Setting a Standard?

[Patrick]

If you had two bikes you could have a situation were under one set of test conditions (say a 2% slope at 10mph) Bike A has a range of 30 miles while Bike B has a range of 40 miles, while under another set of conditions (say a 3% slope at 8mph) Bike A has a range of 40 miles and bike B has a range of 30 miles. In this scenario which has the greater range?

The problem is not only does the answer to the question "what's the range of this bike" depend of the rider and conditions, if you've got two bikes with similar specs the answer to the question "which of these bikes has the greater range" may depend on the rider and conditions as well.

To be able to meaningfully compare two bikes it wouldn't be enough to one standard arbitrary set of conditions, you would need at least a set of graphs showing how the bikes' performances varied as the parameters of the test conditions varied.

 

[flecc]

Lets try and simplify the given task, Bike weighted to simulate rider. Terrain, 2 degree incline. Speed set to 10mph. Battery 8Ah.
Result, this ebike can under x conditions achieve x mile's, at x speed, on an x Ah battery. Apply this test to all ebike's and you have a comparison that is informative for the consumer and a target for manufacturers to match/improve on.

Not so, you are still missing the point that rider input level is necessary to set a power level on the pedelec types I mentioned. The bike cannot merely be weighted to simulate rider, the rider has to be present and working and their input works like a throttle. Using a machine to simulate a rider would still beg the questions of what pedal input pressure should be used, over what degree of the crank arc and what cadence, all of these affecting the motor power out.

I do know these systems, no meaningful comparison is possible for all types.
.

 

[lemmy]

As an observation Flecc's previous post:

I go to the gym 3 times per week for a one hour workout. Some days I sail through with little effort. Other days it is hard work. I cannot predict which it will be from my own feelings but plainly one's strength varies from day to day for reasons which are not obvious or predictable. This is a common observation, by the way.

Accordingly, when I ride into central London on my Kalkhoff, a roughly 22 mile return journey involving one fairly steep and long climb each way, some days I get home having the 2nd handlebar charge lamp still lit, other days well into the first lamp lit. I know which it will be because on the long range days I find myself whistling up the steepest slope in 6th gear way beyond the speed at which asistance has started to attenuate. On the short range days the bike feels sluggish and I am dropping to 4th gear up the same hill and slowing to get full assistance. This is independent of the ambient temerature.

In other words, the same person on the same bike can experience something like a 25% difference in range for no appreciable (or predictable) reason.

This is for the Panasonic/ Yamaha systems, of course. As Flecc says, no standards are possible - even for the same rider and bike on different days.

On the other hand, knowing myself and the terrain I cover and reading Flecc's observations on the Panasonic, I was able to get an good estimate of whether it would do the job I wanted it for. That estimate has turned out to be correct on average with a bit in hand for my feeble days.

 

[onmebike]

Not so, you are still missing the point that rider input level is necessary to set a power level on the pedelec types I mentioned. The bike cannot merely be weighted to simulate rider, the rider has to be present and working and their input works like a throttle. Using a machine to simulate a rider would still beg the questions of what pedal input pressure should be used, over what degree of the crank arc and what cadence, all of these affecting the motor power out.

I do know these systems, no meaningful comparison is possible for all types.
.

I'm not missing the point. The pedelec of any ebike can be bypassed for test purposes. The unknown variable's such as rider weight, ability, input, temperature etc are far too complex, so an average must be used. The intention is to test each combination of motor, controller and battery under a series of identical condition's. The condition's wouldn't be identical if you throw all the unknown variable's into the mix.
It doesn't even need to be done on an ebike. A test bed designed for the purpose would do.

Car driver input with throttle, clutch and gear changes throw up far more variable's than a pedelec ebike, yet manufacturer's can quote pretty accurate specs for them. Why is it so difficult with an ebike?

 

[flecc]

I'm not missing the point. The pedelec of any ebike can be bypassed for test purposes.

You should know the system before saying this. The pedelec integration of the Panasonic system is not bypassable without destroying the system, which would make any measurements invalid.
.

 

[onmebike]

That estimate has turned out to be correct on average with a bit in hand for my feeble days.[/QUOTE]


In any test the unpredictable has to be assumed to be an average of all possibilities. The results won't be pinpoint accuracy but a rough guide.
In other words, useful information.

 

[NRG]

No, it will result in just another set of numbers that are as meaningless as those we have now.

 

[NRG]

....
It doesn't even need to be done on an ebike. A test bed designed for the purpose would do. ....

Then you may as well just forget the bike, ignore the geometry, weight, rolling resistance, free-wheel ability etc. and simply test the motor / controller / battery combination in isolation for all the good it will do.

 

[onmebike]

You should know the system before saying this. The pedelec integration of the Panasonic system is not bypassable without destroying the system, which would make any measurements invalid.
.

I didn't mean bypass as in, modify after production. I'm talking about tests that would take place under laboratory conditions during manufacture.

By not being bypassable do you mean it couldn't be activated by any other means than someone pedalling?

 

[onmebike]

Then you may as well just forget the bike, ignore the geometry, weight, rolling resistance, free-wheel ability etc. and simply test the motor / controller / battery combination in isolation for all the good it will do.

Its impossible to perform identical tests on each and every ebike if your going to throw so many variable's into the mix, so you take an average of the variable's and measure what can be measured, as you say motor, controller and battery performance. Your making a bike with a motor more complex than a 747.

 

[Old Timer]

I agree with you, take the bike/rider element's out of the equation and simply quote battery/motor ability at a given resistance to simulate an average of all the other variable's. It sets a standard measurement under normal/average operating condition's, something the consumer can compare. The intention isn't to measure rider input, ability, weight or fitness, its to measure the ability of different manufacturer's battery/motor combinations at a set task. The set task being an average of all the variable's.

I fully agree, test all bikes mechanically against the same criteria in the workshop. the one that comes out on top is the winner. Meaning, it`s then down to your fitness and location and style of riding to get the best from the bike. At least you will know that you are actually in with a chance and not trying to get a decent performance out of a machine that could never deliver. Simple surely

 

[onmebike]

I fully agree, test all bikes mechanically against the same criteria in the workshop. the one that comes out on top is the winner. Meaning, it`s then down to your fitness and location and style of riding to get the best from the bike. At least you will know that you are actually in with a chance and not trying to get a decent performance out of a machine that could never deliver. Simple surely

Thanks for the above Dave, I was beginning to think I was speaking in a foreign language.
As the title of the thread suggest's, setting a standard doesn't mean throwing a mix of unpredictable variable's in to confuse the issue, you work from a set of average's and measure the measurable.
If the only way to do it is by having a pre programmed robot ride each bike under a preset series of condition's, then thats what it take's, but as you suggest it can be done mechanically on a test bench.
The object for me was not to find a winner but set a criteria that all ebike's could be tested against. The results being informative to the consumer and a target for manufacturers.
We all know that in actual practice the test results would be difficult to match, as are figure's quoted by car manufacturers[far more complex], but it would still be useful to have a comparison of all ebike's even if done under less than realistic condition's.

 

[lemmy]

I fully agree, test all bikes mechanically against the same criteria in the workshop. the one that comes out on top is the winner.

Think of all the money it would cost to set all this up. The endless meetings and squabbles between dealers, bicycle makers, pedelec system makers, our government, the scientists, the EU and the endless b****dy committees set up to oversee it all

There is plenty of information out there for people to make an informed choice. For those too dim or lazy to research it, making a bad choice is the price to be paid for dimness or laziness.

E-bikes are already expensive and over-legislated. We should add to that already?

 

[onmebike]

Think of all the money it would cost to set all this up. The endless meetings and squabbles between dealers, bicycle makers, pedelec system makers, our government, the scientists, the EU and the endless b****dy committees set up to oversee it all

There is plenty of information out there for people to make an informed choice. For those too dim or lazy to research it, making a bad choice is the price to be paid for dimness or laziness.

E-bikes are already expensive and over-legislated. We should add to that already?

It need not be quite so complicated. It need not be compulsory for a start. Manufacturers participate only if they choose to do so and suffer the consequences if they don't. It could help put an end to cheap knock off clone's.
High end manufacturers would benefit from participating while the low end manufacturers would need to up their game before daring take part.
All it need is for someone to devise the test.

As far as government and legislation are concerned, I fully expect insurance and helmets to become compulsory, maybe mot's too. It'll only take a few accidents involving ebike's.

 

[flecc]

I didn't mean bypass as in, modify after production. I'm talking about tests that would take place under laboratory conditions during manufacture.

By not being bypassable do you mean it couldn't be activated by any other means than someone pedalling?

Yes, I did understand you meant in that way, but no, it couldn't be done and Panasonic would refuse to try it because the results would bear no relation to reality. Let me explain.

First the physical problem, the pedelec electronics and controller are integrated into one dual feedback system in a motherboard, so it would be necessary to have a controller separately built to get rid of the pedelec elements. Unfortunately that cannot then be the same since in the original, not only does the pedelec torque sensing system control the output, the motor feedback also controls how the torque sensing system issues power, a two way system in which the pedelec torque sensing is vital and integral.

Much bigger though is the operational problem. Even if a separate controller was permissible, the motor would then run continuously like other bike motors, but the Panasonic unit doesn't work like that. Instead it's a servo motor system that runs in pulses, mimicking the rider's leg muscle actions. When riding, at crank top dead centre the motor is effectively off since there is no pedal torque force. As the crank angle progresses and the leverage gain enables pedal power into the sensor, the motor gains power, reaching a maximum when the cranks are roughly horizontal when the leverage advantage is best. Then as the crank travels to bottom dead centre the power falls and effectively cuts off again, ready for the next down stroke. This is an accurate bionic imitation of the rider muscle action.

You can see that this is roughly a 50% power system, but if the power was graphically presented, it would be neither a triangle wave nor a sine wave. It would appear like a pinched sine wave with a logarithmic rising front and an anti-logarithmic tail.

This is where the more complex elements appear. The rider's pedal action would alter that profile since each person's pedal action differs. The "lead footed" would emphasize the central power peak, losing out elsewhere, while the experienced cyclist who rotated the foot about the ankle to give some power through the top and bottom dead centres would smooth out the response somewhat and gain more power overall.

Now it gets more awkward, bear with me! The torque sensor pressure from pedalling determines the motor power output during each thrust, which is why I said " a maximum" above, not "the maximum", but the output pattern with speed is totally different from hub motor bikes. The hub motor will give falling torque with rising power as speed rises from zero, reaching an optimum combination at about half maximum speed. Thereafter the power and torque will fall away but the efficiency increase up to near the maximum speed.

By contrast, the normally used Panasonic system will give the roughly equal optimum combination of torque, power and efficiency at all speeds from zero to 9.4 mph (15 kph). Thereafter the torque and power will phase down progressively through to zero at 15 mph. Now, to complicate matters even more, the rider has three switchable power levels which operate within each of the control parameters already discussed, and these are continuously and optionally used while riding, high power up a hill, standard, low or off on the easier going, the patterns of use varying infinitely with every rider. For example, on my long range test to give 50 miles per charge, I used normal power and off according to conditions. David Henshaw of A to B magazine on his long range test to get 46 miles on the same bike used the low eco setting on all the time. Every rider will use the combinations of the four possibilities, off, eco, standard and high, randomly in a pattern unique to them.

You can see that this power profile is very different from a hub motor continuously running system so the two could never be comparable. Not only that, the almost infinite complexities of the system in dynamic conditions will defeat any attempt to simulate a meaningful human presence in a test.
.

 

[onmebike]

Yes, I did understand you meant in that way, but no, it couldn't be done and Panasonic would refuse to try it because the results would bear no relation to reality. Let me explain.

First the physical problem, the pedelec electronics and controller are integrated into one dual feedback system in a motherboard, so it would be necessary to have a controller separately built to get rid of the pedelec elements. Unfortunately that cannot then be the same since in the original, not only does the pedelec torque sensing system control the output, the motor feedback also controls how the torque sensing system issues power, a two way system in which the pedelec torque sensing is vital and integral.

Much bigger though is the operational problem. Even if a separate controller was permissible, the motor would then run continuously like other bike motors, but the Panasonic unit doesn't work like that. Instead it's a servo motor system that runs in pulses, mimicking the rider's leg muscle actions. When riding, at crank top dead centre the motor is effectively off since there is no pedal torque force. As the crank angle progresses and the leverage gain enables pedal power into the sensor, the motor gains power, reaching a maximum when the cranks are roughly horizontal when the leverage advantage is best. Then as the crank travels to bottom dead centre the power falls and effectively cuts off again, ready for the next down stroke.

You can see that this is roughly a 50% power system, but if the powe was graphically presented, it would be neither a triangle wave nor a sine wave. It would appear like a pinched sine wave with a logarithmic rising front and an anti-logarithmic tail.

This is where the more complex elements appear. The rider's pedal action would alter that profile since each person's pedal action differs. The "lead footed" would emphasize the central power peak, losing out elsewhere, while the experienced cyclist who rotated the foot about the ankle to give some power through the top and bottom dead centres would smooth out the response somewhat and gain more power overall.

Now it gets more awkward, bear with me! The torque sensor pressure from pedalling determines the motor power output during each thrust, which is why I said " a maximum" above, not "the maximum", but the output pattern with speed is totally different from hub motor bikes. The hub motor will give falling torque with rising power as speed rises from zero, reaching an optimum combination at about half maximun speed. Thereafter the power and torque will fall away but the efficiency increase up to near the maximum speed.

By contrast, the normally used Panasonic system wil give the roughly equal optimum combination of torque, power and efficiency at all speeds from zero to 9.4 mph (15 kph). Thereafter the torque and power wil phase down progressively through to zero at 15 mph. NOw, to complicate matters even more, the rider has three switchable power levels which operate within each of the control parameters already discussed, and these are continuously and optionally used while riding, high power up a hill, standard, low or off on the easier going, the patterns of use varying infinitely with every rider. For example, on my long range test to give 50 miles per charge, I used normal power and off according to conditions. David Henshaw of A to B magazine on his long range test to get 46 niles on the same bike used the low eco setting on all the time. Every rider will use the conbinations of the four possibilities, off, eco, standard and high, randomly in a pattern unique to them.

You can see that this power profile is very different from a hub motor continuously running system so the two could never be comparable. Not only that, the almost infinite complexities of the system in dynamic conditions will defeat any attempt to simulate a meaningful human presence in a test.
.

No matter how I look at this, it cannot be anywhere near as complex as it appears. The variable's involved with car's are far more complex yet the manufacturers have little trouble quoting specification's.
A Robot or machine that simulate's an average rider with average input, under average conditions is all it take's. The object is to devise a standard test, not give everyone a personal data sheet to suit their individual criteria.

 

[flecc]

No matter how I look at this, it cannot be anywhere near as complex as it appears. The variable's involved with car's are far more complex yet the manufacturers have little trouble quoting specification's.
A Robot or machine that simulate's an average rider with average input, under average conditions is all it take's. The object is to devise a standard test, not give everyone a personal data sheet to suit their individual criteria.

I almost despair! A car is not a hybrid of two power sources and is not more complex, it's dynamically less complex than the Panasonic type system.

And the standard test you propose would not give any meaningful comparison with hub motor systems. The question is, where and at what level do you set all the robotic variables. Wherever you set them, they will be wrong since the essence of the working system is continuous dynamic variability, not constants, so any results would not in any way reflect the real world system. Panasonic would reject such results.
.

 

[Patrick]

No matter how I look at this, it cannot be anywhere near as complex as it appears. The variable's involved with car's are far more complex yet the manufacturers have little trouble quoting specification's.
A Robot or machine that simulate's an average rider with average input, under average conditions is all it take's. The object is to devise a standard test, not give everyone a personal data sheet to suit their individual criteria.

If we had a full size version of Murtaca's cycling robot

set to mimic the arthimetic mean cycling ability of able bodied UK adults, then we'd have the "average rider with average input" you need. But how would you then go about establishing "average conditions"?

 

[timidtom]

I can buy a car or a motor bike or even a normal bicycle and expect to be able to travel a hundred miles and arrive at my destination. As yet I can't do that with an electric bike. The arguments about the range, durability, fuel consumption and hill climbing abilities of available pedelecs are quite interesting but about as relevent to reliable journey information as medieval debates of how many angels could dance on the point of a pin were to Christianity.
What ambitious journeys, without close support from manufactures/stockists have members undertaken on pedelecs using only the capacity of their single battery?
Mine? Six-speed Euro Powabyke regular 40 mile round trip to Tatton Park, numerous occasions, average speed 14 mph. Bike allowed to rest while I consumed coffee & cake to nourish my 12 stone bulk ...
I know the Powabyke is an outdated, heavy tank but it's durable, reliable and has a practical range.
Come one - show me what can do better?

 

[Alex728]

I regularly go 27 miles from Ipswich to friends in Mid Suffolk (almost into Norfolk) and managed that on both the Powacycle Salisbury and the Wisper 905SE. On the Wisper that is with a fair amount of throttle usage... (i.e full power) and an uphill journey and I get the impression there is still plenty of power left in the battery afterwards.

however I am on the lighter side (1.65m height and 65 kilos weight) and have been cycling regularly for 10 years, am physically active and have never owned a car....