Minipro intermittently LIFE-THREATENING unsafe, veering right

[trevmar]

11 hours ago, mrelwood said:

It would be interesting to rig extra batteries to the MiniPro.

I have looked at how the miniPRO firmware is tied together, and one problem seems to be in how the control firmware assesses the size of the battery (mini or PRO). There are current measuring resistors on the control board, but they don't seem to do anything (because I made their resistance smaller and there was no difference in performance). I think a primary problem may be how the firmware handles the drop in battery voltage under heavy load or heating. I rode my new Airwheel S5 up to the same point where my miniPRO dislocated my shoulder, and I was amazed to see the battery indicator (the S5 has a built-in 4" LCD display) down at almost zero, until I stopped riding so hard and let everything cool down. The S5 has twice the battery size as the miniPRO (680Wh) so it is clear that my 200lb weight, plus the slope of a normal, if steep, paved road was putting extreme stress on the miniPRO. No wonder it kept complaining on the way up I just wish it hadn't dumped me unceremoniously on the way down. I measured the road at 12-13 degrees with the "angulo" app on Android.

I found that one miniPRO wheel hitting a minor pothole or not-so-big rock can cause the machine to go into a similar shutdown sequence, probably because the miniPRO also senses similar 'overload' from that obstacle. It should, of course, be trying to allow momentary surges. I note that I saw a video of an EUC (a Gotway Nikola, I think) where the rider had connected a battery-current alarm at 90 amps, and it was being frequently triggered up an unpaved hill by the powerful motor in that EUC. As far as I can see, the miniPRO is limited to 20 amps max from its batteries, maybe as high as 30, but nowhere near 90, and it shuts down rather than powering through the momentary overloads.

Edited August 15, 2019 by trevmar

[mrelwood]

5 hours ago, trevmar said:

the rider had connected a battery-current alarm at 90 amps, and it was being frequently triggered up an unpaved hill by the powerful motor in that EUC.

I’m familiar with the video you mention. Gotway wheels don’t report the current it draws from the battery, but the current at the motor phases after a voltage conversion. (Not sure if it is correct even then.) I ride a Gotway MSX daily, and I can reach (reported) 150A and over 8000W at a steady slow speed on a 13 degree uphill. The much less powerful 16S is able to get up the same hill with similiar riding reporting only 30-40A, if my memory serves me correctly.

Anyway, MiniPro shutting down at extreme stress sounds like the issue from the early EUCs BMS, which was soon understood to be the wrong approach. I think none of the EUCs manufactured since 2016 or at least 2017 does that. I seem to recall that MiniPro entered the markets in 2015.

But since the other wheel keeps on spinning, it is not a BMS issue, nevertheless the priority of the safeguards in place sound alike.

[mrelwood]

I got interested so I just made a test. There is a short but steep incline leading to our parking lot. I rode each vehicle up the hill, trying to ride identically. I graphed the current reading in the app DarknessBot for each vehicle. The peak current was:

Gotway MSX (2000W, 1600Wh, 18”): 100A

KingSong 16S (1200W, 840Wh, 16”): 20A

NineBot MiniPro (400W, 310Wh, 10”): 35A

The reason for the high amps on the MiniPro were the acceleration rushes that followed when the overload tilt-back slowed down my speed to a crawl and therefore stopping for a moment just at the few tiny dips in the pavement.

Fastest acceleration bursts I’m willing to do on each vehicle on flat ground drew the numbers 100A, 20A, 10A. Although, I did once reach 18A on the MiniPro, followed by beeping and a tilt-back.

Worth noting is that on MiniPro firmware 1.4.0 I was not able to ride up the incline at all, as the tilt-back prevented me completely from continuing. The test above was made with the more allowing 1.1.7.

[trevmar]

3 hours ago, mrelwood said:

MiniPro firmware 1.4.0 I was not able to ride up the incline at all ...

I got interested so I just made a test .. The reason for the high amps on the MiniPro were...

I agree.  I had to discard Swallowbot because it started to tilt back on the uphill roads around here, due to the 1.3.1 firmware underlying it. I found both 1.3 and 1.4 were unusable to me for that reason.

How do you display current on the MiniPRO? I have tested the 240Wh and 310Wh batteries on my testbench, but have not been able to figure out a simple way of getting data from the controller itself (except via the JTAG programming equipment).

[trevmar]

As for miniPRO BMS: I built a pulsed Battery Tester capable of drawing up to 27 Amps out of the battery pack for half a second, with a 1/16 duty cycle so we don't overheat any components, or blow any fuses. The LG battery cells are rated at 10Amps continuous, there are two in parallel in the battery packs, so I am drawing more than that to make sure that I will see any difference under peak load between the battery packs. Here is a post where I gave the results I saw, there were no cutouts at 20 or 30 amps (pulsed)(>1500W). It looks to me as though shutdown is a control-firmware issue.

https://forum.electricunicycle.org/topic/10991-minipro-batteries-4300mah-or-5700mah-boondoggle/?do=findComment&comment=189387

Edited August 16, 2019 by trevmar

[mrelwood]

2 hours ago, trevmar said:

How do you display current on the MiniPRO?

Incredibly unscientificly, that’s how...  I connected with the app DarknessBot to each vehicle, and it only displays whatever values the vehicle decides to send. The app shows readings and can draw graphs for speed, battery level, voltage, current, power, temperature and mileage, amongst a few others.

I would give the mentioned values a precision of roughly +-30%. Except for Gotway it would be +(-)80%.

I only did the comparison to show that while Gotway reports these huge numbers, it is not possible that their wheels draw five times the power compared to a KingSong or the MiniPro in an identical situation.

But that is not to say that I wouldn’t agree with the MiniPro handling the overload situations badly. Luckily though I haven’t gotten the single wheel cutout on the MiniPro that causes hard veering left or right.

[trevmar]

On 8/16/2019 at 1:04 AM, mrelwood said:

I only did the comparison to show that while Gotway reports these huge numbers, it is not possible that their wheels draw five times the power compared to a KingSong or the MiniPro in an identical situation.

My thoughts exactly, although my feelings are based on the number of NCR18650GA 3500maH Li-Ion cells in the 20-cell 84V Gotway battery pack I can see in the tear-down video, with each '84V x 800W'  bank of cells having an <18Amp max repetitive peak rating (peak not cited for GA series, but PF series is 18A peak for 5-6 seconds). With the 6 parallel cell-packs of GA they would get up to 15 x 6 = 90A, so 90 amps is feasible, especially if they over-rate the cells a little. The cells are reported to get very hot... (63-75 degrees centigrade).

The LG cells in the MiniPRO 310Wh battery are only rated at 15Amp peak (x2) = 30 Amps peak... But that's still 1500 watts peak, which is 2 horsepower. It ought to be enough, I would have thought. I suspect that they cut the motors off at a lower value, to 'protect' the MOSFETs (or whatever). Or perhaps 35A is their cutoff, and they frequently bump into it..  I am pretty certain that v1.1.7 firmware allocates a lower peak current when it senses a 'mini' 234Wh battery pack instead of the 310Wh pack.

..

Edited August 18, 2019 by trevmar
Modified the Gotway battery values based on new data.

[mrelwood]

13 hours ago, trevmar said:

...

I agree. You have obviously thought this through with great detail.

13 hours ago, trevmar said:

I am pretty certain that v1.1.7 firmware allocates a lower peak current when it senses a 'mini' 234Wh battery pack instead of the 310Wh pack.

Lowering the absolute maximum cut-off limit based on the choice of battery wouldn’t make sense for protecting the Mosfets though. Based on your calculations they might have been just overly careful in protecting the battery.

OTOH, switching off only one motor to save the battery (or mosfets, or any component that serves both motors) doesn’t make any sense either. I’m not quite convinced that the one-sided cut-off is a purposeful decision after all.

Wouldn’t it have been nice if the Inmotion variety would’ve had 2x600W motors, a 480Wh battery and 12” tires...  Well, one can dream. At least my credit card might’ve been genorous enough for one.

[trevmar]

21 hours ago, mrelwood said:

Lowering the absolute maximum cut-off limit based on the choice of battery wouldn’t make sense for protecting the Mosfets though. Based on your calculations they might have been just overly careful in protecting the battery...

Wouldn’t it have been nice if the Inmotion variety would’ve had 2x600W motors, a 480Wh battery and 12” tires...  Well, one can dream. At least my credit card might’ve been genorous enough for one.

The MOSFETs are rated at 100V in the MiniPRO, so that is not a factor. But as WEagle95 confirmed, the interrupt handling is very poor in the firmware, and there is no exception vector should a CPU bus fault occur. This is bad programming practice, but something I noticed as soon as I started to dig through the code, as there are too few vectors in the boot table, and therefore many unhandled exceptions.

Inmotion? I went for Airwheel S5. Even their S3 has 1000W of motor power, and twice the battery (520Wh) of the miniPRO (rated 220Kg load). It also has 14" tyres. I chose the S5 model with 19" OD tires, 3.5" wide, 1500W of motors and 680Wh battery. Problem is that the S5 has been removed from their range now. Mine came from April 2019 production (date was stamped on the outside of the box it came in). Still, these have the same brushless motors as the unicycles do (unlike the Segway I2 and X2 with their gearboxes) and maintenance ought not to be a problem. In fact, as I was watching the Gotway strip-down I couldn't help but notice the Nikola motor looked the same as the 12" rimmed motors on my S5

Edited August 18, 2019 by trevmar

[mrelwood]

I checked out the Airwheels earlier, and due to their enormous weight they couldn’t replace my MiniPro. The key benefit and feature for me to even own the MiniPro was that it’s easy to lift and fit into my car.

[trevmar]

Yes, the extra batteries and tyres are heavy! The Unicycle really does present the best 'portable' option, IMO. They are still heavy, but the weight is manageable, especially for those with the handle..

[Arizona Ruta La Apa]

Hi guys this is what happened to me today. I was noticing strange noises of the motors(as well as vibrations) I went later on i felt the strange noise as a "gear was engaging" and begun to go slowly. and eventually got off. Got dow, turned of and on the minipro and  saw it veering on the left on its own and fall. the app error message was "Warning your vehicle is tumbling accidentallly. Please drive carefully"  Minipro on and of a couple of times and as soon as I touched with a finger the strrering bar (knee bar) it begun kinda to correct something and veered again to the left and cut off.

 

More details, many of em ;)

Today i took of the left tire to clean the slime in the interior.

To seal it in a tire shop they put it to 3 bars in order to "fix" itself.

I noticed some strange noises afterwards but I told myself should be for the pressure of the tires.

I've done almost 1000km on the segway mini pro and always had trusted it 100% Never a problem till today.

I took the mini light of my 4 year old daughter and had a ride with her as always we do.

we go to a dirty road( as countless times before)  but i notice it to shaky and doing a strange noise

I begin to check the sensivities through the app ( I remebered there's an otion to reduce vibrations)

Turned off the auto mode that was enabled, tried but nothing changed.

 

At a certin point, it begun to veer on the left and fall on the floor ( I tried to replicate and record it)

Touching the kneebar slightly, the unit begun kinda to correct itself, and abruptly turn on the left.

I turned it off and went to fix the tires pressure to 2.4 bar that is what i used to have 90% of the time.

 

Nothing changed, it steered to the left and seemed crazy.

Did a reset, and kept doing the same a couple of times.

Firm v 1.40

total mileage: 2043 km

Sn N3mpd1614t1349

Battery status: normal

BMS firm v1.1.3

 

Hope all this info helps someone. 

Anyone who might know what I should do next, please speak up :)

[Rider222]

What you are describing is VERY different than the issues described in my original post.  Your issue isn't the very rare intermittent exceeding dangerous issue I was having.  Your unit is showing itself to not be functioning properly.    Obviously something has broken on your unit; maybe the factory can help get your unit fixed

[John1973]

This just happened to me after 2-3 years of use.  All of a sudden you hit a medium bump and it vers 90 degrees to the right.  I was able to replicate this on a bad sidwalk where both tires felt into the crack about the same time and hit pretty hard.  It did a 90 degree to the right . it threw me off and it kept circling to the right slowly.  I had to turn it off for a few minutes to reset it.  I did this again and the same thing happened.  I had to re calibrate it this time and it made it worse.  I shut it off for a bit and then ran over it a third time.  same thing.  hard right turn.  I had to re calibrate it it and then it seemed to make it worse, it would happen over small bumps.  I'm a mechanical engineer.  I'm not an electrical, but with all the vibrations would it be sensor/input/magnet motor related?  if you would like to text 920-397-0302  John

[Timmy0222]

This exact same thing happened to me last week on my nine bot miniPro and was the second time for me.  The first time about 3 months ago, I thought it was a fluke, after the sudden right turn, I flew off and with a couple of big steps managed to crash into a lawn, saving skin damage.  I had decided it wasn't going to happen again, but in fact it did, this time I ended up with a broken toe and a broken finger.  I will never ride it again.  But I do miss the little guy, just can't trust it.  Upon review of both events, there was no rocks, cracks or any other obstacle, just straight riding at medium speed.

[Rocky Romero]

20 hours ago, Timmy0222 said:

This exact same thing happened to me last week on my nine bot miniPro and was the second time for me.  The first time about 3 months ago, I thought it was a fluke, after the sudden right turn, I flew off and with a couple of big steps managed to crash into a lawn, saving skin damage.  I had decided it wasn't going to happen again, but in fact it did, this time I ended up with a broken toe and a broken finger.  I will never ride it again.  But I do miss the little guy, just can't trust it.  Upon review of both events, there was no rocks, cracks or any other obstacle, just straight riding at medium speed.

This is a good reason for ongoing MiniPro maintenance and precautions.

Wearing a helmet is a must and other protective gear optional.

Still I will ride.

➰

[Scott nemo]

I have had this veer right issue 3 times now. No bumps or irregularities. Also after it happened to my kid the unit was spinning in right hand circles for a little bit. There has to be something wearing on these causing the same exact issue. Always seems to he right. I will not allow my kid to ride and I am not sure I will ride it anymore. This is a very serious issue. And I am not an alarmist or really get worked up over things. But throwing you off at 11 mph unexpected onto concrete is a big problem. 

[MetricUSA]

Or it is rider error....where do you place your feet??? Close to the center or as far apart possible??? A simple nudge of the control arm can make unit turn on you...KEEP legs as far apart as possible...when you lean into turn, leg then touches the arm and you turn....some said, turn happens when riding over crack or stone...well dah...you accidentally hit control arm...

[HBoard]

I have a Ninebot S that experienced the same issue about 4 weeks ago which resulted in a broken first metatarsal. I’m 5’8” about ~155 pounds and have ridden over 1700 miles. I had 100% trust in my device until the incident where I was going 8-10 mph on a relatively smooth sidewalk and the device without warning turned 90 degrees to the right throwing me off breaking my foot in the process. I should of seen the warning signs, as my device had been turning 35 degrees to the right unexpectedly but after calibration worked just fine. 5 minutes before the ride the device was slowly veering to the right but one again after calibration gave me no indication not to trust the device. I have sent in the device for warranty repair and waiting for the device back, but I’m hesitant to trust the device again. After the accident and power cycling the device it started going in circles. Over the 1700+ miles and 11 months I owned the device I have had to jump off a a couple of times due to pavement issues but 80% of the time landed on my feet. This issue is very dangerous as your body doesn’t know what’s coming leaving little time to react. 

[Dalbster]

I have had the same problem and actually fractured my Radial head in my fore arm. Mine veered off to the right and threw me onto my elbow I thought of selling it but I didn’t want anybody else to get hurt on it. It has done this a few times before but it did not hurt me but this time it did. Anyway I thought about getting the handle bar attachment and thought that that might keep it from veering to the right if I’m holding the handle bar while riding it.  Any thoughts from anyone on that 

[trevmar]

@DalbsterMy MiniPRO units did the same thing to me. The last time my arm was out of action for 2 months. I was using a semi-flexible steering column, it made no difference to the MiniPRO> I bought Airwheels S5 units (their 1500W off-road units) and have had no trouble. Flawless. Slow (12KM/h), but a good complement to my 2400W scooter and Fat Bear Plus ebike. For a slow and casual ride around the neighborhood.  I have four miniPRO in the garage for much the same reason as you are keeping yours...

Edited December 11, 2020 by trevmar

[Calmclear]

On 7/26/2019 at 11:22 AM, cleanoldman said:

Accomplished rider here. Many happy miles until March 2017. Same sudden turning behavior described here left me with a severely torn shoulder. Surgery was required. Recovery was over a year. Only after Segway read the 'black box' data, did they ship me a new one. This thread is the first I've found to acknowledge the problem in depth. I'm afraid someone is going to get killed. I truly love this machine but now my trust is gone.

I can’t Believe I finally found others experiencing this. We baby our Segway. All 3 of them randomly at one point did the sharp right turn. All on smooth level flat pavement. Segway sent me a brand new replacement model. About 3 mo later the new one did it. It’s not me it was a completely different rider who I was following behind. 
 

it’s such a random bug that’s near impossible to reproduce but I PROMISE you if you have your Segway 9bot it will happen at some point  we have over 4,000 combined KM on our Segway scooters. I’m terrified we won’t ride them any more  it doesn’t sound dangerous but imagine going straight 10mi and suddenly flying sharp to the right  I flew probably 10 feet and almost went over the canyon ledge on one road  I have scars across my face from skidding across the pavement  about a month later it happened to my wife  this was after having the scooter over 15 mo without issues  I promise it’s not user error  it happens suddenly without warning with no change in speed or posture  

 

 

 

 

 

[Calmclear]

I can’t Believe I finally found others experiencing this. We baby our Segway. All 3 of them randomly at one point did the sharp right turn. All on smooth level flat pavement. Segway sent me a brand new replacement model. About 3 mo later the new one did it. It’s not me it was a completely different rider who I was following behind. 
 

it’s such a random bug that’s near impossible to reproduce but I PROMISE you if you have your Segway 9bot it will happen at some point  we have over 4,000 combined KM on our Segway scooters. I’m terrified we won’t ride them any more  it doesn’t sound dangerous but imagine going straight 10mi and suddenly flying sharp to the right  I flew probably 10 feet and almost went over the canyon ledge on one road  I have scars across my face from skidding across the pavement  about a month later it happened to my wife  this was after having the scooter over 15 mo without issues  I promise it’s not user error  it happens suddenly without warning with no change in speed or posture  

 

[Calmclear]

Is there an alternative? I absolutely love the profile of the Segway mini. But I can’t ever ride it again for fear of the violent right turn. If someone could suggest a different brand that works similarly. Riding these self balancing 2 wheel scooters used to be my favorite activity. 

[mrelwood]

23 hours ago, Calmclear said:

Is there an alternative? I absolutely love the profile of the Segway mini. But I can’t ever ride it again for fear of the violent right turn. If someone could suggest a different brand that works similarly. Riding these self balancing 2 wheel scooters used to be my favorite activity. 

Inmotion now makes a direct competitor to the Segway Mini Pro/Plus, model Inmotion E3:

https://www.inmotion-france.fr/en/self-balancing-scooters/312-inmotion-e3-self-balancing-scooter-3760243825514.html