Hi Tim,

That's an interesting point, you've got me worried.

I hope the answer lies in the engine's flywheel. I imagine the flywheel to be smoothing out the power strokes, maintaining a steady-ish rpm at the output by storing and releasing energy as it spins. There's a bit more smoothing in the torsion spring and the brake disk as well. If I'm right the power pulses average out sufficiently to give a reasonable result.

What I don't know is just how effective the engine's flywheel is at damping out the piston's stops and starts. I hope the inertia of the system is sufficient but haven't proved it!

I have a busy few days ahead but as soon as I get the chance I'll put lots of black/white stripes around the flywheel and have a go with an optical detector at measuring how much the acceleration of the flywheel alters within each cycle. If it varies a lot I shall have to bribe Tim to keep my embarrassment a secret!

Dave

Edited By SillyOldDuffer on 15/05/2017 21:16:41

Done some swotting on flywheels this morning using 'Applied Mechanics for Engineers, Duncan, Macmilllan, London 1949' , Chapter XX. The book draws a distinction between disturbances due to variations of energy supply rate during a cycle, and disturbances that take place over several cycles. Variations inside the cycle are typically caused by the piston accelerating and decelerating, and these are smoothed with a flywheel. Variation over several cycles is due to the engine speeding up or slowing down and are smoothed with a governor. You really need both.

In considering the smoothing effect of a flywheel, the argument starts with the premise:

Energy Supplied = Energy Abstracted + Energy Wasted in the Machine

Then, supposing the energy provided by the engine to be increasing (as the piston accelerates), then the excess energy has to be disposed of by increasing the engine's speed. This adds kinetic energy to the parts of the machine, and the reverse is true when the piston decelerates. The outcome is a jerky action.

Putting a suitably sized flywheel on the output shaft allows the energy to be stored in an object having a large moment of inertia. Because the moment of inertia is large, changes to the flywheel's speed are 'comparatively small' compared with those of a piston.

The book continues with a mathematical analysis before describing how to estimate the size of flywheel needed by a given engine. Just for laughs it concludes with an analysis of the centrifugal forces noting that if excessive, the flywheel will burst. Not likely on model a Stirling Engine I hope!

So, in theory the flywheel should reduce impulses considerably. In practice I don't know if the flywheel on my version Jan Ridder's engine is fully effective or not. Nor do I know quantitatively what my books means by "changes to the flywheel's speed are 'comparatively small' …". And of course the engine is ungoverned.

Nonetheless, I think there's enough for me to claim that my measurements were reasonable, but not to claim that they are accurate. Duncan Webster's suggestion of adding more detectors is a good one and should be easy enough to do provided I can find all the bits (the work was done almost a year ago.) It also occurs to me that an Arduino project to measure fluctuations in the angular velocity of a flywheel might be a more useful way to spend the time.

If anyone has the time and interest to pursue it, I'm sure there are many ways in which the prototype dynamometer could be improved.

One thing I haven't done yet is to use what I've learned about the Coffee Cup Stirling to see if I can improve it. One example: in the first article on taking Indicator Diagrams I noticed that the pressure inside the engine varies above and below atmospheric. Is that a good thing or not? I believe it is, but it would be interesting to test the effect on power output of adding a relief valve to the cylinder so that the piston never sees a vacuum. I'm sure it's all been done before but society benefits. Tinkering in a garage keeps me off the streets…

Dave

Flying adjectives! I got the Coffee Cup engine out of it's box for fun and although there are signs of life it won't run. I'm hoping it just needs a good clean.

To say these little engines are pernickety would be an understatement. Did someone say that Flame Lickers are much more reliable…

Dave

Pressurising the engine should increase power as long as the heat exchangers can put in and take out the extra heat. The Phillips engines built in the 1960s used air at about 200 psi according to the interweb. There used to be a lot of stuff about hot air engines in ME, I seem to remember the name Rizzo, and for some reason I think he lived in Malta

We made a hall effect backplate for a 2.5 cc engine for F2A, runs from 39k to 40 k on our test prop and rpm readings from the model. We were really surprised at the RPM variation from shot to shot and the in cycle variation. From 1 rev to the next, can be as much as 400 rpm difference , or effectively a plus and a minus of 200 rpm per rev, Then it can run for a short duration of being within 2 or 3 RPM. What was more interesting was that the model plane flies as fast as it does due to the pulse of power as the engine picks up in rpm at the power pulse. So we could see the model flying in mini jumps. So it pulses every 120 to 140 mm on the circumference. So when compared to a backplate that was made of brass , more fly wheel effect and an Ali one, it was actually faster on the Ali one, but easier to set on the brass one. So in short , it makes more usable power with less flywheel, so it must be actually making more total power. Trying to dyno the power output is very difficult on model engines. I think a brake type dyno is the way to go. For model plane motors, I think that a calibrated propeller is the way to go. So a prop run up on an electric motor and rpm, torque are measured to know it's power absorption. The tachometers sold for model engines take averages of the pulses of light to give a readable average value.

Neil

This thread has quite a bit about a dyno used for model engines in those sort of speed ranges, dyno starts on the second page. You will need to join the forum though if you want to see the attached pictures.

Apart from comparative willy waving down the pub / shed / club, the accuracy isn't exactly critical as long as it's consistent. After all, what is the ultimate purpose of the measurement?

One old chestnut that always comes up is the difference between a metric horsepower ("PS&quot and a true HP. That difference is about 1.5% – but just step back and think for a moment.

We are talking about engines here, where you'll not find 2 "identical" engines the same and even the same engine measured on 2 different occasions will give different results. And these are going into vehicles, where the difference couldn't possibly be noticed. Almost identical difference between a "ton" and a metric "tonne". Makes no noticeable difference, unless recounted by a lunchtime legend.

When you look at the tolerances in a dyno system (even a top end professional one) – plus the variations in the engine side, you'll see that many of the final numbers are fairly nominal.

Murray

I used to think that engines were simple beasts, as in squirt in some petrol, make it go 'bang' and then watch the wheels go round. Easy but wrong! Now I'm finding that the closer you look the more complicated it gets.

When it comes to accuracy, it's difficult to compare two engines of the same type; its difficult to compare the same engine in two different runs; and now I've learned that it's difficult to compare two cycles of the same piston in the same cylinder!

No wonder it costs a few hundred million quid to develop a new car engine.

Dave