💪 EUC pull force test results - reviewers please do this!

[alcatraz]

I'd love to see the RS HS and EXN HS added to this list. 

I feel like the HS is the continuation of the MSX motor. (speed increases linearly with voltage)

Does that mean that the HS versions have similar pull with the good ol' MSX?

[alcatraz]

Would anyone else like to see a 126V KS-16X?

[yoos]

11 hours ago, Denny Paul said:

Anecdotally, having ridden both I can say they feel about the same. If there's a difference, it isn't a huge and glaringly obvious one.  

I would not be surprised if the differences are superficial. Unlike gas engines, the power is mostly limited by efficient cooling and robustness of parts. Both motors had around 5kw peak power (if I am not mistaken) so going from 2kw to 2.2kw in nominal power could be down to just using thicker wiring, different glue and screws etc, without any substantial differences. Perhaps the only real difference is better cooling (or less heating in the first place) which is great but does not translate into better performance (until the "weaker" rated motor overheats which is seems a rare scenario in EUCs, board usually blow first).

Anyway there's no clarity whether the published ratings concern the motors per se or if they are meant in conjunction with the particular board+battery setup. iirc, the 16XS was advertised as 2kw while the 16X as 2.2kw. I bet they use identical motors though. So all those ratings are partially about the whole setup, partially pure marketing.

[RagingGrandpa]

22 hours ago, supercurio said:

Is the Tesla motor a lower Kv rating?

Yes certainly! Because it has a lower freespin speed than 84V Nikola.

16 hours ago, alcatraz said:

RS and EXN

Does that mean that the HS versions have similar pull with the good ol' MSX?

Precisely. 
It's one of the important points demonstrated by the results- the C30 motor (wound with 3 turns) will give the same torque for the same current, in any EUC. And if the tire size is the same, it means the same pull force too.
The only remaining difference is the max current, which is a property of the controller and its firmware.

Voltage does not change the story for low-speed acceleration; instead, higher voltage lets you sustain that pull force at higher speeds. 

[jtm94]

Finally someone that understands voltage does not increase torque! I would like to see HS wheels tested regardless just to see the numbers.

[conecones]

This is a fantastic test - thanks for the photos showing the setup.

Any plans for Inmotion V12? I am considering doing this to stress test the new board and get some useful data at the same time.

[supercurio]

I have a V12 now @conecones and wonder how I could complete this test, modifying the methodology without compromising accuracy.
Essentially: without having to build the handles.

[RagingGrandpa]

13 minutes ago, supercurio said:

wonder how I could complete this test <> without having to build the handles.

Build some handles!
Hockey sticks with hooks screwed on sound very practical...

(This test requires you have some way to overlean the EUC in a sustained and controlled way. There's no getting around it- you'll need some sort of handle.)

[supercurio]

@RagingGrandpa okay I ordered the scale already, I found that searching for "mini crane scale" and disabling automatic translation on Amazon.se allowed to find the same as the one you showed in the video.

There's a common alternative here which has 3xAAA batteries instead. In doubt, I got the same since it works for you.

Yeah I'm not sure yet what could be an easy to replicate way to make these handles. I think that it's the main blocker for now for most (including reviewers) to replicate the test.

[RagingGrandpa]

Yup.

Perhaps you could clamp the pedals with something like this, and attach a pipe to the wooden clamps (with brackets or clamps).

I think any handyman or fabricator can make handles according to their favorite materials and tools... these "maker"-people are around us everywhere, we just need to discover them  

[supercurio]

@RagingGrandpa I'm designing this model (work in progress) to 3D print, since for me it's easier than getting metal work done.

At the moment it's suited to insert a 25mm square profile in the front. Do you think a 25mm steel section would be strong enough at the desired length?

Also for now the max pedal length is 285mm (Nylon Nove XL are 280mm) and minimum approx. 205mm. Pedal height, 30mm.
Do you think the dimensions would work with almost all configs?

[RagingGrandpa]

47 minutes ago, supercurio said:

for now length 285mm, height 30mm.
Do you think the dimensions would work with almost all configs?

My handles have an interior space of 280x35mm.
Maybe add a little height in there, to accommodate studded pedals.

49 minutes ago, supercurio said:

Do you think a 25mm steel section would be strong enough at the desired length?

Yes, 2.5mm wall (or thicker), 25mm square steel tube should be enough.

Beef up your print around the 'tube receptacle'- that's where the stresses are greatest. Your image above would certainly tear out the plastic wall...

Also consider making the tube receptacle angled forward (not exactly vertical) for easier ergonomics during testing.

[xorbe]

Interesting data -- the problem is that I don't ride with pull sticks on my EUC.  It needs to be standardized on a particular lean force applied, which would be pretty hard for Joe Random to duplicate.

For instance, MCM5v1 and MTen3 don't have high numbers in this test, but those two wheels give it their all much easier than other wheels.  Nothing really communicates that in spec numbers so far.

We need graphs of ft-tb lean to horizontal force produced, not a single peak number.

Edited July 11, 2022 by xorbe

[alcatraz]

Until someone builds an euc-dyno, this is what we have. (it's way better than nothing)

[supercurio]

6 hours ago, RagingGrandpa said:

My handles have an interior space of 280x35mm.
Maybe add a little height in there, to accommodate studded pedals.

Yes, 2.5mm wall (or thicker), 25mm square steel tube should be enough.

Beef up your print around the 'tube receptacle'- that's where the stresses are greatest. Your image above would certainly tear out the plastic wall...

Also consider making the tube receptacle angled forward (not exactly vertical) for easier ergonomics during testing.

Thanks for the good advice and dimensions! I made changes.

I am constrained by the max print size of my printer bed and a 1kg spool of material, and mostly managed to make it fit that way 😁
With reinforcements (100% infill) in the sensitive areas.
It means each bracket will ends up weighing 1kg, if the design is correct that will be plenty strong. But I didn't run any simulation or test yet.

(Here's how it would look inside.)

 

Edited July 12, 2022 by supercurio

[dycus]

Thank you @RagingGrandpa for compiling this data and encouraging others to test their wheels! I extrapolated some extra info based on the results (maximum wheel torque and maximum acceleration).

https://docs.google.com/spreadsheets/d/1VRuyx-0Z2JEgDD7eHrQMTZVxvqD1sQYg/edit#gid=981762669

Please feel free to take the formulas and data to add to the spreadsheet!
The torque is just the pull force times the wheel radius. For the 16X, 203lb * (16 / 2)in = 135 lb-ft.
Acceleration is a = F/m. We have the force, 203lb, and the mass (rider weight plus wheel weight). Some unit conversions were needed.

Disclaimer:
Electric motors only have full torque at 0 rpm. Once you start moving, your acceleration will drop. The values here are only the theoretical peak you could experience, relative to other wheels, at the moment you take off, if pushing the wheel exactly to its limit. The exact values aren't necessarily accurate or useful but it should help for relative comparisons.
It also doesn't tell how torquey a wheel will be at speed, but this data might be useful to compare wheels for hill-climbing use.

And I didn't take into account any effect from the handles being used to pull on the EUC.

It does reinforce why the RS and MSP are among the favorite wheels of those who seek high performance, and that they've been sleeping on the 16X (though it's not that fast)!

 

The weight of the wheel sometimes makes a big difference - you can see the Sherman is uncharacteristically slow for its pull force because even for a light rider, the wheel is so heavy that its weight becomes significant. The MCM5 is right up with there with the high-acceleration wheels while also being the lightest (non-MTen) EUC on the list!

Edited July 12, 2022 by dycus

[yoos]

Nice acceleration numbers (though a bit useless since we don't know the full F(v) curve). Note that the image in your post shows wrong data for accelerations (probably a wrong overall factor). I even wrote up a long calculation but then noticed that the actual google sheet shows correct numbers. I suggest to update the image here since many will reference it instead of the actual sheet.

[dycus]

You may have commented right as I was actually fixing that very thing - I realized I did the lbf to N conversion wrong. Acceleration numbers are I think more accurate now (and the image is updated). I agree they're not very useful as absolute values, but should be good for comparing wheels to each other.

[supercurio]

Nice contribution @dycus!

What about a "responsiveness index" column? That would be pull force vs solely wheel weight.

[RagingGrandpa]

13 hours ago, supercurio said:

I am constrained by the max print size of my printer bed and a 1kg spool of material

My poor-man's FEA says: 

(red is compression; blue is tension; green is volume you could expand into to help both)

Print one and break it and iterate, sounds like fun!

And one last point: remember that you're going to need to step on this thing, when it's installed on the pedal, to put a big vertical load on the tire to maintain traction while testing... hopefully your foot doesn't crush it.
 

Edited July 12, 2022 by RagingGrandpa

[supercurio]

Oh haha good suggestions. I decided to make it narrower (I was initially aiming for 100mm width) in order to get headroom to reinforce what's most stressed. You're right I can adjust the shape there in the most stress area!

I didn't think about stepping on it tho, which of course is necessary for grip. I understand better the shape of your design now. I'll need to  accommodate for that better as well.

Then yes there will be some slow trial and error... because prints are gonna take a long time. Luckily I'll have a 16X to try it, since it has the best results so far and has very little risk to die in the process 😅

BTW once it works I'll give a setup to Adam / Wrong Way who's keen on testing this in his review process (we talked about it since he's visiting Stockholm this week)

Edited July 12, 2022 by supercurio

[dycus]

6 hours ago, supercurio said:

Nice contribution @dycus!

What about a "responsiveness index" column? That would be pull force vs solely wheel weight.

Thank you!

I thought about it, but it doesn't really give more information. The wheels would still be in the same order in terms of acceleration, just with higher numbers. Plus, who's using their wheels with nobody on them?

Edited July 12, 2022 by dycus

[supercurio]

I believe that (aside from soft/medium/hard mode pedal programming) the pull force vs wheel weight will be what determines how "responsive" or "reactive" the wheel feels. Not only how the wheel transports you, but also how it moves under you when driven mostly by the feet instead of your whole center of gravity.

You'll hear people describe it as nimble as well, and as a nice to have for city riding and some slow speed technical off-road.

[yoos]

41 minutes ago, supercurio said:

I believe that (aside from soft/medium/hard mode pedal programming) the pull force vs wheel weight will be what determines how "responsive" or "reactive" the wheel feels.

Indeed, reactivity is how readily the wheel moves in response to your input. However еhe test in this thread is about the extreme case: what is the max force you can squeeze out by tilting as much as you can. Reactivity is usually judged by the opposite case: modest forces in response to gentle commands. Depending on the algorithm and overall firmware you can have very different response curves (force vs tilt) and that's assuming there is indeed just a single curve and no memory effect involved (I think descent and ascent assistance in inmo wheels can only work if there is a memory effect: the wheel recognizes it is going at a constant speed despite a strong torque and adjusts the tilt/torque offset accordingly). In other words, this test is about the max force-max input end of the input-response curve while perceived reactivity is about the low-input part of that curve.

 

[alcatraz]

Aha, now I understand what (I think) you mean by responsiveness. That's interesting. I call it rider input. 

Some wheels need very little of it and some need tons of it.

I'd love to see that in the graphs. Like, how much force was applied to rotate the wheel to produce a certain (baseline) force forward.