Creative Commons License

Model Engine Development

Part 9

by Gordon Cornell

    Performance Evaluation
    Power Curves
    Fuels and Performance
    Engine Test propellers
    Power Factors
    Noise and Pollution

Click on images to view larger picture.
Hover for a description.

Performance Evaluation

Constructors should by now have assembled their engines and need to determine how successful they have been in producing a working engine. Hopefully along the way some knowledge will have been gained as to how the engine should function. There is bound to be some variation in performance achieved, the most likely reason for this will be variations in fit and finish. I deliberately avoided specifying tolerances on the drawings because these need to be established as part of the development program. Whilst the experienced engineer will have some idea of the requirements, practical examination of a number of examples by performance testing soon reveals unsatisfactory items. Defects will be exposed for which solutions have to be established. Until you start testing you will not be aware of how the engine performs at a range of operating temperatures. An example of this is does it start easily in both hot and cold situations? In most circumstances this is determined by piston cylinder fit however there may be other contributory factors.

During this testing it can be an advantage to observe the change in fuel consumption as the engine reaches peak performance. The engine should hold maximum RPM when run for an extended period on a relatively lean setting. If the engine starts up at high RPM and fades away there is either to much Isopropyl Nitrate in the fuel, or the compression setting is too high. Most diesel engines are at there best when running with a slight misfire.

The next stage is to continue the test work by installing the engine in an appropriate model. You need to establish contact with your particular interest group; model aircraft, car or marine. All I am sure will be willing to help. There will be somebody able to help you test your engine in the user environment.

Power Curves

The principal purpose of performance testing is to determine the shape of the BHP and Torque curves. A simple way to achieve this is to record the RPM on a range of what are known as calibrated propellers. From this data power curves are produced. Unfortunately the BHP absorbed by a propeller generates a cubic curve—Torque is presented by a quadratic. The curves presented here are generated from predetermined constants hence they are theoretically correct. Because the curves are difficult to read, the constants used in their generation are specified in the table. Use these constants to determine the performance when RPM differences are small. A pocket calculator is applied to achieve this—an approximate curve can be established directly from the graphs.

Precise power outputs will be difficult to confirm between methods and testers. If the shape is not similar then there is something wrong with the calibration of the test rig or propellers (refer to Part 1 for problems arising during Ron Warring's tests).

There can be significant differences between Performance Prediction and measured results. Normally a computer program such as ICE does not calculate differences in power output due to a thrust component from the propeller. Similarly the power output determined on a Dynamometer should not be subjected to a fore and aft crankshaft thrust. Hence prediction and dynamometer tests should be similar. Just how well an engine will handle the thrust component of a propeller will depend on the crankshaft bearings, and what provision has been made for this. It therefore follows that testing with airscrews has to be part of the test procedure.

It is important to establish how the test propellers were calibrated—there are bound to be errors in calibration. In many cases there is more than a single mould, or the design has been updated. Mark your test propellers as masters, otherwise you may not be able to replicate you test results later All propellers have RPM range limits over which results are consistent; do not operate these outside the recommended limits.

Fuels and Performance

Currently there are no rules regarding fuel mixtures for compression ignition engines. It is important to establish what fuel mixture is applied for the tests. A single pre-determined fuel mixture may not be suitable for the engine under review (it is unlikely to establish the performance limits of all engines).Ron Warring failed to do this when testing the ED Super Fury. In Part 7 an overview of how the fuel mixture may affect the performance was indicated. What this implies is that this can vary substantially due to differences in Mechanical Efficiency etc.

When specifying propeller RPM figures the fuel mixture should be indicated together with the rate of consumption in cc per sec/or minute. This rarely occurs.

Engine Test propellers

The data in Table 6 may be used by constructors to compair their work against my prototypes. It shows the RPM achieved for a range of readily available APC propellers. An engine which is tight may be down as much as 2,000 RPM in the initial running stage. If the engine is free and the RPM is down, check for out of round cylinder and other defects.

Engine7x6 APC7x5 APC7x4 APC7x3 APC
ED Super Fury and Replica12,500/13,00013,900/14,50015,500/16,20018,000 Plus
Drum Replica12,00013,000/14,00015,000/15,60017,000 Plus

Table 6: Performance Comparison.

Note that these figures are for standard engines, not tuned examples, so constructors should be able to achieve these figures. My best motor from 1969/1962 turned the Frog 7x6 nylon propeller at 14,500 RPM. The fuel mixture used for these tests is given in Table 7 below. The component percentages are by volume and total 102%, the Isopropyl Nitrate being added as a proportion of total volume of the other constituents.

Castor Oil15%
Isopropyl Nitrate+2%

Table 7: Test Fuel Composition.

This is a general purpose fuel used for initial testing. No fuel consumption data was taken because the engines were in new condition. 15 seconds per cc is likely at 14,500 RPM.

Power Factors

There appears to be considerable confusion as to what a calibrated propeller is and its function. What purpose does it serve?

For engine test purposes it is a device used to apply a resisting load to crankshaft output. It's rate of change of load with increasing RPM having been established previously. It is not used directly to determine propeller efficiency or output. Neither does it indicate that propeller A is better than B. In our case the most important feature is that its geometry should not change significantly when subjected to a wide RPM range. To ensure repeatability its form should not change with ageing or temperature. The torque absorbed by a propeller can be predicted by:

TORQUE=K x RPM 2 (where K is a constant for the propeller)
BHP=P x RPM 3 (where P is a constant for the propeller)

Obviously 'K' only remains constant whilst propeller deflection under centrifugal and aerodynamic forces is minimal. The efficiency effects of the propeller can be ignored since we are only concerned with the BHP and TORQUE, required to achieve the specified RPM. Both engine and propellers are tested under similar conditions outside a wind tunnel. Once 'K' deviates from a constant the propeller is being operated beyond useful and safe limits. In other words performance is not predictable. The upper speed limits are indicated by some makes.

There are distinct advantages to be gained by the application of theoretically correct curves at a range of 'K' values. To introduce a new propeller all you have to do is check this against others and determine the nearest curve and K value, adding a new curve if the difference is too great. It is not necessary to check this over a large RPM range. The system ensures that all examples are correct relative to each other. Correct shape engine power curves can only be produced when this is so. It also eliminates the so called rogue propeller. Initial determination of the base curves can be made from prior engine reports and propeller RPM data. Any errors are rapidly exposed when the data will not fit these curves. The greater the number of propellers tested the more precise the curves become. If the predictions over a period of time are found to be inaccurate giving either high or low BHP figures, a fresh set of curves with revised K values can be plotted or the propellers allocated alternative existing curves. You do not have to re-test the propellers, you simply move the datum. This is a well known procedure used in industry when trying to predict performance.

Propeller Absorption Constants
COX4.5 x 26.600001E-96.545220E-15
GRA5 x 21.006347E-89.979939E-15
APC5.7 x 32.008828E-81.992155E-14
COX6 x 32.800000E-82.776760E-14
APC7 x 34.411617E-84.375000E-14
COX7 x 3.55.746866E-85.699166E-14
APC7 x 46.488218E-86.434365E-14
APC7 x 61.029467E-71.020922E-13
COX8 x 41.030000E-71.021451E-13
APC8 x 41.091020E-71.081964E-13
GRA8 x 51.433387E-71.421490E-13
APC8 x 61.888325E-71.872652E-13
APC9 x 41.771248E-71.756546E-13
GRA9 x 52.400000E-72.380080E-13
APC9 x 62.462209E-72.441773E-13
GRA10 x 32.074834E-72.057613E-13
APC10 x 42.857579E-72.833861E-13
APC10 x 63.734702E-73.703704E-13
APC11 x 65.737196E-75.689577E-13
APC12 x 68.287743E-78.218954E-13

Table 8: Propeller Absorbtion Constants.

Noise and Pollution

Many model engine users have neglected these issues, the result being a loss of flying sites. Two strokes tend to produce more noise than the equivalent Four strokes due to the increased number of firing cycles. Early attempts at fitting silencers caused disputes because no specific noise level was defined. Bill Wisniewski developed tuned pipes for his record breaking speed engines in the 1960's. These increased output, in the Free Flight applications these were banned and the silencer rule abandoned because it did not appear feasible to distinguish the difference between a tuned pipe and a silencer. The reality is that any properly designed exhaust extension should increase power output. The problem with this is the size and weight of the object. Whether silencers/pipes are fitted now depends on the class of model being flown—check the rules.

Inexpensive noise meters are now available so there does not appear to be any logical reason for not specifying maximum noise limits for engines. Ever increasing outputs will eventually lead to further restrictions. Development time must therefore be allocated to dealing with this issue. A silencer will have to be part of the engine specification if the engine is to be used in most applications.

A standard set of test procedures needs to be established to ensure universal acceptance. The problem is that accurate measurement of noise levels is difficult to achieve. However this should not stop us investigating solutions to reduce noise levels, there are commercial implications to be resolved. The problem is to get your competitors to do the same thing.

The first thing to do is to determine the noise level at increasing RPM without a silencer fitted. Then establish what is required to reduce this by 5 DB increments. It will most likely be very difficult to achieve 90 DB. There is no point at the outset in setting a level which cannot be achieved. Reduction in noise levels will have to take place over an extended period.

12,1001 m106
13,5001 m107
14,5001 m108
16,7001 m110

Table 9: Noise Recordings—ED Super Fury (Open Exhaust)

Note—Old fuel was used for these tests, what this indicates is how noise increases with escalating RPM—better fuel more noise. The readings were deliberately taken close up because this is where maximum levels occur. We need to know what this is. To reduce it will require a silencer, setting a maximum RPM will not be a satisfactory solution.


    Model Engine News Home

This page designed to look best when using anything but IE!
Please submit all questions and comments to