One for the BMS-o-philes

How do you store and manage your electricity?
User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 00:24

The BMS is a source of much effort and bandwidth. I happen to think they have a place, especially in cars, but that place is specific and the benefits should outweigh the risks.

I don't want a BMS which needs a laptop to be plugged into it before it will work properly.

I don't want a BMS which occupies valuable real estate.

I don't want a BMS which unnecessarily risks flattening a cell, even if this means no balancing (which makes a centralised one more attractive).

I DO want a BMS which is distributed, rather than centralised (however, I concede that manual balancing of cells would be difficult).

I DO want a BMS with HVC and LVC features that I can control (sound an alarm, throttle back the controller etc)

So brainstrust, is it possible to do all of this? And is it possible to do it with a module which is no bigger than 40 mm long, 20 mm wide and 4 mm deep?
LiPosml.JPG
LiPosml.JPG (51.67 KiB) Viewed 1867 times
I reckon I could stick with my G10-FR4 board, but instead of fitting the modules to the copper bus, you could use the same 18 gauge wires I'm using for balancing and fix the module to the board with double sided tape. Any ideas?

Chris


(I have started this thread so Ian's thread isn't jammed up with more bandwidth on the merits or otherwise of a BMS with all the bells and whistles)
AEVA National Secretary, WA branch chair.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Wed, 31 Oct 2012, 00:44

My suggestion is to redesign the PCB you already have there, and put the BMS parts on the underside. I assume there's some space there where the two sides of the pouch are welded together before the tabs emerge from it. Is that right?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 00:52

I have been looking at that for months now, and hoping it might materialise into a good idea. But unfortunately the only space available is the 4 mm wide strip between the bus bars. I can't see this being enough room...

And I was thinking of filling this space with silicone as the cells aren't quite touching like I wanted them to. They need a bit of compression but when they say the cells are 10 mm thick, it's more like 9.9 mm.

There is about a 4 mm by 3 mm space between the tops of the cells on the underside of the PCB. I can send you a Gerber of the PCB if you wanted to play around with it. Here is a side view of my latest pack:
74V17Ahpack001.JPG
74V17Ahpack001.JPG (208.97 KiB) Viewed 1870 times
Last edited by jonescg on Tue, 30 Oct 2012, 13:58, edited 1 time in total.
AEVA National Secretary, WA branch chair.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Wed, 31 Oct 2012, 03:16

The busbars are only on the top, correct? So what have they got to do with the amount of space on the bottom? Why aren't there strips on the underside that are 10 mm wide less the width of the slots the tabs go thru? 9.5 mm? Can you post a screenshot of the gerber and a pic of the top of a bare cell?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 05:41

Yes, the busbars are on top an solder onto the board below, with the tabs of the cells sandwiched in between.
20s3p5800mAhpack001.JPG
20s3p5800mAhpack001.JPG (252.34 KiB) Viewed 1868 times
The underside is the slots where the tabs go in:


The reason why the space between them matters is because they are plated through. The tabs only just fit through the slots too - many of them need trimming down a bit on either side so that they fit through the slot. All of them are trimmed by about 4 mm to avoid overlapping cells creating a high spot.
AEVA National Secretary, WA branch chair.

Faz
Groupie
Posts: 96
Joined: Tue, 05 Jul 2011, 16:43

One for the BMS-o-philes

Post by Faz » Wed, 31 Oct 2012, 06:19

I've been thinking about the bms issue for ev's for a little while now.
Do we really need a LVC (or HVC) per cell in "drive" operation? If the pack is built and conditioned properly then monitored (maybe centrally) whilst on charge, is not cell monitoring, outside of charging times redundant?

User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

One for the BMS-o-philes

Post by Johny » Wed, 31 Oct 2012, 14:22

Faz wrote:If the pack is built and conditioned properly then monitored (maybe centrally) whilst on charge, is not cell monitoring, outside of charging times redundant?
In a perfect world, that's true. Unfortunately cells vary enough that the charge cycle can't gaurantee that all cells have equal capacity. You are almost putting the case for bottom balancing. Even so, per cell monitoring allows you to stop before a huge amount of damage is done - maybe as the cells age, maybe because they are not all the same to begin with.

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 16:28

Monitoring is great. Management, i.e. active balancing can be problematic. Any feature which involves a circuit across the terminals of a cell risks over discharge. If BMSs are so 'failsafe' why have so many folks endeavoured to come up with their own systems??

Manually balancing cells from a central point is a technichal challenge in itself, but the risk of something going wrong is lower, and the versatility higher.
AEVA National Secretary, WA branch chair.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Wed, 31 Oct 2012, 16:50

It's not so much the plating-thru that's the problem. It's that the slots are connected together by PCB copper on the underside. This is not necessary (and neither is the plating-thru). This space could instead be used for surface-mount BMS components. That's what I've been trying to say all along, only I didn't know you thought you needed tab-connecting copper there.

A problem I see with your busbars is that you can't inspect the solder joints to the tabs. I assume you at least tinned the underside of each busbar before sweating it on. Some small holes for feeding additional flux-cored solder through the busbar to the middle of each tab would be an improvement. And maybe use a gas torch instead of electric irons to get the soldering over with quickly to avoid cooking the cells, although you'd still need to apply pressure somehow.

The two-iron method should be fine if the irons are powerful enough. I could be wrong, but the solder puddles, where the irons were, have a "didn't get hot enough" look about them, to me.

How about, instead of a complete busbar bridging 6 cells, you had 5 small pieces that just fill in the gaps between tabs and you leave the tabs full length and solder them _on_top_ of the busbars, and overlapping each other. i.e. the pieces of copper bar are first soldered to the PCB to thicken the copper that's already there. But actually, there doesn't need to be any copper already there. It would be easier to just epoxy the copper to the FR4.

[Edit: Spelling]
Last edited by weber on Wed, 31 Oct 2012, 05:52, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
PlanB
Senior Member
Posts: 465
Joined: Sat, 16 Jan 2010, 15:24
Real Name: Kris McLean
Location: Freemans reach
Contact:

One for the BMS-o-philes

Post by PlanB » Wed, 31 Oct 2012, 17:07

I've never been able to get my head around theory & practice when it comes to BMS in different applications. The EV scene seems to favour monitoring max cell volatge on charge with bypass resistors (top balance) & low voltage cutout on discharge during driving.

The aero modellers (who drain their batteries even quicker than EVs do) all have Turnigy style chargers that seem to actively discharge the cells for a bottom balance.

Meanwhile laptop computer batteries often have 3S or 4S packs with no attempt at cell balancing at all, just a little circuit board that limits overall pack max & min voltage & max discharge current.

I guess all these different conventional wisdoms are borne of experience in that particular field of endeavour?

bga
Senior Member
Posts: 492
Joined: Mon, 01 Sep 2008, 19:27
Real Name: Bruce Armstrong
Location: Perth WA

One for the BMS-o-philes

Post by bga » Wed, 31 Oct 2012, 17:34

hey Chris,
Like the bus bar arrangement!


The 9.9mm could be accommodated by shimming every 2nd or 3rd cell with a thin sheet of plastic such that the cells line up nicely.

Cheers
Last edited by bga on Wed, 31 Oct 2012, 10:43, edited 1 time in total.
It's not the end of the world, but I can see it from here.

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 18:28

weber wrote: It's not so much the plating-thru that's the problem. It's that the slots are connected together by PCB copper on the underside. This is not necessary (and neither is the plating-thru). This space could instead be used for surface-mount BMS components. That's what I've been trying to say all along, only I didn't know you thought you needed tab-connecting copper there.
Fair call. It would be a unique PCB, that's for sure! I got a friend to make the PCB for me, and the plated through bits are really to ensure the balance tap holes are connected to the paralleled cell. In my case I plan on running all of these balance wires up to a centralised serial connector which allows voltage monitoring as well as ~1 A balance charging. It's not ideal, but a lot cheaper than getting BMS PCB made.
weber wrote:A problem I see with your busbars is that you can't inspect the solder joints to the tabs. I assume you at least tinned the underside of each busbar before sweating it on. Some small holes for feeding additional flux-cored solder through the busbar to the middle of each tab would be an improvement. And maybe use a gas torch instead of electric irons to get the soldering over with quickly to avoid cooking the cells, although you'd still need to apply pressure somehow.
Never been a problem. I thought about adding holes, but really, there is enough solder on both the tinned busbar and the cell tabs, it melts down beautifully. Gas torches just oxidise the copper too much, and it's too easy to burn something you shouldn't have. Both my irons are 200 W, so with the little puddle of solder there, it's done in about 5 seconds total. The 4 kg mass bearing down on the copper ensures a really good joint.
weber wrote:The two-iron method should be fine if the irons are powerful enough. I could be wrong, but the solder puddles, where the irons were, have a "didn't get hot enough" look about them, to me.
No, it has that look because after soldering I put a damp cloth on the copper to suck the excess heat out of the copper. Just to minimise potential heat damage to the cells below. The tops of the cells are a bit warmer for sure, but within 30 seconds they are at room temp.
weber wrote:How about, instead of a complete busbar bridging 6 cells, you had 5 small pieces that just fill in the gaps between tabs and you leave the tabs full length and solder them _on_top_ of the busbars, and overlapping each other. i.e. the pieces of copper bar are first soldered to the PCB to thicken the copper that's already there. But actually, there doesn't need to be any copper already there. It would be easier to just epoxy the copper to the FR4.


Way too much trouble. There is enough solder on the tabs and on the copper that a good joint is self-evident. I could probably have got away with 1.6 mm copper, but 2 mm copper is useful for other things. I have dragged 300 A from a 7.4 V battery and these connections don't even roll over in their sleep. The cells start to warm up after 5 minutes of constant 150 A, but by then the pack is nearly empty. The copper is still ice cold.
AEVA National Secretary, WA branch chair.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Wed, 31 Oct 2012, 18:45

I understand that in the terms "LVC" and "HVC" the "C" stands for "cutout". I understand that's terminology from the hobby world. Full size EVs can't usually afford to just disconnect the battery when a cell is in distress. They must instead politely request that the motor controller or charger (or in some scary cases, the driver) do something about it. So they are more like alarms than cutouts.
jonescg wrote: Monitoring is great. Management, i.e. active balancing can be problematic. Any feature which involves a circuit across the terminals of a cell risks over discharge.
That is true.
If BMSs are so 'failsafe' why have so many folks endeavoured to come up with their own systems??
Who says BMSs (in general) are failsafe. There is some total crap out there, designed by complete amateurs who don't have a clue. One of the worst I've seen was held on to the cell tops with double-sided tape (which lets go when it gets warm) and had miles of exposed leads on its bypass resistors which were flopping about off the side of the boards, held by screw terminals. It was a fire waiting to happen, and when it _did_ happen (although it looked to me like it was actually started by a loose high-current link between cells), other morons chime in to say "see I told you all BMSs were dangerous". What a farce.

The lesson _we_ took from that was to redesign our cell-top battery management units (BMUs) to be able to _survive_ a high-current link going high resistance under load or charge. They were already bolted to the cell terminals and were already _monitoring_ link voltage to hopefully pick up a link going high-R _before_ it gets that bad.

And we switched to using low-cost industrial optic fibre to take the signals in and out of each battery box.

The main reason we designed our own was because, when we started 3.5 years ago, it was hard to find a BMS that even _specified_ what maximum system voltage it was designed for, let alone the 750 volts of our battery. The assumption seemed to be that a BMS only had to be rated for the voltage of one cell, and never mind that the signal wires passing over the tops of the whole pack were referenced to chassis potential.
Manually balancing cells from a central point is a technichal challenge in itself, but the risk of something going wrong is lower, and the versatility higher.

You need to multiply the risk by the cost of the consequences if it does go wrong.

We ourselves experienced a cell going flat (0.86 V) due to an errant microcontroller allowing bypass to turn on when it shouldn't. But this was on the test bench, not in a car.

So more redesign. We added a pulldown resistor to the gate of the MOSFET that switches the bypass resistors, and we chose a MOSFET with a threshold such that even if the micro was actively turning it on, it could not be maintained on after the cell voltage fell below about 2.0 V.

But what if the MOSFET fails short-circuit drain to source? Then the BMU should raise a low voltage alarm. But why didn't that happen with our cell. It did, but the cell with faulty BMU was sitting all by itself on a shelf, so no one was looking at its LEDs and nothing was monitoring its signals. So we added a beeper to every BMU.

But if the micro isn't working properly, it won't raise an alarm either. What then? If it's on its own on a shelf there's not much we can do, but we made the software so, if it's part of a communicating system, as it usually will be, then the BMU that comes _after_ it, will report that it hasn't heard from its neighbour for some time. In which case a human should investigate.

Oh, and we added the ability to reset all microcontrollers remotely via the existing optic fibre. We already had the capability of reloading their software remotely over the fibre.
Last edited by weber on Wed, 31 Oct 2012, 07:57, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Faz
Groupie
Posts: 96
Joined: Tue, 05 Jul 2011, 16:43

One for the BMS-o-philes

Post by Faz » Wed, 31 Oct 2012, 20:06

Johny wrote:
Faz wrote:If the pack is built and conditioned properly then monitored (maybe centrally) whilst on charge, is not cell monitoring, outside of charging times redundant?
In a perfect world, that's true. Unfortunately cells vary enough that the charge cycle can't gaurantee that all cells have equal capacity. You are almost putting the case for bottom balancing. Even so, per cell monitoring allows you to stop before a huge amount of damage is done - maybe as the cells age, maybe because they are not all the same to begin with.

I may not have been clear. I was suggesting that the charging setup was performing a balance charge (ie:each cell monitored and balanced) every charge.
Have groups of 8 or so cells with large enough wire from each cell going to a single (multi-pole) connector. It is effectively a bms built into the charger but not connected whilst driving. Much in the same way all hobby rc batteries are charged. The benefit is it doesn't need to take up space in the battery pack. (though people using ts style cells probably don't need to worry about space concerns)

User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

One for the BMS-o-philes

Post by Johny » Wed, 31 Oct 2012, 21:01

Fair enough - good idea. The vehicle system only needs the monitor side.

Edit: Assuming a racing vehicle. A road EV normally has the charger on board so nothing saved.
Last edited by Johny on Wed, 31 Oct 2012, 10:02, edited 1 time in total.

User avatar
Richo
Senior Member
Posts: 3678
Joined: Mon, 16 Jun 2008, 00:19
Real Name: Richard
Location: Perth, WA

One for the BMS-o-philes

Post by Richo » Wed, 31 Oct 2012, 21:17

Last time I spoke to you I was suggesting it go on the PCB you already have.
I have to get mine on a PCB on the end of a headway cell (38mm OD) with a fat hole in the middle.
So the short answer is NO but the long answer is YES.
Help prevent road rage - get outta my way!

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Wed, 31 Oct 2012, 22:08

The other problem I see with having the BMS circuitry surface mounted on the back is soldering the bus down might heat up the board too much.

At this stage, a compact, robust BMS which fits on top of the insulating layer of G10-FR4 board would suffice, and probably take up as much room as populating the underside with components.
AEVA National Secretary, WA branch chair.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Wed, 31 Oct 2012, 22:49

jonescg wrote: The other problem I see with having the BMS circuitry surface mounted on the back is soldering the bus down might heat up the board too much.
Good point. It wouldn't do to have the solder melting on the underside and parts dropping off. And it would be a bit difficult to solder the bus bars on with the whole assembly upside-down. Image
At this stage, a compact, robust BMS which fits on top of the insulating layer of G10-FR4 board would suffice, and probably take up as much room as populating the underside with components.

It would take up more room, but it's probably the best idea.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jackhyq
Groupie
Posts: 101
Joined: Tue, 22 Jul 2008, 23:42
Real Name: Jack
Location: China
MSN: jackhyq3@hotmail.com
Contact:

One for the BMS-o-philes

Post by jackhyq » Thu, 08 Nov 2012, 22:50

Nice assembling .
Portable AC/DC/Solar UPS;
3D BirdView,RearView,A.I System;
1C~90C LFP/LiPo Battery ;
BatteryShow@126.Com;
www.Ponilion.Com

Faz
Groupie
Posts: 96
Joined: Tue, 05 Jul 2011, 16:43

One for the BMS-o-philes

Post by Faz » Fri, 09 Nov 2012, 06:14

The issue I see with most bms's I've seen is the per cell shunt current is typically 1% or so of the cell capacity. Which is great if the pack was already commissioned cell by cell or your charger is of a ridiculously low current. Otherwise what is the point of a full blown bms?
I have the same thought on "lvc" if a cell voltage is dropping while driving, well either you stop or drive till the pack makes you stop. The end result is similar, but one way gets you closer to home.
Having a (generally) high current draw from 10 cells paralleled (through who knows what inter-cell resistance) to me negates the level of function current market bms's offer.
Saying all that I believe in cell monitoring and proper cell by cell charge control.
Last edited by Faz on Thu, 08 Nov 2012, 19:23, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Fri, 09 Nov 2012, 06:47

Faz wrote: The issue I see with most bms's I've seen is the per cell shunt current is typically 1% or so of the cell capacity. Which is great if the pack was already commissioned cell by cell or your charger is of a ridiculously low current. Otherwise what is the point of a full blown bms?
The point is to use a charger with controllable current and to have a BMS that will control it down to the shunt current when necessary. Our most recent comissioning ran for several days and nights that way, and didn't require any human intervention.
I have the same thought on "lvc" if a cell voltage is dropping while driving, well either you stop or drive till the pack makes you stop. The end result is similar, but one way gets you closer to home.

How can you say the end result is similar? In the second case a cell may be destroyed and even a fire started. Our BMS not only monitors cell voltage, but also cell temperature and connecting-link voltage.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Faz
Groupie
Posts: 96
Joined: Tue, 05 Jul 2011, 16:43

One for the BMS-o-philes

Post by Faz » Fri, 09 Nov 2012, 08:15

weber wrote:
Faz wrote: The issue I see with most bms's I've seen is the per cell shunt current is typically 1% or so of the cell capacity. Which is great if the pack was already commissioned cell by cell or your charger is of a ridiculously low current. Otherwise what is the point of a full blown bms?
The point is to use a charger with controllable current and to have a BMS that will control it down to the shunt current when necessary. Our most recent comissioning ran for several days and nights that way, and didn't require any human intervention.
I have the same thought on "lvc" if a cell voltage is dropping while driving, well either you stop or drive till the pack makes you stop. The end result is similar, but one way gets you closer to home.

How can you say the end result is similar? In the second case a cell may be destroyed and even a fire started. Our BMS not only monitors cell voltage, but also cell temperature and connecting-link voltage.

Waiting several days to charge the battery is exactly my point "why bother"?!?!?!

Um destroyed cell? Well it is clearly already stuffed if a properly commissioned pack has a corrupt cell.A fire started? What battery chemistry are you talking about?

You refer to "our bms" I wasn't referring to indi built bms's.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

One for the BMS-o-philes

Post by weber » Fri, 09 Nov 2012, 17:55

Faz wrote:Waiting several days to charge the battery is exactly my point "why bother"?!?!?!
I'm not sure what you're missing here, so I'll try explaining the whole thing in more detail.

Commissioning (initial balancing) only has to be done once. From then on, a full recharge from flat can be done in around 8 hours (in our case) because our charger can then run at its full 5.5 A for most of that time. The balancing phase at the end, where the current tapers down to 0.4 A, only takes a few minutes. But it's exactly the same charger and exactly the same BMS-controlled charging algorithm in both cases.

Our BMS shunt capability is exactly the 1%C or C/100h that you mentioned as typical (0.4 A for a 40 Ah pack). Of this, you wrote:
Which is great if the pack was already commissioned cell by cell or your charger is of a ridiculously low current. Otherwise what is the point of a full blown bms?
So I've explained that even with only 1%C shunt capability you do not need to have a charger that's only capable of a ridiculously low current, and you do not need to commission your pack cell by cell.

If I had commissioned my 228 cell pack cell-by-cell it would have taken far longer. But even if you only have say 45 cells and so it would take less elapsed time to commission them cell by cell, that requires a different charger and a lot of your time spent moving it along from cell to cell. Whereas my BMS-controlled initial charge did not require any attention from me and so let me work on the many other jobs required to get an EV conversion on the road. So from my point of view it took zero time.
Um destroyed cell? Well it is clearly already stuffed if a properly commissioned pack has a corrupt cell.
You didn't previously say anything about it being "corrupt". You wrote:
I have the same thought on "lvc" if a cell voltage is dropping while driving, well either you stop or drive till the pack makes you stop. The end result is similar, but one way gets you closer to home.

Of course all cell voltages are "dropping while driving" but I assume you meant "dropping much faster than the others" or "dropping below the manufacturer specified minimum voltage". This doesn't require the cell to be corrupt in any way. It may simply be the lowest capacity cell in the pack. Every pack has a lowest capacity cell. In a top-balanced pack, if the whole pack is nearing empty, it will go empty first. Even with a bottom-balanced pack, if it has been a long time since the last bottom-balance, some cell may go empty before the others.

If you "drive till the pack makes you stop" you will by then have completely flattened that cell and begun to "charge" it in reverse. This will lead to it going high resistance (where "high resistance" here may still be less than one ohm) and thereby ending up with a large fraction of the battery voltage across it in reverse, while the remaining cells continue to force current through it.

Both jonescg and I have packs with around 200 cells. Imagine one of those cells with say 100 volts across it (in reverse) and 100 amps being pushed through it. That's 10 kilowatts being dissipated in one cell! How long do you think it will last?
A fire started? What battery chemistry are you talking about?
Any Lithium-ion chemistry. They all use flammable solvents, e.g. DMC and DME, that will begin to react with the anode, break down and vent from the cell, if the cell's internal hotspot temperature goes over about 80°C. See http://www.nfpa.org/assets/files/pdf/re ... hazard.pdf

Of course it still needs to be mixed with air and ignited by something at around 600°C. But that won't be long in arriving with 10 kW being dissipated in the cell and hundreds of volts available to make sparks. So you see, hazardous-voltage packs have some additional considerations compared to extra-low-voltage packs.
You refer to "our bms" I wasn't referring to indi built bms's.

For the purpose of this argument, I don't see how it matters who builds them. You're talking about two of the most basic and necessary functions of a BMS for Lithium-ion batteries. You seem to be saying that an undervoltage alarm is not worth having, and that a balancing shunt is not worth having unless it can bypass far more than 1% of C/1h. I'm trying to explain why you might be mistaken about those.

There is a third basic and necessary function of a BMS for Lithium-ion batteries, and that is an overvoltage alarm (or a single combined over/undervoltage alarm).
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

One for the BMS-o-philes

Post by Johny » Fri, 09 Nov 2012, 18:05

Faz wrote:I have the same thought on "lvc" if a cell voltage is dropping while driving, well either you stop or drive till the pack makes you stop. The end result is similar, but one way gets you closer to home.
That reminds me of a cartoon I saw once in a Readers Digest Magazine.
It was a single frame with a hot and sweaty guy carrying a petrol can walking toward a lone car parked on the side of the highway. There was a women with her head hanging out of the driver's side window saying "Look dear, I made it another 400 meters on the starter motor.".

Just how close to home do you want to be before it catches on fire?

User avatar
jonescg
Senior Member
Posts: 2935
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

One for the BMS-o-philes

Post by jonescg » Fri, 09 Nov 2012, 21:03

It's true that you need some kind of warning that a cell might be behaving strangely - strangely enough that it might be dissipating more heat than electrons, and hitting 2 volts or less while you are operating the vehicle. It's easy enough to have an on-board monitor which warns you that a cell could be too low (let's call this LVC). It need not do anything other than sounds a buzzer.

A warning system for an over-voltage cell (let's call it HVC) need not be on-board, especially for a racing vehicle. It will be of no use at any time other than charging. So it can be off-board, but will require a pretty serious serial plug if you have more than 40 cells in series. It also needs to be able to handle the plug-unplug mating cycles. So for a racing vehicle which doesn't see much frequent use, it's not a bad idea to leave this feature off-board. But for a road going vehicle, it's probably just as easy to have a distributed system with a HVC and LVC built into it.

Now, as for balancing, I am of the opinion that it need not do this on board. Any circuit which seeks to draw current from across the terminals risks flattening a cell completely if you're not careful. Sure, you can put another feature onto it so the cell cannot be drained any lower than 2.4 V or something, before going completely open circuit to save it's life. But is it worth the complexity? If you have the room for it, sure. On a race bike, I can't see the appeal.

So what I plan on doing for my 168 cells in series is to have 4 x 44 pin serial plugs for each isolated sub-pack which normally has a centralised LVC/HVC circuit plugged into it. It warns me if a cell is too low during operation, or too high during charge. If a cell does go HVC/LVC, I can unplug the monitors and plug in a balance charging harness to even things out.

Best compromise for the circumstances, I think.
AEVA National Secretary, WA branch chair.

Post Reply