Preview of version 2 of the hub motor / ebike simulator

[Lemlux]

This thread just motivated me to purchase what seems to be a very effective $29.95 digital angle measurement device that measures only 2" x 2" x 1 1/8" to measure the slopes on my normal riding paths.

http://www.harborfreight.com/digital-angle-gauge-95998.html

Unfortunately I have a 0.1 mile 20% slope on the road down to primary local shopping to the east of me and extended 12% 13% slopes on the 1.3mi main road down toward the ocean which lies 397' below to the West.

If I take circuitous longer alternative routes I can limit slopes to 6% to 8% grades.

Unfortunately, the only relatively flat surfaces around here are the heavily traveled Route 101 and frontage roads on the other side of RR tracks from 101.

Given my 260# weight my best single hub choice for avoiding overheating per the simulator would seem to be putting a BMC V2T in a 16" wheel BOB Trailer pusher. With a 20S LiFePo4 pack and a 20A controller top end on the flats simulates to be 20.9 mph which corresponds nicely with what ES people observe is the maximum pusher speed before trailer wobble becomes a problem.

Unfortunately, it appears that the only way to limit wheel spin on this configuration would be to slowly throttle up after pedalling to at least 5mph with minimal trailer tire pressure.

Various dual hub approaches with 24" wheels without a trailer could accomplish the same without limiting speed to 20.9 mph.

Accordingly, I am very interested in the developing discussions on simulator slope accuracy and passive heat management without drilling holes into hub covers.

 

[justin_le]

Find it at http://www.ebikes.ca/simulator

I've written a fairly comprehensive explanation of the features and details in the text under the graph.

Justin

 

[adrian_sm]

Very nice, but I can't seem to find the custom motor input.

- Adrian

 

[miuan]

Justin, the page layout in Firefox 6 is all over the place, please have a look at it

 

[Skippic]

This simulator is the best thing ever!

Justin could you add an option "Constant speed" x = [mph], y1 = [efficiency], y2 = [power] at constant speed?

 

[TylerDurden]

Justin, the page layout in Firefox 6 is all over the place, please have a look at it

Looks fine in FF 5.0...

 

[miuan]

This simulator is the best thing ever!
Justin could you add an option "Constant speed" x = [mph], y1 = [efficiency], y2 = [power] at constant speed?

Equally interesting would be to plot a relation of Output Watts vs. speed at a given efficiency figure.

 

[The Mighty Volt]

just had a look for myself, love it love it love it.

 

[Skippic]

It looks like you are using a different formula for the torque for each graph. Just try the same values in both systems, a high enough voltage and you will see different torque values.



If you really have a different function/method for each case I strongly advise using just one, it's simpler and any modifications you make will be directly applied to both cases.

 

[justin_le]

It looks like you are using a different formula for the torque for each graph. Just try the same values in both systems, a high enough voltage and you will see different torque values.

Intresting. It looks to me like there is a bug in the display graphs being updated rather than the equations. You'll notice that the numeric values showing the actual torque underneath are correct and the same for both systems. And in the legend it says "Torque A" and "Torque A" rather than A and B. Also, you are not using the correct simulator, the image you showed (version 3-2) is one as we were developing the improvements, there were several releases after that and the most current is just at http://www.ebikes.ca/simulator/ .

Also, anything over about 85 N-m of torque is bogus since the hubs start to saturate around then, and I can't vouch at all for how well it models when you put 4KW of power input into a motor designed and tested at only 1/5th of that.

-Justin

 

[Skippic]

http://www.ebikes.ca/simulator/ doesn't work well with Firefox 6.0. I also checked on the official version and it gave the same results. You are right, that these values are too high for the motor anyway. I just wanted to help debugging

 

[John in CR]

Hey Justin,
Auraslip brought up an interesting point in another thread. Since you include overheat times, does that take into consideration the increasing resistance of copper as temp increases?

 

[justin_le]

Hey Justin,
Auraslip brought up an interesting point in another thread. Since you include overheat times, does that take into consideration the increasing resistance of copper as temp increases?

I think that was answered pretty thoroughly in my 3rd post on the 1st page of this thread.

 

[justin_le]

http://www.ebikes.ca/simulator/ doesn't work well with Firefox 6.0. I also checked on the official version and it gave the same results. You are right, that these values are too high for the motor anyway. I just wanted to help debugging

OK, Michael will look into what's up with FF6 next week sometime.

I got so frustrated with all the newer firefox releases and their inability to disable tab scrolling that I downgraded everything back to FF3.6 here, so we didn't catch this!

 

[John in CR]

Hey Justin,
Auraslip brought up an interesting point in another thread. Since you include overheat times, does that take into consideration the increasing resistance of copper as temp increases?

I think that was answered pretty thoroughly in my 3rd post on the 1st page of this thread.

Sorry about posting without reading everything. My time is very short these days, but I have played around with the new simulator. With it I've been unable to recreate a situation of an overheating motor in tens of seconds, yet I have seen that effect first hand with a big heavy hub and a controller feeding it only 25A. ie Bog a motor down to very low pedal speeds on a steep hill, and you may very well have less than 30 seconds before you better be back on flat ground cooling that motor off with some rpms and low load. I wouldn't have a clue as to how make a simulator reasonably estimate that, but it's a repeatable result in the real world.

Let me know if there are some simple measurements I can gather to help in some small way. I'm one of the guys that bangs on your simulator quite regularly to help point out stuff those graphs help the less experienced grasp.

John

 

[justin_le]

ie Bog a motor down to very low pedal speeds on a steep hill, and you may very well have less than 30 seconds before you better be back on flat ground cooling that motor off with some rpms and low load. I wouldn't have a clue as to how make a simulator reasonably estimate that, but it's a repeatable result in the real world.

Yeah, this comes down to how comprehensive of a thermal model you make. If you dump the power in the motor REALLY fast, then it's just the heat capacity of the copper that comes into play to suck it up as there isn't enough time for the stator itself to absorb much of the energy, and effectively you have I^R going just into the copper alone, resulting in a very rapid temp buildup. The fact that the resistance of the copper nearly doubles by the time the motor starts to cook makes it worse.

If the heat goes in at a lower rate, then you get close to a thermal equilbrium between the copper and all the steel in the stator as well, adding a substantial amount of heat capacity, while the motor shell still isn't doing much since the air gap between stator and rotor is something of a thermal barrier.

At lower input rates still, even the thermal mass of the shell ends up absorbing an appreciable amount of energy.

In the first order model that is used, the entire hub hub is treated just as a single mass with a heat capacity with a single thermal resistance to ambient, using some kind of average parameters of the above 3 cases. Once I get a chance to do more empirical testing with multitudes of temp sensors in a spinning motor at low and high power inputs, then it will be possible to have enough data for a model that goes:

[Thermal Mass of Copper/Steel Stator] --->(thermal resistance to shell)--> [Heat Capacity of Motor Shell] -->(Thermal Resistance to Ambient) -->[Infinite Heat Capacity of Ambient]

Let me know if there are some simple measurements I can gather to help in some small way. I'm one of the guys that bangs on your simulator quite regularly to help point out stuff those graphs help the less experienced grasp.

OK, appreciate that. It's always good to hear first hand evidence that either corroborates or throws into question the modeling, so once we have something a bit more comprehensive on the thermal end it will be good to see how it matches up with your real world experience.

Justin

 

[alsmith]

Justin, a big thanks for the simulator- it gave me the first clue that my Crystalyte 5304 isn't a 5304- the performance seen didn't match reality. I tried the other models and found the 5303 was an almost perfect match. After some emails back and forwards with Crystalyte Europe they have accepted that is a 5303 after measuring coil resistances. Without the simulator I'd never have known.

 

[auraslip]

Justin -

We were talking in another thread about the efficiencies of dual motor setup.

Basically what it came down to was two motors of the same winds run at WOT on the flats will be ~5%, BUT it should actually be much more efficient on the motors end. One thing I found playing with your simulator is that controller losses make up a big portion of losses as power levels go down (as they would would two hubs running together and nearing no load speed)

Is there a way you could specify losses in the controller vs. in the motor? (I know I ask too much )

Also, one thing that was brought up was JUST HOW MUCH heat makes a difference in efficiency! Compared to room temp, copper is 130% more resistive at 100c and 150% more resistive at 150%! This is a huge difference in power, range, efficiency, and motor heat. I really don't envy who ever has to work out a way to model that!

Still though, it's a big deal and I think your clients would have a lot to gain from realizing how heat effects a motor. Perhaps an option to graph heat production in the motor? Perhaps a slider to adjust the temperature of the motor?

 

[Skippic]

It looks like I found another error in the simulator

Changing the motor in system A also modifies values in system B.

The differences are not huge, but seem more than just a rounding error.

 

[miuan]

Another suggestion for the simulator: introduce a setting of phase/battery current ratio with 2.5x default value. It should be quite easy to implement into the existing simulator.
As it stands now, the simulator will ramp up motor current without any limit, creating a sensation of higher than actual torque and power consumption at very low speeds and 100% throttle. It also appears the similarly powered differently wound motors perform similar at these speeds, while they don't.

 

[chroot]

Justin- Can you update the CroMotor and xlyte 40mm motor on ebike simulator please.

 

[justin_le]

Another suggestion for the simulator: introduce a setting of phase/battery current ratio with 2.5x default value. It should be quite easy to implement into the existing simulator.
As it stands now, the simulator will ramp up motor current without any limit, creating a sensation of higher than actual torque and power consumption at very low speeds and 100% throttle. It also appears the similarly powered differently wound motors perform similar at these speeds, while they don't.

Hey Miuan and thanks for the feedback! Indeed, phase current limit has been on the agenda, and should get included as one of the parameters you can put into the "custom controller" option box. The next round of upgrades will have this implemented.

Without a phase current limit, it is true that different wound motors do perform similar at low speeds with the same voltage and controller. But with a phase current limit this is no longer the case as you say.

-Justin

 

[auraslip]

I bet he would love to get his hands on a cromotor.

 

[rwright]

The simulator was the reason I just placed an order from ebike.ca for an eZee system with CA and other accessories. It was an extremely valuable tool for estimating what I needed.

A while back there was some discussion on adding the human power pedaling contribution. That was the biggest thing I found missing from the simulator as I am a non-electric bike rider just looking for an e-assist. Currently, I fudge the aerodynamics to change the load at a certain speed, but that is a bit unwieldy.

In my case, I am used to riding with PowerTap hubs and have a pretty good feel for my power output. For instance, if I could run the motor at 100W and add 100W from my legs, I would be quite happy over long distances. So, just being able to specify my pedaling contribution in a fixed number of Watts would be great.

Again, great job on the simulator to both Justin and Michael. This kind of effort is what places you above your competition.

Robert Wright
Oregon, USA

 

[psycholist]

Welcome to the forum rwright. To simulate the addition of 100w of leg power, just click on the load line at the point where it would intersect 100w higher.