Hub motor steady acceleration using Phaserunner?

thunderstorm80 · May 25, 2017

Hi,
I love the Phaserunner so much not only because it's small, powerful and fully programmable - It (finally) maps your throttle signal to phase current (and therefore your direct torque output).
Now let's imagine a take-off from stand still scenario, on a 10%+ uphill, using motor power alone. Let's assume we neglect the effects of wind resistance as I am asking about the takeoff performance where the wind force is still negligible.
According to kinematics&dynamics, if I fully throttle and get the maximum programmed maximum torque, then as long I hold it that way - I will have the same acceleration, right? From takeoff, and until wind resistance would start taking it's toll. I also assume we have a battery voltage high enough so we are not back-EMF limited, battery current limited, or any other electronic limitation.
I ask all this because I have the feeling that on such standstill takeoff - the motor first "bog down", and once you are in the 15km/h zone - you can feel a dramatic and much better acceleration, but this doesn't make sense at all!
I think it's because we are so used to driving cars where you use the mechanical gear reduction (or overdrive) all the time, same as we use gear shifting when pedalling:
We are so used that upon startup we get a greater torque output (and acceleration) because of those gear-benefits (and as we up-shift the maximum possible acceleration and torque go down), that when we use a constant DD machine - it feels it "bogs down", while in fact - in such machine the maximum acceleration is always the same across the speed spectrum. (neglecting the effects of wind resistance). Am I correct about this?
This is bugging me in the recent days..

Roy

MarkSea · Jun 3, 2017

One hypothesis, if I'm reading your question right: If you look at the Efficiency curve at http://www.ebikes.ca/tools/simulator.html, you will see that at low speeds, you have very low efficiency, so even though you are at 100% throttle, and generating a lot of torque, the efficiency is low, which presumably means you are feeling "bogged down". With the 25A controller, as you hit 15kph you pass 60% efficiency peaking at over 80% efficiency.

I'm not sure how the phaserunner effects that, as it is not an option on the simulator.

thunderstorm80 · Jun 6, 2017

MarkSea said:
One hypothesis, if I'm reading your question right: If you look at the Efficiency curve at http://www.ebikes.ca/tools/simulator.html, you will see that at low speeds, you have very low efficiency, so even though you are at 100% throttle, and generating a lot of torque, the efficiency is low, which presumably means you are feeling "bogged down". With the 25A controller, as you hit 15kph you pass 60% efficiency peaking at over 80% efficiency.

I'm not sure how the phaserunner effects that, as it is not an option on the simulator.

Hey Mark,
No, that's not how the Phaserunner works at all.
The performance curve vs eff is the same yes, but this is being expressed in higher battery current input.
The Phaserunner maintains a maximum phase current through the speed spectrum (until you reach the back-EMF limit speed), and this is why it's such a brilliant product!

Alan B · Jun 6, 2017

Hi folks.

The simulator is a wonderful tool, however it is easy to misinterpret what it is showing.

From Google:
Efficiency: the ratio of the useful work performed by a machine or in a process to the total energy expended or heat taken in.

Efficiency only has the usual meaning at ONE POINT on that sim graph, the point of equilibrium, where motor power equals the load power.

The simulator is a graph of what would happen if the given system was placed at the chosen throttle level and allowed to accelerate until equilibrium and beyond (which it would not do, but this is just a calculation). At all the points to the left of the equilibrium point the system is accelerating, so motor power is producing acceleration in addition to doing work required by the loading. So the efficiency is "low" because more energy is being produced than is required to move the load at the given speed (this extra energy is going to acceleration). It is similar to looking at the miles per gallon display of a vehicle while accelerating, it may drop to a single digit value, but this is not a reflection of the average freeway MPG efficiency. You have to set the throttle and reach equilibrium at a speed to measure the efficiency at that speed. You cannot read it from the efficiency curve on the sim graph at more than the one equilibrium point.

The PhaseRunner is a very sophisticated controller, despite the small size. It has a number of programmable limits, and they include motor current (sometimes called phase current), as well as battery current. The power flow through the PhaseRunner will be limited when any of these parameters exceeds the limiting values. The 6 FET PhaseRunner will generally limit motor current at low speeds, and then at higher speeds the battery current limit will tend to kick in, and of course back EMF will also reduce the power transfer as speed rises at some point.

What I have discovered in my experiments and analysis is that for good power transfer with the PhaseRunner the "torque wind" motors and higher voltages work better (see the Bonanza linked below) because the "speed wind" motors require so much current at low voltage/speed that the modest current capacity becomes the power transfer limit. It also prefers geared motors because the higher RPM allows the motor to produce good power at lower torque and current. In trying to match the torque from the rear DD motor with the front geared motor it is difficult to find a rear DD motor that develops similar torque before reaching the motor current limits of the PhaseRunner.

Another characteristic of the simulator is that it does not model the motor current limiting feature of any controller. So the acceleration at the low end that it shows only matches a theoretical controller that only limits on battery current. Most controllers don't measure motor current, but they have some way of estimating it based on the information they have combined with the battery current, which they do measure. And they do limit motor current based on their estimate (which is often inaccurate). So the low end of the acceleration curve of a real system may be reduced compared to the simulator's graph for most real controllers.

The controller is a power converter, which converts battery voltage to motor voltage (a down conversion). In the process it increases the current, since power output equals power input (minus small losses). To drive current into the motor the output voltage first must equal the back EMF, then it must increase above that to force current into the resistance of the FETs, wiring and motor. So at low speeds this voltage is low, and the current gets "mulitplied" significantly by the conversion. As speed rises the back EMF rises and the current multiplication drops. So the torque would drop as well since it is proportional to motor current. Of course any limiting taking place may reduce the power transferred by reducing the voltage and current output of the controller.

Bottom line is - the interactions of all the factors make something that is "simple" into something that is not so easy to understand.

thunderstorm80 · Jun 11, 2017

Alan B said:
Hi folks.

The simulator is a wonderful tool, however it is easy to misinterpret what it is showing.

From Google:
Efficiency: the ratio of the useful work performed by a machine or in a process to the total energy expended or heat taken in.

Efficiency only has the usual meaning at ONE POINT on that sim graph, the point of equilibrium, where motor power equals the load power.

The simulator is a graph of what would happen if the given system was placed at the chosen throttle level and allowed to accelerate until equilibrium and beyond (which it would not do, but this is just a calculation). At all the points to the left of the equilibrium point the system is accelerating, so motor power is producing acceleration in addition to doing work required by the loading. So the efficiency is "low" because more energy is being produced than is required to move the load at the given speed (this extra energy is going to acceleration). It is similar to looking at the miles per gallon display of a vehicle while accelerating, it may drop to a single digit value, but this is not a reflection of the average freeway MPG efficiency. You have to set the throttle and reach equilibrium at a speed to measure the efficiency at that speed. You cannot read it from the efficiency curve on the sim graph at more than the one equilibrium point.

The PhaseRunner is a very sophisticated controller, despite the small size. It has a number of programmable limits, and they include motor current (sometimes called phase current), as well as battery current. The power flow through the PhaseRunner will be limited when any of these parameters exceeds the limiting values. The 6 FET PhaseRunner will generally limit motor current at low speeds, and then at higher speeds the battery current limit will tend to kick in, and of course back EMF will also reduce the power transfer as speed rises at some point.

What I have discovered in my experiments and analysis is that for good power transfer with the PhaseRunner the "torque wind" motors and higher voltages work better (see the Bonanza linked below) because the "speed wind" motors require so much current at low voltage/speed that the modest current capacity becomes the power transfer limit. It also prefers geared motors because the higher RPM allows the motor to produce good power at lower torque and current. In trying to match the torque from the rear DD motor with the front geared motor it is difficult to find a rear DD motor that develops similar torque before reaching the motor current limits of the PhaseRunner.

Another characteristic of the simulator is that it does not model the motor current limiting feature of any controller. So the acceleration at the low end that it shows only matches a theoretical controller that only limits on battery current. Most controllers don't measure motor current, but they have some way of estimating it based on the information they have combined with the battery current, which they do measure. And they do limit motor current based on their estimate (which is often inaccurate). So the low end of the acceleration curve of a real system may be reduced compared to the simulator's graph for most real controllers.

The controller is a power converter, which converts battery voltage to motor voltage (a down conversion). In the process it increases the current, since power output equals power input (minus small losses). To drive current into the motor the output voltage first must equal the back EMF, then it must increase above that to force current into the resistance of the FETs, wiring and motor. So at low speeds this voltage is low, and the current gets "mulitplied" significantly by the conversion. As speed rises the back EMF rises and the current multiplication drops. So the torque would drop as well since it is proportional to motor current. Of course any limiting taking place may reduce the power transferred by reducing the voltage and current output of the controller.

Bottom line is - the interactions of all the factors make something that is "simple" into something that is not so easy to understand.

I only use the Trip Simulator on Grin's website to simulate the results with Phaserunners. There you can directly control your phase current limits, and directly choose your working point (speed, grade, etc...). Its a great app!

d8veh · Jun 11, 2017

MarkSea said:
One hypothesis, if I'm reading your question right: If you look at the Efficiency curve at http://www.ebikes.ca/tools/simulator.html, you will see that at low speeds, you have very low efficiency, so even though you are at 100% throttle, and generating a lot of torque, the efficiency is low, which presumably means you are feeling "bogged down". With the 25A controller, as you hit 15kph you pass 60% efficiency peaking at over 80% efficiency.

That's it.

justin_le · Jun 12, 2017

Alan B said:
Hi folks.
The simulator is a wonderful tool, however it is easy to misinterpret what it is showing.

Thanks Alan for the very clear and correct explanation on the motor simulator's function, limitations, and common misinterpretations. We get people all the time seeing like 40-50% efficiency numbers at low speeds and assuming that the motor efficiency is terrible at low speeds, without realizing that this is at way higher power/torque levels than needed to move the bike at said speed. When you correctly back off on the throttle % slider so that the actual steady state cruising speed is this low speed, then the efficiency is right up in the ~80% territory. I see soo much misinformed conclusions being made here on ES from simulator curves and I appreciate attempts at correcting the record.

Anyways, onto the OP's question:

thunderstorm80 said:
I ask all this because I have the feeling that on such standstill takeoff - the motor first "bog down", and once you are in the 15km/h zone - you can feel a dramatic and much better acceleration, but this doesn't make sense at all!

You say you have a "feeling" to this effect, but do you have a data log of your acceleration? If you have a CA on the bike and can log the serial data stream and then plot it, I'm quite sure that what you will find is a very nice an linear velocity increase up to, as you say, the point where wind drag starts to have an effect, and then the acceleration will decrease. But from a perception point of view, it will feel pretty anemic at the low speeds region of take-off simply because the human brain knows intuitively that it there is substantially more power required to accelerate from 15 to 18 kph than from 5 to 8 kph. Anyone can accelerate from 5 to 8 kph in a few seconds just by pedaling, but to accelerate from 25 to 28 kph or say 35 to 38 kph in the same time frame takes way more power output, and so when a motor does that we really appreciate it much more than when it does the 5-8kph acceleration.

That's why many people would say that a power throttle rather than a torque throttle is the most natural at feeling the same over a wide speed range. Half throttle means a given number of watts into the motor, both at low speeds and at high speed, and that feels more consistent than half throttle giving substantially more power into the motor when moving fast than when moving slow as happens with a direct torque throttle.

d8veh said:
If you look at the Efficiency curve at http://www.ebikes.ca/tools/simulator.html, you will see that at low speeds, you have very low efficiency, so even though you are at 100% throttle, and generating a lot of torque, the efficiency is low, which presumably means you are feeling "bogged down". With the 25A controller, as you hit 15kph you pass 60% efficiency peaking at over 80% efficiency.
That's it.

In this context this is completely incorrect. The motor efficiency is a non-factor in the equation as the controller is running a direct phase current control and the torque->phase current relationship is almost perfectly linear. So the motor will be putting out the same X number of newton-meters just the same regardless of whether it is 40% efficient or 90% efficient.

thunderstorm80 · Jun 12, 2017

justin_le said:
Alan B said:

Hi folks.
The simulator is a wonderful tool, however it is easy to misinterpret what it is showing.

Click to expand...

Thanks Alan for the very clear and correct explanation on the motor simulator's function, limitations, and common misinterpretations. We get people all the time seeing like 40-50% efficiency numbers at low speeds and assuming that the motor efficiency is terrible at low speeds, without realizing that this is at way higher power/torque levels than needed to move the bike at said speed. When you correctly back off on the throttle % slider so that the actual steady state cruising speed is this low speed, then the efficiency is right up in the ~80% territory. I see soo much misinformed conclusions being made here on ES from simulator curves and I appreciate attempts at correcting the record.

That's what I say as well!

justin_le said:
Alan B said:

Anyways, onto the OP's question:

thunderstorm80 said:

I ask all this because I have the feeling that on such standstill takeoff - the motor first "bog down", and once you are in the 15km/h zone - you can feel a dramatic and much better acceleration, but this doesn't make sense at all!

Click to expand...

You say you have a "feeling" to this effect, but do you have a data log of your acceleration? If you have a CA on the bike and can log the serial data stream and then plot it, I'm quite sure that what you will find is a very nice an linear velocity increase up to, as you say, the point where wind drag starts to have an effect, and then the acceleration will decrease. But from a perception point of view, it will feel pretty anemic at the low speeds region of take-off simply because the human brain knows intuitively that it there is substantially more power required to accelerate from 15 to 18 kph than from 5 to 8 kph.

Click to expand...

This is because of the much higher acceleration you have in your lower mechanical gearing when you start to accelerate, giving you advantage: I am almost sure that this "bogging" impression is because we are used to drive variable-gearing vehicles, notably private cars, buses, and even our pedalling drivetrain on our bicycles: Over there we actually DO get a much higher acceleration on the lower gears due to the proportional higher torque, and our brain learn to perceive it that way. If we all just used fixed gear bicycles and used only trains, then I have the impression we wouldn't be "conditioned" to perceive it that way. This is in addition, as you said, to the higher and higher delta power demand in order to increase your speed, while the Phaserunner just pumps through an increasing amount of power as you accelerate so you would still have a constant acceleration.
I have the Analogger and it would be interesting to log the steady acceleration graph with the Phaserunner and post it here, and perhaps it can be used as a tool to show the difference between a torque mapped controller like the Phaserunner and a standard PWM controller like the Grinfineon.

justin_le said:
Anyone can accelerate from 5 to 8 kph in a few seconds just by pedaling, but to accelerate from 25 to 28 kph or say 35 to 38 kph in the same time frame takes way more power output, and so when a motor does that we really appreciate it much more than when it does the 5-8kph acceleration.

If we assume a fixed gear pedalling bike, and we assume you are comfortable pedalling at 30kph, then you will have a MUCH harder time accelerating through the 5-8kph due to the narrow effective RPM bandwidth of the human leg, and therefore you will accelerate way slower in that zone.
If we assume a variable gear pedalling bike, then you will accelerate those 5-8kph in few seconds because of the gear advantage, giving you a much higher torque. The higher speeds zones would be accelerated slower, assuming you apply a constant leg force and steady RPM.
This is from forces&torques perspective. If we discuss from energy point of view, then yes - Due to the kinetic energy equal 0.5*m*V^2, then increasing your speed will require a higher and higher delta power with each pedal stroke.

justin_le said:
That's why many people would say that a power throttle rather than a torque throttle is the most natural at feeling the same over a wide speed range. Half throttle means a given number of watts into the motor, both at low speeds and at high speed, and that feels more consistent than half throttle giving substantially more power into the motor when moving fast than when moving slow as happens with a direct torque throttle.

I never thought of it, and it makes sense! This come in addition to the mechanical gearing advantage I was talking earlier, since the human leg is also limited by the continuous power it can produce.
Interesting discussion!

justin_le · Jun 13, 2017

thunderstorm80 said:
If we all just used fixed gear bicycles and used only trains, then I have the impression we wouldn't be "conditioned" to perceive it that way

That may be true. If only ebike technology had matured in the late Victorian era!

Dumsterdave · Aug 23, 2018

I've been running an infineon CAv3 combo from em3ev and it's been loud, but runs great. I recently bought a phaserunner and swapped out the controller, went through the setup and now my bike accelerates painfully slow (but amazingly quiet). I'm only running 52v 40A on a 1500w leaf motor. What should my phaserunner settings be at to add a bit more torque? Or is it the CA that needs adjustment. If I lift the wheel of the ground and apply throttle, it accelerates very fast, but if I put it on the ground... Slow

I'm not doing regency (yet), have a grin slim half twist and running the latest firmware on the CA

fussler · Nov 15, 2019

I have the exact same problem: leafbike 1500w and phaserunner be extremely slow acceleration and no torque.

justin_le · Dec 18, 2019

fussler said:
I have the exact same problem: leafbike 1500w and phaserunner be extremely slow acceleration and no torque.

You want to make sure that you have the maximum phase current set to 96 amps, and that you also have your throttle voltages mapped so that at full throttle the PR's output is at 100% as well. If you find that the acceleration is still slower than what you want, then it means you need a controller with high phase current capabilities. Typically that would mean a 12fet or 18 fet device.

lioni05 · Sep 25, 2020

I have been having the same problem with my Grin All axle motor and baserunner L10. I have tweaked some settings and have been able to get slightly better acceleration, but compared to what I have been used to, it feels lacking. I have a feeling that I am not understanding the relationship between the CA3 and the Baserunner and need to do more tweaking. Justin - if you see this, can I share my CA3 settings and Baserunner settings with you? I bought a ready to roll kit so I am hoping that you might have some suggestions.

justin_le · Sep 26, 2020

Hey Lioni, for sure feel free to post the CA3 and Baserunner settings here to have a look, but most likely what you are experiencing is just the actual behavior of a system that has a phase current limit. Right off the line the baserunner_L10 will do 80 amps (assuming you have 80A max phase setting), but this will be more like 50 amps when the controller is hot. A lot of the more generic ebike controllers don't have any phase limits in them at all and so people get used to having like 100+ amps of phase current off the line, even if it's just like a 30A motor controller.

If you hook up the baserunner connected to a computer and on the dashboard tab, then apply full throttle on the CA3. Does the throttle % setting go up to 100% while the throttle voltage you see on the dashboard go to about 3.55V? If so, and if you've set the max phase current on the basics tab to 80 amps, then you're already at the torque limit for this controller and motor pairing.

The 80A maximum set on the Baserunner is a bit conservative since it's the same mosfets as the phaserunner that does 96A, but it's being limited further more as a result of capacitor ripple current concerns than mosfet frying concerns.

Hub motor steady acceleration using Phaserunner?

thunderstorm80

1 kW

MarkSea

1 µW

thunderstorm80

1 kW

Alan B

100 GW

thunderstorm80

1 kW

d8veh

1 GW

justin_le

Administrator

thunderstorm80

1 kW

justin_le

Administrator

Dumsterdave

100 W

fussler

1 mW

justin_le

Administrator

lioni05

1 µW

justin_le

Administrator

Similar threads