10 kW
deardancer3 said:
So the 28 in the rear is 2:1 gearing (28:14)?
I am assuming in 1st, the bike would have climbed the 8% since it climbed it in 2nd. However, the speed would have been under 7 mph. Is that correct?
Standard hill-climbing test suggestion
- Thread starter pdf
- Start date
So the 28 in the rear is 2:1 gearing (28:14)?
I am assuming in 1st, the bike would have climbed the 8% since it climbed it in 2nd. However, the speed would have been under 7 mph. Is that correct?
deardancer3
10 kW
pdf said:
Yes that is correct. It seems that when the cyclone is over reving, the power drops pretty fast.
But if i did that hill in first and pedaled a significant amount it would do 9mph and be easy on the battery. At least till the end of life for the NIMH ---which cooked itself recharging due to a loose thermistor which delayed sense of temperature rise during recharge.
dogman dan
1 PW
I took the new motor, an E-bikeKit 36v up the same hill today. Same bike, Same hill, Same battery, different motor and controller. This is the motor kit designed for 20 mph on 36v batteries.
So 36v pingbattery, E-BikeKit ( 9 continents) front hub motor, 22 amps E-BikeKit controller, 180 pound rider, 85 pound bike. The hill is about 1.5 miles, starting out 5% with the last half 7% . Ambient temp 69F, wind about 3 mph tailwind.
Starting out from a dead stop, the motor reached a top speed of 15 mph on the 5% section. The slowest speed was 12.5 mph on the steepest part, and most of the 7% grade went at 13-13.5 mph.
Motor heat was guessimated at about the same as the aotema motor test, nice and toasty by the top. This is really a torture test for a motor.
A second run was about 1 mph slower, since the battery was near 50% discharged by then. Top speed on the second run was about 14 mph, and 12 mph on the 7%
So 36v pingbattery, E-BikeKit ( 9 continents) front hub motor, 22 amps E-BikeKit controller, 180 pound rider, 85 pound bike. The hill is about 1.5 miles, starting out 5% with the last half 7% . Ambient temp 69F, wind about 3 mph tailwind.
Starting out from a dead stop, the motor reached a top speed of 15 mph on the 5% section. The slowest speed was 12.5 mph on the steepest part, and most of the 7% grade went at 13-13.5 mph.
Motor heat was guessimated at about the same as the aotema motor test, nice and toasty by the top. This is really a torture test for a motor.
A second run was about 1 mph slower, since the battery was near 50% discharged by then. Top speed on the second run was about 14 mph, and 12 mph on the 7%
dogman said:
Thanks for the test. My simulator seems to predict about 11.5-12 mph on a 7% grade with the conditions you inputted, so that might suggest that pdf's noticed discrepancy might be related to the accuracy of the ezee's dataset or possibly there was some other condition with the bike such as a manually lowered current limit.
By the ways, 13 mph at 7% is pretty good for a hub motor. :wink:
mwkeefer
1 MW
Guys,
This is all great, standardizing the "testing" protocol would help with the apples to apples comparisons and evaluating what is the best solution for a given purpose!
One suggestion - couldn't we just use a variable resistance stationary stand to "simulate" a 0 wind hill of practically any grade or length? I ask merely because although I have plenty of killer hills here, taking them on an ebike with traffic sort of precludes me from accurately reproducing test conditions as Ive got vehicular traffic to be concerned with and a mirror to watch when I'm on the road... With an excercise bike / stationary stand we could just mimic the effect of the hill and therfore be able to reproduce the test on any bike setup without deviation for temperature, wind, etc.
Just my .02 for what it's worth, I don't even know what a stationary resistance type stand costs... I made mine, it charges SLA batteries or powers halogen headlamps for load... I haven't figured the mechanical loss but assuming 30% (friction) so if I can generate 1500w under full load (driving generator) then I assume the motor is producing approx 30% more power than I am recording. Honestly I haven't figured out the simulation all that well... I analyzed an eagle tree log of one of my BIG hills and then attempted to vary the load on the stand to mimic the power consumption and speeds I saw in the EagleTree logs... this gave me a platform for base testing which is actually quite good at relative measurement... with calibration and proper loss calculations it would be a simple, scalable dynometer for eBikes.
-Mike
This is all great, standardizing the "testing" protocol would help with the apples to apples comparisons and evaluating what is the best solution for a given purpose!
One suggestion - couldn't we just use a variable resistance stationary stand to "simulate" a 0 wind hill of practically any grade or length? I ask merely because although I have plenty of killer hills here, taking them on an ebike with traffic sort of precludes me from accurately reproducing test conditions as Ive got vehicular traffic to be concerned with and a mirror to watch when I'm on the road... With an excercise bike / stationary stand we could just mimic the effect of the hill and therfore be able to reproduce the test on any bike setup without deviation for temperature, wind, etc.
Just my .02 for what it's worth, I don't even know what a stationary resistance type stand costs... I made mine, it charges SLA batteries or powers halogen headlamps for load... I haven't figured the mechanical loss but assuming 30% (friction) so if I can generate 1500w under full load (driving generator) then I assume the motor is producing approx 30% more power than I am recording. Honestly I haven't figured out the simulation all that well... I analyzed an eagle tree log of one of my BIG hills and then attempted to vary the load on the stand to mimic the power consumption and speeds I saw in the EagleTree logs... this gave me a platform for base testing which is actually quite good at relative measurement... with calibration and proper loss calculations it would be a simple, scalable dynometer for eBikes.
-Mike
10 kW
swbluto said:
I agree, something about the test was not the same as the simulator in my case. I want to replicate this as soon as I get a chance.
10 kW
mwkeefer said:
This is in essence what has been done by swbluto and ebikes.ca. ebikes.ca has a simulator that creates power and torque curves as a function of speed. They have been calibrated against a dynamometer. swbluto has gone a step further and generated a simulator that gives the time dependent speed for vaious environmental conditions; grade, wind, etc. I was simply interested in how people's experience matched up against the simulators; in a few tirials I did, the simulation results did not match mine for speed up a grade. So I suggested an emperical approach involving actual hills.
Take a look at the simulator of swbluto and ebikes.ca. I think they have done what you suggest for several motors. It looks like with the few results we have here, my results might be anomolous. If so, the focus should shift to generating motor characteristics files for more motors.
dogman dan
1 PW
Lotsa stuff can effect real world results. I bet both motors I tested would have been more perky at 85F ambient temp. Condition of the battery is most likely the biggest variable between two bikes. My ping lost a lot of zip by the second trip up the hill. I bet with a 10 ah ping, it would be a lot slower, and slower still on 8 ah nicads.
And yeah, I was pretty amazed either direct drive motor could do this hill at all with no pedaling at all. I was suprised by the Aotema, but the E-BikeKit motor simply amazed me. It may be only a 20 mph motor at 36v, but it rocks on hills for a direct drive hubmotor.
And yeah, I was pretty amazed either direct drive motor could do this hill at all with no pedaling at all. I was suprised by the Aotema, but the E-BikeKit motor simply amazed me. It may be only a 20 mph motor at 36v, but it rocks on hills for a direct drive hubmotor.
mwkeefer
1 MW
PDF,
I know swbluto has designed a software simulator but I am talking about building our own hobby dyno meters using a modified stationary bike stand driving a fair sized (must be rated for a > capacity than you want to test) DC motor used as a generator... I meant for comparison of data sampled by individuals...
For instance... a simple atmel avr interface to control a current limiting FET bridge would provide a variable load to the eBike or variable resistance, no need to reinvent the wheel here, just use a basic controller in reverse and combine with a CA or EagleTree to monitor current, voltage, temperatures (ambient for instance), etc. This would allow a simple PC app to control a diy dynometer, I don't even know what a true automotive dyno costs but it can't be cheap... I figure 2-300.00 and we can build a dyno suitable for standardized testing of eBikes or LEVs.
Auto dyno is way overkill for eBikes and heck, most of us DIY everything we can anyway even if just for fun... This would allow automated and standardized measurements, including controlling the load presented to the ebike (effective resistance) which would simulate climbing a particular grade.. I am suggesting we actuate this in the real world.
To me, the numbers are everything... pretty much any particular type of incident or issue with a hub motor can be detected in the data, when I began collecting logger data was when I left total ignorance behind in favor of scientific approach (at least to testing and measuring effect of various system configuration changes). If we could agree on a 3,5 or 10kw dyno design using off the shelf parts/plywood, etc... for a reasonable price then I would be happy to write the control software to manage the test simulations. This will give us the advantage of being able to reproduce any physical environment (I got this idea watching a video about the design of Mile's bike) down to temperature (I have AC and heat, do you?)... I just think we could begin to collect much more accurate engineering data with such a project and that would benefit the end consumer, swbluto's modeling and our design and building direction.
-Mike
-Mike
I know swbluto has designed a software simulator but I am talking about building our own hobby dyno meters using a modified stationary bike stand driving a fair sized (must be rated for a > capacity than you want to test) DC motor used as a generator... I meant for comparison of data sampled by individuals...
For instance... a simple atmel avr interface to control a current limiting FET bridge would provide a variable load to the eBike or variable resistance, no need to reinvent the wheel here, just use a basic controller in reverse and combine with a CA or EagleTree to monitor current, voltage, temperatures (ambient for instance), etc. This would allow a simple PC app to control a diy dynometer, I don't even know what a true automotive dyno costs but it can't be cheap... I figure 2-300.00 and we can build a dyno suitable for standardized testing of eBikes or LEVs.
Auto dyno is way overkill for eBikes and heck, most of us DIY everything we can anyway even if just for fun... This would allow automated and standardized measurements, including controlling the load presented to the ebike (effective resistance) which would simulate climbing a particular grade.. I am suggesting we actuate this in the real world.
To me, the numbers are everything... pretty much any particular type of incident or issue with a hub motor can be detected in the data, when I began collecting logger data was when I left total ignorance behind in favor of scientific approach (at least to testing and measuring effect of various system configuration changes). If we could agree on a 3,5 or 10kw dyno design using off the shelf parts/plywood, etc... for a reasonable price then I would be happy to write the control software to manage the test simulations. This will give us the advantage of being able to reproduce any physical environment (I got this idea watching a video about the design of Mile's bike) down to temperature (I have AC and heat, do you?)... I just think we could begin to collect much more accurate engineering data with such a project and that would benefit the end consumer, swbluto's modeling and our design and building direction.
-Mike
-Mike
deardancer3
10 kW
I lived a life of hi tech stuff from 1965. Lots of time getting paid to write software, work on hardware, saw some great computer rooms in real nice cities. Managed lots of very nice S/W Support engineers.
One of our Best consultants was once asked what the job was of a Senior Consultant; he said --
"Broke, Fix, Get Banana."
Will now leave designing and building dyno's and simulators to others. I would rather ride bikes/ebikes up hills than do any more computer stuff. If I break something , I'll fix it.
Best of luck and success on the dyno.
As one member says:
"Several days of work can save you hours of planning"
and another:
"One test is worth a thousand opinions"
d
One of our Best consultants was once asked what the job was of a Senior Consultant; he said --
"Broke, Fix, Get Banana."
Will now leave designing and building dyno's and simulators to others. I would rather ride bikes/ebikes up hills than do any more computer stuff. If I break something , I'll fix it.
Best of luck and success on the dyno.
As one member says:
"Several days of work can save you hours of planning"
and another:
"One test is worth a thousand opinions"
d
10 kW
mwkeefer said:
Wrote a long reply, managed to lose it somehow. Gist is that what I should have written is that the simulator at ebikes.ca has been validated with a dynamometer and swbluto's simulator has been found in agreement with that simulator for flat terrain. The reason for hill climbing trials was to validate swbluto's model for elevation changes.
So, in summary, what you propose has been done, partially. Look at the page on ebikes.ca for a picture and description of dynamometer validation. What would help is dynamometer data for more hub motors and tests on controllers because controller variability is pretty common from what I've read here. Also, I am working with a group of students on a light electric vehicle design and would be interested in a testing apparatus for small electric motors also.
Possible DIY Dyno:
http://www.sci-spot.com/Mechanical/dyno.htm
http://www.sci-spot.com/Mechanical/dyno.htm
dogman dan
1 PW
I agree, a dyno couldn't really simulate hill climbing in any way I can think of. You could add resistance, but climbing a real hill lifting the real weight of the rider and bike is the only way to truly get the hill performance data. When I did my tests the other day, I found only about .1 mile was needed to get the motor to settle in at equilibrium speed. Surely most of us can find a hill with .1 to .2 mile at a steady grade to do real world no pedaling tests to post here. Even the flat landers might have enough hill on a freeway overpass somewhere.
deardancer3
10 kW
Yeah, but can you make a decent inclinomoter with 2 yardsticks a foot ruler and a level?
I thought this was going to be tough, till I did it;
Cruise area looking for consistent hills. measure hills for incline and length. (writing stuff down helpful)
If writing is not your forte, take a voice recorder.
Charge battery, pump up ebike tires. Typical bike check.
look at trees and flags for wind; if none to minor, proceed. if significant, record data.
ride test ebike to hill, obtain desired speed and hit the uphill, full throttle no pedaling.
If the ebike cant keep speed at decent level, abort; find different hill.
Watch speedometer and road. at end - feel/measure temp of motor. again, write stuff down.
weigh bike and self.
repeat.
Now that temperatures are well below 100F (for most of us) its a nice time to do this. at least for the ebike.
But I sooo wish I had my meters hooked up for watts and amps.
The weather is nice however- I have a bum knee today; good to know the cyclone can handle the load if I need to do much.
I thought this was going to be tough, till I did it;
Cruise area looking for consistent hills. measure hills for incline and length. (writing stuff down helpful)
If writing is not your forte, take a voice recorder.
Charge battery, pump up ebike tires. Typical bike check.
look at trees and flags for wind; if none to minor, proceed. if significant, record data.
ride test ebike to hill, obtain desired speed and hit the uphill, full throttle no pedaling.
If the ebike cant keep speed at decent level, abort; find different hill.
Watch speedometer and road. at end - feel/measure temp of motor. again, write stuff down.
weigh bike and self.
repeat.
Now that temperatures are well below 100F (for most of us) its a nice time to do this. at least for the ebike.
But I sooo wish I had my meters hooked up for watts and amps.
The weather is nice however- I have a bum knee today; good to know the cyclone can handle the load if I need to do much.
dogman dan
1 PW
I used to use a homemade protractor with a peice of wire attached to measure slopes in degrees while out backcountry skiing. The piece of wire would swing to vertical like a plumb bob, and then indicate the degrees of slope. One could be made that showed % grade. It only needs to be about 9 inches long to be accurate.
Russell
1 MW
What I find odd about climbing a hill with a motor is how much it will slow down when used by itself and then how much faster it is with just a little input from me. I'm not talking a bit faster with a little pedal input I'm saying on my motor alone I may be able to climb an 8% grade at 5 mph but I just had to turn the pedals a little and the motor actually spun the front tire as it surged ahead with only a small input from me required to stay at 12 mph up the grade.
Obviously then we're not dealing simply with power here but the thrust the motor in the wheel can develop. So when looking at hill-climbing ability I started paying attention to the THRUST as well as the POWER curves in the simulator plots at ebikes.ca. Justin talks about a simple way to figure the approximate thrust required as (weight x percent grade) however this power calculator will give a more precise value in the "Average Traction Force" box.
http://www.mne.psu.edu/lamancusa/ProdDiss/Bicycle/BIKEcalc1.HTM
-R
Obviously then we're not dealing simply with power here but the thrust the motor in the wheel can develop. So when looking at hill-climbing ability I started paying attention to the THRUST as well as the POWER curves in the simulator plots at ebikes.ca. Justin talks about a simple way to figure the approximate thrust required as (weight x percent grade) however this power calculator will give a more precise value in the "Average Traction Force" box.
http://www.mne.psu.edu/lamancusa/ProdDiss/Bicycle/BIKEcalc1.HTM
-R
Rassy
1 MW
Russell wrote:
Exactly. I have about 1/4 mile hill leading to my house with spots that get up to 15% grade. I've had brushed and brushless, Direct and geared, 16"/20"/26" wheels, 36V and 48V, and on the 15% grade they all take a nosedive unless I give a little help. One exception is my little Bafang in a 16" wheel, when I quit pedaling it keeps on going, until I hit the 20% grade in my driveway. Friends ask my how fast I can go up the hill without pedaling. My answer is "I don't know, because when the motor starts lugging I can't stand not helping it. So if I have a motor that goes up the hill at 12 MPH with moderate pedaling and I stop pedaling until it slows to about 8 MPH (and still dropping), It comes right back to 12 MPH when I resume moderate pedaling. It puzzles me how a 48V motor sucking about 25 amps (well over 1000 Watts) can respond this way to my meager input.
dogman dan
1 PW
Let me rephrase that. " I can't think of any way a dynomometer could PHYSICALY simulate a hill. " Drag on the wheels can be simulated by putting friction on the system. But simulating lifting weight up a hill would need to lift a weight. Mabye you could ride up a treadmill and re create it, but something in my dog brain says it would be different somehow. My knowledge of physics is a 12 grade class, sooo. In a purely computer simulator, hills should be doable, but the comfirmation of that simulator would best be done on a real hill.
The reason your pedaling input helps so much is because you do that imput through a properly selected gear. This gives you great tourqe even with low wattage. When the motor is stalling, your pedaling can apply many times the tourque the stalled motor is supplying. Cyclone owners know this, and it's why the long term future of ebikes is going to be bottom bracket drives in hilly places. Cyclone just puts the motor in a bad location in my opinon, but they have the right idea driving the chain.
I allways get a chuckle out of the ones that want a motor to get em up a hill without any pedaling. A few burnt motors will learn em. This stuff wasn't really designed to do that on steep hills. No pedal all you want on the flats or mild hills, but above 4%, start pedaling dude.
The reason your pedaling input helps so much is because you do that imput through a properly selected gear. This gives you great tourqe even with low wattage. When the motor is stalling, your pedaling can apply many times the tourque the stalled motor is supplying. Cyclone owners know this, and it's why the long term future of ebikes is going to be bottom bracket drives in hilly places. Cyclone just puts the motor in a bad location in my opinon, but they have the right idea driving the chain.
I allways get a chuckle out of the ones that want a motor to get em up a hill without any pedaling. A few burnt motors will learn em. This stuff wasn't really designed to do that on steep hills. No pedal all you want on the flats or mild hills, but above 4%, start pedaling dude.
10 kW
A couple of quick thoughts:
First, my interest in excluding pedaling is so that the comparison between people is consistent, not because I think this is an appropriate way to operate a 400W motor. I think everyone understands that, but just so everyone is on the same page.
Second, I think it is completely possible to simulate a hill by loading up a motor on a dynamometer. The reason I suggested actual trials is that my personal experiments did not agree with the simulators so I wanted to try to figure out if my experience was an anomoly or not. I don't see any reason why a simulator calibrated with dynamometer data would not, in theory, agree with experimental data. However, it is possible that there are significant effects (and I think dogman alludes to this) that are not accounted for in simulators that give time dependent speed curves. For example, it is possible that a simulator assumes a constant temperature but that in reality, the temperature changes a lot during a trial. It might be quite difficult to model the temperature rise by electrical heating for every motor, in fact I suspect it is. So while I think it is quite possible to model this (phenomena that are MUCH more complicated have already been modelled highly accurately in some cases, like fluid turbulence for example) it may well be that the state of the models is not yet mature. The same was true of fluid mechanics modeling not 25 years ago but now a large fraction of all fluid transport applications can be modelled highly accurately.
My 2 cents.
First, my interest in excluding pedaling is so that the comparison between people is consistent, not because I think this is an appropriate way to operate a 400W motor. I think everyone understands that, but just so everyone is on the same page.
Second, I think it is completely possible to simulate a hill by loading up a motor on a dynamometer. The reason I suggested actual trials is that my personal experiments did not agree with the simulators so I wanted to try to figure out if my experience was an anomoly or not. I don't see any reason why a simulator calibrated with dynamometer data would not, in theory, agree with experimental data. However, it is possible that there are significant effects (and I think dogman alludes to this) that are not accounted for in simulators that give time dependent speed curves. For example, it is possible that a simulator assumes a constant temperature but that in reality, the temperature changes a lot during a trial. It might be quite difficult to model the temperature rise by electrical heating for every motor, in fact I suspect it is. So while I think it is quite possible to model this (phenomena that are MUCH more complicated have already been modelled highly accurately in some cases, like fluid turbulence for example) it may well be that the state of the models is not yet mature. The same was true of fluid mechanics modeling not 25 years ago but now a large fraction of all fluid transport applications can be modelled highly accurately.
My 2 cents.
Similar threads
- Question
