EUC Treadmill

[IamIan]

1st trial => I need to balance the rollers more.

1st Trial Video

Edited February 24, 2021 by IamIan
spelling

[.....]

Wheel hop is gna be a challenge. Lookin good! Maybe you could wrap the rollers in a dense rubber mat or some form of material that will help with traction while mitigating the hop of the euc tire? Im excited to see how you end up developing it all and how you're gna handle downforce and tilt. Looks like you're making a micro dyno..

Edited February 25, 2021 by ShanesPlanet

[IamIan]

In an effort to improve the rolling balance I took lots of measurements and found 3 points of potential improvement.

1 = The previous 4 brackets were about ~3/16" out of square.

2= When I hand mounted the layers of plywood to the center roller shaft , they were a little non-perpendicular to that center shaft.

3 = When I hand drilled the center holes there was a roughly ~3/32" variation from center layer to layer.

- - - - -

For 1

I rebuilt the base out of some scrap 18mm thick plywood .. cut down to 1' x 2' .. center for roller/wheel cut out .. 4 bracket bolt through points drilled .. Now those 4 are as parallel , perpendicular , and square as I can get them by hand .. within about ~1/2mm.

 

For 2

I covered a scrap piece of fiberglass angle with DA sand paper .. and positioned it parallel .. same distance from each side center shaft .. then spin the wheel to sand off the high points .. started this , but not done yet.

 

For 3

I set a block perpendicular to the center shaft and spin the wheel to sand off extra .. one is done .. within ~1/2mm .. the 2nd wheel not yet done , is still at about ~2mm of high/low variation.

 

 

[IamIan]

Decided the cheap fishing yoyo's will not be enough force for side to side balance .. upgraded to ~21Lb springs from McMaster 9044K209 $5.26 for 3pk.

Glued the plywood base together , so that the roller brackets are pretty much parallel and square.

My brother in law was replacing the belt on his treadmill , so he gave me the old one .. I might try to cut it to size and glue it to the rollers .. It might help with wheel grip/slip.

 

[Tawpie]

wow. That's getting very impressive. Don't give yourself a hernia.

[IamIan]

The treadmill belt was a bust .. the EUC torque on the surface was stronger than the spray adhesive , and once at a faster speed the belt few off the roller I glued it to.

 

The upgraded springs are better .. but , not good enough.

They do correct for side to side position changes as intended .. to keep the EUC centered on the rollers.

.. but ..

They only allow for very slow walking speed like testing .. I'm not sure at what the specific frequency is .. but at one specific point , a feedback loop seems to establish .. EUC too far to left , left spring isn't pulling but right spring is to correct it .. as EUC begins to 'turn' slightly to the right , once past center the right spring isn't pulling but the left spring starts to pull to correct back to center .. but .. at slow speeds it works correctly .. but at faster speeds there is enough momentum that the side to side motion grows instead of diminishes

 

'Back to the drawing board'.

[.....]

would a groove or channel in the wheels help? Im thinking baseball pitching machine could be an item to look at for ideas. Im enjoying just reading on your trial and error/design phases. Obviously you are doing it because you enjoy it. Hell, you may have reached the point of "oh Ima figure this out no matter WHAT!!" too. Keep after it man. Just be minful that the tolerances needed to acomodate an euc at speed, is catch 22. These things dont roll flat or straight, even on a good day. Im betting youll need pnuematic or foam tires on the wheels. You need so much pressure but also room for how en euc hops. The euc tire could be deflated a little, but it really depends on what you're testing for. Having a downforce of 150lb (avg man), could be a decent stat. Id COULD be easier to apply upwards force via a jack, on your rig. And mount the euc solid as hell. Only need to change the front/rear angle on the euc. You could leave the jig holding it, solid as hell up and down and side to side. Use the rollers themselves to apply the testing force to accomodate weight you are trying to mimic? Im just thinking you will definitely need to either have tires on the machine or springs on its feet. Im a little unsure that even springs will react fast and smooth enough to prevent wheel hop on hardwood. Maybe some kind of rubber/foam bushing on the axles of the wheels? Interesting project for sure!

Edited March 15, 2021 by ShanesPlanet

[IamIan]

I did a baseline .. pretty much OEM settings .. turned on (self balancing active) , not moving , no fan , daytime running lights on .. V11 for 12hrs used ~7% .. If accurate , If linear , If 1.5kwh total .. than ~7% = ~105wh / 12hr = ~8.75w

 

I also did some more roller/wheel balancing work :

Perpendicular to axis of rotation is about as good as I can get it geometrically balanced with my simple tools .. not perfect , but somewhere less than ~1/2mm.

By flipping it upside down on my table saw .. equal thickness blocks on each side so the wheel/roller is free to rotate just above table surface .. then slowing raise the table saw blade up little by little .. repeat for each side of each roller/wheel.

 

Made a little more progress (but not done yet) on the OD / parallel to axis of rotation geometric balancing .. I'm probably around ~1mm now .. I'd like to get that also to under 1/2mm .. I've been doing this by putting sand paper on a piece of scrap angle .. the angle has a strong resistance to deflection .. both right and left side of that angle are measured to be the same distance from the axis of rotation , and clamped down into place .. then just move / spin the roller / wheel by hand .. low spots just miss by a hair , high spots get sanded a little each time .. tedious , but seems to be working.

 

Once the geometric balancing is done I think I might also try to do a little mass balancing as well .. I'll see how close the mass is balanced after the geometric balancing is done.

 

Of course a good quality CNC .. or lathe .. could probably do a better / faster / easier job of all this geometric balancing .. but I don't have those tools .. The irony is they have them where I work (I see them every day) , but I'm not allowed to use them (for personal use) .. oh well , I'll make do with what I have .. hopefully it will be good enough for my application.

[IamIan]

On 3/14/2021 at 7:51 PM, ShanesPlanet said:

would a groove or channel in the wheels help?

I have a little more balancing I can do .. which should help.

After that balancing effort .. if it still isn't good enough .. then maybe I'll try that idea .. a slight inward groove / curve in the roller wheel might help .. to entice the EUC to stay pretty centered / aligned on the roller/wheel.

 

[.....]

I had a thought the other day. Wouldnt large rollers need spin slower? Of course, then you may also be changing traction demands... The devil is in the details! Maybe work would let you pay for them to cut something. You know, at a nice 99.5% discount? Most rollers I've seen, have an inward bevel at the least, unless they are on a flat bed of them. Looks like a lathe would be your best bet.

 

Edited March 21, 2021 by ShanesPlanet

[IamIan]

5 hours ago, ShanesPlanet said:

I had a thought the other day. Wouldnt large rollers need spin slower? Of course, then you may also be changing traction demands... The devil is in the details! Maybe work would let you pay for them to cut something. You know, at a nice 99.5% discount? Most rollers I've seen, have an inward bevel at the least, unless they are on a flat bed of them. Looks like a lathe would be your best bet.

 

Yes .. larger dia rollers have a larger circumference .. and thus slower RPMs would be needed to test the same EUC speed .. that's why I added the wood to the initial center rollers I got from Graingers .. The larger diameter means instead of a max of around ~2000 RPM @ ~11MPH .. it's a max ~2000 RPM up around ~47MPH .. soo a bit of included safety margin faster than I ever expect to use it.

Soo .. 1st I need to finish the balancing I can do .. geometric and mass .. if that isn't enough .. then add the bevel you mention .. and then I'll have to rebalance everything again , because the bevel will change it .. if that still isn't good enough .. then yes .. I could start over from scratch and try an even larger diameter wheel, thus even slower RPMs .. I'm not quite ready to pitch this one in the garbage just yet.

I can see (all else being equal) as the wheel / roller gets bigger and bigger .. closer and closer to a flat road surface the total contact surface area could increase some small amount .. I haven't yet had a traction issue .. I haven't done anything with a high enough torque for it to come up yet .. Soo far just trying to get is to run smooth enough for an entire battery charge top to bottom .. ideally up as high as 35 MPH (~1,470RPM @ 8"Dia) .. but I'd probably be satisfied if I at least got up to around ~25 MPH (~1,052 RPM @ ~8" Dia).

Well .. the 1st part of using the machines where I work to do it .. would be the machine operator to agree to do it on his own time (not company time) .. then I could submit a request for permission .. The production manager usually doesn't mind as long as it doesn't interfere with his schedule .. soo , the approved machine operator would have to agree to do it when that machine and that person are not already scheduled to be doing something else .. and I already a month ago asked the operator .. before I used my own table saw to cut the wheels and such .. needless to say he has not yet been willing .. not as a favor .. not if I pay him for his time .. etc .. he doesn't want to do it .. not that he can't .. I've seen him do far more complicated work .. and I've seen him do off the clock work for some other people sometimes .. he just doesn't want to do this for me .. soo , I just started trying to make it myself with the tools I have .. I'm far too small of a fish to be able to go through full company channels .. pay for shop time and have his boss force him to do it on the clock as part of the production schedule .. The company doesn't usually take jobs unless they are 7 or 8 figure jobs .. something as small as a 6 figure job is very rare .. yeah I've gone a little above my original cost projection .. but I'm still well under less than $200 (3figures) out of pocket .. like I said , way too small of a fish to be able to go through channels to have his boss force him to do it on the clock.

Edited March 22, 2021 by IamIan

[.....]

Lawmower/wheelbarrow rims? I wonder if they are round? Maybe something more narrow like a bicycle rim? Its too bad we dont all own our own multi-million dollar tooling shops...

Edited March 22, 2021 by ShanesPlanet

[IamIan]

I have finally finished getting the roller wheels as geometrically balanced as I can with the tools I have .. and .. still not quite good enough .. I can manually hold it steady by hand easily enough .. but that won't work for the testing I wanted.

 

Some other additional mechanism will be needed in order to prevent the positive feedback loop.

 

I'm currently considering :

#1> making the roller wheels curve in slightly toward the center , instead of the flat shape they are now .. major con of this is that it will require me to again spend many many ours re-balancing the roller wheels again .. soo , I might start with other options 1st.

#2> Instead of springs on the sides .. maybe I'll make a rigid bracket for the sides .. scrap aluminum with a slot in it .. EUC can then tilt and move up and down in the slot , but not side to side.

#3> Sense it is a tilting side to side that is part of the positive feedback loop .. perhaps if I extended a rigid piece from the top center .. kind of like how Marty showed attaching a bicycle seat .. then have the side side motion restricted with that .. the longer length would create a longer leveler to act on.

#4> Maybe put a spring in each of the 4 corners .. front right , front left , back right , back left .. instead of just pulling from the center of the pedal bracket... This would have to be calibrated to still allow for the needed tilting function.

 

[IamIan]

On 3/21/2021 at 4:47 PM, IamIan said:

I did a baseline .. pretty much OEM settings .. turned on (self balancing active) , not moving , no fan , daytime running lights on .. V11 for 12hrs used ~7% .. If accurate , If linear , If 1.5kwh total .. than ~7% = ~105wh / 12hr = ~8.75w

I redid a more complete 100 - 9% powered on but static and non-moving consumption rate test .. to better establish the base electronics self consumption.

The battery % reported is only a little bit off from a truly linear 7% / 12hr rate.

The V11 started beeping and did the extreme forward tilt at 9% after 153.5 hours.

It seems to me .. Either one of 3 things explains the variation from linear.

#1> The electronic's own rate of consumption does vary with battery %.

#2> The electronic's own rate of consumption does not vary .. instead it is an artifice of software.

#3> Some combination of 1 & 2

[.....]

I wouldnt think you would need rebalance the wheels if you are merely sanding/cutting a bevel/groove in the center. DO you not have anyone handy with a lathe? I admire your dedication for sure!

[Chriull]

12 minutes ago, IamIan said:

1> The electronic's own rate of consumption does vary with battery %.

+1. Efficiency changes with input = battery voltage.

As battery efficiency changes with delivering some constant power with decreasing voltage and increasing current.

Battery % could already be some not linear from battery voltage computed value from inmotion? Maybe battery voltage makes some change?

Speed was sufficient stable over time?

[IamIan]

3 minutes ago, Chriull said:

+1. Efficiency changes with input = battery voltage.

As battery efficiency changes with delivering some constant power with decreasing voltage and increasing current.

Battery % could already be some not linear from battery voltage computed value from inmotion? Maybe battery voltage makes some change?

Speed was sufficient stable over time?

The above was non-moving .. no speed .. a base case for the electronics alone .. still hoping to iron out the details for stable at speed over time.

 

One of the curiosities I hope to eventually see is how the SoE (energy based %) compares to app reported battery available %.

Energy (wh, joules, calories, etc) is the capacity to do work .. heat something , light something , moving something , sound , switch a transistor , etc.

Some battery devices use simple Ah counting , thus a SoC (Ah based %) .. 1Ah @ 84v = 84wh capacity to do work .. but the same 1Ah at 70v is only 70wh capacity to do work .. ie. same number of Ahs but different ability to do work .. ie move less , less sound , less light , etc .. some devices to try and report as close to a real energy based available % , adjust the available reported % with other sensors , voltage , amps , temperature, etc .. Soo , one thing I would like to eventually see is .. how many wh of available energy does that battery % indicate ? .. is 50% actually 1/2 .. or is a reported 50% actually something more like 60% or 40% .. etc .. what does that reported % curve look like at 100w rate .. or a 500w rate .. etc.

  

4 minutes ago, ShanesPlanet said:

I wouldnt think you would need rebalance the wheels if you are merely sanding/cutting a bevel/groove in the center. DO you not have anyone handy with a lathe? I admire your dedication for sure!

It might not effect the balance .. however .. it is possible that any time one removes material for any reason , that doing so can effect the rotational balance.

Right now they are as balanced as I can get them (with the tools I have) .. geometrically less than 1/2 a mm , and close enough in mass to not cause the bearings to allow rotation. Soo , I am more inclined to 1st try those other things that don't have that kind of potential risk to adjust the balance (geometric or mass).

I could be wrong in my chosen sequence .. and maybe wasting my time .. and eventually after other things fail , just coming back and doing that bevel/groove anyway .. only time will tell.

 

[IamIan]

Turtle steps continue.

4 corner points added.

I got anchor bolts .. #6 wood screw thread on one side , to go into the corner post .. and 8-32 thread on the other side.

Got four 30lb rated turn buckles .. this will allow me to 'calibrate' the 4 corners to try and get the tension properly balanced to keep the EUC stable enough for testing.

The honey comb V11 pedals use m3 size allen wrench bolts for the 'nubs' / 'spikes' .. replacing one on each pedal with a longer m3 bolt allows for a reasonably secure connection to the EUC.

Short Video   https://youtu.be/sQIRtrh_ZzU

 

[Tawpie]

Dad gum you ARE getting there! Very cool. Bravo.

[Rehab1]

15 hours ago, IamIan said:

Turtle steps continue.

Amazing work! 

[IamIan]

Turtling

The small oscillations cause the threads to 'self turn' .. attempt with using electrical tape on threads failed.

Attempt to replace free weights with clamps on trolley handle worked well .. not only stays put , but also allows me to more precisely control forward bias (throttle).

Switched all the nuts to nylon lock nuts .. and added a set screw for the turn buckle.

 

Recharge , reset .. try again ..

So far ~24hours latter of continuous walking speed operation , has seemed to be stable.

 

[IamIan]

Just a quick note 1st :

Precision , Accuracy , Resolution .. are 3 different things .. these numbers from a ~$150 home made tinkering project should only be considered within a reasonable +/- range for each of those 3 concepts .. Or said another way .. 1 ~ 1 , not 1 = 1.

I think I'm satisfied with the 1st test series for 'trolley' mode .. I think the ~760g of clamps used for creating a forward bias is small enough for that.

Total distance didn't change allot .. a little bit of a curve up and down .. but at least on this scale so far nothing big .. not that I would ever expect someone to go in trolley mode for ~270km.

Rolling Resistance on it's own is linear with speed or weight .. All else being equal 10% faster would = 10% more rolling resistance .. but actual speed seen in the test was a little slower for a given increase in watts .. thus the increased watts at higher speed was not entirely from rolling resistance .. I suspect internal efficiency curves for the electronics, motor , etc. .. as was previously pointed out , this test doesn't include real world things like hills , acceleration , wind , etc .. soo , one should not expect to go this fast with this many watts.

If one prefers another way to see that is that the wh energy use rate per km goes up as watts of power goes up.

The motive Wh output is a better metric for a ball park real world expectation .. at some __Watt rate to get some net __ Wh out .. and that watt rate in the real world will be made up of a combined total of lots of things , speaker sounds, lights , gravity , wind , weight , tire pressure , etc.

It was also interesting to see how the initial idle load test ran for 153.5 hours no movement just idle .. but as a faster rates more and more of the total shown wh shows up as motive wh .. eventually I would expect at high enough watt rates to see this peak and then reverse .. which is also an indication that the displayed watts and wh are not 1st thing from the battery , as they did not show or count the watts/wh from the light , the base electronics, etc during that 1st 153.5 hour test.

Also reasonable confirmation of a roughly true ~1500wh battery being in the device .. using 90% (100-10) of 1500 is 1350wh total (for motive , lights, elctronics, etc).

 

 

- - - - - -

Soo , that's the trolley testing ..

Now I have to think a little how I'd like to try and mount balanced weights to it for weighted testing ... hmmm , maybe secured to the pedals with 4 of the honeycomb M3 pedal stud bolts .. ideally attachable in smallish increments, for handling and for incremental testing.

 

 

 

Edited April 18, 2021 by IamIan

[Chriull]

43 minutes ago, IamIan said:

Soo , that's the trolley testing ..

Now I have to think a little how I'd like to try and mount balanced weights to it for weighted testing ... hmmm , maybe secured to the pedals with 4 of the honeycomb M3 pedal stud bolts .. ideally attachable in smallish increments, for handling and for incremental testing.

Afai see weight is only needed to keep up traction between the tire and the rolls.

Resistance to simulate air drag, inclines, accelerations, etc comes from the rolls. By some kind of braking mechanism, an electric "braking" motor or just some "higher mass" of the roll to be accelerated.

[IamIan]

1 hour ago, Chriull said:

Afai see weight is only needed to keep up traction between the tire and the rolls.

Resistance to simulate air drag, inclines, accelerations, etc comes from the rolls. By some kind of braking mechanism, an electric "braking" motor or just some "higher mass" of the roll to be accelerated.

I haven't decided exactly what my approach will be ..

Yes a braking mechanism is an option to introduce additional watts of drag .. it could be used on it's own .. or it could be used in conjunction with added weight.

Weight is not just for traction (static Friction) .. Rolling Resistance increases linearly with both speed and weight (that is perpendicular to the surface being acted on) .. 2x as much weight will double the watts of rolling resistance , just as 2x as much speed with will double the watts of rolling resistance .. 60lbs V11 +60 Lbs = 2x watts rolling resistance (compared to 60Lbs V11 alone) .. 60+120 = 3x .. 60+180 = 4x .. 60+240 = 5x .. etc.

 

[caumargu]

Wow! I love people who can do something like this with their own hands! I'm thrilled!