Preview of version 2 of the hub motor / ebike simulator

[amberwolf]

At a glance to the results, it appears to be somewhat close to what I see on CrazyBike2 on typical cruising setups...but the Wh/mile is pretty far off, because my rides include LOTS of stops and starts for traffic controls. I am not sure about the watts used at that speed, because I keep forgetting to set the CA for watts and MPH on the same screen (I just went and fixed that), but the Amps seems about right from my admittedly poor memory.


One thing I wish was possible with simulators like this was to include those, so one could pick how many complete stops and starts are typically encountered on a typical route, and have that inefficiency included in the final simulation. It'd give a MUCH better estimate of power usage and range than a "simple" :lol: single-speed simulator does, for most folks, as I am sure most of us are in cities with a fair bit of traffic on our roads, and lots of complete stops necessitated by red lights or stop signs (or other traffic conditions).


For CrazyBike2, for instance, it estimates 13.2Wh/mile at about 68% throttle, which equates to cruising at 19.8MPH (close enough to what I keep it at). But in reality I get about half that efficiency (twice the Wh/mile), or worse, depending on the number of stops I must make.

 

[ptd]

I would sure like to see a few other windings of the 9c in there though. Since I'm so stoked on the slow windings for dirt riding.

+1... the 9C 2808, 2810, and 2812 would be fantastic. and thinking non hubs could be represented on an x axis being rpm instead of velocity.

and if more is better, and in this case it is (two comparisons on the same chart, which is AWESOME, for fiddling with power settings of the same motor, damn), a third would really help. imagine you have two points, and want to extrapolate data. all you really have is straight line formula. add one more point, and you get a MUCH better idea of what's going on.

just a preference here, but up and down toggles (buttons) might be nice instead of sliders (or maybe both). like +/- 1 and +/- 10 for throttle, and +/- .5 for grade.

on range, is that assuming 100% dod ? might be nice to have a setting for that too. and it might be nice to have a background color on those lines of data, instead of clear.

and sorry to repeat myself, but are you ever planning on outputing in spreadsheet form ?

lastly, the bike selection doesn't seem to change load performance, is there a missing link ? (and still wondering if there's any thought for another "enter preset values" kind of thing there).

 

[auraslip]

I would sure like to see a few other windings of the 9c in there though. Since I'm so stoked on the slow windings for dirt riding.

Third-ed

Question: When you put the throttle at a low setting, the efficiency stays pretty much the same. Like, the motor is turning slower, but the efficiency curve slows down with it. If that's correct, it makes me wonder the point of a slow wind motor: Since you can just decrease the throttle, and the it is suddenly efficient at slow speeds. I understand that this is more stressful on the controller, but it looks as if it doesn't make difference to the motor WHAT speed it's turning. I know this is because it's dependent on voltage and that the voltage starts off low and rises to pack voltage as you get to speed, but It just goes against the common advice here that running a fast motor slowly will necessarily be inefficient. Sorry if this doesn't make sense - I'm a bit out of it right now.

 

[amberwolf]

I finally remembered to change the main screen to Watts instead of Amps on the lower left of teh CA, and tested this on today's non-work trip: it was a bit windy, headwinds mostly, on the outbound part, and maintaining around 19.5-20MPH took 400-500W.

On the homebound part, there was nearly no wind, so it only took about half that for the same speed, which shows that the new simulator is pretty accurate.

Only exception appears to be the Wh/mile figure--but that may well be off because of the many stops/starts I have to make, which are very inefficient and eat up probably as much power as the rest of the trip does.


Oh, and before I forget, AGAIN, it'd be nice to add another feature: the ability to customize the starting and/or ambient temperature the simulation runs at, so the "overheats at:" function has more meaning in really hot climates like Phoenix, Las Cruces, etc., where the ability to shed the heat is a bit less than it is in colder climates. I'm not sure how much difference it'd make, but would be interesting to see. I don't plan on empirical testing of the result, either way. :lol:

 

[justin_le]

Is a 2wd ( 2hubs) catagory possible?

Not as such. But it isn't too hard in that case to double all of the motor related and energy consumption figures since they are shown numerically, while the load line stays the same. If you want to find out in that case what your cruising speed etc. will be, then just move the cursor line to the right until the value for Load Power in watts is exactly twice the Mtr Power. Then the values for Battery Amps, Wh/km and such will need to be doubled, while the range figure will be cut in half. The motor efficiency will be correct for each hub.

Justin

 

[justin_le]

Hello Justin,
I thought maybe my own Mad Max data could help to get a more and more acurate value on wider dynamic range of speed :wink:

OK Doctorbass, just bear in mind that I've only dyno tested the hub motors up to about 500 RPM and ~1kW power levels. So no guarantee about how things would extrapolate to the kind of abusive inputs that you subject these hubs. At very high RPM's the eddie current losses in the laminations could grow significant, and at the high currents we know already that saturation effects come into play. With a 53XX I'd be skeptical of anything over 90-100 Nm.

My measured data with cycle analyst and as well calibrated shunt and CA value the best as i can give me an electrical power consumption of 7500Watts at 100kmh continuous.

Well, that matches pretty well with what the simulator shows, which is 92.3V at 76.9A = 7100 Watts input power

What i can not understand is why the new simulator load curve does realy not correspond to the actual measured value i have at 100km/h

The simulator value at 100kmh say 11400 watts and my own measured value are 7500 watts. That's 35% error.

I'll be changing the CdA value for the mountain bike option from 0.8 to 0.6 m^2, so the default load curve will now a fair bit less power. Naturally the actual load depends quite a bit on the aerodynamic shape of your bike and your position while riding it, so once we have it so you can manually input custom CdA and Crr values then you can make a more exact match. Chances are that the Race/tuck option is close to what you'd have with your system.

Are the value of the black curve the power at a stabilized speed or while accelerating?é.. Because according to the performances value, i guess it's during accelreration beause at 20kmh at wot the Wh value are 393 Wh per km

Obviously if the motor output power is showing higher than the load power, then the vehicle is accelerating.

Also, why i can't enter 100A in the custom the current value and why tyhe graph dissapear if i enter 100A but dont at 90A?
Also the watt and efficiency scale dissapear at those 90 and 100A current.. is it fixable?
Doc

I don't seem to have any problems like that. Can you tell what browser version you are using?

 

[justin_le]

One thing I wish was possible with simulators like this was to include those, so one could pick how many complete stops and starts are typically encountered on a typical route, and have that inefficiency included in the final simulation.

Hmm, what we had made first was a motor simulator, what we are discussing is a full ebike simulator, and what you are asking for would be a trip simulator! So different objectives. But I'm surprised that you are seeing such a disparity between the cruising speed Wh/km and what your trip result is in practice. If anything I was finding that the predicted Wh/km values were higher than most of my own experiences for similar setups. I would say that your crazybike probably has a drag coefficient more inline with a mtb than a semi recumbent so that could be one factor that is compounding things a bit.

You can easily scroll the cursor down and see the Wh/km during various stages of accelleration. As well, when you hold your 68% throttle or whatever, is that holding it very steady in partial throttle or do you engage the system on and off? It's generally quite a bit less efficient with an electric motor to run full power / no power / full power / no power than it is to hold a steady partial power.

Justin

 

[amberwolf]

Hmm, what we had made first was a motor simulator, what we are discussing is a full ebike simulator, and what you are asking for would be a trip simulator!

That's me, never satisfied. :lol: On to version 4, then, right?

But I'm surprised that you are seeing such a disparity between the cruising speed Wh/km and what your trip result is in practice. If anything I was finding that the predicted Wh/km values were higher than most of my own experiences for similar setups. I would say that your crazybike probably has a drag coefficient more inline with a mtb than a semi recumbent so that could be one factor that is compounding things a bit.

Could be, especially with those boxes on the back end. I never have done a real coastdown test to see about that; I should.

You can easily scroll the cursor down and see the Wh/km during various stages of accelleration.

I know; I guess I am wishing for more than is practical. I do that a lot; rarely is it possible to do all of what I'd like to have.

As well, when you hold your 68% throttle or whatever, is that holding it very steady in partial throttle or do you engage the system on and off?

I hold it steady during cruising at that level. I'm not sure if it's actually 68% throttle on teh bike itself, but using that number on the sim gets the right readings.

When accelerating, I try to hold the throttle just above what would be cruising at that speed, to gradually speed up, rather than full-throttle. Only exception is when I am in front of a car at a traffic light, where I need to get across and out of their way as quickly, so I ramp it up a lot faster than normal, and end up at 100% throttle in a couple of seconds, to try to push it across faster (it makes only a little difference, but it helps).

It's generally quite a bit less efficient with an electric motor to run full power / no power / full power / no power than it is to hold a steady partial power.

I found that out via experimentation over time; if I pulse and glide I typically wind up with a few Wh/mile more than if I hold it steady.

 

[Lemlux]

Justin: Thanks so much for the updated simulator.

I'm a heavy 260# and was very interested in simulating overheating scenarios on hills. I was surprised to notice how much more hill tolerant the BMC V2t was when run at 20A and 66V than various HT3525 voltage / current options were. I also was impressed with how much more heat tolerant smaller wheel setups were than larger.

The new simulator has motivated me to strongly consider putting a 24" V2t on the back of my 29er even though I'd have to run my older 2004 frame without rear brakes.

 

[justin_le]

Justin: Thanks so much for the updated simulator.

I'm a heavy 260# and was very interested in simulating overheating scenarios on hills. I was surprised to notice how much more hill tolerant the BMC V2t was when run at 20A and 66V than various HT3525 voltage / current options were.

I wouldn't read TOO much into this at this stage, since you are comparing a geared system vs. a direct drive motor and their ability to withstand internal heating may not be all that comparable. It's true that the BMC windings have about half the resistance as a similar V/RPM 205mm direct drive motor. But in the geared BMC hubs the rotor magnet ring is not coupled to the motor shell and exposed to ambient like it is in a DD setup. So in a scenario with convection cooling on the motor shell like you'd have on an actual ebike, the DD motors can almost for sure tolerate more heat generation in the stator than a similar sized geared hub. Plus the nylon / composite gears are not as heat tolerant as the copper windings and Neo magnets.

I also was impressed with how much more heat tolerant smaller wheel setups were than larger.

This will always be the case, as a hub in a small wheel diameter can generate the same vehicle thrust with less phase current and motor heating than it can in a larger wheel.

-Justin

 

[justin_le]

The plan is to have a base option at the bottom of each dropdown menu where you can select "custom", and then via a pop-up window enter the specific model parameters. The way it's done now (forcing all fields to require custom parameters) is rather clutzy.

Done. Mike's got most of the final details we wanted implemented, which you can now preview from here:

http://www.ebikes.ca/simulator3-4/

In a nutshell:

• Individual "custom" field for each item at the bottom of the drop-down list
• Addition of a custom CdA and Cr field in the vehicle type drop down (though they seem swapped at the moment!)
• Ability to show/hide graphs from the plot (circles on the legend dual purpose as radio buttons)
• Reorganization of numeric data under the graph grouped into into 3 columns rather than one long row
• Addition of battery power and vehicle acceleration fields
• Graph options organized into drop-down tables rather than radio buttons for consistency

There are a few cosmetic things to tidy up but that's almost all the base functionality done. Oh, and still working on the best way to include a human power element to the simulator.

Judging from the way that people were using and sometimes misinterpreting the vertical cursor bar, I'm wondering if it could be worth having an option where the throttle is automatically adjusted to make the load power / motor power crossover point happen at the cursor position. Say perhaps if you right click and move the cursor it would attempt that, while left clicking would only move the cursor while leaving the throttle value the same?

Thinking outloud, Justin

 

[Lemlux]

Justin:

Appreciate your detailed comments re DD motor stator convection heat dissipation ability being greater than that of geared motors. Do you intend to tweak any simulator parameters that would better enable apples to apples heat dissipation comparability?

 

[ptd]

still hoping for the extra 9C variants and a system C. maybe something of a motor's parameter's? strands, turns, wire diameter, stator diameter, magnet strength, and on and on, lol. could become quite the developers toolkit though, from start to finish. you might have to sell it. i got 5 on it.

if you're not going to do a spreadsheet version, cookies might be a way to keep track of settings. although the spreadsheet would be ideal, for something like undo, vs, recalculate, or ANYTHING, vs recalculate, lol. automatic refresh is a nice thing.

bike selection, fixed, but i might lose the weight and put it in with the custom bike parameters. along with splitting cDa into it's two components, frontal area and Cd. and Cr seems off, maybe it's a matter of units. just used to .00x values. might be nice if the four values were displayed, as well. maybe have the bike selection as the start point. then able to tweak the parameters individually.

love the acceleration values. awesome. a graph for seconds to speed, would be incredible. you're basically there, aren't you? just calculate the formula for acceleration and take the integral, right? and just treat the before and after peak power as two functions? and maybe have the y maxima correspond to the power/load intercept, and only label the y axis up to that point. and then you're just one step away from calculating wh used for a leg. and then, and then, lol. i like the thinking out loud thing too.

love how you added the individual volts and amps. could always add graphs of those.

throttle and grade sliders still quite twitchy. speaking of throttle, how difficult would it be to do something like progressive throttle?

i still can't see system B efficiency. i'm sure it's just because my background color is the same. if those had white backgrounds, might help.

lastly, you might add programmable controller features, like amp limiting and soft start, parameters. ok, maybe in the toolkit.

all in all, awesome job. thank you. seriously, paul

 

[Starson]

Weird thing with my I.E. Browser. If I use the new simulator at 100% (normal zoom level) the Miles per hr is distorted in the resulting graph. But if I change to 92% zoom, it seems to come up accurate in the browser. Took me awhile to figure out what was happening. Anyone else experiencing this?

Gary

 

[kev82]

I'd love to see lipos in the battery list.

 

[ptd]

I'd love to see lipos in the battery list.

good point, but with the proliferation of lipos, maybe it would be easier to just allow a C rate in the battery specs.

 

[auraslip]

I love you.

udging from the way that people were using and sometimes misinterpreting the vertical cursor bar, I'm wondering if it could be worth having an option where the throttle is automatically adjusted to make the load power / motor power crossover point happen at the cursor position. Say perhaps if you right click and move the cursor it would attempt that, while left clicking would only move the cursor while leaving the throttle value the same?

Great idea. Just a little tick box to turn it on and off would work and it'd be obvious.

Edit: Make it like the women you want to marry; beautiful and easy to have a good time with, but deep as the pacific.

 

[nicobie]

This thing is sucking up way too much of my time.

Many thanks Justin :2 thumbs up:

 

[Toorbough ULL-Zeveigh]

Ja, das ist gut!

 

[amberwolf]

Regarding the battery selection (actually all of the selections), it'd be nice if first selecting any non-custom item from the dropdown would then "populate" the custom selection with it's default values, so the user would have a place to start if they don't actually know for sure what theirs is.

For instance, my LiFePO4 pack is 16s, not 15s, but since I have no idea what the internal resistance is for it, I just picked the 15s LiFE on the listing.

Also, the battery listings don't all show an Ah rating, but the custom dropdown has one (that defaults to 8Ah), so I don't know if that makes a difference in the simulation itself. I presume there *is* an Ah number used in all the calculations, regardless of battery type, but that it simply isn't given for those batteries in the dropdown.


FWIW, I tried the Mountain Bike option instead of semi-recumbent, and it still appears drastically optimistic for Wh/mile consumption, about 20Wh/mile, rather than what I get in reality. I guess I need to do the coastdown test.

 

[ptd]

the link for version 3/4 now throws windows into some kind of infinite calculation loop. version 2 still works fine.

EDIT... fixed, thank you

 

[lizardboy]

Hi I've been using the simulator and as a guy who rides up 20% slopes 4 or 5 times a day I can say that something is way off once you start adding hills to the simulator.

I ride a mac standard wind with 36v lipoly pack and I get almost exactly twice the speed uphill compared to the simulator at the same current and voltage readings it seems to get more inaccurate the more slope is added.

The settings I'm using are; 30 amps battery current 38v for battery voltage and 100kg for vehicle weight up a 20% slope (I ride more than that but it only goes up to 20)

Thanks for the simulator good luck with the final details!

 

[miuan]

I'm afraid it may have something to do with the method of % calculation. Some take 45 degrees as 100%, some 50%.

 

[justin_le]

I'm afraid it may have something to do with the method of % calculation. Some take 45 degrees as 100%, some 50%.

% grade is defined as the tangent of the angle, so indeed a 45 degree angle is 100% grade and this is how it is done in the simulator.

Lizardboy, are you sure that the grades you climb up 4-5 times a day are indeed 20%? You can go out with a level and ruler to measure it easily enough. Generally what most people call a really really steep hill is like a 10-12% grade. 20% is very uncommon except for short steep sections on private driveways or whatever.

Justin

 

[liveforphysics]

I'm afraid it may have something to do with the method of % calculation. Some take 45 degrees as 100%, some 50%.

45deg IS 100% slope. 100% slope is when vertical gain matches horizontal movement.