Help! My Ninebot One E+ won't start, and the wheel is very stiff / jerk-ish.

[veggies]

I was standing on it leaning back against a wall, waiting for a light to turn green, when it suddenly died. 

 

Symptoms: 

-Emits a single beep when button is pressed, then nothing. No bluetooth, no power. 

-If you disconnect and reconnect the power on its side, it gives a long series of beeps until uprighted, and then goes totally silent again.

-The wheel spins, but it’s incredibly stiff and kind of goes through a ‘bump’ every couple of inches. So if I turn the wheel by hand I feel a rhythmic ‘buh-buh-buh-buh’, extremely stiff, like the brakes are on.

 

Someone on the internet [Edit: dmethvin, and thanks for pointing me in the right direction) said I probably fried my MOSFETs that I might need a new control board. I disconnected the battery and then the board to verify this, but the behavior in the wheel is unchanged.

Thanks to anyone who might know what's going on!

Edit / Update: Issue resolved. Jake (who is awesome) at forwardca just had me send him a brief video of the issue and then diagnosed the issue and literally mailed me a replacement control board within hours. I can't speak highly enough of that business.

(I'm new here so I'm not sure if standard etiquette is to post updates in the comments or the original post, so my apologies for doing both.)

[dmethvin]

I talked with @veggies over on the EUC section of Reddit and sent him here because there are probably more people who could help to diagnose it. The symptoms sounded like MOSFETs to me, any other thoughts?

 

[AceFace]

Yup, sounds like a blown mother board.  Unfortunately there have been lots of cases of this happening dependent on your NB1 version, mother board version, software version and how you use (abuse) the bot.   My first NB1 (Serial N10) blew two MB, once from uploading a faulty software update and the second time during only gentle use, but I've have had no trouble at all with my later N20 serial number.  I'd suggest you contact your retailer to find a resolution.

Post all the details of your NB1 and what you think caused the issue (i.e. heavy braking) for others to see.  There are numerous threads on the Ninebot forum to post these details to.

 

[esaj]

Certainly sounds like one or more mosfets have short circuited, especially when the motor brakes even without any power. Probably deed a new mainboard, of course you could try to replace the mosfets (with the same model there's already), but you can't be sure if something else has blown also.

[Hunka Hunka Burning Love]

How old is the wheel?  Any warranty?  That's strange when the battery is disconnected that the wheel exhibits the same resistance behaviour.  I wonder if there could be something jammed around the axle like some hair or string?  If there's no warranty, maybe taking it apart for a check might be worthwhile.  There also could be something broken inside the motor?  Maybe a magnet has come loose or there could be some warpage of the rim?  Do you do any tricks or drops?  Do you keep the air pressure high near 40 psi?  The jam may have fried the control board or maybe it senses the problem and is just shutting down for safety?

[veggies]

Update: Issue resolved. Jake (who is awesome) at forwardca just had me send him a brief video of the issue and then diagnosed the issue and literally mailed me a replacement control board within hours. I can't speak highly enough of that business.

(I'm new here so I'm not sure if standard etiquette is to post updates in the comments or the original post, so my apologies for doing both.)

[Hunka Hunka Burning Love]

Darn those Forwardca guys... So fast to help repair people's problems.    Give the rest of us EUC repair wannabees a chance at helping too! 

[esaj]

10 hours ago, HunkaHunkaBurningLove said:

That's strange when the battery is disconnected that the wheel exhibits the same resistance behaviour.

The motor brakes when two or all three phases are shorted together. I've tried it with the Firewheel motor, and with all three phases, the braking is really strong. Basically the back-EMF of the motor itself (caused by turning it) will feedback to the phases, causing the motor to resist its own movement   I believe the regenerative braking works by switching between shorting the phases through the low-side bridges to build up the magnetic fields in the coils and then allowing the coils to discharge towards the batteries through the high-side bridges. There's a couple of posts in the electronics stackexchange that explains how the regenerative braking works, and how half-bridges act like boost converters to raise the voltage, so the current actually changes direction when the batteries are being recharged during braking:

http://electronics.stackexchange.com/a/56187

http://electronics.stackexchange.com/a/56171

So, I'm still claiming it's caused by phases shorted together through burnt mosfets

 

EDIT: Oh, I also tried with the motor + home-made half-bridge and did some measurements, regenerative braking in action with oscilloscope-screenshots:

 

[Hunka Hunka Burning Love]

Thanks for the info!  I haven't played around with these types of motors before, but I can see from your explanation how the energy generated by spinning the motor feeding back can create resistance to spinning when some of the phases are shorted together through the mosfets.  Very interesting.  On a normally functioning EUC the wheel spins freely and I suspect that the voltage created goes towards charging the batteries even with the control board turned off?  In that one Solowheel video on how to recover a dead battery, they spin the wheel on the ground to try to get some voltage going through the circuit, and I think the charger takes over from there?  So in veggies case, if he were to disconnect the motor wires from the control board the wheel should spin freely once more.

[esaj]

32 minutes ago, HunkaHunkaBurningLove said:

Thanks for the info!  I haven't played around with these types of motors before, but I can see from your explanation how the energy generated by spinning the motor feeding back can create resistance to spinning when some of the phases are shorted together through the mosfets.  Very interesting.  On a normally functioning EUC the wheel spins freely and I suspect that the voltage created goes towards charging the batteries even with the control board turned off?

Not actually sure, basically there's a "body diode" in the mosfet that allows the current to travel through it in "reverse", if the voltage on the source-side (for N-channel mosfet) goes high enough, but the problem is that on the low-side, the mosfet diode is against the ground, but I guess the voltage could become negative on another phase at the same time, to allow the current to flow..? Ie. on one phase, the voltage is higher on the source-side (motor phase) than drain-side (battery positive) of a high-side mosfet, and at the same time, on another phase, the voltage on the drain-side (motor) is lower (negative) than battery ground on the source-side of a low-side mosfet. If you can follow my train of thought...

Or the energy just vanishes through friction in the bearings etc.  

 

Quote

 In that one Solowheel video on how to recover a dead battery, they spin the wheel on the ground to try to get some voltage going through the circuit, and I think the charger takes over from there? 

I think I've seen that video, but don't remember it that good. My best guess is that the idea is to "wake up" the BMS, which could have turned off if the battery voltage has gone low enough? But again, not that sure on these things...   At least I've read that protected cells (the kind which have a very small BMS inside them) can "die" if the voltage drops so low that the BMS inside them turns off.

 

Quote

So in veggies case, if he were to disconnect the motor wires from the control board the wheel should spin freely once more.

Yes, if nothing is shorting the phases and the motor is otherwise mechanically ok, it should spin freely when it's disconnected from the mainboard (again, assuming the mosfets on the mainboard are shorting the phases).

[Hunka Hunka Burning Love]

It seems like MOSFETs are the weak link on the control board as they seem to burn out so commonly.  I wonder if there is some other electronics component that could be substituted which would be more durable in various situations, or are they the only choice available that works?  My electronics knowledge ended back in high school when I took a couple of basic courses, but you would think there would be some other options to choose from.

 

[esaj]

1 minute ago, HunkaHunkaBurningLove said:

It seems like MOSFETs are the weak link on the control board as they seem to burn out so commonly.  I wonder if there is some other electronics component that could be substituted which would be more durable in various situations, or are they the only choice available that works?  My electronics knowledge ended back in high school when I took a couple of basic courses, but you would think there would be some other options to choose from.

Some people have suggested IGBT's, but I know very little of them. They're some sort of "hybrid" between BJT's (bipolar junction transistors, ie. the "traditional" kind) and mosfets. Apparently they're used with very large currents and can take heck of a beating before being destroyed... But probably also not a cheap option.

[Hunka Hunka Burning Love]

Just browsing a bit, but it looks like some cost about 7 british pounds that can handle 120 A.  What would be a good rating to ensure that they don't burn out during any circumstances?   I don't know if these are directly swappable, but I was just looking at some prices...

http://goo.gl/EFhfQI

[esaj]

19 minutes ago, HunkaHunkaBurningLove said:

Just browsing a bit, but it looks like some cost about 7 british pounds that can handle 120 A.  What would be a good rating to ensure that they don't burn out during any circumstances?   I don't know if these are directly swappable, but I was just looking at some prices...

http://goo.gl/EFhfQI

Well, I know next to nothing about IGBT's, but I doubt you could just throw them in place of mosfets. Probably the entire mainboard needs to be redesigned to work with them.

[MetricUSA]

With the MOSFETs...they are all in parallel, right? For them to spread the current/power across all equally... how do you find all the bad shorted ones, if not for overheating/burned or discolored board? Do you need to desolder one at a time and test? There is some ten+ to check...

[steve454]

Wish I could help, but when I was in BEE (basic electrical and electronics, I woke myself up from sleep by saying aloud, But where does the current go? 

[Hunka Hunka Burning Love]

It looks like there's no easy way to test mosfets in circuit as it seems the recommended method is to remove and test.

http://www.electro-tech-online.com/threads/testing-fets-in-circuit-with-an-ohm-meter-and-an-external-dc-power-supply.139192/

That's one of the reasons why I think it would be nice to have these MOSFETs on a daughterboard that has a wire connector that plugs into the control board.  That way they can be positioned more ideally for cooling and testing would be easier as well as you have access to the connector's end terminals.   In addition, changing out a MOSFET daughter board might be less expensive than the entire control board.  @Polpus successfully moved his MOSFETs off the control board and into the battery compartment with a separate heatsink I believe without affecting the EUC's normal function.

[veggies]

For what it's worth, I disconnected the board and battery both, and the same resistance was given. I did get my replacement control board in the mail today, so later today I will install it and see if that fixes it. Fingers crossed.

[Hunka Hunka Burning Love]

In your other post you mentioned you had a rubbing noise issue.  Did you add a washer?

If you disconnected the board and battery from the motor leads, shouldn't the wheel spin freely unless there are some wires shorted out from the motor somewhere?  I checked mine just powered down, and it spins freely with a woo woooo sound so I still wonder if there is a physical issue (or electrical short) related to the resistance you are feeling.  It's too bad these wheels don't have an error blink code that can help narrow down problems.

A bit over my head, but interesting reading if you're into electronics:

http://electronicdesign.com/power/igbts-or-mosfets-which-better-your-design

[veggies]

1 hour ago, HunkaHunkaBurningLove said:

In your other post you mentioned you had a rubbing noise issue.  Did you add a washer?

If you disconnected the board and battery from the motor leads, shouldn't the wheel spin freely unless there are some wires shorted out from the motor somewhere?  I checked mine just powered down, and it spins freely with a woo woooo sound so I still wonder if there is a physical issue (or electrical short) related to the resistance you are feeling.  It's too bad these wheels don't have an error blink code that can help narrow down problems.

A bit over my head, but interesting reading if you're into electronics:

http://electronicdesign.com/power/igbts-or-mosfets-which-better-your-design

I'm sure that's not over your head. If you watch the whole video he just takes it apart (just screws and sockets) and adds a washer to increase the space between the two sides by a millimeter or two. I thought that video might be what I needed to go along with, but I don't have the socket wrench and crank, so couldn't try that solution - fortunately Jake from forwardca wound up letting me know I had just clumsily re-inflated my tire, so the valve was scraping against the shell. I basically just removed / re-added the inner tube and tire and then was very careful that the valve stay at the correct angle. That fixed the rubbing noise perfectly.

Now to see if I can get it to turn on! ?

[Chriull]

11 hours ago, esaj said:

Not actually sure, basically there's a "body diode" in the mosfet that allows the current to travel through it in "reverse", if the voltage on the source-side (for N-channel mosfet) goes high enough, but the problem is that on the low-side, the mosfet diode is against the ground, but I guess the voltage could become negative on another phase at the same time, to allow the current to flow..? Ie. on one phase, the voltage is higher on the source-side (motor phase) than drain-side (battery positive) of a high-side mosfet, and at the same time, on another phase, the voltage on the drain-side (motor) is lower (negative) than battery ground on the source-side of a low-side mosfet. If you can follow my train of thought...

You could look what thecircuit simulator says if one has this situation: Battery with ~60V the two body diodes (one from the plus side, one from the negative side) and the circuit closed with a second battery with ~70V (simulating the motor).

Edit: The longer i look at the bridge circuit, the unsafer i get about this statement - If one puts both of this diodes "on one side" they should just start conducting, but with one diode in every path i just get some headache  i hope your simulator gives as an result.

I'd assume without the low side beeing switched on, there is no way for current could to flow from the motor to the battery. (if the simulater does not say it otherwise...;) ) There is always a (body) diode antiparallel... Just if the motor voltages goes above the breakdown voltage of the mosfets (75V for the 75NF75) - but if this can happen it would mean an uncontrolled bridge behaviour which could easily lead to a bridge shoot-through...

Quote

Or the energy just vanishes through friction in the bearings etc.  

This should be the way to go for a turned off EUC. If breakdown voltages are reached this should be an extreme design fault...

11 hours ago, HunkaHunkaBurningLove said:

It seems like MOSFETs are the weak link on the control board as they seem to burn out so commonly.

Imho not. Its the inproper cooling that makes them the weak link - properly cooled they have absolutely no problem driving an EUC. Given that there are no design flaws in the circuit - but BLDC motor drivers exist already long enough and should (hopefully) be no prob for someone working in this field.

11 hours ago, esaj said:

Some people have suggested IGBT's, but I know very little of them. They're some sort of "hybrid" between BJT's (bipolar junction transistors, ie. the "traditional" kind) and mosfets. Apparently they're used with very large currents and can take heck of a beating before being destroyed... But probably also not a cheap option.

IGBT's have at the on state a minimum Collector Emitter Voltage (Vce sat) like Bipolar transistors, as Mosfets have an minimum resistance while on (Rds on).

10 hours ago, HunkaHunkaBurningLove said:

Just browsing a bit, but it looks like some cost about 7 british pounds that can handle 120 A.  What would be a good rating to ensure that they don't burn out during any circumstances?   I don't know if these are directly swappable, but I was just looking at some prices...

http://goo.gl/EFhfQI

So the IGBT in this example has a Vce sat (minimum collector emitter voltage) of 2.1V. Assuming an quite average (a bit burdened) current for a Ninebot One E+ of 15A this would mean a power dissipation of 2 times (upper and lower side of the bridge) 2.1V*15A~30W.

With the 75NF75 one has an Rds on of ~11mOhm. With 2 in parallel that's ~6mOhm. So this leads to 2 times (again lower and upper bridge side) 6mOhm*(15A)²=1,35W.

So with these IGBT's one needs to get rid of 2*30W=60W power dissipation while quite normal driving - with the Mosfets one need only to get rid of 2*1,35W~2,7W power dissipation. (ignoring the switching losses for this example, which could be about the same for both drivers)

The advantage of IGBTs is once the (much) higher possible "blocking voltage" and on the other side that by the "fixed" collector emitter voltage dissipated power gets less with high currents then with an fixed resistance (if this "intercept" point is still within the specifications) . In the comparison of this IGBT and the 75NF75 this would be true for currents above 350A 

So IGBT's are imho just no choice for EUCs.

9 hours ago, MetricUSA said:

With the MOSFETs...they are all in parallel, right? For them to spread the current/power across all equally... how do you find all the bad shorted ones, if not for overheating/burned or discolored board? Do you need to desolder one at a time and test? There is some ten+ to check...

The Mosfets with an internal short circuit should be easy to be identified just with an ohm meter, if you disconnect the motor from the controll board. Just between each two in parallel one cannot distinguish. But if one mosfet of two in parallel burned one should exchange both in any case.

(With this ohmmeter test one can only see if the MOSFET is dead - not if the mosfet is still in good shape)

Imho exchanging all Mosfets after such an "accident" is not the worst idea. Also checking all the cables, connectors and "high current PCB tracks"

[esaj]

1 hour ago, Chriull said:

You could look what thecircuit simulator says if one has this situation: Battery with ~60V the two body diodes (one from the plus side, one from the negative side) and the circuit closed with a second battery with ~70V (simulating the motor).

Edit: The longer i look at the bridge circuit, the unsafer i get about this statement - If one puts both of this diodes "on one side" they should just start conducting, but with one diode in every path i just get some headache  i hope your simulator gives as an result.

I'd assume without the low side beeing switched on, there is no way for current could to flow from the motor to the battery. (if the simulater does not say it otherwise...;) ) There is always a (body) diode antiparallel... Just if the motor voltages goes above the breakdown voltage of the mosfets (75V for the 75NF75) - but if this can happen it would mean an uncontrolled bridge behaviour which could easily lead to a bridge shoot-through...

Simulating a motor behavior isn't exactly easy (atleast LTSpice doesn't have a ready-made component for that), but maybe it could be tested with couple of sine-wave (AC) voltage sources... if you consider there are two phases, wouldn't the two phases become opposite polarities at the same time? Of course it needs to be two phases/half-bridges, because the voltage cannot be simultaneously both negative and positive on the same phase

This picture would seem to indicate so (but it's actually just a simple representation of a three-phase system in general, not specific to motors)

So the situation I meant was that there are two different phases, one being at Battery voltage + Mosfet body diode forward voltage + a little more to get current going and the other at battery ground level - body diode voltage - a little more to get current going  Shouldn't that cause current to flow through the phases in "reverse" (in regards to battery polarity), even if the wheel was otherwise powered off? Assuming there's a direct path to battery outside the half-bridges, and nothing in the mainboard that cuts the circuit totally between the battery / half-bridges / motor...

 

Quote

This should be the way to go for a turned off EUC. If breakdown voltages are reached this should be an extreme design fault...

At least as long as any phase doesn't reach battery voltage + body diode forward voltage -voltage levels, I'd expect no current can flow.

 

Quote

Imho not. Its the inproper cooling that makes them the weak link - properly cooled they have absolutely no problem driving an EUC. Given that there are no design flaws in the circuit - but BLDC motor drivers exist already long enough and should (hopefully) be no prob for someone working in this field.

IGBT's have at the on state a minimum Collector Emitter Voltage (Vce sat) like Bipolar transistors, as Mosfets have an minimum resistance while on (Rds on).

So the IGBT in this example has a Vce sat (minimum collector emitter voltage) of 2.1V. Assuming an quite average (a bit burdened) current for a Ninebot One E+ of 15A this would mean a power dissipation of 2 times (upper and lower side of the bridge) 2.1V*15A~30W.

With the 75NF75 one has an Rds on of ~11mOhm. With 2 in parallel that's ~6mOhm. So this leads to 2 times (again lower and upper bridge side) 6mOhm*(15A)²=1,35W.

So with these IGBT's one needs to get rid of 2*30W=60W power dissipation while quite normal driving - with the Mosfets one need only to get rid of 2*1,35W~2,7W power dissipation. (ignoring the switching losses for this example, which could be about the same for both drivers)

The advantage of IGBTs is once the (much) higher possible "blocking voltage" and on the other side that by the "fixed" collector emitter voltage dissipated power gets less with high currents then with an fixed resistance (if this "intercept" point is still within the specifications) . In the comparison of this IGBT and the 75NF75 this would be true for currents above 350A 

So IGBT's are imho just no choice for EUCs.

It seems they can handle much higher power dissipation (as long as properly cooled, of course), but don't know how low the Vce(sat) then can be in some models. Probably not "low enough" for the dissipation power to be lower than in a mosfet, at least if you look only at conduction losses. Switching losses of a mosfet (passing the linear region) could turn the tables, though? But I really don't know...

 

Quote

The Mosfets with an internal short circuit should be easy to be identified just with an ohm meter, if you disconnect the motor from the controll board. Just between each two in parallel one cannot distinguish. But if one mosfet of two in parallel burned one should exchange both in any case.

(With this ohmmeter test one can only see if the MOSFET is dead - not if the mosfet is still in good shape)

Yes, if it has zero resistance, definitely in short-circuit (destroyed), but even if it wasn't, you can't be sure if the mosfet is ok based on resistance alone.

 

Quote

Imho exchanging all Mosfets after such an "accident" is not the worst idea. Also checking all the cables, connectors and "high current PCB tracks"

Worth a try probably, if the Ninebot One -mainboards are still using 75NF75, they're pretty dirt cheap (I bought 25 pieces at some point at something like 6€). Not that big monetary loss, even if it doesn't work out, yet in the best scenario, you get (another) working mainboard.

[Chriull]

12 hours ago, Chriull said:

You could look what thecircuit simulator says if one has this situation: Battery with ~60V the two body diodes (one from the plus side, one from the negative side) and the circuit closed with a second battery with ~70V (simulating the motor).

...

 

10 hours ago, esaj said:

Simulating a motor behavior isn't exactly easy (atleast LTSpice doesn't have a ready-made component for that), but maybe it could be tested with couple of sine-wave (AC) voltage sources...

To see if current can flow a simple dc source is sufficient - i found partsim (an online spice simulator):

http://imgur.com/1hvY5p7

So it seems, that if the generated motor voltage is above battery voltage and two times the forward voltage the circuit is closed and both body diodes are conducting.

Quote

if you consider there are two phases, wouldn't the two phases become opposite polarities at the same time? Of course it needs to be two phases/half-bridges, because the voltage cannot be simultaneously both negative and positive on the same phase

This picture would seem to indicate so (but it's actually just a simple representation of a three-phase system in general, not specific to motors)

...

So the situation I meant was that there are two different phases, one being at Battery voltage + Mosfet body diode forward voltage + a little more to get current going and the other at battery ground level - body diode voltage - a little more to get current going  Shouldn't that cause current to flow through the phases in "reverse" (in regards to battery polarity), even if the wheel was otherwise powered off? Assuming there's a direct path to battery outside the half-bridges, and nothing in the mainboard that cuts the circuit totally between the battery / half-bridges / motor...

I think i like Spline  So here is are the three motor coils as phase shifted AC sources (65V peak each) and the Mosfets reduced to their body diodes:

http://imgur.com/a/ayt8Z

(Edit: These graphs are wrong - did not rename the networks after moving the GND...;(...)

In the diagramm the Phase 1,2,3 are the black,green and blue lines, yellow is Vbat. "Battery" Current is the turquoise line and the rest the three diode currents...

 

[AceFace]

19 hours ago, veggies said:

For what it's worth, I disconnected the board and battery both, and the same resistance was given. I did get my replacement control board in the mail today, so later today I will install it and see if that fixes it. Fingers crossed.

Don't forget you may need to add some heat transfer paste (heat sink compound) between the board and heatsink.

[Rehab1]

@veggies @HunkaHunkaBurningLove @FORWARD california Yes the forwardca people are excellent!