and, thus, cracking. However, I have completely cured this problem by inserting a heat-shield between the prop and the upper pulley.
(Continued on Next Page) SPORT AVIATION 39
REDUCTION DRIVES . . .
(Continued from Preceding Page)
The belts at about £ 5-£7 a set will easily last 50 hours at pretty near maximum power. What their ultimate life is, I don't know, but at the worst a belt change every 50-200 hours is not too expensive and is not difficult to accomplish. There were no detectable signs of pulley groove wear, but for an additional cost of £ 3 per pulley they can be hard-anodised by an A.R.B. Approved Firm.
WEIGHT. The total weight of my drive is about 29.4 Ibs. thus bringing the weight of the engine up to 168 Ibs. (including 4 carbs, Scintilla magneto, exhaust stacks and prop). There is a little room for paring some weight off the casting which currently weighs 7Vfe Ibs. COST. The cost of the bearings, oil seals, raw materials and casting (excluding X-ray and amortization of pattern costs) is about £60 ($148.50) and the number of hours labor required after the required jigs and fixtures have been made is about 80 hours, thus bringing the total cost to about £ 220 ($544.50) if a reasonable outside machine shop is used. (Editor's Note: These figures were quoted before the most recent devaluation of the dollar.)
SUMMARY. If a VW won't give enough thrust with a direct driven prop, then it is really a waste of effort to increase its power output without at the same time using a bigger diameter prop, which will then necessitate a reduction drive between the engine and prop. The increase in thrust greatly offsets the increase in weight of the engine. The initial cost of the reduction drive, if properly designed and manufactured, will be about the same as the bare engine itself (carbs and mags extra), but it means that one can end up with a reliable and efficient power package capable of powering 2 seat fixed wing aircraft engine of the requisite power. The extra running costs of the reduction drive will be offset many fold by the reduction in fuel consumption and wear and tear on the engine.
(Photo Courtesy of the Author)
An aft view of the Ekin modification. The author did not state what the device at the rear/top is. Notice the use of a carb for each cylinder.
40 MAY 1973
NOTES:
1) W. H. Ekin (England) Co. Ltd. do NOT supply drawings, but can supply complete engines with the reduction drive and propeller or can fit their reduction drive to customer's own engines. 2) The reduction drive may be used for either pusher or tractor installations. 3) According to Propellerwerk Hoffmann GMBH and Co. of Germany, our figures, to use their words, "quite excellent" and changing the ratio from 1.5:1 to 2:1 using props of 57" and 60" in diameter would have no advantage. However, the ratio could be changed to 2:1 if necessary, say for a prop of about 75". 4) The height of the prop above the crankshaft center line is 7Vfe". Therefore, if the reduction drive is added to a VW already installed in a conventional fixed wing aircraft the ground clearance will be same if the diameter of the new prop is not 15" greater than the old one. Of course, raising the thrust line will increase the nosedown moment on the aircraft. 5) Because the crankcase flange is used for the reduction drive, it can't therefore be used for mounting purposes. However, we can supply the engine mounted on a sub-frame which eases installation problems. 6) Because there is not a constant angular relationship between the propeller and the crankshaft, hand swinging of the propeller is more difficult. However, a hand starting lever and an electric starting system have been designed and tested to overcome the above problem.
