Z10 BMS FW 1.1.7 - Vampire Drain

[houseofjob]

Interesting, the new 1.1.7 BMS FW claims better/lesser vampire drain, but after leaving my Z10 unridden indoors for 1+ day (42 hours), I'm seeing what translates into .14V drop per day, roughly 1% per day.

Edited March 8, 2019 by houseofjob

[Marty Backe]

Which means essentially no improvement 

I'm waiting to check mine on Sunday. That will have been a full week sitting after a charge.

If it were one of my Gotway or KingSong wheels, it will read 100-percent.

[houseofjob]

17 minutes ago, Marty Backe said:

Which means essentially no improvement 

I'm waiting to check mine on Sunday. That will have been a full week sitting after a charge.

If it were one of my Gotway or KingSong wheels, it will read 100-percent.

Yeah. 

Maybe something was improved(?)

[stephen]

Maybe it goes into deep sleep on more battery percentage if  they can't fix the drain completely 🤔

[Rehab1]

2 hours ago, Marty Backe said:

I'm waiting to check mine on Sunday. That will have been a full week sitting after a charge.

[fbhb]

5 hours ago, stephen said:

if  they can't fix the drain completely 🤔

Ninebot will never be able to fix the battery drain completely because both batteries status are always going to be constantly monitored, as evidenced by the "Blue/Red" flashing led on each battery!  The very best we can all hope for is that the real world standby consumption is actually reduced, as stated in the BMS 1.1.7 release notes. 

Unfortunately, by design, power is always being drawn from both batteries for the smart/safety reasons deemed necessary by Segway-Ninebot's engineers so software changes can only go so far!  Best solution is to charge to 100% then go ride/enjoy the wheel as often as is humanly possible, that's what I fully intend to do with my Z10!

Edited March 9, 2019 by fbhb

[stephen]

3 hours ago, fbhb said:

Best solution is to charge to 100% then go ride/enjoy the wheel as often as is humanly possible, that's what I fully intend to do with my Z10

I can't keep off it 😁 it's an amazing machine😁

[Filalapat]

Interesting.

I made the same kind of observations more than a month ago. These had shown a voltage loss of about 0.0038V per hour. It's not enough but it's enough to empty a fully charged battery in about 4 months. So for low charged batteries as they are when shipping from China, it is normal that they are delivered in deep sleep state.

All my readings were done with the BMS FW v1.1.2 ;  Z10 FW v1.0.5

The value of 0.0038V / h is an average value obtained on the two Z10 batteries by several readings on stand-by times each time greater than 4 days, on two different Z10 wheels of the 1832 series.

17 hours ago, houseofjob said:

Interesting, the new 1.1.7 BMS FW claims better/lesser vampire drain, but after leaving my Z10 unridden indoors for 1+ day (42 hours), I'm seeing what translates into .14V drop per day, roughly 1% per day.

Over 42 hours, we can expect a loss of 0.0038 x 42 = 0.16V.

0.14V is better than 0.16V, but the difference does not seem to me significant enough to really show an improvement in the power loss management with regard the various inaccuracies made during the readings. (as well as the short time taken into consideration)

When I go to version 1.1.7 I may see it more clearly.

I created a topic on the French forum to highlight the current of stand-by wheels

 

https://www.espritroue.fr/topic/7758-courant-de-stand-by-de-nos-roues/

 

I showed that :

The Z10 absorbs standby current 1500 times higher than an MS3 through its XT60 connectors.

The total stand-by current flowing through the XT60 connectors is 150 micro-amperes.

This current of 150 uA is not enough to discharge the battery as fast. I have not done any measurement on the other outgoing connections of the battery pack, but it is very likely that the loss is internal to the battery itself (BMS structure)

[Smoother]

Welcome @Filalapat to the English speaking forum.  Keep up the good work.

[Filalapat]

Thank you for your welcome here 

[EUCMania]

This is exactly what 9B wants: Do not forget your BabyZ, care it, ride it everyday. Else it will suicide.

[Marty Backe]

When I updated to the latest firmware 7-days ago, I left the wheel alone to see what affect this update had to the phantom power drain.

Not much. The battery level was showing 96%  At that rate it would be down ~20-percent after ~1-month of inactivity.

Maybe the firmware stops the power drain when it gets down to the single-digits? But if that's the case I don't know why it couldn't be turned off all of the time.

Oh well.

[ruohki]

Out of curiosity... Why is 1.0.5 the latest fw for me? Are we poor eu people getting mobbed?

[US69]

16 minutes ago, ruohki said:

Out of curiosity... Why is 1.0.5 the latest fw for me? Are we poor eu people getting mobbed?

There is a difference to BMS firmware and General wheel firmware numbers...just saying...

[ruohki]

Oh okay, so it comes along with the normal patches trough the ninebot app?

[Filalapat]

On 3/9/2019 at 12:49 PM, Filalapat said:

When I go to version 1.1.7 I may see it more clearly.

Since my last posts, so I updated the firmware BMS to version 1.1.7.
And I did the same kind of long-term readings to get reliable data and to understand how this standby consumption is.

After turning off the wheel, I first waited for 24 hours so that the voltages of the batteries are well stabilized.
The next day, I started the readings. (for each set of measurements on the ninebot app, the wheel did not stay on for more than a minute to avoid altering the results)

12/03 20h37 
battery 1    50,11V    43%
battery 2    50.08V    43%

7 days after ... (exactly 169h)

19/03 21h37 
battery 1    49,12V    34%
battery 2    49,15V    34%

voltage difference (average) = 0.96V →  divided by 169h = 0.00568V/hour

2 days (47 hours) after I retry to make a measure and .... the wheel does not start. This is not the first time, I'm used to this strange behavior with my Z10. Below about 30% of battery, the wheel can refuse to start. Although not all Z10s have this flaw, other people than me have exactly the same problem. I do not know if this is due to the motherboard firmware version (1.0.5 for mine) or the production series of the wheel (1832). To restart, I can obviously recharge the battery, but it is also possible to very long press the ON/OFF button, as I did also on this video. The sounds that come from the wheel during this startup are really amazing : jerking clicks louder and louder, then whistles whose frequency increases, much like the old electronic flashes... Listen for yourself :

I was optimistic that the BMS firmware update v1.1.7 corrects this failure to restart the Z10, but no, it is not.

After pressing the button for 45 seconds, my wheel started up :

21/03 20h37 
battery 1    48,75V    31%
battery 2    48.81V    32%

voltage difference (average) = 0.355V →  divided by 47h = 0.00755V/h

I notice that these two values of consumption in stand-by state are much higher than those which I had obtained before (see my post above). I don't know why.

I decide to continue my investigations to find where will be lost this current of stand-by. I unplug all the connectors of the battery pack: the two XT60s, the 4-pin charging connector and the 2-pin communication connector.
After exactly 8 days (192 hours), I reconnect connectors, restart the wheel by pressing the ON / OFF button for about a minute and measure :

29/03 20h37 
battery 1    48,65V    30%
battery 2    48,73V    31%

voltage difference (average) = 0.09V →  divided by 192h = 0.000468V/h

We see that the battery pack alone absorbs in standby mode 8 to 16 times less than the electronic circuitry integrated in the wheel (essentially the motherboard)

I also took advantage of the update of the BMS firmware v1.1.2 to v1.1.7 to make another test : the assymetrical regeneration of the two batteries. For some time I had read on the French forum that the firmware v1.1.7 corrected the problem that only one of the two batteries recharges in a long descent. So I did this test myself by noting the percentage of the charge of the two batteries before and after going down a slope of about 90m (280 ft) over a length of 1200m (0,75 mi) (Roule mountain, Cherbourg). This test, I did it twice: once with the firmware BMS v1.1.2, and I did it again in the same conditions a few days after update v1.1.7.
In both cases, the battery 1 remained with a percentage unchanged, and the battery 2 increased by 5% the first time, 6% the second time.
So I think that v1.1.7 does not fix this bug. I think it is logical because the regeneration current enters the XT60 connectors, and therefore comes from the motherboard. Nothing is less certain that this motherboard, even if its firmware was different, allows to distribute differently the energy returned by the inverter bridge. This possibility of regeneration must depend on the electronic design of the motherboard and not its software firmware or that of the BMS.

The good thing is that fortunately all the time that these measurements took me, I could still ride on my Gotways MS3 and Tesla! 

For information, I have just finished the same kind of measurement readings on my Ninebot S2 in standby mode : for 9 days and 7 hours, the voltage drop is 1,005V, thus 0,0045V/h… It's the same order of magnitude that I measure on the Z10 !

Finally, vampire drain, compared to a gotway for sure, but compared to another Ninebot, I'm not so sure ...

[Marty Backe]

12 minutes ago, Filalapat said:

Since my last posts, so I updated the firmware BMS to version 1.1.7.
And I did the same kind of long-term readings to get reliable data and to understand how this standby consumption is.

After turning off the wheel, I first waited for 24 hours so that the voltages of the batteries are well stabilized.
The next day, I started the readings. (for each set of measurements on the ninebot app, the wheel did not stay on for more than a minute to avoid altering the results)

12/03 20h37 
battery 1    50,11V    43%
battery 2    50.08V    43%

7 days after ... (exactly 169h)

19/03 21h37 
battery 1    49,12V    34%
battery 2    49,15V    34%

voltage difference (average) = 0.96V →  divided by 169h = 0.00568V/hour

2 days (47 hours) after I retry to make a measure and .... the wheel does not start. This is not the first time, I'm used to this strange behavior with my Z10. Below about 30% of battery, the wheel can refuse to start. Although not all Z10s have this flaw, other people than me have exactly the same problem. I do not know if this is due to the motherboard firmware version (1.0.5 for mine) or the production series of the wheel (1832). To restart, I can obviously recharge the battery, but it is also possible to very long press the ON/OFF button, as I did also on this video. The sounds that come from the wheel during this startup are really amazing : jerking clicks louder and louder, then whistles whose frequency increases, much like the old electronic flashes... Listen for yourself :

I was optimistic that the BMS firmware update v1.1.7 corrects this failure to restart the Z10, but no, it is not.

After pressing the button for 45 seconds, my wheel started up :

21/03 20h37 
battery 1    48,75V    31%
battery 2    48.81V    32%

voltage difference (average) = 0.355V →  divided by 47h = 0.00755V/h

I notice that these two values of consumption in stand-by state are much higher than those which I had obtained before (see my post above). I don't know why.

I decide to continue my investigations to find where will be lost this current of stand-by. I unplug all the connectors of the battery pack: the two XT60s, the 4-pin charging connector and the 2-pin communication connector.
After exactly 8 days (192 hours), I reconnect connectors, restart the wheel by pressing the ON / OFF button for about a minute and measure :

29/03 20h37 
battery 1    48,65V    30%
battery 2    48,73V    31%

voltage difference (average) = 0.09V →  divided by 192h = 0.000468V/h

We see that the battery pack alone absorbs in standby mode 8 to 16 times less than the electronic circuitry integrated in the wheel (essentially the motherboard)

I also took advantage of the update of the BMS firmware v1.1.2 to v1.1.7 to make another test : the assymetrical regeneration of the two batteries. For some time I had read on the French forum that the firmware v1.1.7 corrected the problem that only one of the two batteries recharges in a long descent. So I did this test myself by noting the percentage of the charge of the two batteries before and after going down a slope of about 90m (280 ft) over a length of 1200m (0,75 mi) (Roule mountain, Cherbourg). This test, I did it twice: once with the firmware BMS v1.1.2, and I did it again in the same conditions a few days after update v1.1.7.
In both cases, the battery 1 remained with a percentage unchanged, and the battery 2 increased by 5% the first time, 6% the second time.
So I think that v1.1.7 does not fix this bug. I think it is logical because the regeneration current enters the XT60 connectors, and therefore comes from the motherboard. Nothing is less certain that this motherboard, even if its firmware was different, allows to distribute differently the energy returned by the inverter bridge. This possibility of regeneration must depend on the electronic design of the motherboard and not its software firmware or that of the BMS.

The good thing is that fortunately all the time that these measurements took me, I could still ride on my Gotways MS3 and Tesla! 

For information, I have just finished the same kind of measurement readings on my Ninebot S2 in standby mode : for 9 days and 7 hours, the voltage drop is 1,005V, thus 0,0045V/h… It's the same order of magnitude that I measure on the Z10 !

Finally, vampire drain, compared to a gotway for sure, but compared to another Ninebot, I'm not so sure ...

Great post.

I will confirm (learned the hard way) that the Z10 will not turn on with a relatively low battery. So if any Z10 owners are trying to get home and the battery is getting low, do not take a break and turn off the wheel while you're taking a sip of water. You'll be walking the rest of the way home.

Another stupid design decision by Ninebot.

[Chriull]

3 hours ago, Filalapat said:

 

 12/03 20h37 
battery 1    50,11V    43%
battery 2    50.08V    43%

7 days after ... (exactly 169h)

19/03 21h37 
battery 1    49,12V    34%
battery 2    49,15V    34%

That numbers are amazing every time again!

~10% of ~1000Wh in 7 days! Gives about 0,6W of continous power dissipation for nothing! Really crazy!

Some 180mA at 3.3V - the current consumption of about the top ST32 models at full speed. The low power consumption models are somewhere below the mA range...

By using sleep/standby modes while the wheel is turned off there could be a good chance to get consumption down many magnitudes with a firmware update!

... if they made not crazy design faults like low resistance voltage dividers to measure battery voltage , or maybe more probable inefficient all time on power supplies for everything...

3 hours ago, Filalapat said:

21/03 20h37 
battery 1    48,75V    31%
battery 2    48.81V    32%

...

I unplug all the connectors of the battery pack: the two XT60s, the 4-pin charging connector and the 2-pin communication connector.
After exactly 8 days (192 hours), I reconnect connectors, restart the wheel by pressing the ON / OFF button for about a minute and measure :

29/03 20h37 
battery 1    48,65V    30%
battery 2    48,73V    31%

voltage difference (average)

This very very roughly 1% in 8 days give some 15mA continous current at 3,3V. Much nicer, but still crazy for a Battery Managment System .... But at least this would be numbers one could live with...

Ps.: 1% in 8 days should give ~2years from 100% downto 0. This could be about the range of self discharge of Li Ion cells?

Edited April 2, 2019 by Chriull

[Filalapat]

11 hours ago, Marty Backe said:

I will confirm (learned the hard way) that the Z10 will not turn on with a relatively low battery. So if any Z10 owners are trying to get home and the battery is getting low, do not take a break and turn off the wheel while you're taking a sip of water. You'll be walking the rest of the way home.

 

At first, the fact that the wheel with low battery may not restart also worried me ... not even possible to stop for a moment while leaving the wheel running since it turns off automatically after a few minutes.
And in the exceptional unpredictable cases of a loss of control or a runaway of the wheel which forces to turn it off and on again?

Since I realized that we can still restart the Z10 with a weak battery, I am more serene. I know it will not let me down.
I have already tested the method I give in the video of my previous post to restart the Z10 with about 10% of battery. And it works fine. The wheel starts again and is as functional as it was before it was turned off.
Obviously with low battery speed is limited, but this could still allow to go home.

[Chriull]

12 hours ago, Filalapat said:

12/03 20h37 
battery 1    50,11V    43%
battery 2    50.08V    43%

7 days after ... (exactly 169h)

19/03 21h37 
battery 1    49,12V    34%
battery 2    49,15V    34%

voltage difference (average) = 0.96V →  divided by 169h = 0.00568V/hour

Somehow i don't get the numbers really together.... Or Ninebot is calculating the Charge % in a very "own" way?

A battery with 4.2V has 100% charge, with 2.5V 0% charge (according to the datasheets!). This would be the full capacity (~998Wh) for the whole pack, and more than is used by the wheel and reported as capacity.

So the usable Voltage Range from 2.5V to 4.5V are 1.7V difference. 0.00568V per hour should make 0.00568V/1.7V = 0.33% charge used per hour.

The difference in reported charge is 9%, divided per 169 hours make 0,05%! Which is about a tenth of the result looking at the voltages...

[Filalapat]

22 hours ago, Chriull said:

Somehow i don't get the numbers really together.... Or Ninebot is calculating the Charge % in a very "own" way?

A battery with 4.2V has 100% charge, with 2.5V 0% charge (according to the datasheets!). This would be the full capacity (~998Wh) for the whole pack, and more than is used by the wheel and reported as capacity.

So the usable Voltage Range from 2.5V to 4.5V are 1.7V difference. 0.00568V per hour should make 0.00568V/1.7V = 0.33% charge used per hour.

The difference in reported charge is 9%, divided per 169 hours make 0,05%! Which is about a tenth of the result looking at the voltages...

When I trace the points I have taken on a graph, I find them finally very coherent.

During the 3 weeks of my readings, the voltages and the percentages obtained were not finally so distant, which introduces necessarily an imprecision on the placement of the curve (which I suppose to be a straight line ...)

Just a note on cell tensions. If it is obvious that a fully charged 4.2V battery is 100%, nothing says that the app (via the data that the wheel returns with bluetooth ...) will not already indicate 100% then that the voltage has not yet reached the 4,2V. This is common: the app already says 100% while it still loads ...
Regarding discharge of a cell, none of my wheels go down to 2.5V per cell even with the wheel completely empty and fortunately. We know that 2.5V is the minimum voltage below which chemical deterioration of the cell can can occur. This value is therefore a part of the 'absolute maximum ratings' given in the manufacturer's datasheet for these batteries.
But in practice, on the wheels that I own (MS3, Tesla, Z10), none goes below 3.3V per cell. Discharged until the final tiltback as shown in the photo below, my tesla was 2% just below 67V, which is more than 3.3V per cell.
The only time I completely drained (roughly) the battery of my Z10 after a 64 km ride at 0°C, I could hear the very low battery beeps. I did not know there was a specific sound for this. It was no longer the beeleep beeleep beeleep that we usually hear. These had given way to the same sounds that are heard at the very first start of the Z10, when it is not yet linked to the Ninebot account. This surprised me. To date I have heard in these conditions of very low battery only once. Probably you will have made the same observation with your Z10. That day I was over 46V, so about 3.3V per cell.
This voltage (arbitrary?) 3.3V is not imposed by the BMS but by the motherboard to still allow the balancing of the wheel when pushed with the trolley, and especially to remain at a voltage greater than the limit value of 2.5V.

To get back to the graph, on the right line obtained (roughly), the discharge at 0% corresponds to 45V or 3.2V. We are close to 3.3V which is consistent. The maximum value crosses approximately 57V, instead of 58.8V. Possible explanations: imprecision in the construction of the graph, non-linearity of the treatment voltage -> percent ... or other

In the end, with the experience I have, I never rely on the percentage of battery given by the app. I know (very roughly) how many miles of autonomy I will have. I adjust this according to the season (temperature) and the average speed that I will try to respect. I'm rarely wrong. It must be believed that it works: in 16000 km, I finished the ride by walking because of the empty battery only twice. Most of the time, I empty the battery and go home with the beeps.

[Chriull]

2 hours ago, Filalapat said:

The maximum value crosses approximately 57V, instead of 58.8V. Possible explanations: imprecision in the construction of the graph, non-linearity of the treatment voltage -> percent ... or other

57/14~4.1V is imo with most wheels already 100%, as the ~3.3V are 0% for most wheels.

[esaj]

Those drain numbers are pretty crazy... to put things into perspective, I put my wheels to winter rest in late September, didn't check the voltages (or maybe I did, but at least didn't write them down), but both were showing 3 leds (out of 9) for the battery percentage. When I checked them last week by turning both on (over 6 months unused and without charging), they're still both showing 3 leds. The drain of a BMS alone is somewhere in the µA -range, Ninebot must be running the MCU at all times...

Edited April 3, 2019 by esaj

[Girth Brooks]

I just had a battery issue (first of any issue) with my Z10 just now on lunch. Said "bad communication with battery 2". Battery one showed 78% and battery 2 showed 20% before not showing up at all. I luckily have my charger with me. I'm going to charge it and see if it persists. I have not updated the BMS firmware. I'm still on v1.1.2.

[Girth Brooks]

I didn't see any prompt or option to update the BMS firmware in the Ninebot app either. Is it hiding somewhere I'm not looking? I went to settings < battery to look but just showed current BMS without choice to update.