Brands with / without unexpected shutdowns (new buyers look here)

[Chriull]

The problem is that at the instant where the EUC lacks the torque to go up a hill it also lacks the torque to tilt back. In fact, the wheel is trying to tilt back in order to keep the pedals level as you go up the hill! It's just that there's no torque left, so the pedals lean forward as you try to drive it up the hill.

By the time the wheel detects this situation it is too late. The only solution is a larger margin of safety that holds back more performance so that it can tilt back the pedals in this situation. To make this variable rather than a large wasteful static margin of safety, the wheel would need terrain mapping to detect that a steep hill is coming up in the next few meters and thus tilt back even though the situation at this instant is okay. Even then, we all know that it's possible to suddenly hit a bump and be knocked off balance, those can happen even when the terrain seems smooth and level. 

I just made a rough estimation, which power is necessary to go uphill. To move 90kg (rider+EUC+...) up a 10° hill (17% slope) with 10 km/h one needs 432 W! This is just to "lift" the weight up - than there is still power needed for the air resistance, rolling friction and so there should not be to much/anything left to accelerate, pedal tilt back, etc... ;(

So driving from the flat with about this speed or more onto a 10° slope brings the EUC immedeatly to its limits... 

ps.: i hope my estimation is right:

Energy needed to lift something: delta E=m*g* delta h

Power: P=delta E/ delta t

delta t= delta s/ v

so finally one gets: P=v*m*g* delta h / delta s

where delta h / delta s is (almost) der slope percantage

[hobby16]

it seens like the EUC misses this point to tilt back/warn the driver, the available power is not sufficient for balancing and the EUC tilts forward and you fall off..

I don't know where you get the idea that because the wheel lacks powers for balancing, the rider will fall off but sorry, it's false.
When the battery is near empty (nearly everyday), my Firewheel lacks power but I never fall. Never. There is no tilt-back warning, I simply feel the pedals can't be maintained horizontally and establish the balance myself. It's not any more difficult than keeping side balance, as simple as that ! All I need to keep my balance, even in very steep hills with a deeply discharged battery, is to NOT have this bloody power shutdown.

I repeat, any contrived justification for the power shutdown on a monowheel is nonsensical. There is absolutely NO justification for it. All the on-the-fly out-of-the-hat speculations I've heard so far on some hypothetical software or hardare limitation or whatever are nonsense. Really. Any motor control engineer would laugh at it.

[esaj]

I just made a rough estimation, which power is necessary to go uphill. To move 90kg (rider+EUC+...) up a 10° hill (17% slope) with 10 km/h one needs 432 W! This is just to "lift" the weight up - than there is still power needed for the air resistance, rolling friction and so there should not be to much/anything left to accelerate, pedal tilt back, etc... ;(

So driving from the flat with about this speed or more onto a 10° slope brings the EUC immedeatly to its limits... 

ps.: i hope my estimation is right:

Energy needed to lift something: delta E=m*g* delta h

Power: P=delta E/ delta t

delta t= delta s/ v

so finally one gets: P=v*m*g* delta h / delta s

where delta h / delta s is (almost) der slope percantage

That sounded high, so I did some of my own calculations... it's been a long time since I've done any physics, but here goes (I'm not sure if this is completely right ):

If using the equation from here: http://www.dummies.com/how-to/content/calculating-the-force-needed-to-move-an-object-up-.html
it is
  F_push = m*g*sin(theta)+µ_s*m*g*cos(theta)

Where F_push is the force (in newtons, not watts), m is the mass to move, g is the gravitational constant (9,80665 m/s²), sin(theta) is the sine function of the uphill angle, the part after the sum (+) is the static force of friction (which I'll leave out, as I don't know the coefficients, and probably it should be rolling resistance and airdrag instead of static friction to get the refrigerator moving in the example?)...

Plugging in the values of m = 90kg and 10 degree slope, we get:

90kg * 9,80665m/s² * sin(10 degrees) = 153,2616... N

1 Newton (N) = 1 (kg * m) / s²

Now we add the speed, 10km/h, which needs to be changed to basic units (meters per second), to get the watts (W, power), 1 watt = 1 J/s, 1 J = joule = 1 N*m (newton meters), so

N*m/s  =  J/s = W (watt)

10km/h is 10000 meters / 3600 seconds = 2,77777..., round it to 2,778 meters per second (m/s)

So:

153,2612... N * 2,778 m/s = around 425,7W to move 90kg uphill at constant speed of 10km/h without taking the frictions, rolling resistance etc. into account.

I was wondering how the Firewheel can then easily take me up such slopes with even higher speeds (and friction/rolling resistance/whatever!). But then again, I'm lighter so let's say Firewheel (about 13kg) + me (57k) + safety gear and other stuff (3? kg) comes in around 73kg:

73kg * 9,80665m/s^2 * sin(10 degrees) = 124,312... N
124,312... N * 2,778 m/s = around 345,3W

So the mass matters a lot here. The real power needed will be even higher due to losses between battery, wiring, mainboard, motor etc + rolling resistance, air drag etc.

[hobby16]

 

So the mass matters a lot here. The real power needed will be even higher due to losses between battery, wiring, mainboard, motor etc + rolling resistance, air drag etc.

Indeed, it's a highly simplifed calculation because if the slope is 0% (flat road), calculated power would be... 0W !

However, with a direct drive motor, anything upstream (battery...) doesn't count, the W motor (the one display in the motor's torque/power curve) is the W available for rolling.

[esaj]

Indeed, it's a highly simplifed calculation because if the slope is 0% (flat road), calculated power would be... 0W !

However, with a direct drive motor, anything upstream (battery...) doesn't count, the W motor (the one display in the motor's torque/power curve) is the W available for rolling.

True, the losses and voltage sag etc. only affect the current needed from the battery etc... another thing that occurred to me afterwards was that as you already have some forward momentum, the power needed from the motor to maintain the speed could be lower..?

[Planetpapi]

esaj, How do you know all that technical stuff? Fascinating! Of-course I didn't understand any of it but still fascinating.

[esaj]

esaj, How do you know all that technical stuff? Fascinating! Of-course I didn't understand any of it but still fascinating.

Well, I've always had the kind of mind that somehow remembers all sorts of (mostly useless ) trivia from various fields and can combine and apply it into many situations... I did attend vocational school in the turn of the millenia and became a computer technician, after spending 6 years to get 3 year school completed... the usual reason why my classes failed was that while I had high marks from the tests, I was absent around 90% of the hours of mandatory classes when the maximum allowed was around 10% . But there were some electronics and electricity -related classes, so at least I got some basic knowledge about electronics from there (but I must admit that I've forgotten most of it, since I never really needed it).

After finally graduating in the spring of 2005, I applied and got into University of Applied Sciences (fancy name for a polytechnic school), and got my "Bachelors degree in Software Engineering" in 2010. I've been a hobbyist programmer since I was maybe 9 years old, so you could say it was kind of a calling. Since it's an engineering school, there were lots of mandatory mathematics and physics courses and few electronics and embedded systems courses, but mostly the classes were just programming related.  So those plus elementary school are basically where most of my (very lacking) understanding of maths/physics/electronics comes from... 

You've probably noticed I use words like "maybe", "probably", "could...", "I assume" etc., meaning most of the time I'm just throwing in best guesses and hunches, not real pure knowledge. I apply what I know to the situation to try and understand it, but I'm not always right and could go very wrong from time to time...

[hobby16]

. another thing that occurred to me afterwards was that as you already have some forward momentum, the power needed from the motor to maintain the speed could be lower..?

For a flat road, yes, if speed doesn't change, kinetic energy (1/2 mv^2 ) doesn't change so required power  is (theoretically) zero .

But on a slopped road, even at constant speed, you still need to  augment potential energy (mgh), so the lowest estimate of power is as calculated above (both Chriull's and your formula are right, his numerical application is slightly off). Real power, accounting for rolling resistance is of course higher.

[Jason McNeil]

For a flat road, yes, if speed doesn't change, kinetic energy (1/2 mv^2 ) doesn't change so required power  is (theoretically) zero .

Yeah, this is the problem with the gradient power requirements that are found on the Internet for bikes, such as http://www.gribble.org/cycling/power_v_speed.html, they assume steady-state, whereas a EUer has highly dynamic power-profile. In reality, a slight increase in acceleration on the above example is going to cause the power to spike. 

My new in-line power-meter is going to hopefully help illustrate this with good quality empirical data. 

[esaj]

Yeah, this is the problem with the gradient power requirements that are found on the Internet for bikes, such as http://www.gribble.org/cycling/power_v_speed.html, they assume steady-state, whereas a EUer has highly dynamic power-profile. In reality, a slight increase in acceleration on the above example is going to cause the power to spike. 

My new in-line power-meter is going to hopefully help illustrate this with good quality empirical data. 

I'd at least assume that accelerating needs a whole lot more power than just keeping the speed steady. Telemetrics for the wheel would be really cool, I'd love to be able to see the graphs for power usage vs. change in altitude (ie. climbing/descending/going level) vs. speed etc. I actually was thinking of modding a voltage meter (and asked Vee73 if he knew of any good) to the front/topside of the wheel for the the new batteries, so I could see the change in battery voltage during climbs and accelerations (plus it would probably be more reliable battery meter than the 0-99% display of Firewheel ). Just needs to be able to show up to around 70V (more wouldn't hurt, in case I ever decide to try higher voltages), with at least 1 decimal after the point accuracy... plus I'd have to figure out how to power the damn thing, as it would probably still use different voltage, unless there are some that can use the voltage they're measuring, but that's going to change a lot.

[Jason McNeil]

It's such a shame none of the manufacturers have yet implemented an easy to view bright LED panel that can display all this data. There's some decent power meters you can buy for about £25, even with graphing functions built-in, but are limited to 60v!

[esaj]

It's such a shame none of the manufacturers have yet implemented an easy to view bright LED panel that can display all this data. There's some decent power meters you can buy for about £25, even with graphing functions built-in, but are limited to 60v!

Most riders aren't probably that interested in all that data, so just a simple battery meter is enough for them. Same goes for pretty much everything, cars, cellphones, amplifiers, whatever... most normal users just want to turn the thing on and use it, without caring how it works or what's going on, as long as it works. And there's nothing wrong with that. But us enthusiasts want to take peek "under the covers", to understand and to see what happens, change things... It's probably not enough of an incentive for the manufacturers to do this, so we have to do it ourselves.   Who knows, maybe over time we'll see actual self-built custom wheels (the safety of such contraptions is a whole another issue, though... )... Vee73's mods of the Gotway are probably the most modded so far (he already had separate battery heating & battery temperature displays in the Firewheel with the old shells for winter riding, before selling it to me).

Edit: For my purposes, something like this should be adequate: http://www.amazon.com/dp/B00IG1AYEW?psc=1  Looks like it can use the voltage which it measures directly to power itself   Now I'd just need to finally get a credit card or Paypal or something to actually order one, I'm old fashioned and usually just go with cash & wire transfers...

[Chriull]

I don't know where you get the idea that because the wheel lacks powers for balancing, the rider will fall off but sorry, it's false.

My EUC did not arive till now - hopefully next week *keepfingerscrossed*. So i am lacking practical riding expirience - that was a wrong asumption to which i got by the articles regarding overleaning. I will stop such assumptions until i got my practical expirience

That sounded high, so I did some of my own calculations... it's been a long time since I've done any physics, but here goes (I'm not sure if this is completely right ):

Was also my first thought - that's to much. So i double checked it by the definition of HP (1 HP means lifting 75 kg 1 m in 1s and equals to 745 W)

 

If using the equation from here: http://www.dummies.com/how-to/content/calculating-the-force-needed-to-move-an-object-up-.html
it is
  F_push = m*g*sin(theta)+µ_s*m*g*cos(theta)

Where F_push is the force (in newtons, not watts), m is the mass to move, g is the gravitational constant (9,80665 m/s²), sin(theta) is the sine function of the uphill angle, the part after the sum (+) is the static force of friction (which I'll leave out, as I don't know the coefficients, and probably it should be rolling resistance and airdrag instead of static friction to get the refrigerator moving in the example?)...

The rolling resistance is "estimated" with the same formula as the static friction - they just write cr instead of µ_s. From the tables i choose 0,02 for cr (Motorcycle tyre on asphalt) as nearest approximation. This should lead to about 50 Watts for 90 kg and 10 km/h. Air Drag should be somewhere around 10 Watt at 10 km/h.

 

For a flat road, yes, if speed doesn't change, kinetic energy (1/2 mv^2 ) doesn't change so required power  is (theoretically) zero .

But on a slopped road, even at constant speed, you still need to  augment potential energy (mgh), so the lowest estimate of power is as calculated above (both Chriull's and your formula are right, his numerical application is slightly off). Real power, accounting for rolling resistance is of course higher.

I choose to disregard the length difference of the hypotenuse and the adjacent side of the right-angled triangle - so there is a systematic error of a couple of percent for low slopes. Which is enough accuracy for this and one does not have the trigonometic functions in the formula.

For accelerating this about 430 W would be enough to get to the 10 km/h in a little bit more than a second.

 

Most riders aren't probably that interested in all that data, so just a simple battery meter is enough for them. Same goes for pretty much everything, cars, cellphones, amplifiers, whatever... most normal users just want to turn the thing on and use it, without caring how it works or what's going on, as long as it works. And there's nothing wrong with that. But us enthusiasts want to take peek "under the covers", to understand and to see what happens, change things... It's probably not enough of an incentive for the manufacturers to do this, so we have to do it ourselves.   Who knows, maybe over time we'll see actual self-built custom wheels (the safety of such contraptions is a whole another issue, though... )... Vee73's mods of the Gotway are probably the most modded so far (he already had separate battery heating & battery temperature displays in the Firewheel with the old shells for winter riding, before selling it to me).

Edit: For my purposes, something like this should be adequate: http://www.amazon.com/dp/B00IG1AYEW?psc=1  Looks like it can use the voltage which it measures directly to power itself   Now I'd just need to finally get a credit card or Paypal or something to actually order one, I'm old fashioned and usually just go with cash & wire transfers...

Or you adopt a charge doctor for it. You just need to add an additional shunt resistance parallel to the current measurement points, so that the measured max 3A then would mean 30A, 60A or whatever... I assume hobby16 knows the internal resistance the charge doctor uses for current measurement?

Would be the more or less the same price as your link. Its just the question if it makes sense (data logging of the charge doctor is every 10 secs - is this the frequency of the measurements, too? Does the charge doctor average out the values over the intervall or gets the actual voltage/current at the moment of measuring?)

So you could see the used up Wh from the battery pack, too

[esaj]

Or you adopt a charge doctor for it. You just need to add an additional shunt resistance parallel to the current measurement points, so that the measured max 3A then would mean 30A, 60A or whatever... I assume hobby16 knows the internal resistance the charge doctor uses for current measurement?

Would be the more or less the same price as your link. Its just the question if it makes sense (data logging of the charge doctor is every 10 secs - is this the frequency of the measurements, too? Does the charge doctor average out the values over the intervall or gets the actual voltage/current at the moment of measuring?)

So you could see the used up Wh from the battery pack, too

I've already got the Charge Doctor on the way, but I don't think I'm going to install it into the wheel itself, nor am I going for "full" telemetry logging (at least for now)... the voltage display is so I could see the voltage drops during acceleration and hill climbing, and to act as "secondary" battery meter, as the Firewheel's 0-99% display is not very accurate, unless you check it stationary (well, the same probably goes for the voltage display), and I don't know what voltage the 0% or 99% means. It would allow me to then see the more accurate voltage where the mainboard stops you from riding anymore, and probably would serve better to "guess" the remaining battery than the current battery display (the Firewheels' current display can show 0% while riding for the last kilometers, at best I've ridden over 4km over large hills with it showing nothing but 0% while riding, and the battery wasn't even empty when I got home, but somewhere around 15-20% stationary).

[hobby16]

Or you adopt a charge doctor for it. You just need to add an additional shunt resistance parallel to the current measurement points, so that the measured max 3A then would mean 30A, 60A or whatever... I assume hobby16 knows the internal resistance the charge doctor uses for current measurement?

Would be the more or less the same price as your link. Its just the question if it makes sense (data logging of the charge doctor is every 10 secs - is this the frequency of the measurements, too? Does the charge doctor average out the values over the intervall or gets the actual voltage/current at the moment of measuring?)

So you could see the used up Wh from the battery pack, too

The shunt resistor of the Charge Doctor is 25 milli-ohms. Sampling frequency is 0.5s in order to calculate Wh values with best accuracy, but indeed, data are send only every 10 s for logging since they are so boringly stable

For embedded real time measurements, to capture all the interesting transients, e.g. at acceleration or braking, I think minimum  100hz sampling is required.

I have a small-board to do just that (with shunt resistor for currents up to 40A) and store data on a 2Gb sd-card so no lack of memory, even for hours of logging. I have the project to test and log all possible wheels, just lack time to do. Maybe it can be a collaborative project (with vee perhaps) but I need to finish my mobile charger first, a lot of people want to buy it.

[Daan]

This is a great thread with very important advice; can we update it? 

1. Can we pin it to the top of the forum so it doesn't get lost
2. @KaleOsaurusRex: Can you put all good/bad brands in 'bold', and color the back ground of the bad ones red, and the good ones green so it stands out more. Perhaps move that list to the top of the post so it is immediately visible -- and then follow with the explanation. Also, perhaps put the uncertain ones in orange -- looks like huanxi will not respond is probably unsafe 
3. Perhaps we can extend the list with how the warnings are issued -- i.e. for Solowheel overcharge/low battery/emergency all do the pedal vibration; hispeed warning does tilt-back.

Btw. I did contact inventist about the Solowheel and it is definitely "safe", eg. on overcharge/low battery or emergency it will do pedal vibration. Because it has a custom BMS 'emergency' is something like when a cell fails or something. On hi-speed it will tilt-back.  

Also, note that even a 'safe' wheel like the Solowheel can still cut-off: the mainboard could fry, or if you keep over-leaning against the tilt it may fail to balance you etc. This is of course clear when reading the thread -- this thread is really about making manufacturers produce wheels with safe designs that are made to protect the rider-- not the battery.

[dpong]

I want to thank the contributors who are offering anecdotal reports of safety. We are all happy to hear you haven't had an unexpected shutdown yet.

However, please take into account confirmation bias. In the near future, when the rider numbers are much, much larger, when most riders do not read the forums, unaddressed unsafe electrical designs will cause absolutely avoidable injuries. Now is our chance to change directions and hold the manufacturers' and distributors' feet to the fire.

The argument of "there will never be an electric unicycle that's perfectly safe" is an absolute copout.

The BMS shutdown design will go away first; that design is already on it's way out of style.The next problem going forward is going to be mainboard thermal overload / shutdown.

There is no reason the controller/mainboard couldn't modulate motor performance to avoid mainboard thermal overload. There's no reason the computer couldn't force a safe and gradual complete low battery shutdown while maintaining enough charge to avoid an absurd unexpected shutdown.

Yes, these solutions might add some cost. Yes, the solution might reduce the usable range of a given battery size. However, this is not optional. It is our responsibility.

There will be other problems to be addressed, but for now the two main known issues are entirely solvable. We will solve them by using the power of the almighty dollar, to make sure the manufacturers and distributors can ignore this problem no longer.

What do you think of the BMS issues reported against Ninebot One E+?   By @hobby16 ?  We can start the discussion here.  I should hope we get an answer from Ninebot soon, as perhaps my front teeth depend upon it?  Sure summer is fine, but the fall and winter is coming here in north america, so temperatures will be cooler and I will be closer to face-plant?  Can you report back what Ninebot thinks of the problem?  

[KaleOsaurusRex]

@dpong

I want answers too. Let's do this.

@hobby16

I want to make sure that I have all the ammunition that I need when I approach Ninebot about the BMS issue. Could everyone help by linking here to the most specific reports and analyses? I'm going to go through the thread and the forums to compile everything I can when I have a minute, but I can use all the help I can get on this one.

[esaj]

I don't recall hearing any reports where the Ninebots BMS would clearly have actually cut power, but of course if the protection circuits are there, it can happen. Maybe they just use more "sane" under voltage/over current limits than other manufacturers. When I was looking through some BMS's (which had configurable undervoltage protection), the default was 3.3V (from what I've read, the safe limit for Li-Ion is around 2.5V, and can even go temporarily lower under load without damage)! So if using such BMSs and leaving the protection to default settings, that probably is high enough to trigger on strong acceleration/uphill/more depleted battery (a gut feeling), considering that wheels seem to let you to ride them until 3.5V or below voltage before stopping you, and the voltage will go much lower temporarily during riding (especially under high load like acceleration or hill climbing, or cold environment). I'm planning on adding a voltage / current -display on my wheel with the new batteries, so I can see the voltage & current in real time while riding... the BMSs in my packs should have undervoltage protection around 2.9V, which might still be a bit high, but if need be, I'll shunt at least one or two of the packs but lose the warranty on them then.

 

[high'tems]

@KaleOsaurusRex,

We will be distributor of KingSong in Europe. Could you please update your post at the top? 

here is our link,http://hightems.eu/ 

Thank you mate

[KaleOsaurusRex]

@high'tems , I think you're looking for the "where to buy thread." Esaj is the one to contact.

[high'tems]

@KaleOsaurusRex,

cheers mate, I already update the file and inform her. 

[BlackM]

Ninebot has a similar issue as the unexpected shutdown except it happens because you charge while running downhill a feature only Ninebot has (to the best of my knowledge), it only happens if you attempt to start a ride while having a fully charged battery with going down hill immediately because then it will be charging a fully charged battery which is bad. Easily avoidable if you start off running on flats but definitely worthwhile to note down.

[dpong]

Ninebot has a similar issue as the unexpected shutdown except it happens because you charge while running downhill a feature only Ninebot has (to the best of my knowledge), it only happens if you attempt to start a ride while having a fully charged battery with going down hill immediately because then it will be charging a fully charged battery which is bad. Easily avoidable if you start off running on flats but definitely worthwhile to note down.

I started a trip on 100% battery going down a steep grade and I could feel vibrations in the NB1 E+ pedals alerting me to the regenerative braking problem..  So it was too steep to turn around.. so for safety I simply ceased regenerative braking and let her rip down the hill.   It felt like the safest thing to do given the circumstances and worked out fine.  

[KaleOsaurusRex]

Major update to main post at the beginning of the thread. At your leisure, smart people please rip it to shreds for honesty, completeteness, clarity, etc!