Headway battery pack FIRE! <lot pics>

[etriker]

Been searching and can not find any info on a Dewalt DC9360 burning up. If one did I can not find it.

There must be thousands out there in use every day with cells made in 2006.

Surely a design worth study.

Sure, but we're talking about folk building packs out of what they have, and then trying to calibrate a fusible link that's encased in whatever housing they've come up with to see if it blows at a sensible current and doesn't set fire to anything.

DeWalt will have done this with their packs, they will have spent hundreds of hours doing product design testing, simulations of heat build up in the packs, testing for a multitude of failure modes and will have fixed the design once they were confident it was safe. DeWalt have the advantage of knowing that every pack they make is the same, uses the same clearance around the fusible link, the same materials etc, so their link will behave as they have designed it when, and only when, it's inside one of their packs. Take that link outside one of their packs and it will behave differently, wrap it up in something better insulated than a DeWalt case and it'll get hotter and blow at a lower current. Leave it out in open air and it'll run cooler, and blow at a higher current.

A fuse has all this testing done for us, unlike a fusible link that is very installation and environment sensitive.

I think it is a last chance fail safe device put near the middle of the pack. I would guess there is a fuse in the BMS also.

I don't have a problem with putting fuses between all the cells in a headway pack. The increased sag would not be that bad.

If it is add an extra cell in series.

 

[Jeremy Harris]

I think it is a last chance fail safe device put near the middle of the pack. I would guess there is a fuse in the BMS also.

I don't have a problem with putting fuses between all the cells in a headway pack. The increased sag would not be that bad.

If it is add an extra cell in series.

I agree about it being a last chance device, although when discussing battery limiting and rating with the technical people at Makita they did say that there's no current limiting or overload protection on any of their cordless tools, except for their BGA452 cordless grinder. My guess is that they usually rely on the motor resistance being high enough to limit max current, then use the fusible link as a last ditch measure.

Putting fuses on each cell has serious problems, especially if all those fuses end up inside the pack. The power loss from even 16 fuses is going to create a fair bit of heat, and add to the heat load right where you don't want it. Even 20 to 30 watts of extra heat inside a pack that's well protected is going to heat it up, and lose power. The risk of failure from the added heat from the fuses might well be far greater than the risk of failure without them. This is a classic case where a proper FMEA will show whether many fuses, a single fuse, or no fuses is the safest option.

My guess is that, for a well-padded and secured pack, made from suitable materials, with the cells, BMS, terminations and wiring properly specced and secured, a single fuse in the high side, maybe connected to the pack mid-point, and fitted outside the pack padding might well be the safest option.

 

[etriker]

I think it is a last chance fail safe device put near the middle of the pack. I would guess there is a fuse in the BMS also.

I don't have a problem with putting fuses between all the cells in a headway pack. The increased sag would not be that bad.

If it is add an extra cell in series.

I agree about it being a last chance device, although when discussing battery limiting and rating with the technical people at Makita they did say that there's no current limiting or overload protection on any of their cordless tools, except for their BGA452 cordless grinder. My guess is that they usually rely on the motor resistance being high enough to limit max current, then use the fusible link as a last ditch measure.

Putting fuses on each cell has serious problems, especially if all those fuses end up inside the pack. The power loss from even 16 fuses is going to create a fair bit of heat, and add to the heat load right where you don't want it. Even 20 to 30 watts of extra heat inside a pack that's well protected is going to heat it up, and lose power. The risk of failure from the added heat from the fuses might well be far greater than the risk of failure without them. This is a classic case where a proper FMEA will show whether many fuses, a single fuse, or no fuses is the safest option.

My guess is that, for a well-padded and secured pack, made from suitable materials, with the cells, BMS, terminations and wiring properly specced and secured, a single fuse in the high side, maybe connected to the pack mid-point, and fitted outside the pack padding might well be the safest option.

Right ? I am thinking the cheaper the cells the more fuses maybe ?

We are talking about Heatways.

 

[Jeremy Harris]

Right ? I am thinking the cheaper the cells the more fuses maybe ?

We are talking about Heatways.

If anything I'd take the opposite view, as cheap cells are likely to have a higher internal resistance and so run warmer than more expensive cells with a lower internal resistance. The last thing you want to do with a pack that may already have a propensity to run warm is add something that makes it run warmer still.

Cell failure, on it's own, is very rare. Pretty much every failure we see is caused by something external to the cell, be it over charging, wiring faults, wires or cell cases chafing, BMS failure or whatever. Even cheap cells seem to fail open circuit or develop a very high internal resistance when they do fail.

 

[etriker]

Right ? I am thinking the cheaper the cells the more fuses maybe ?

We are talking about Heatways.

If anything I'd take the opposite view, as cheap cells are likely to have a higher internal resistance and so run warmer than more expensive cells with a lower internal resistance. The last thing you want to do with a pack that may already have a propensity to run warm is add something that makes it run warmer still.

Cell failure, on it's own, is very rare. Pretty much every failure we see is caused by something external to the cell, be it over charging, wiring faults, wires or cell cases chafing, BMS failure or whatever. Even cheap cells seem to fail open circuit or develop a very high internal resistance when they do fail.

These cells failed and became open.

http://www.youtube.com/watch?v=XoIpsuCDkl0

That is what I want them to do when they fail.

They are not cheap cells.

They have temp sensors on them. The Dewalt packs also sense temp.

The only shorted cells I have seen were from Hobby King and only had the connecter wires for fuses. They were cheap.

http://www.hobbyking.com/hobbyking/store/__14609__Turnigy_nano_tech_5000mah_10S_25_50C_Lipo_Pack.html

 

[Jeremy Harris]

[

These cells failed and became open.

http://www.youtube.com/watch?v=XoIpsuCDkl0

That is what I want them to do when they fail.

They are not cheap cells.

They have temp sensors on them. The Dewalt packs also sense temp.

The only shorted cells I have seen were from Hobby King and only had the connecter wires for fuses. They were cheap.

http://www.hobbyking.com/hobbyking/store/__14609__Turnigy_nano_tech_5000mah_10S_25_50C_Lipo_Pack.html

As I've already said, it's incredibly rare for a lithium cell to fail short circuit and cause a problem without there being some external cause (damage etc). I'm sure there may well have been the odd case where this has happened, but it's one of those extremely improbable events that can't really be countered by any form of protection, other than mounting the cells inside a robust and fire resistant case, with gas escape vents.

In the case of packs (not cells) that are assembled with varying degrees of quality assurance (and pretty much all RC packs fall into this category) then the failure modes seem to almost always be either a cell going open or high resistance due to the normal cell failure mechanism, or some sort of failure associated with the pack construction (rather than the cells), like an internal connection short, loose or failed balance tap wire or somesuch.

99.99% or more of lithium cell failures will be either open circuit of the cell will suddenly develop a very high internal resistance. There is a mountain of accrued evidence that this is the case.

 

[etriker]

[

These cells failed and became open.

http://www.youtube.com/watch?v=XoIpsuCDkl0

That is what I want them to do when they fail.

They are not cheap cells.

They have temp sensors on them. The Dewalt packs also sense temp.

The only shorted cells I have seen were from Hobby King and only had the connecter wires for fuses. They were cheap.

http://www.hobbyking.com/hobbyking/store/__14609__Turnigy_nano_tech_5000mah_10S_25_50C_Lipo_Pack.html

As I've already said, it's incredibly rare for a lithium cell to fail short circuit and cause a problem without there being some external cause (damage etc). I'm sure there may well have been the odd case where this has happened, but it's one of those extremely improbable events that can't really be countered by any form of protection, other than mounting the cells inside a robust and fire resistant case, with gas escape vents.

In the case of packs (not cells) that are assembled with varying degrees of quality assurance (and pretty much all RC packs fall into this category) then the failure modes seem to almost always be either a cell going open or high resistance due to the normal cell failure mechanism, or some sort of failure associated with the pack construction (rather than the cells), like an internal connection short, loose or failed balance tap wire or somesuch.

99.99% or more of lithium cell failures will be either open circuit of the cell will suddenly develop a very high internal resistance. There is a mountain of accrued evidence that this is the case.

I hear you. I bought those packs because people on this forum bragged them up and I got the very, very rare ones.

I would suggest to you that maybe more of them short than you know but they burned after they shorted.

My Hyperion alarm went off while cycling a new pack and we took it apart. I am not the .001#

 

[Jeremy Harris]

I hear you. I bought those packs because people on this forum bragged them up and I got the very, very rare ones.

I would suggest to you that maybe more of them short than you know but they burned after they shorted.

My Hyperion alarm went off while cycling a new pack and we took it apart. I am not the .001#

Just to be clear, I'm not specifically talking about cells failing whilst being charged, but cells failing whilst just sitting in a pack or being discharged.

Charging is ALWAYS going to carry a risk of something failing, 99.99% of the time because the charger did not, or could not, adequately control the charge to every cell in the pack. These aren't cell failures, as such, they are failures induced by the treatment of the pack by the charger, or perhaps because the wiring configuration caused an imbalance between cells (very easy to do with high current charging, as I found out a couple of years ago).

If you over-charge a cell, then you run a high risk of it failing in one of several rather unpleasant ways, depending on where the heat builds in the cell and how the cell is held in the pack.

No amount of fusing, good pack construction, careful wiring etc is going to protect from an over-charge event. Only a reliable charger with the correct charge profile for the cell chemistry and, ideally, cell temperature monitoring, is going to help prevent this sort of failure.

 

[etriker]

I hear you. I bought those packs because people on this forum bragged them up and I got the very, very rare ones.

I would suggest to you that maybe more of them short than you know but they burned after they shorted.

My Hyperion alarm went off while cycling a new pack and we took it apart. I am not the .001#

Just to be clear, I'm not specifically talking about cells failing whilst being charged, but cells failing whilst just sitting in a pack or being discharged.

Charging is ALWAYS going to carry a risk of something failing, 99.99% of the time because the charger did not, or could not, adequately control the charge to every cell in the pack. These aren't cell failures, as such, they are failures induced by the treatment of the pack by the charger, or perhaps because the wiring configuration caused an imbalance between cells (very easy to do with high current charging, as I found out a couple of years ago).

If you over-charge a cell, then you run a high risk of it failing in one of several rather unpleasant ways, depending on where the heat builds in the cell and how the cell is held in the pack.

No amount of fusing, good pack construction, careful wiring etc is going to protect from an over-charge event. Only a reliable charger with the correct charge profile for the cell chemistry and, ideally, cell temperature monitoring, is going to help prevent this sort of failure.

Right ? It is the users or chargers fault 99.99# of the time. Stupid newbees !

Are you the person we talked to at Hobby King ?

 

[Jeremy Harris]

Right ? It is the users or chargers fault 99.99# of the time. Stupid newbees !

Are you the person we talked to at Hobby King ?

It's nothing to do with "newbees" at all, but a fair bit to do with chargers and some of the poor wiring and construction that is prevalent in many Far Eastern products in this market sector. Take a look at the way that a lot of chargers are put together (I've stripped maybe five or six now and they were universally pretty dire in terms of design and build quality). Look at the instances we've seen here over the years of charger-related cell failure. Look at the instances of charger failure we see (two out of three of my HK purchased chargers have died, one died the moment I first connected it up).

Even the supposed "high end" RC type chargers aren't that great (witness the recent thread here about chargers giving results for cell voltage that were way off when they were measured with a decent meter) and the cheaper ones that are cloned and sold by many of the Far Eastern suppliers with various badges on are so poor in terms of quality and reliability that I'm not convinced they do any testing on them at all before they leave the factory.

I've been on this forum since long before HK became a regular supplier to the ebike community, BTW, and have no connection with them.

 

[etriker]

Right ? It is the users or chargers fault 99.99# of the time. Stupid newbees !

Are you the person we talked to at Hobby King ?

It's nothing to do with "newbees" at all, but a fair bit to do with chargers and some of the poor wiring and construction that is prevalent in many Far Eastern products in this market sector. Take a look at the way that a lot of chargers are put together (I've stripped maybe five or six now and they were universally pretty dire in terms of design and build quality). Look at the instances we've seen here over the years of charger-related cell failure. Look at the instances of charger failure we see (two out of three of my HK purchased chargers have died, one died the moment I first connected it up).

Even the supposed "high end" RC type chargers aren't that great (witness the recent thread here about chargers giving results for cell voltage that were way off when they were measured with a decent meter) and the cheaper ones that are cloned and sold by many of the Far Eastern suppliers with various badges on are so poor in terms of quality and reliability that I'm not convinced they do any testing on them at all before they leave the factory.

I've been on this forum since long before HK became a regular supplier to the ebike community, BTW, and have no connection with them.

For real. I have been working on TVs and such for a long time and the quality right now is the poorest I have ever seen.

Sales are down too. Even big names like Sony. Junk.

 

[chroot]

Hi Lyen,

Yes, I always come see you when I am coming to SF.

Umm 2 different ampere 3A or 100A same violet and I see it says "Mini" or "Reg" size fuse. It look like "Mini" or "Reg" size to me. I will email Jim Delgado from headway and find out what is his standard fuse ampere and size fuse. He is the one professional handmade the pack and BMS integrated.

I will ask him to look up this thread and maybe he would answer lot question with his experience for some customers incident or not in the past.

You are very welcome Chroot. Let me know if you need help or the next time you come to the city (San Francisco).

Judging from the blade fuse color, it would be 120A if it is purple color, or 100A for the violet color. See below:

Blade fuses use a common coloring scheme for the low-profile mini / mini / regular size fuses, and a partial color similarity with the maxi size fuses. The following table shows the commonly available fuses for each size group.

Regards,
Lyen

 

[Alan B]

Judging by the appearance of the fuse itself that is likely a MaxiFuse.

The MaxiFuse is also the only one available in a high enough voltage rating for this application.

 

[Jeremy Harris]

Judging by the appearance of the fuse itself that is likely a MaxiFuse.

The MaxiFuse is also the only one available in a high enough voltage rating for this application.

I agree, plus the maxi is quite a good fuse in terms of minimal resistive loss - just 1.2 mohms for the 50 A rated one. Add in that they are readily available pretty much all around the planet and they seem to be a good choice. The car audio fuse holders for these fuses (despite their high price) might also be a good choice, as they tend to have pretty big terminals and are likely to have a lower resistance at the connections that some of the cheaper stuff. Being gold plated they are also unlikely to corrode with time and increase the contact resistance with the fuse.

 

[JimmieD]

Hi chroot,

Got your email this morning and although a Sunday, after church I had to get on ES and find out what happened!

I am so sorry to hear this happened and I praise God nobody was hurt and that you and your family are safe!

Looking at the pictures of the 'before' battery I remember building that battery, I thought it odd that the order was to have it built with the PCM connected but not affixed to the battery, so it kinda stuck in my head. The zip ties were used for shipping purposes which is the reason they were tied together, no black electrical tape was used in the original build, was shipped with 2 mil clear heat shrink around entire pack to eliminate possible short circuiting during transit as a prototype Li-ion battery.

I have read through most (minus the chatter about A123/DeWalt) of the posts and will try and answer some of the speculative questions as best I can, or at least give my 2 cents worth!

Great job on documenting with pictures, do you happen to have any of the battery and it's connections pre-fire? This may also help out in trying to determine what possibly could have occurred.

FUSES: The fuses that we use when we fuse our batteries are MaxiFuse. It is simply put into the circuit to protect the controller or whatever else the battery was being attached since it is DC. Just a little bit of added protection since most (not all) other small sized (60V and lower) battery packs that are built (from companies not DIY'ers) do not install any type of fuse at all.

SPACER BLOCKS/BUSS BARS: The 2 hole spacer blocks were the newer ones that Headway offers. Most recognizable feature is the triangle shaped stand-offs, the older ones were round for the #40 spacer blocks. (By #40 I am referring to the cell diameter ie. 40160SE, 40152SE, 40120SE, etc.). Also the newer (last 18months or so) standard Headway buss bars have one side that is the proper diameter for attachment while the other side is slightly oval in shape to allow for proper connections (sliding into place) without the need to force the connection in place due to 'absolute measurement placement' of holes.

WIRING: The wiring that we use is http://www.jscwire.com/ except for the balance wire harness, that was what came with the SignaLab PCM. So not sure what circuit considers as cheap wire for 'all of the other wiring', but JSC Wire and Cable have always been good for me for anything 4AWG and under! :wink:

I am not sure how this could have happened to be honest with you. Everyone has given some valid input, although some is stretched a bit farther than what my logic could imagine, but then again truth is sometimes stranger than fiction! I am not a fire science guy or a fire forensic guru, but I am tending to think that it was a short that caused this to happen, not so much the fault of the chemistry which, from what you said, was not being discharged or charged so that speculation for me would need to go on the back burner for awhile until all other scenarios are exhausted!

In looking at the first post with the pictures, if breaking it down by separating the PCM pictures from the cells, it appears to me that the short possibly could have started where the stand-off screw was. In the picture there is a 'puddle' of polycarbonate muck with a screw and stand-off in it, and if looking at the rest of the PCM unit, it appears to me that approximately where this screw and stand-off was originally, it started burning. The polycarbonate has on the underside of it 2mm foam to create some protection from the screws that were used to mechanically connect B- and P- to the PCM and to protect the screws for the stand-offs from accidentally causing a short on the cells. It is plausible to me that the screw for the stand-off my have worn through the foam, come in contact with a cell, rubbed through the plastic coating of a cell and caused a short. Then again this is just another theory on what occurred.

I thought of this possible scenario first because of two reasons, first is that the PCM was not attached to the battery pack securely (not sure how the PCM or the battery pack was secured, so again only speculations here) allowing it to move about freely, and because this scenario is similar to what had happened to me when I first started building and seeing what would work best to build a quality battery pack. I had built a 36V10Ah 'prototype' test battery, and after a few days of various vibration tests that I was performing, I notice that the battery began smoking, but it was coming from the middle of the pack and underneath the PCM. There was no fire yet, but I did the proper procedure for Li-ion fires according to the DOT regulations. What I found was that the B- screw attachment had worn through the PVC coating of the cells underneath and started a small short. After removing the PCM, I also noticed that and stand-off screws had begun to wear through the PVC coating of the cells. This was the reason we use 2mm fire resistant (not proof!) foam to eliminate the chance for shorting of the cells through their PVC casings, and have not had this issue since that time until now with concerns to any of our battery packs that we have to have smoke or fire. Since this occurred to me with the PCM attached securely to my test battery for vibration tests (done with a neighbors paint can mixer), then with it not being attached securely to the battery may have caused this to happen.

Then again, in looking at the picture of the battery on fire, it looks almost as if the fire started near the most positive and most negative cells since the PCM is not quite 'on fire' yet.

It very well could be something as simple as vibrations over time causing abrasions to occur on the wiring and have nothing to do with anything else.

Right now I do not have any other plausible opinions on how this could have happened, but will be pondering it for sure!

Again, I am so thankful to hear that your family and home are okay.

 

[etriker]

Ouch ! A battery pack builders nightmare.

Are you recalling any packs ?

 

[knoxie]

yes sad to see this but happy that it wasnt a lot worse :?

Everyone needs to understand that any chemistry / pack is volatile, people running lipo get so much flack about fire all the time mainly from people running life packs under the assumption they are totally safe they just arent we should all treat these high capacity packs just like the bombs waiting to go off that they are, treat them like that and you will have no problems apart from the self destruction of you battery of course.

I hope this thread puts to bed the belief that life lovers somehow have that they can sleep safe with their batteries in house...not so and as Luke and myself have seen with our own eyes these fires are common amongst all chemistries.

I still think there is so much to be said for keeping packs as simple as you can, no permanently connected BMS, conservative charging and discharging and absolutely no leaving LVC or HVC kit connected, you are just introducing more points of failure in to the system.

Whilst running without a BMS is not for newbs, I consider for the technical amongst us its the safest way forward.

 

[JimmieD]

Ouch ! A battery pack builders nightmare.

Are you recalling any packs ?

Well, I do not consider this a nightmare, as far as being a battery builder, but I am curious as to what caused the fire.

I do not see a reason to recall any packs, especially since it is not yet specified on how this occurred. We have changed a few components in how we build our batteries (since the time this battery was built) but they were changed due to the components performance rather than for safety reasons. As I mentioned in my earlier post, this is the first experience that I have had where a battery that we had built has had an unfortunate incident such as this.

I agree with the many points that knoxie brought up, even the 'simple and humble' Pb can be volatile if not properly taken care of. There was some chatter earlier about putting a fuse at each cell sensor wire, between each cell, etc. amongst some other good ideas and thoughts, but again I agree with knoxie, keep packs as simple as possible. The more connections, the more possible failures.

I just hope that chroot has a few pictures of his set-up pre-fire to see if that can shed any light on this subject.

 

[Alan B]

Thanks for giving us your comments Jim.

Lyen posted some photos on page 2 of chroot's pack before the fire, in case you missed those.

 

[etriker]

Ouch ! A battery pack builders nightmare.

Are you recalling any packs ?

Well, I do not consider this a nightmare, as far as being a battery builder, but I am curious as to what caused the fire.

I do not see a reason to recall any packs, especially since it is not yet specified on how this occurred. We have changed a few components in how we build our batteries (since the time this battery was built) but they were changed due to the components performance rather than for safety reasons. As I mentioned in my earlier post, this is the first experience that I have had where a battery that we had built has had an unfortunate incident such as this.

I agree with the many points that knoxie brought up, even the 'simple and humble' Pb can be volatile if not properly taken care of. There was some chatter earlier about putting a fuse at each cell sensor wire, between each cell, etc. amongst some other good ideas and thoughts, but again I agree with knoxie, keep packs as simple as possible. The more connections, the more possible failures.

I just hope that chroot has a few pictures of his set-up pre-fire to see if that can shed any light on this subject.

Hey, It could have been caused by a rat chewing the wires ?

And we do all know what bms stands for, right ?

May never know.

 

[JimmieD]

Thanks for giving us your comments Jim.

Lyen posted some photos on page 2 of chroot's pack before the fire, in case you missed those.

You're welcome, and thanks Alan, saw those which were the catalyst for my memory on which pack it was, I was referring to chroot's set-up on his ride to see how he had things set-up and how he had the battery mounted.

 

[Alan B]

You're welcome, and thanks Alan, saw those which were the catalyst for my memory on which pack it was, I was referring to chroot's set-up on his ride to see how he had things set-up and how he had the battery mounted.

That would be helpful - if chroot has some photos of the battery setup before the conflagration it would be great to see them.

 

[grindz145]

Huge bummer man, glad nobody was hurt. Thanks for sharing. Sorry to hear!

In the end IMHO, it's the BMS' fault. One was or another, the cells were abused, and the BMS aught to not allow it to happen, with multiple tiers of protection, or else why even bother.

 

[circuit]

Huge bummer man, glad nobody was hurt. Thanks for sharing. Sorry to hear!

In the end IMHO, it's the BMS' fault. One was or another, the cells were abused, and the BMS aught to not allow it to happen, with multiple tiers of protection, or else why even bother.

Haha, BMS-haters gonna hate. Whatever the true reason.

 

[amberwolf]

The power loss from even 16 fuses is going to create a fair bit of heat, and add to the heat load right where you don't want it. Even 20 to 30 watts of extra heat inside a pack that's well protected is going to heat it up, and lose power.

I suppose all this boils down to "resistance is fusile"?

(sorry, it's not a perfect joke, but I'll not likley get as good a chance to use it any time soon)