Charging speed, any conclusions?

[null]

Deleted because I do not wish to contribute to a forum that accepts racism.

Edited July 6 by null

[bryon01]

Here was my post of how fast the cells CAN charge and the exact time it takes for the 24S10P Sherman pack: 

 

The cells/wires can take charging at 16A, which is more wattage than you can likely pull on a normal household plug. That said, the BMS/Controller likely cannot safely handle that.  You need the BMS to be able to bypass a set of cells at a significant fraction of the charge current in case the pack is out of balance, or you have a bad cell.  If the BMS can't shunt that current, then the pack takes it, and a set of parallel cells can go well over 4.2V and start a fire.  That is why I suspect they limit the charging to only 8A.  I believe someone stated the control board on the Veteran acts in place of a separate battery BMS, meaning it now likely has a MOSFET that can cutoff the charge current and the limits of that mosfet and the PCB traces need to also be taken into account.  Regardless, an 8A charge will get you a 4 hour recharge time from absolute dead, assuming that 8A means 8A really.  I would not go over what they stated as a lithium battery fire of multiple kWh is no joke and will absolutely burn your house down.

[bryon01]

You can save an easy 30 min by just stopping when the voltmeter first reaches 100V.  You will have put in almost 95% of the energy it is going to take at that point and can cutout 13% of the time of the charge.  I do agree it seems that if they are shipping with 5A chargers, it all but makes sense that people are likely to plug 10A into it.  If I was them I'd upgrade it to support that before someone burns their house down.  

 

Also, I am not 100% sure the limit was 8A, I think it was in Marty's video and assume he got it from Jason or Veteran directly.

[MrRobot]

5 hours ago, bryon01 said:

Here was my post of how fast the cells CAN charge and the exact time it takes for the 24S10P Sherman pack: 

 

The cells/wires can take charging at 16A, which is more wattage than you can likely pull on a normal household plug. That said, the BMS/Controller likely cannot safely handle that.  You need the BMS to be able to bypass a set of cells at a significant fraction of the charge current in case the pack is out of balance, or you have a bad cell.  If the BMS can't shunt that current, then the pack takes it, and a set of parallel cells can go well over 4.2V and start a fire.  That is why I suspect they limit the charging to only 8A.  I believe someone stated the control board on the Veteran acts in place of a separate battery BMS, meaning it now likely has a MOSFET that can cutoff the charge current and the limits of that mosfet and the PCB traces need to also be taken into account.  Regardless, an 8A charge will get you a 4 hour recharge time from absolute dead, assuming that 8A means 8A really.  I would not go over what they stated as a lithium battery fire of multiple kWh is no joke and will absolutely burn your house down.

8a even with the first batch? 

[Chriull]

20 hours ago, null said:

The next batches have GX16-5 four pin which official specs give as 5A so 10A with doubled pins (8A according to veteran?)

So the first batches can use 2x5A, (maybe 2x7A?)
Next batches 2x8A (?)

Is the total charging limited to 10A as I seem to have seen here and there, or could we target 14-16?

The gx16-4 plug is said to take 10A if all 4 pins are used - mine gets cosy warm with 8A. From 8A to 10A dissipated power raises to 10*10/8*8~1.56 times. So about 56% more power to dissipate!

 

18 hours ago, bryon01 said:

You need the BMS to be able to bypass a set of cells at a significant fraction of the charge current in case the pack is out of balance, or you have a bad cell.  If the BMS can't shunt that current, then the pack takes it, and a set of parallel cells can go well over 4.2V and start a fire.  That is why I suspect they limit the charging to only 8A.

All in EUCs used BMS i've seen so far have single cell overvoltage cutoff for charging. As you stated balancing resistors do not limit cell voltage, they just slow down the "rising voltage".

Normaly this cut off was directly done by one (or more) mosfets on the BMS, like here for the gw nikola (1)

Here are 4 mosfets and a 5 mOhm resistor between C(harge) - and B(attery). The cannot be seen in real detail here - it seems 2 are nChannel Mosfets in parallel for switching (cell overvoltage cutoff). The other two could be for reverse polarity protection (just diodes, body diodes of mosfets or even pChannel?)?

The only picture i found from the sherman BMS is (2)

These are here the 2 BMS from one side handling each 12s5p.

They are definitely going a different way here as GW. Balancing circuitry can he seen, communication lines between the BMS and some ICs. Also one communication wire ?to the control board/fuse box?

But there is definitely no (power) mosfet to cut off charging in case of single cell overvoltage!

Quote

 I believe someone stated the control board on the Veteran acts in place of a separate battery BMS, meaning it now likely has a MOSFET that can cutoff the charge current and the limits of that mosfet and the PCB traces need to also be taken into account.

It seems "BMS funcionality" is still on the BMS boards and the mosfet switches for charge cutoff moved to the "fuse box": (2)

Seems the control wire ends at the "fuse box", but it's not to be seen definitely on the photos?

edit: is confirmed that the control wires go to the main board

"Each assembly has its own separate fuse. In case of overcharging or over-discharge (over-discharge) of an individual element, the BMS “sees” this state of affairs and an appropriate signal is sent via the signal wire in case of a problem (two options: over-charge or over-discharge). As you probably noticed, there are no keys on the BMS (i.e. transistors that would open the circuit), so the controller will directly deal with blocking the battery." (google translations from (2))

Hopefully they did not implement a single cell undercharge cut-off (as could be (wildly) guessed from what @EcoDrift mentions here?) - would be a step back to the beginning of EUC riding were they cut-off at to low voltages :(. Or maybe they implemented a single cell undervoltage warning mechanism - which on the other side would be a big leap forward!

This should be a great (thermal) improvement to move the switching mosfets out of the plastic wrap into a seperate box! Unfortionately hard to see if this black thing below the ?transparent plastic? is the switching mosfet, mounted on metal as heatsink?

Maybe @Marty Backe can take a closer look? Or some other tester/owner?

 

From these points i'd be (a bit) optimistic that full nominal charging current (2x8A) could maybe work out.

A switching mosfet with about 10mOhm R_DS_on would dissipate at 10A (rounded the 8A per side up, just in case the current is not divided equaly) 0.01Ohm*10A*10A = 1W. This should be no real problem in this external box if there is some cooling provided (?as it seems?).

One pack (24s5p) has an internal resistance of ~0.2Ohm. So that would make 0.2Ohm*10A*10A=20W. Cell temperatures should be closely monitored the first time(s). At discharging (riding) the battery packs are more burdened, but here it is an over 3h constant burden (without air drag - so worse cooling?)

The battery wires from the control board have to take this 8 (10A) easily as they also have to take the discharge current while riding.

So last uncertain point would be the wires from the charge connectors to the control board and the PCB traces to the battery connectors.

8 hours ago, null said:

I suppose the board is the same, so if Veteran think 8A should be the total max thats it. As they upgraded the ports to take 8A that would be one charger, there is a nice fairly compact one I was eyeing on Ali (sold as for gotway) but then Id have to replace the v1 port with the new one.

That's for sure the best and safest way. And there are by far too many uncertainties as to recommend anyone ignore to this limit.

Also 2*8A could work out, that couple of open points should be double (triple) checked and best the first charge attempts closely supervised (as @Seba did on a different wheel with his faster charger and a thermal camera)

18 hours ago, null said:

As for wall sockets fast foods have become the gas station of PEVs in france, they have several sockets intended for their vacuum cleaning, so upwards 3000W+ I guess.

Normal "circuit brakers" for mains plugs are 10A and 16A. So ~2.3kw and ~3.6kW should be about the limits.

But one never knows if this mains plug is the only thing attached to this "circuit braker" - so together with other consumers the limit could be anywere below...

And the charger only draws the full power at the end of constant current phase and beginning of constant voltage phase - once the full ~100V are available and the maximum current flows...

(1) from https://ecodrift.ru/2019/12/02/gotway-nikola-na-elementah-21700-razbiraem/

(2) from https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/

@EcoDrift - I hope it is ok for you that i directly link photos from your teardown (together with the information of the source)?!

Edited July 23, 2020 by Chriull

[Marty Backe]

5 hours ago, Chriull said:

The gx16-4 plug is said to take 10A if all 4 pins are used - mine gets cosy warm with 8A. From 8A to 10A dissipated power raises to 10*10/8*8~1.56 times. So about 56% more power to dissipate!

 

All in EUCs used BMS i've seen so far have single cell overvoltage cutoff for charging. As you stated balancing resistors do not limit cell voltage, they just slow down the "rising voltage".

Normaly this cut off was directly done by one (or more) mosfets on the BMS, like here for the gw nikola (1)

 

Here are 4 mosfets and a 5 mOhm resistor between C(harge) - and B(attery). The cannot be seen in real detail here - it seems 2 are nChannel Mosfets in parallel for switching (cell overvoltage cutoff). The other two could be for reverse polarity protection (just diodes, body diodes of mosfets or even pChannel?)?

The only picture i found from the sherman BMS is (2)

 

These are here the 2 BMS from one side handling each 12s5p.

They are definitely going a different way here as GW. Balancing circuitry can he seen, communication lines between the BMS and some ICs. Also one communication wire ?to the control board/fuse box?

But there is definitely no (power) mosfet to cut off charging in case of single cell overvoltage!

It seems "BMS funcionality" is still on the BMS boards and the mosfet switches for charge cutoff moved to the "fuse box": (2)

Seems the control wire ends at the "fuse box", but it's not to be seen definitely on the photos?

"Each assembly has its own separate fuse. In case of overcharging or over-discharge (over-discharge) of an individual element, the BMS “sees” this state of affairs and an appropriate signal is sent via the signal wire in case of a problem (two options: over-charge or over-discharge). As you probably noticed, there are no keys on the BMS (i.e. transistors that would open the circuit), so the controller will directly deal with blocking the battery." (google translations from (2))

Hopefully they did not implement a single cell undercharge cut-off (as could be (wildly) guessed from what @EcoDrift mentions here?) - would be a step back to the beginning of EUC riding were they cut-off at to low voltages :(. Or maybe they implemented a single cell undervoltage warning mechanism - which on the other side would be a big leap forward!

This should be a great (thermal) improvement to move the switching mosfets out of the plastic wrap into a seperate box! Unfortionately hard to see if this black thing below the ?transparent plastic? is the switching mosfet, mounted on metal as heatsink?

Maybe @Marty Backe can take a closer look? Or some other tester/owner?

 

From these points i'd be (a bit) optimistic that full nominal charging current (2x8A) could maybe work out.

A switching mosfet with about 10mOhm R_DS_on would dissipate at 10A (rounded the 8A per side up, just in case the current is not divided equaly) 0.01Ohm*10A*10A = 1W. This should be no real problem in this external box if there is some cooling provided (?as it seems?).

One pack (24s5p) has an internal resistance of ~0.2Ohm. So that would make 0.2Ohm*10A*10A=20W. Cell temperatures should be closely monitored the first time(s). At discharging (riding) the battery packs are more burdened, but here it is an over 3h constant burden (without air drag - so worse cooling?)

The battery wires from the control board have to take this 8 (10A) easily as they also have to take the discharge current while riding.

So last uncertain point would be the wires from the charge connectors to the control board and the PCB traces to the battery connectors.

That's for sure the best and safest way. And there are by far too many uncertainties as to recommend anyone ignore to this limit.

Also 2*8A could work out, that couple of open points should be double (triple) checked and best the first charge attempts closely supervised (as @Seba did on a different wheel with his faster charger and a thermal camera)

Normal "circuit brakers" for mains plugs are 10A and 16A. So ~2.3kw and ~3.6kW should be about the limits.

But one never knows if this mains plug is the only thing attached to this "circuit braker" - so together with other consumers the limit could be anywere below...

And the charger only draws the full power at the end of constant current phase and beginning of constant voltage phase - once the full ~100V are available and the maximum current flows...

(1) from https://ecodrift.ru/2019/12/02/gotway-nikola-na-elementah-21700-razbiraem/

(2) from https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/

@EcoDrift - I hope it is ok for you that i directly link photos from your teardown (together with the information of the source)?!

That control wire from the BMS does not go to the fuse, it connects into the control board. And there is another control wire from the other pair of batteries that plugs into the control board.

[Chriull]

43 minutes ago, Marty Backe said:

That control wire from the BMS does not go to the fuse, it connects into the control board. And there is another control wire from the other pair of batteries that plugs into the control board.

Thanks!

Then the mosfet for cutting off charging has to be on the mainboard

(Picture again from https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/)

Seems that on the right side just aboce the battery connectors is one mosfet mounted (lying) on the PCB. One of the control wires is just crossing above his legs.

If that's the one and it's the only one this could have some 40-60K/W thermal resistance?

So with 10mOhm R_DS_on and 8A charging current power dissipation that'll be 0,64W and so some nice 25-40°C above ambient temperarute.

With 16A charging current it'll be 2,6W and 100-156°C above ambient letting him melt away...:(

@Marty Backe - would you mind taking some detail pictures of this area?! With the print on the mosfet visible? Were the charging port wires and the battery wires arrive at the port and this mosfet(s?) can be seen?

[Esbu]

30 minutes ago, Chriull said:

Seems that on the right side just aboce the battery connectors is one mosfet mounted (lying) on the PCB. One of the control wires is just crossing above his legs.

[Esbu]

[Chriull]

1 minute ago, Esbu said:

Great! There seem to be four mosfets between charge and battery wires! Or two switching, two reverse voltage protection?

[bryon01]

Like I said, I would not shove 16A into this thing, even if the mosfet could take it, the traces on the board could easily let go well below what the mosfet could handle.  If someone takes some temp measurements at 5/8A charging you could likely extrapolate what would happen without actually destroying your board in testing.  

Also, if the BMS is in the battery, and it can maybe only shunt 1A per cell bank, you are taking risks that the overvolt logic works in the first place to shutoff the controller mosfet. If there is any sort of failure in this loop, like let's say the controller continues to allow current to flow if the control wire is broken/gone, 16A will start a fire if your pack is unbalanced for some reason whereas something lower may just get hot.  Typically your charger should also be monitoring the charge and cell bank health such that the charger itself stops charging, and the BMS is an additional failsafe, not the only failsafe.  As far as I know none of the EUC chargers do this so that the charger can be cheap.  This isn't unique to the Veteran by any means, but the consequences of high power charging accidents is quite severe.  As a long time LiPo user, I typically only charge my batteries in a firebox and always within sight.  There are too many batteries in a EUC to realistically do this as you'd need to build something enormous.  If anyone is actually planning on pushing the limits, do not do this IN your house.

[Marty Backe]

3 hours ago, Chriull said:

Thanks!

Then the mosfet for cutting off charging has to be on the mainboard

(Picture again from https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/)

Seems that on the right side just aboce the battery connectors is one mosfet mounted (lying) on the PCB. One of the control wires is just crossing above his legs.

If that's the one and it's the only one this could have some 40-60K/W thermal resistance?

So with 10mOhm R_DS_on and 8A charging current power dissipation that'll be 0,64W and so some nice 25-40°C above ambient temperarute.

With 16A charging current it'll be 2,6W and 100-156°C above ambient letting him melt away...:(

@Marty Backe - would you mind taking some detail pictures of this area?! With the print on the mosfet visible? Were the charging port wires and the battery wires arrive at the port and this mosfet(s?) can be seen?

Here's a high resolution picture of the brand new control board that I just received. Click for full-resolution view.

The control wires from the BMS connect to those two connectors to the left of the battery charging input cables.

Edited July 23, 2020 by Marty Backe

[bryon01]

The set of 4 are 65A, 100V Mosfets.   The larger Fairchild mosfet is 300V, 14A.  Really hard to tell what the charging lines are hooked up to with just a photo.  I'd be curious to know what the microcontroller is, couldn't see it in the photo myself.

[Bridgeboy]

According to Darknessbot my Sherman only reaches 97% or 98% charge (it keeps flipping back ands forth between those two values) with the 5A charger that came with it.

I'm using latest DarknessBot 6.1 (287) through TestFlight.

All other EUCs of mine (Kingsongs) correctly register 100% charge in Darknessbot when the charger finishes.

Is this cause for concern? 

[Chriull]

14 hours ago, Bridgeboy said:

According to Darknessbot my Sherman only reaches 97% or 98% charge (it keeps flipping back ands forth between those two values) with the 5A charger that came with it.

I'm using latest DarknessBot 6.1 (287) through TestFlight.

All other EUCs of mine (Kingsongs) correctly register 100% charge in Darknessbot when the charger finishes.

Is this cause for concern? 

Which voltage is shown in the app? Can you measure the charger output voltage with some voltmeter?

 

[Bridgeboy]

9 hours ago, Chriull said:

Which voltage is shown in the app? Can you measure the charger output voltage with some voltmeter?

 

It fluctuates between 100.2 to 100.3 Volts. I can measure with a volt meter when I get a chance, but I haven't yet.

[Chriull]

1 hour ago, Bridgeboy said:

It fluctuates between 100.2 to 100.3 Volts. I can measure with a volt meter when I get a chance, but I haven't yet.

 

On 10/11/2020 at 7:02 PM, Bridgeboy said:

According to Darknessbot my Sherman only reaches 97% or 98% charge

That'some strange numbers - 100.8V is full charge. Some 75-79V should be empty.

100.2V - 100.3V should still be in the 100% charge range ...

Measurements are not as accurate and battery voltage of li ion cells settle.

Could be some strange way of sherman to report charge % or some customized charge % from darknesbot not really appropriate for the sherman?

 

[Bridgeboy]

1 hour ago, Chriull said:

 

That'some strange numbers - 100.8V is full charge. Some 75-79V should be empty.

100.2V - 100.3V should still be in the 100% charge range ...

Measurements are not as accurate and battery voltage of li ion cells settle.

Could be some strange way of sherman to report charge % or some customized charge % from darknesbot not really appropriate for the sherman?

 

Yeah, it could be an inaccurate reading from Darknessbot. All my Kingsongs reach 100% no problem.

What does this mean?: "Some 75-79V should be empty."

[Chriull]

2 hours ago, Bridgeboy said:

Yeah, it could be an inaccurate reading from Darknessbot. All my Kingsongs reach 100% no problem.

Kingsongs report 100% for per cell voltages from around 4.1xV up to full 4.2V.

Maybe Sherman reports 100% only for exactly 4.2V (100.8V for the 24 cells in series)?

2 hours ago, Bridgeboy said:

What does this mean?: "Some 75-79V should be empty."

About 3.15V to 3.3V are normally 0% charge = empty. Times 24 cells gives about 75-79V.

Could fit with about 100.2V shown leading to ~98% if one does the math...

[Chriull]

3 hours ago, null said:

Do the batteries charge and discharge by the same thick wires, the thin ones being signal wires? Or are the thin ones the charge cables?

From https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/ it seems as charging and discharging uses the same wires.

The thin ones are for communication.

[Chriull]

4 minutes ago, null said:

The "is the charging speed limited to 8A" (leaperkim specifies 10A)

I'd be worried about the bms and maybe also mainboard mosfet taking this charging current.

As power dissipation grows vy the square of the current one has to compare 8*8=64 and 10*10=100! 64 and 100 are different magnitudes!

No ideas how these mosfets are placed, designed or choosen - if one insists on charging with 10 A one should supervise the Bms/battery pack and mainboard very carefully!

[Chriull]

8 hours ago, null said:

thin cables where accused of being a bottle neck

My answer 

8 hours ago, Chriull said:

charging and discharging uses the same wires.

is just half the truth

Remembered there were better hi res photos of the motherboard: (1)

 

Here one sees the 4 thick cables (2 red, 2 black on 2 XT60 connectors) going to the battery.

Below these are thinner cables leading to the GX connector

12 hours ago, null said:

- Anybody knows if they started using 4x pins on the GX charge plug like it was talked about?

which have to be connected on all 4 pins to stay (somewhat) within the specification.

So - the charging cables are still thinner as the battery cables. But hard to say if they are sufficient for 10A? At least there are 2 pairs in parallel - so could be enough, and since most presumably none of the 4 wires is leading to nowhere, this could be a strong sign that all 4 pins of the connector are used?!

BTW: beside this charging cables are two small connectors which, if i remember right are for the (really thin) communication wires with the two BMS.

 

Another nice detail to be seen is, that there is no (power) Mosfet on the BMS PCB; (2)

Which imho is really great!

Were should such a power transistor transfer his heat to within the plastic wrap... There should be no heating element together in a plastic pack with the Li Ion cells!

But there are 4 Mosfets between the charge and the battery wires on the motherboard - they all seem to be KIA3510A and presumably 2 in parallel per battery pack.

They are specified with 9mOhm on resistance - so 4.5mOhm for 2 in parallel makes

4.5mOhm * 8A/2 * 8A/2 ~ 0.07W

or 4.5mOhm * 10A/2 * 10A/2 ~0.11W

power to dissipate while charging. And this two times (once for each pack) - so 0.22W which should not really pose any problem?!

Although i have no idea were they are located - If there is any convection possible or the fans are active while charging and cooling these mosfets?

 

(1) from 

(2) from https://ecodrift.ru/2020/07/05/razbiraem-veteran-sherman-iz-pervoj-postavki-mnogo-udivitelnyh-otkrytij/

[JetSerge]

I did a test with the fast charger where the amps can be controlled. With 13 amps Sherman stops charging in about a minute and displays the “E CHAr” message. With 12.5A it charges without stopping. I know some people who can charge at 15A. So it may depend on the specific unit. Mine stops charging with the same error within several seconds if I set the charger to 15A.

[JetSerge]

I didn't open my Sherman yet to verify how the connectors are wired. I've made the adapter with 2 pairs of wires for each connector, but not sure if only 1 pair is actually used.

As for the charging speed, there is a theory that Sherman limits the watts, so the max amps that can be used would depend on the state of charge (voltage).

Quote

it looks like it limits the watts, so at zero battery it can charge at 15A x 75V = 1125W. But at half the battery and higher it will stop and you will need to lower the amps. 95V x 12A = 1140W. My unit seems to stop at 1250-1300W.

My Sherman is from the one of the most recent batches (the one from the eWheels container that was on fire during transit because of the faulty Begode RS battery, still has the nasty smell).

Edited January 25, 2021 by JetSerge

[JetSerge]

4 minutes ago, null said:

I guess it could be checked easily by looking for continuity between pin 1-2 and 4-5 (Not saying you should do it, just mentioning it) 

Oh, indeed. Just tested, there is continuity between 1-2 and 4-5 pins on both connectors. So it looks like 4 pins are wired on each connector.