Msuper v3 top speed crash - wrecked shell

[OliverH]

18 hours ago, EricGhost said:

1rst there are motors and there are motors, electrical or not and they have their own curve not all the same, but all of them must follow the power definition, I.e: power=torque x angular speed, as power has a limit Named max power it is clear that if you run at max power the higher the speed the lower the torque. But when you start your run you do not start at max power therefore torque & speed can increase both

This does not mean that your available torque increase with speed, it means you're reaching the max power point once you reach it on plain terrain at a certain speed go uphill and try to keep the speed let's see if torque will increase  

http://lancet.mit.edu/motors/motors3.html

P= M x n

[Chriull]

48 minutes ago, Paco Gorina said:

In enclose s .pdf explaining why sometimes wheels fall down with no apparent cause which is related to the torque / speed question. There is also a Mathematica file so you may play a little.

 

english.pdf

Model_2.nb

 

29 minutes ago, Paco Gorina said:

In word format t in this link. Better graphics

http://www.gorina.es/9BMetrics/physics/english.docx 

Pfff.... It's still hard to follow/understand  Is the main point of your paper, that once the actual torque/speed leaves reaches the limits of the torque/speed graph of the motor the torque decreases and no balance can be obtained anymore? And this shown/backed up by the simulation of the whole system?

[Dancer]

If @Jane Mo could post the program flow chart of the firmware, we could stop guessing.

[Paco Gorina]

The point is that 

1.- Out of balance happens very fast. It is not something slow and predictable.

2.- When you have low torque available anything may send you out of balance.

3.- Speed reduction also may send you out of balance. Attention in descents.

4.- You may be OK but voltage reduction may also send you out of balance

5.- Voltage reduction may happen because how battery responds loads

The simulation just solves the wheel equations and you may change the related data or external input but for the moment doesn't includes batteries which are a big part of the picture

Really you need to have a BIG security margin to be sure 

6 minutes ago, Dancer said:

If @Jane Mo could post the program flow chart of the firmware, we could stop guessing.

That would be very interesting

[Chriull]

32 minutes ago, Paco Gorina said:

...

1.- Out of balance happens very fast. It is not something slow and predictable.

...

You think you could implement some (maybe simplified) simulation like this in 9BMetrics to predict (possible) overleans and release a warning in time? Or is this already the reason you made the simulation, to identify "dangerous states" which will rise warnings in 9BMetrics? 

[zlymex]

1 hour ago, Dancer said:

Now I am confused. I just went for a test.

My battery is at 60 %. When I accelerate hard the alarm goes off at 30 km/h. When I accelerate slow I can go to 35 km/h without alarm.

That warning at 30 km/h is probably the 2nd level speed warning, three beeps every 1 second.

[Dancer]

8 minutes ago, zlymex said:

That warning at 30 km/h is probably the 2nd level speed warning, three beeps every 1 second.

1st and 2nd level speed warning were turned off. I stopped pushing after the 2nd beep.

[zlymex]

I agree with this very much:  Out of balance happens very fast.
I've seen many faceplants directly or indirectly. I classify them as:
1. obstacle encountered, such as a ditch, a ramp, the curb, a piece of brick on the road. If the rider is unaware of the obstacle, 9 out of 10 will out of balance.
2. accelerate too hard
3. over speeding.
There off balance happens fast in the first two cases, no time for the wheel to give warning or tilt-back. In the third case, there should already a warning of overspeed but the rider ignore it or shut it down before hand.

[US69]

5 minutes ago, Dancer said:

1st and 2nd level speed warning were turned off. I stopped pushing after the 2nd beep.

The Battery sag/Voltage drop is so much/great, that you are getting a warning Kind of overload/low Batterie!

Do not push to hard near 50% Batterie!

[Cloud]

5 hours ago, Chriull said:

 

The "used" torque is direct proportional to the acceleration of the wheel. So if one "asks" the wheel for a specific acceleration by leaning forward the controller lets a current flow through the motor which generates this torque. So current is direct proportional to the torque which is direct proportional to the acceleration. (The controller can regulate the current by the duty cycle of the PWM pulses).

This is true as long the torque (~current) and speed of the wheel is within the borders of the torque/speed limit chart. Once one "hits" this limit one start to overlean!

If you look at the last "D.C. motor torque/speed curve" @zlymex posted (with the red dot) and imagine you start driving by leaning forward constantly so the wheel starts a constant acceleration with exactly the torque of this red spot. Then one "moves" in this chart on a constant line (parallel to the "rotational speed"/x-axis) until one reaches the speed indicated by the red spot. This is exactly the last moment where the motor can deliver this torque. After this spot the torque decreases and as one still leans forward this means the wheel does not accelerate enough anymore to balance the driver and a faceplant happens (if the rider cannot react fast enough and balance himself again by leaning back)

 

Yes, this is exactly correct. I still would like to see the real life curve showing torques ate different speeds after the wheel reaches the constant speed after acceleration. The windloading will increase the requested electrical power supplied to the motor and will consequently mke the torque be slightly about the torque/ speed curve.

5 hours ago, Chriull said:

The graph was created by accelerating a ?1 ton? barrel by the wheel. From the acceleration of this barrel the computer calculates power and torque over the speed. The wheel was "pitched" forward to accelerate this barrel. Some "frame" helped holding the wheel in position. The initial acceleration (something like up to ~10 km/h) was not the maximum the wheels could "deliver", since quite some fuses burned/overpower warnings where given.

I don't know exactly what happened after the max power was reached - either the "dynamometer" stopped, the tester stopped accelecating/testing or the wheel cut-off?

Anyway - it was not a measurement of the power/torque over speed but only the max power the wheel could produce - and this is not limited by the motor but by the electronics. (quite sure but still imho;) ) This is imho also shown by the "undulating nature" of the graph - this could result from the regulating mechanisms in the firmware. Or could also be from the "soft" tires coupling the wheel with the barrel - or maybe a combiation of both?

This max power measurement has (as good as) nothing to do with the torque capabilities when riding at around 40/45 km/h.

ive now watched the video showing the dyno test the russians have conducted. They used the motocycle rig with a very heavy barrel, not really designed to test eucs. The wheel had to find the inertia of the barrel rather than the inertia of accelerating the weight of the rider. Therefore, their placing the rider's weight on the pedals was almost completely useless and inconsequential!  All that did was increase friction between the tyre and the surface of the barrel. Like @Chriull correctly said, if anything, the test will show  the wheel's max power, but maybe for comparison purposes only. 

The wheel went into the overload mode at around 20kmh because the wheel was leaned forward while having to accelerate the super heavy barrel! So the spoed at which the wheel "crashed" and the torque was going down rapidly has nothing to do with the maximum wheel's speed, as in real life conditions, the wheels torque and power will be reacting to accelerate ( fight the inertia) of the rider, not the wheelbarrel.

They mentioned they will be publishing more thorough tests results in the few days.

[zlymex]

10 minutes ago, Cloud said:

Yes, this is exactly correct. I still would like to see the real life curve showing torques ate different speeds after the wheel reaches the constant speed after acceleration. The windloading will increase the requested electrical power supplied to the motor and will consequently mke the torque be slightly about the torque/ speed curve.

The torque is direct proportional to the armature current which in turn proportional to the input current of the battery. I have tested by a precision current sensor plus an external recorder to make some chart while I was riding my IPS T260 at different scenarios. Here is the chart riding up slope/down slop. Red curve is the current in A, blue voltage in V, green power in Watts/10. X-axial is time in seconds.

[Greg Spalding]

16 minutes ago, Cloud said:

Yes, this is exactly correct. I still would like to see the real life curve showing torques ate different speeds after the wheel reaches the constant speed after acceleration. The windloading will increase the requested electrical power supplied to the motor and will consequently mke the torque be slightly about the torque/ speed curve.

ive now watched the video showing the dyno test the russians have conducted. They used the motocycle rig with a very heavy barrel, not really designed to test eucs. The wheel had to find the inertia of the barrel rather than the inertia of accelerating the weight of the rider. Therefore, their placing the rider's weight on the pedals was almost completely useless and inconsequential!  All that did was increase friction between the tyre and the surface of the barrel. Like @Chriull correctly said, if anything, the test will show  the wheel's max power, but maybe for comparison purposes only. 

The wheel went into the overload mode at around 20kmh because the wheel was leaned forward while having to accelerate the super heavy barrel! So the spoed at which the wheel "crashed" and the torque was going down rapidly has nothing to do with the maximum wheel's speed, as in real life conditions, the wheels torque and power will be reacting to accelerate ( fight the inertia) of the rider, not the wheelbarrel.

They mentioned they will be publishing more thorough tests results in the few days.

I'm glad you just posted that

I am very appreciative for anybody doing studies that can help us make decisions about what we would like to purchase, but the truth is there are no conclusions that can be drawn from that particular set of tests. At first it might be difficult to discern what was going on because of the language barrier, but the truth is they ruined several wheels in that test because of the way in which they had their motors/cbs taxed inappropriately. I hope they didn't have to pay for those because it was a shame

most importantly, I like those guys a lot and I appreciate their efforts, but basically it was just a demonstration for no reason

[Cloud]

4 minutes ago, zlymex said:

The torque is direct proportional to the armature current which in turn proportional to the input current of the battery. I have tested by a precision current sensor plus an external recorder to make some chart while I was riding my IPS T260 at different scenarios. Here is the chart riding up slope/down slop. Red curve is the current in A, blue voltage in V, green power in Watts/10. X-axial is time in seconds.

The sections of the graph where the current/ power was negative, was that regenerative braking/downhill riding?

[Cloud]

@Greg Spalding i dont think the tests were totally useless. I think they do show max power which can somehow be used to compare between different wheels. I also think its a great start and they will develop the tests in the future to show more useful parameters

[Chriull]

12 minutes ago, Cloud said:

Yes, this is exactly correct. I still would like to see the real life curve showing torques ate different speeds after the wheel reaches the constant speed after acceleration. The windloading will increase the requested electrical power supplied to the motor

In my charts in this thread (http://forum.electricunicycle.org/topic/5541-msuper-v3-top-speed-crash-wrecked-shell/?do=findComment&comment=64654) you see the power needed to overcome the air drag and the friction. If you add both powers (P Used) and divide it by the velocity (in m/s) you get the Force of the air drag and the friction. If you divide the force by the radius of the wheel (16 inch /2 in m) you get the needed torque to counter the force of air drag and friction. Around this torque you will have something plus minus needed for balancing the rider.

12 minutes ago, Cloud said:

and will consequently mke the torque be slightly about the torque/ speed curve.

The torque/speed curve is the absolute maximum - the torque can be at this curve or near, but never above.

 

12 minutes ago, Cloud said:

ive now watched the video showing the dyno test the russians have conducted. They used the motocycle rig with a very heavy barrel, not really designed to test eucs. The wheel had to find the inertia of the barrel rather than the inertia of accelerating the weight of the rider. Therefore, their placing the rider's weight on the pedals was almost completely useless and inconsequential!  All that did was increase friction between the tyre and the surface of the barrel. Like @Chriull correctly said, if anything, the test will show  the wheel's max power, but maybe for comparison purposes only. 

Exactly how i understood it. This "machine" and how this test was performed is only aimed at determining the wheel'max power and nothing else. Putting the riders weight is/was absolutely useless - the load is solely determined by the inertia of the barrel.

12 minutes ago, Cloud said:

The wheel went into the overload mode at around 20kmh because the wheel was leaned forward while having to accelerate the super heavy barrel! So the spoed at which the wheel "crashed" and the torque was going down rapidly has nothing to do with the maximum wheel's speed, as in real life conditions, the wheels torque and power will be reacting to accelerate ( fight the inertia) of the rider, not the wheelbarrel.

They mentioned they will be publishing more thorough tests results in the few days.

You speak russian if i remember right? Could you maybe ask them to also test the braking abilities in their test?

If you look at their charts the curve at the rightmost end crosses the x-axis and goes in the negative region and back to lower speeds - that shows the torque/power with which the wheel "brakes" the barrel. If they "pitch" the wheel backwards after the max power was reached they could test the max breaking power of the wheel, too. I am still very very conservative while braking with my wheel - i always remember the burned mosfets reported here after "emergency braking". Would be interesting to see the abilities of the tested wheels! 

And you should warn them, that it could easily be, that while performing this test they'll burn some more fuses and motherboards... (as they already did with their max power test...)

[zlymex]

1 minute ago, Cloud said:

The sections of the graph where the current/ power was negative, was that regenerative braking/downhill riding?

Yes, down hill.

For hard braking here is another chart. The first part was doing quick swing(small distance forward/reverse). Later part, I was doing accelerate-hard brake to complete stop, and repeat.

[Greg Spalding]

9 minutes ago, Cloud said:

@Greg Spalding i dont think the tests were totally useless. I think they do show max power which can somehow be used to compare between different wheels. I also think its a great start and they will develop the tests in the future to show more useful parameters

Perhaps the word was a bit harsh, but there was no reason to burn out three different wheels Which might give someone the impression there was something wrong with them... which was not true.

Also, there was really nothing repeatable about the way those tests were preformed and therefore while the numbers that came out were interesting, I'm pretty sure that the values Obtained would not necessarily lead to data that could be extrapolated in such a way that proper comparisons between them could be made. Don't misunderstand me, I found the results and the testing itself fascinating, especially because they let us in on how it was all done, I'm just not sure we can draw any conclusions from those results.  This is mostly because, as was mentioned above, the use of that heavy barrel, which does not really approximate the way a wheel pushes the ground as someone does or does not stand on it.

As I said before, I do like those guys very much and I appreciate their efforts and I will look forward to subsequent testing in the future.

I really don't mean to offend anyone in what I said. Let my take-home message be I'm looking forward to further studies and I appreciate the efforts and the fact that the data and testing methods and even the video supporting all of that was shared with us here

[Chriull]

 

26 minutes ago, Greg Spalding said:

I'm glad you just posted that

I am very appreciative for anybody doing studies that can help us make decisions about what we would like to purchase, but the truth is there are no conclusions that can be drawn from that particular set of tests. At first it might be difficult to discern what was going on because of the language barrier, but the truth is they ruined several wheels in that test because of the way in which they had their motors/cbs taxed inappropriately. I hope they didn't have to pay for those because it was a shame

most importantly, I like those guys a lot and I appreciate their efforts, but basically it was just a demonstration for no reason

 

13 minutes ago, Greg Spalding said:

Perhaps the word was a bit harsh, but there was no reason to burn out three different wheels Which might give someone the impression there was something wrong with them... which was not true.

The wheels where overpowered at low speeds. If this situation cannot be handled by a wheel, there is something wrong... Also a load like this is not likely to happen in real life driving.

Or if one thinks "wrong" is to harsh for this case then at least it is not "state of the art" for electronic ?desing/engineering?

Quote

Also, there was really nothing repeatable about the way those tests were preformed and therefore while the numbers that came out were interesting.

Really? The only "degree of freedom" was the way he tilted the wheel forward to accelerate.

But thats imho more or less the same as the max power measurement with motorcycles - they use exactly this machinery and also pull the throttle by hand... (imho)

Quote

I'm pretty sure that the values could be extrapolated between the different wheels because the way in which the test was set up. Don't misunderstand me, I found the results and the testing itself fascinating, especially because they let us in on how it was all done, I'm just not sure we can draw any conclusions from those results.

 

Imho thats the way max power measurements are performed and i can not really see how this could be improved. Maybe by a special firmware which controlls the acceleration of the wheel and so noone has to tilt the wheel forward.

Edit: Maybe the weight of the barrel could be lowered, and by this the test could deliver more refined results?

[Greg Spalding]

CHRIULL,

I appreciate all of your comments above in helping explain to me some of the things I was seeing. I am reminded once again of how we are really in the infancy of this technology and advancements will only occur when testing is done. Perhaps, most of all, I was saddest to see those three wheels destroyed during the testing done in such a way that may not simulate, as you said, a real life load situation

[Jose Otal]

lo siento por el traductor . amigo my buscando trabajo para poner a prueba sus límites ruedas y buscar el poder, no tiene miedo caída cara faceplan, broma
Estoy muy atento a este post, porque estoy esperando que mi nueva msupe rueda 820, y aquí y encontrado algunas respuestas para fururo, pero tengo miedo, yo y caído de mi ninebt, al principio, porque yo no sabía el sistema tildbak seguridad, y de haber salido de la rueda, pensando que había sido desconectado, no tengo dolor en la rodilla, ya que, y después de tres meses todavía tienen miedo de escuchar un bip bip, Gotway tiene diferentes tipos de alertas, pero cuando miro a vídeo, que casi no puede distinguir y memorizar 9 o 10 alertas diferentes, que un buen vídeo con los consejos más importantes, se quiere ser mejor para aquellos que no entienden gráfico

[zlymex]

There is an easier way to test the max peak power.
Two points determine a line. Stall torque at zero speed and no-load speed determine the speed/torque line.
Max peak power = Stall-torque * no-load-speed / 4.
It's easier to test Stall-torque and no-load-speed.

[Cloud]

26 minutes ago, Chriull said:

In my charts in this thread (http://forum.electricunicycle.org/topic/5541-msuper-v3-top-speed-crash-wrecked-shell/?do=findComment&comment=64654) you see the power needed to overcome the air drag and the friction. If you add both powers (P Used) and divide it by the velocity (in m/s) you get the Force of the air drag and the friction. If you divide the force by the radius of the wheel (16 inch /2 in m) you get the needed torque to counter the force of air drag and friction. Around this torque you will have something plus minus needed for balancing the rider.

The torque/speed curve is the absolute maximum - the torque can be at this curve or near, but never above.

Exactly how i understood it. This "machine" and how this test was performed is only aimed at determining the wheel'max power and nothing else. Putting the riders weight is/was absolutely useless - the load is solely determined by the inertia of the barrel.

You speak russian if i remember right? Could you maybe ask them to also test the braking abilities in their test?

If you look at their charts the curve at the rightmost end crosses the x-axis and goes in the negative region and back to lower speeds - that shows the torque/power with which the wheel "brakes" the barrel. If they "pitch" the wheel backwards after the max power was reached they could test the max breaking power of the wheel, too. I am still very very conservative while braking with my wheel - i always remember the burned mosfets reported here after "emergency braking". Would be interesting to see the abilities of the tested wheels! 

And you should warn them, that it could easily be, that while performing this test they'll burn some more fuses and motherboards... (as they already did with their max power test...)

Yes, youare right, the torque wont go above it due to the additional loading. my bad

Yes, i can talk to them about testing the braking. i also burned the mosfets braking sharp a while ago.

[Chriull]

9 minutes ago, zlymex said:

There is an easier way to test the max peak power.
Two points determine a line. Stall torque at zero speed and no-load speed determine the speed/torque line.
Max peak power = Stall-torque * no-load-speed / 4.
It's easier to test Stall-torque and no-load-speed.

Imho thats not possible. These actual high power wheel limit the current/power at low speeds to prevent the mosfets from dying. So they never reach the stall torque of the motor.

On the other end of the chart (at least some of them) seem to never reach the no load speed. Someone here (imho @KingSong69) reported from one wheel (?ACM16 or Msuper V3?) that the lift-cut-off speed stays constant for full battery and battery with low charge - and this cannot be true for the real no load speed. So imho this wheel at least cut of by a fixed speed programmed in the firmware...

[Cloud]

7 minutes ago, zlymex said:

There is an easier way to test the max peak power.
Two points determine a line. Stall torque at zero speed and no-load speed determine the speed/torque line.
Max peak power = Stall-torque * no-load-speed / 4.
It's easier to test Stall-torque and no-load-speed.

Yes this will be a very good approximation.

But  torque/ speed curve is not a perfect straigh line. And its always good to test in conditions close to real what kind of max power the wheel can reach while in motion.

 

i think they have to either reduce the weight of the barrel or install a motor to assist the barrel , so that only additional torque equivalent to the torque required to finght the inertia of the riders weigh will be needed to be applied to the barrel.

[Cloud]

5 minutes ago, Chriull said:

Imho thats not possible. These actual high power wheel limit the current/power at low speeds to prevent the mosfets from dying. So they never reach the stall torque of the motor.

On the other end of the chart (at least some of them) seem to never reach the no load speed. Someone here (imho @KingSong69) reported from one wheel (?ACM16 or Msuper V3?) that the lift-cut-off speed stays constant for full battery and battery with low charge - and this cannot be true for the real no load speed. So imho this wheel at least cut of by a fixed speed programmed in the firmware...

Why cant they gradually ( but still quickly) bring the wheel into the stalled condition ( by applying additional loading) until the wheel/ magnets start slipping but the fuse wont burn yet as it requis a certain time to burn? Alternatively why cant they bypass the fuse for the purpose of the test?