RKOLI1983

Hi guys, I found several interesting topics on how people want to increase power and torque on their motorcycle engines by changing intake systems, improving combustion by changing plugs, portwork, etc. But there seems a fundamental flaw in everyone's approach. In order to gain an increase in performance you first have to be able to measure performance first!
Before someone jumps to conclusions, I am not asking you buy million dollar dyno setups. But at the same time, your mind cannot be a reliable device for measurement. What I'm proposing is a simple in expensive backyard dyno. Basically a roller with a fixed rotational inertia. All you have to do is measure the time it takes to spin the roller from Idle rpm to peak power rpm. You can do steady RPM measurements of throttle % vs RPM. You can do interval sweeps of X rpm to Y rpm in order to judge difference in power in that particular RPM range. I know it would have many flaws as it wouldnt give you a perfect corelation of RPM to power, but it will give you a baseline benchmark against which you can run after doing the mods and see if you actually gained anything. As unbelievable as it sounds all you need from an electronic perspective is a timer! Fellow Mechanical ENgineers and Automotive enthusiasts may have something to contribute here.

ken cool

Interesting proposition. Someone like OF can probably even develop it. Other members probably have already developped it!

Approved

__________________
The Wheel was a great invention; Two Wheels with a Motor in between was even better!


BMW Motorrad Days 2011

Xbhp's Indo-French Kashmir-Ladakh Tour

TnT

A very interesting idea.. but there's quite a lot of human modulation and feel involved, don't you think? Maybe we can evolve this into something pakka..

Old Fox

The world around is not that simple though we can perceive it to be so. Performance benchmarks are always made before performance enhancing changes are made to the engine. Provided the work is being done in an organised and focused (read professional) manner. Gauging improvements through comparison is the easiest way out and thats whats usually preferred. Right from those impromptu traffic-light drag races to expensive dyno comparisons.

As for your fledgling concept of having a basic 'dyno' sort of set-up and compare roll-up times as a measure of performance, well, there's a certain naivety in the concept though undoubtedly thats probably how the concept of the 'rolling road' dyno originated. Naive as we have hindsight to be wise with. And there's no better way to ascertain the workability of a system than by a detailed analysis of the assumptions it is based on.

Basic assumption: Reduce performance to time. Compare timings to compare performance. Your proposed 'measured spin up time of roller being directly proportional to engine output' assumes:

1. Uniform contact between the 'power transferring wheel' and the 'roller surface'.
There will be some slip between the tyre and the roller but if we assume the slip to be identical both before and after performance enhancing work, it can be removed from the equation. The fact is, it will not be the same. There are innumerable variables that define the magnitude of 'tyre slip'. Ambient temperature, tyre pressure, weight on tyre, tyre condition to name a few not to mention the 'human element variables' like how quickly or late the test rider releases the clutch and so on.

2. A quicker spin-up doesn't necessarily mean increased performance as its vice-versa might also be true. Loads of torque at low rpm's allied with tall low-gear ratios will not translate into quick rpm build-up but could well enough mean improved on-road performance.

3. That errors thus induced would be tiny enough and consistent enough to be safely ignored.
The need for complexity and therefore monetary expense arises from the need for reducing the unknowns and minimizing approximations. In short, perfection costs, and usually a bomb at that. Considering the production/stock performance figures of the bikes one would be testing on such an apparatus (our homegrown Pulsars, ZMA's etc etc), small approximation errors could show up amazingly skewed results. Lets get into figures for this. Take the ZMA for example. The proposed 'simple roller apparatus' would give one an approximate measure with probably an error somewhere in the region of 5% (see point #1 for reasons here). Now, an average enthusiast would be looking for a 2-3bhp increase from his engine to be happy with whatever work he has done on it. And this is a big increase mind you. (ask people like Joel on the forum about that) 2-3 bhp is about 15-20% of the base power figure of the bike. Even an error of plus-minus 5% (^assumed above)would take the truth of the work far from what it would be. And to get closer to the truth, go to a 'million dollar dyno'.

In my opinion, what matters to the average 'Joe-racer n grease-monkey combined' is whether his work translates to better on-road performance? Can my bike beat the stock performance figures (and that arrogant Joe racer in the process too)? Compare the bike pre and post engine work through its 'on-road' times. 0-60, 0-80, 0-100 whatever you like. The RTR probably has a stopwatch thingy built into its instrumentation that gives one time measures of 0-60 kph. Something like that would give a far better and practical measure though it would be a tad unglamorous as one just cannot see and flaunt the underlying physics and technicality involved. But then even those hot-shots in the industry bow before the KISS principle, don't they?

__________________
I don't let my motorcycles interfere with my motorcycling...

Join xBhp On

abhijeet080808

A simple rpm timer would be good enough for the job, if all prevailing conditions like rider weight, road inclination/elevation, wind speed etc remains constant.

What a rider will do is to measure the time taken to go from x rpm to y rpm in a fixed gear along the same stretch of the road. The lower the time, the better is the performance in that rpm segment.

What the rider is measuring is basically acceleration. And as we know, acceleration is proportional to power. And in this case the other variable, mass will be constant. (F = M x A)

__________________
Your biking tells a lot about the person you are!

RKOLI1983

The whole point of putting this proposition forth was to remove the cost involved in measuring power accurately. From personal experience, I know how difficult it is to measure how much work's being done by the engine, having repeatability and reliability in the test cell is an expensive ordeal. Lets not get into the discussion of how much money one can spend on a test cell because the simple answer is 'not enough'.
Answer to question 1: We are talking about a fixed inertia roller whose surface can be fabricated in such a way that the friction coefficient between tyre and roller stays high enough for the given range of normal load force(mass at the rear*g) on the tyre. If you can maintain that high friction, between the roller and the tyre, problem solved. A simple rpm comparison between wheel and roller can give you the exact longitudinal slip ratio.

ANswer 2: Yes, a quicker spin up between two rpm points means the engine did more work than before and that extra work done translated into more brake torque, since the inertia of the system will remain constant at a given speed for any state of engine tune. The more the brake torque, the more the increment in rpm because that extra work has to translate into increased kinetic energy, which btw is higher engine speed in the same period of time. Dont forget we are assuming fixed constants in place of other variables like throttle, transmission ratio, normal load on tyre for both the tests.

Answer 3: No offence to you sir, but I would never throw numbers around without hard evidence or proof of any kind. I know the point you are trying to make about percentage errors induced can throw the reliability of the test out the window. However, if you maintain the mechanical and physical consistency in the system from test to test(ambient air pressure, humidity, temperature, fuel lower heating value, oil temperature and all the friction coefficients in the system) you can get very reliable results because if you eliminate the variables, there is simply no way for the engine torque to be lost anywhere except in the design changes made to the engine itself!

Old Fox

(bold mine)

I guess there is a basic difference in our perspective. You believe that a ‘single roller dyno’ (would need at least two rollers in tandem btw), if theoretically well-designed, can produce a predictable comparative relation between its spin-up time and engine performance. I have approached it from an engineering/fabrication perspective and see prospective problems that can be conveniently assumed away while doing a purely theoretical analysis.

The world of engineering is littered with innumerable examples of sound theoretical design that fails to take shape either due to cost over-runs, metallurgical/materiel deficiencies, lack of equipment to fabricate/engineer the needed components etc etc. The Wankel engine is a glaring example as such. The initial dead-end were the rotor seals. There was no suitable metal/materiel available then to make those rotor end-seals. Seal issue surmounted, the excess heat generation became a bottle-neck. The engine became more and more complex; the exact opposite of what its greatest attraction was vis-à-vis simplicity.


Getting back to our dyno, you can no doubt get the ‘longitudinal slip ratio’ between the roller and the tyre but have you factored in its variation through the RPM spectrum? For example, even a slight off-centre rotation of the roller will make the tyre bounce, esp at the higher RPM’s and that puts a question mark on the ‘assume constant longitudinal slip ratio’ element. To ensure good tyre to roller contact throughout the rpm range, we need a roller with a very high level of roundness, an axle that runs true to a very high accuracy, bearings that do not alter the alignment etc etc. All factors that increase costs. And if the argument put forth is that we can assume all such inaccuracies as negated since they remain the same for each run if the rider, tyre and the bike remain the same, then, again re-check the ‘assumption’ whether they actually do remain the same.


Accuracy costs and perfection in accuracy costs a bomb. Keeping costs low has a tradeoff in terms of reduced accuracy. And that’s the point I have tried to make through my post. Comparing spin-up times to compare performance, at the level of accuracy needed and desired and within cost constraints that are assumed, is plain-jane stuff on paper but seemingly well nigh impossible in practice. Remove the ‘within cost constraints’ qualifier and you get your system no doubt. Developing a system through pure theory and then transposing the results thus arrived at onto a physically engineered machine, while wishing away the errors/variations that shall be inevitable with the physical engineering and use of the system, is contradictory to this whole discussion as an intellectual exercise.


I am not loathe to getting into the physics of the thing but as an engineer, I am compulsively trained to assess the idea on the basis of its ‘assumption framework’, to judge whether the assumptions will hold tight when it finally comes to sheet bending and welding, to putting in the money on the idea. And if the ‘assumption framework’ is weak or too contrived just to make the theory look water-tight, my enthusiasm for the idea looses steam.


I would be more than happy to see such simplicity put into practice and probably will make use of it myself. I shall be eagerly looking upto you and others to make up for my possible assessment and understanding deficiencies present in my words above as those from another frail human being.


Italics mine

Some confusion reigns about defined concepts of physics. For example, work done is the result of torque and not vice-versa as stated. (see italicized bold phrase above) Torque is a force and a force applied over a distance is work done. And work done over time is power but then we don't need to go that far here.
Also, there's apparently a lot of flitting between 'systems' here. "Increasing 'brake torque' causing an increase in RPM as the work done by the torque adds to the kinetic energy??"
Adding energy to what? The piston, the engine, as output at the rear wheel? Would be great if you could please define what your 'system boundaries' are in this comment and relate them to the conclusion. Looking forward to inputs here from you.

Ref to comment above: Offense not taken but the numbers were not stated in a ‘cast in stone’ manner but rather as a descriptive analogy with the purpose of explaining a certain aspect of the discussion, with the hope that it would be taken in that context. Apparently it wasn’t.

__________________
I don't let my motorcycles interfere with my motorcycling...

Join xBhp On

RKOLI1983

About question 1: I understand your hesitation to accept that this can be implemented with ease. Implementation has its own problems and I dont think there is any machine in this world whose initial proposal was not mocked or laughed upon by anyone. Anyway to answer some of the implementation issues you have pointed out:
1> "For example, even a slight off-centre rotation of the roller will make the tyre bounce, esp at the higher RPM’s and that puts a question mark on the ‘assume constant longitudinal slip ratio’ element. To ensure good tyre to roller contact throughout the rpm range, we need a roller with a very high level of roundness, an axle that runs true to a very high accuracy, bearings that do not alter the alignment etc etc." - Tyre bounce problem can lead to a loss of friction between roller and tyre only IF the intermittent 'bounce' reduced the friction coefficient to such a magnitude which wasnt enough to sustain the longitudinal propulsion force of the tyre on the roller. With sound calculations, this can be eliminated by having a high 'preload' and normal force on the rear tyre which doesnt let the tyre 'bounce' out of the limit(friction). Roller with high level of roundness will be required, I agree with that. Also the axle will have to run atleast as true as the vehicle's axle. Remember, there will be a suspension on the rear tyre which is designed to keep the tyre in contact with the road/roller at all times regardless of small undulations. As far as the choice of bearings goes, its not a problem which can be solved by doing some math based on normal load forces exerted radially on the bearings, service life, etc(read other factors I might not know). As far as I know there isnt a single automobile/motorcycle which I would ride if its bearings changed alignment on its own.

2> "All factors that increase costs. And if the argument put forth is that we can assume all such inaccuracies as negated since they remain the same for each run if the rider, tyre and the bike remain the same, then, again re-check the ‘assumption’ whether they actually do remain the same." - Cost is an issue. I never argued about this system not requiring some expenditure. The effort is to minimize it and obtain a reliable measurement system based on sound physics. Again, what You have mentioned about the 'variables' can also be assumed within a range. For example. Tyres can come in all shapes and compounds. Obviously you dont wanna use a stone hard 4 year old tyre for the dyno. There will aways be a longitudinal slip ratio indicator which will tell you about how badly the tyre's losing friction. With appropriate normal force load on the tyre its another problem which can be taken care of in a jiffy before the actual test run. (I guess that is why they have huge guys sitting on the hayabusas while doing dyno runs).

About question 2: Engine modifications are done in order to increase the thermodynamic work inside the cylinder integral(P*V) every 4 stroke cycle. This work is also called the IMEP(Indicated Mean Effective Pressure) of the engine. However, what you see at the crankshaft is BMEP (Brake Mean Effective Pressure) of which Brake torque at the crank is a linear function. BMEP is always lesser than IMEP because of FMEP (Friction Mean Effective Pressure). BMEP=IMEP-FMEP. So the whole point of increasing engine torque from a thermodynamic perspective is to increase IMEP, which is increased only by increasing the work done inside the cylinder. By work I mean the work done by expansion of gases caused by combustion exerting pressure on the combustion chamber walls and the piston crown. Also there is work done by the piston crown on the combustion chamber gases during intake stroke. this work is called the pumping work and it has to be minimized. So I was talking about the improvement of the net work or IMEP in my previous post. Since our dyno is a simple constant load type, any extra work performed by the engine, which will act on the load as Brake torque will translate as increased speed because there has to be an equilibrium between potential energy (Fuel and air inducted) and kinetic energy of the roller, crankshaft, transmission, wheel, piston, etc.(Ignoring heat losses)