MiniPro batteries - 4300maH or 5700maH. Boondoggle?

[trevmar]

I keep getting thrown forward by my US 2018 MiniPro versions. Haven't walked for the last 2 weeks because of a badly sprained ankle apparently caused by a small loose rock on a dressed-dirt path. Sigh So I now have real incentive to figure out what new Segway safety paranoia has led to this unsafe behavior (cutoff at small obstacles).

I saw the videos on the MiniPlus negotiating steps, etc, and noted that their battery is considerably larger than the US 2018 MiniPro (Plus is more than 6000maH), so I went looking for a MiniPro battery larger than the 4300maH fitted to the US-2018 models.

I just bought another MiniPro from Frys Electronics for $329 (on special this week only). It is Certified Refurbished, and out of the box the battery is 3 bars, and it is working great. But the reason I bought it was for its 5700maH battery. Yes, the refurb battery is 33% larger than the 2018 US MiniPro! Unbelievable...

My US models are running 1.1.9 firmware (I JTAGgged the downgrade) and the new Refurb is still running 1.4.0 out-of-the-box

The hypothesis I am working on is that the Battery electronics is cutting off the motor power at a lower level than the firmware expects. There is a little on-board storage capacity to keep the main CPU going for a second or two after the battery cutout is triggered. I suspect I am getting thrown because I am 200lb, at the limit of the device capability. Incidentally, the refurb MiniPro is only rated at 85Kg maximum

Has anybody already tested the maximum (cutoff) current of these different batteries? I am going to make up a small pulse-discharger to check it out, and see if the peak output current from the US-2018 model has been reduced to match its reduced battery capacity, thus reducing peak motor torque. I expect it has... Grrrr....

UPDATE 7 Oct 2018: My suspicion was incorrect

Rider throwing at obstacles appears to be associated wit the first 50Km run-in period for the MiniPRO. It appears they not only limit the maximum speed a little, but also limit the ability of the MiniPRO to ride-through marginally dangerous situations. Now that I have units which have been run-in, I can see that it is the new (younger) units which get  into the most difficulties. I continue to experiment and observe to figure out exactly what is happening, but I can confirm that on my most mature vehicle there is no observable difference in slope-climbing capability when using the 4300maH N3M260 battery or the 5700maH N3M320 battery, and that the test equipment I built shows that either battery pack can produce more than 1500 watts peak when new...

[jojo33]

@trevmarN3M260 is the hybrid modele existing since the first day into official FCC agreement. This modele was released in Europa in 2016, (Mediamarkt store belgium netherland germany...),

Make that US & European consumers believe that this model was new was really a scam...

Below extract from topic "Do you know all about the MINI? Here some interesting facts" (22 june 2016)

"
Today, a model that has the same specificities as Xiaomi MINI appears in an official document: the MINI PRO N3M260 (see document).The differences with the Xiaomi MINI: only adjustable bar and hubcaps. The engine and the battery will be the same as the Xiaomi MINI.

In pictures below, you can see:

- 3 specs sheets for xiaomi mini and 2 mini pro modele (on the right side it's from austria seller)

- offical FCC where you 'll see the 3 modele name (autorized to be sold to USA and Europa since 2015 !)

- a shematic plan for explain "how to make a mini pro n3m260 (from me   )

 

  

 

 

 

[trevmar]

Excellent info, JoJo! The model being sold as "MiniPro 2018" by Amazon, and others, has the designation "N3M260."  My interest is not so much in 'scam' as whether any safety short-cuts were taken when it was released.

I took the 310Watt battery pack (5700maH) off the refurb (which is N3M320 production date week 23 of 2016) and put it onto the chassis which has knobbly off-road 90/65 tyres on it and firmware 1.1.9  (N3M260, week 33 of 2017).  It works fine, and the correct 5700maH is registered by the firmware (according to the app).

My guess is that the N3M320 model's 5700maH battery is made from LG MG1 28650 18650 cells, while the 4300maH battery probably has LG MF1 cells in it. The biggest question in my mind is whether the motor-drive hardware was changed to reduce the maximum current being drawn from the lower-cost MF1 battery-cells, and if so, I suspect that explains why some MiniPro owners have problems with small obstacles, and some don't.

I noted that the current sensing resistors in the NM260 were 0.04 ohms, which seems a strange value to use for a battery whose cells are capable of 20 amps continuous discharge (less an arbitrary 'safety factor'). I had been assuming that the current level would be sensed by an ADC in the CPU, but it is always possible that the 0.8 volts is intended to trigger a junction transistor for hardware protection...   I need to take out the control board of the N3M320 and see whether its sensing resistors are the same value, and then build a test dummy load which will pulse high currents out of each battery, to find out where the battery hardware cuts out. I am hoping the battery will auto reset after overload, as long as I don't blow the fuse from too much average current. I am aiming for a 1:16 pulse duty cycle and about 0.5 second discharge pulse length.

[jojo33]

 

i think it's more the BMS firmware version who affect 

Here (from french forum) @smallexis dismantle Mini N3M240 battery, https://forum.urban360.com/index.php?/topic/720-lintérieur-de-la-batterie-dun-ninebot-xiaomi-mini/

and here you have a video showing inside new serie battery (blue bms, old series were green bms) at 03:27mn https://v.youku.com/v_show/id_XMTg4NTQ4NTk0NA==.html

old series were assembled with laser, now it's just assembled with simply welding and it's more easy to replace a cell...

The best for a safety use of the battery is purchase a charge doctor  jason mc neil sale it in USA https://www.ewheels.com/product/chargedoctor-duo-for-v5f-king-song-gotway-inmotion-v8/

5 hours ago, trevmar said:

5700maH battery is made from LG MG1 28650 cells

18650 see specsheet, (the mini plus have only 11865 LG ! ) No one of Ninebot Mini device have 28650 cells 

https://www.gearbest.com/scooters-and-wheels/pp_772057.html?wid=1433363

[trevmar]

Interesting, Jojo33! I have fixed my typo (I wrote 28650 instead of 18650).

With 18650 LG cells the important thing is not the "11865" on the battery, but the letters MG or MF just in front of that number.

Looking at the Chinese battery pack at 3:24 into the Chinese video you linked, the color of the cells can be seen through gaps in the black plastic, and as I suspected, it is the pinkish color of the MG1 cells (see MG1 and MF1 images below). It is 2850mah per cell, so two in parallel in the MiniPro battery correctly gives 5700maH, the rated capacity of the N3M320. Similarly, the use of the cheaper MF1 would give 4300maH, the rated capacity of the N3M260.

The 'Charge Doctor' looks interesting, but I achieve the same result by using a true-watt meter to plug the power pack into. I have a '63V More4mini' charging brick which pumps 220 Watts into the battery most of the charge cycle. It must be putting out about 4 amps! This certainly charges the battery as quickly as possible, but for extended battery life I stay with the standard Ninebot 59.5V charger, which only pumps out 120 Watts of charging power, except when I am expecting a long excursion and need the 15% extra miles range from using the More4mini 63V charger.

As a hardware/firmware designer myself, I would expect the BMS firmware not to be used to limit the maximum output current of the battery pack. There are some MOSFETs installed on the circuit board in the battery which could protect the battery more quickly with hardware detection, rather than using the ADC in the BMS. But maybe I am wrong, and that is why I am building a special pulse testing circuit which will help me identify the exact peak current at which each battery pack shuts down. This should never happen in normal riding, as the motor controller on the main control board should be in control of the power at all times. But I suspect there is a design fault on the circuit board of the 'newer' batteries in the N3M260. At least, that is my suspicion right now... Thanks for the links you sent, they were very useful...

[FreeRide]

1 hour ago, trevmar said:

.....  I have a '63V More4mini' charging brick which pumps 220 Watts into the battery most of the charge cycle. It must be putting out about 4 amps! ....

The more for mini Charger for the Mini series is only a 126W 2A charger.  It's one of the best readily available for the Mini, but it's claim to fame is 63V output and decent amperage.  There are less powerful mini chargers, but I don't think they ever carried a 4A charger.

[trevmar]

22 minutes ago, FreeRide said:

 I don't think they ever carried a 4A charger.

I have the More4Mini standard charger, AFAIK. Yes, it is marked as putting out 126W only.

220W is what my AC Wattmeter measures as the power it is pumping into the battery throughout most of the charge. Obviously its charge current drops gradually as the battery gets towards the 100% mark :)

(I have allowed for 20W estimated heat loss in the charger circuitry, the meter actually measures 240W)

[FreeRide]

22 minutes ago, trevmar said:

I have the More4Mini standard charger, AFAIK. Yes, it is marked as putting out 126W only.

220W is what my AC Wattmeter measures as the power it is pumping into the battery throughout most of the charge. Obviously its charge current drops gradually as the battery gets towards the 100% mark

(I have allowed for 20W estimated heat loss in the charger circuitry, the meter actually measures 240W)

I think there is a problem with your meter, the switching noise or a poor PF of the supply must be confusing it.  Next time I charge my miniI will plug in my power meter and see what it says.  

[jojo33]

1 hour ago, FreeRide said:

The more for mini Charger for the Mini series is only a 126W 2A charger.  It's one of the best readily available for the Mini, but it's claim to fame is 63V output and decent amperage.  There are less powerful mini chargers, but I don't think they ever carried a 4A charger.

Unfortunetly I think you can't use a 4A charger because it wont be accepted by the BMS it wont let charge more than 2A

[smallexis]

4 minutes ago, jojo33 said:

Unfortunetly I think you can't use a 4A charger because it wont be accepted by the BMS it wont let charge more than 2A

@jojo33 is right, I have a 3A charger that I can use on my 9B1E but my mini don't take the load with this charger.

There is a "security" that causes the BMS to be cut off if the input current is greater than 2A. And it's the same for 9B1 Z series.

[trevmar]

OK, 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 (The MiniPRO N3M320 Rated 1000Watt peak load is around 22 Amps).

The tester uses a 2Hz signal generator and CMOS logic gates to drive two huge 40 amp MOSFETs which pull the current through the heatsinked Dummy Load. Each resistor on the load is rated at 50 Watts continuous, wirewound so they can take the huge peak powers. Current measured with a 200A Fluke DC current probe. The scope photos are showing first the 310Wh NC1502B bettery from the original MiniPRO and second is the 234Wh .NC1502A battery from the N3M260 "2018 US" model . Measured battery voltage is shown on the yellow trace, ranging between 60V and 50V during the pulse, while the output from the current probe is shown on the blue trace, 5A per division. I calibrated the current probe before the experiment.

RESULTS: Essentially no hardware differences. The two battery packs can both provide well over the rated peak power and have essentially the same internal resistance, whether using the MG or MF cells. Since the battery firmware is the same in each, any difference in the MiniPRO performance must come from changes in the way the Control Firmware (in my case 1.1.7 and 1.1.9) treats the Mini and the MiniPRO battery packs differently.

Since the "US 2018" version repeatably loses power and throws me forward at a minor (1") obstacle and the N3M320 original version of the MiniPRO handles that obstacle without difficulty (beeps, but no pushback or rider-throwing) whomever decided to release the cheaper NC1502A battery back without examining the impact of that decision on the MiniPRO firmware has made a terrible error... and made the 'New' MiniPROs "unsafe at any speed"...

[Scansin]

New member here! I was trying to follow this thread, but I'm not terribly mechanically minded, so I was hoping if someone could dumb it down for me.

Trevmar, you mentioned that you took a 320 model's battery and stuck it in a 2018 260 Minipro. I guess what I want to ask is whether that would "fix" the 260 by turning it into a 320? Or would it still have the firmware limitations of the 260, except with better mileage?

In other words: You said that as it is, the 260 is unsafe at any speed due to the smaller battery's impact on the firmware. So if you swapped out the battery, would that solve that problem? Or would it still have firmware issues?

[trevmar]

5 hours ago, Scansin said:

Trevmar, you mentioned that you took a 320 model's battery and stuck it in a 2018 260 Minipro. I guess what I want to ask is whether that would "fix" the 260 by turning it into a 320? Or would it still have the firmware limitations of the 260, except with better mileage?

It seems that if the MiniPRO control firmware senses the smaller battery it switches the motor power down to the N3M260-rated 700Watts, rather than the 800W with 1000W peak available from the N3M320. But I am still devising and executing more tests, and will have definitive data within 2 weeks.

[Scansin]

Thank you for doing this, trailblazing for the rest of us! Can't wait to see what happens!

[trevmar]

I took two of my MiniPRO up to a steady, steep slope this morning - one that I know is at the edge of the MiniPRO capabilities.

The slope measured with 'Angulo' from the F-Droid repository was 13/2 degrees (I used the phone-rotation method to get good accuracy). The Angulo screen also correctly showed 23% (which is twice the value, due to the measurement method). So the slope comes out at  12.5%, or 1 in 8 (approx)(tangent of 6.5degrees).

I weigh 205lbs, 93kilo.  With the N3M310 body and the 310 watt battery pack pushback started at 10km/h. I could run steadily at 9km/h, but at 10km the beeping and pushback starts. I am using firmware v1.1.9, but this is about the same performance as when I was using the 1.4.1 firmware, and the 1.1.7. I then went back to the slope with my N3M260 body with a 310 watt battery pack fitted to it, and pushback occurred at almost the same point, maybe 1km/h faster. This unit has Knobbly tires on it, a full 3/4" bigger than the smooth off-roads fitted to the N3M310 body.

I still have to change back to a 243W battery pack and measure again. Maybe tomorrow! :) I am focused on this slope as it is not the steepest road around here, but it is long and steady, giving good repeatability.

[FreeRide]

On 8/3/2018 at 4:11 PM, smallexis said:

@jojo33 is right, I have a 3A charger that I can use on my 9B1E but my mini don't take the load with this charger.

There is a "security" that causes the BMS to be cut off if the input current is greater than 2A. And it's the same for 9B1 Z series.

 

Quote

And it's the same for 9B1 Z series.

FYI: This has been shown not to be true on the Ninebot Z Series. A Z has been shown charging at 4A, so further investigation may be required.  

[trevmar]

UPDATE on testing the slope-climbing ability of my various MiniPRO and tire combinations

Slope climbing ability appears to be associated with the first 50Km run-in period for the MiniPRO. It appears they not only limit the maximum speed a little, but also limit the ability of the MiniPRO to ride-through marginally dangerous situations. Now that I have units which have been run-in, I can see that it is the new (younger) units which get  into the most difficulties at kerbs. I continue to experiment and observe to figure out exactly what is happening, but I can confirm that on my most mature vehicle there is no observable difference in slope-climbing capability when using the 4300maH N3M260 battery or the 5700maH N3M320 battery, and that the test equipment I built (earlier in this thread) shows that either battery pack can produce more than 1500 watts peak when new...

[smallexis]

On 10/6/2018 at 5:58 PM, FreeRide said:

 

FYI: This has been shown not to be true on the Ninebot Z Series. A Z has been shown charging at 4A, so further investigation may be required.  

Yes I have made a mistake about Z series. First attempt at 3A was not a success but it was due to a bad charger. I have tested at 5A and the load is ok on my Z10. At 7A, after 1 hour loading, the BMS security turns on.

So there is a security but it's allow to load over 2A but under 7A for a full load.