Last fire.. :-(

[Pyrofx]

There have been at least two fires.

The guy from Asia who sat on his battery and suspects most of his weight pressed on one high terminal which might have caused damage and an internal short.

Then there is this fire of florst’s where she suspects mice/rodents might have caused a short or a wire coming loose or whatever.

As I said earlier, it really doesn’t matter.

Whether from pests, poor connections that come loose, or internal shorts resulting from dendrite growth, the bottom line is that the BMSs (and fuses) we all use do a pretty good job detecting and protecting against excessive battery discharge / short, but do nothing to detect and protect against an internal short (other than shutting down when a cell discharged to LVD if the discharge rate is slow enough).

Especially in this era of the Chevy Bolt (‘don’t park in the garage, park 50’ away from any other vehicle’), I’m surprised to see this subject being dismissed as quickly as it is.

Fire can only result from above-normal temperatures and monitoring battery temperature and even individual cell temperatures is not only easy, but relatively inexpensive.

If we conclude that a reliable precursor of a combustion event must always be enough heat to warm up adjacent cells to the point they bloat and create clamping fixture pressure outside of the normal range, that’s an even easier and cheaper thing to monitor at the whole-battery level.

Of course, if we know that the risk of a cell failure / internal shorting even is <0.01%/year, each of us can decide whether we’re comfortable with that level of risk or want to take additional steps to protect against it.

That’s not the point.

The point is are there any reliable precursors to combustion events escaping detection by BMS and fuse that can be used to add additional levels of protection?

Same guy I think

 

[Gazoo]

The guy from Asia who sat on his battery and suspects most of his weight pressed on one high terminal which might have caused damage and an internal short.

Then there is this fire of florst’s where she suspects mice/rodents might have caused a short or a wire coming loose or whatever.

Same guy... florst had two fires. There are three other fires I know of. One is a vague memory and the other fire totaled the guys van. The cause of that fire was never determined.

If a battery is properly constructed I don't believe there is anything to worry about. I always suggest using insulation spacers between the cells and around the cells if using a metal fixture, and braided busbars. Some battery manufacturers don't do either as can be seen in Will's tear down videos.

On a scale of 1 to 10 I would rate faulty connections being a 10 as the most likely cause of fires. Of course there "could" be other things involved. We have seen cells severely over charged and bloated and they did not cause a fire, Same with severely overly discharged cells.

I am not convinced any of these fires was caused from a cell suddenly shorting internally. Also keep in mind the old timers are referring to cells made a decade or more ago. While I value their input, chemistry and construction has changed since. There is a lot of data missing as well.

 

[fafrd]

Same guy... florst had two fires. There are three other fires I know of. One is a vague memory and the other fire totaled the guys van. The cause of that fire was never determined.

Wow, don’t know why I assumed florst was female (pic?) but assuming you are correct he is the same guy that caused a fire by sitting on his cells, that is certainly something to take into consideration…

If a battery is properly constructed I don't believe there is anything to worry about. I always suggest using insulation spacers between the cells and around the cells if using a metal fixture, and braided busbars. Some battery manufacturers don't do either as can be seen in Will's tear down videos.

I’ve got 6” battery cables made from 2/0 welder’s cable, so think I’m a step ahead of even braided busbars, but I don’t have any enclosure yet. I was thinking of making an enclosure from 3/4” 1x12 (for the insulation) but this whole discussion is causing me to ditch that plan.

If I end up making a metal enclosure, I’ll probably line it with Sheetrock (and include a vent).

On a scale of 1 to 10 I would rate faulty connections being a 10 as the most likely cause of fires. Of course there "could" be other things involved. We have seen cells severely over charged and bloated and they did not cause a fire, Same with severely overly discharged cells.

If cell failure generally involves gradually increasing rates of self-discharge, it should be easy to catch a failing cell well before it can reach dangerous rates of self discharge.

But as the recent Chevy Bolt fiasco has made clear (and again, yes, I understand it’s different chemistry), dendrite formation coupled with a manufacturing defect can have catastrophic consequences.

From what I’ve understood, there's is not yet enough history with these aftermarket cells being used in DIY applications to assess the risk (especially given that many/most of these cells are factory rejects and they are being used in non-standard configurations (threaded aluminum fittings).

So while I agree faulty connections is a 10 and a much higher risk, that does not help establish the remaining risk after being confident wiring has been done properly.

I am not convinced any of these fires was caused from a cell suddenly shorting internally. Also keep in mind the old timers are referring to cells made a decade or more ago. While I value their input, chemistry and construction has changed since. There is a lot of data missing as well.

Certainly agree with that last statement.

When the one-in-ten-thousand cells fails, we really don’t have much of an idea of what that looks like.

Seems like combustion is impossible without generating heat first, so if their are BMSes or other multi-probe temperature sensors they allow individual cell temps to be monitored, that would go a long way towards adding an almost perfect additional layer of security.

Fireproof enclosures is another complementary way to go about reducing the risk of combustion.

And I’m still curious enough to want to understand what the temperature / pressure curve for these 280Ah cells looks like in case that may provide another easy an effective early-warning signal…

 

[toms]

So can I assume the fault/damage is limited to the offending cell within the parallel arrangement, meaning that is the only cell that would be required to be replaced. This assumes the other parallel cells (along with a replacement for the actual short) are brought back to a SOC commensurate with the rest of the array.?

Every time i’ve seen this happen the cell in parallel with the faulty cell was also destroyed, the parallel set continues to discharge to zero volts even after pack disconnect.

 

[Posplayr]

Every time i’ve seen this happen the cell in parallel with the faulty cell was also destroyed, the parallel set continues to discharge to zero volts even after pack disconnect.

That's what I would have been concerned about. So the 2P parallel cell configuration would take out the parallel set whereas in a series configuration it would only be the single cell that shorted. I'm glad to hear that the parallel configuration was still contained. How much insulation and fire protection was being used?

 

[toms]

How much insulation and fire protection was being used?

In some cases the cells were in metal cases, mostly in timber cases.

I’ve never seen an off-grid system using an active fire protection method.

 

[Posplayr]

In some cases the cells were in metal cases, mostly in timber cases.

I’ve never seen an off-grid system using an active fire protection method.

Thanks, what I'm most interested in is if a passive metal enclosure with some appropriate inflammable material can safely contain a cook-off.

I think for OP, while he took precautions, without a BMS LVD ended up dumped only 12% SOC into the single point failure (shorted post) cooking off the entire installation.

I'm interpreting your statements to mean that it should be relatively easy (with some care) to contain the damage of a pair of cells (200-600 amp-hrs??) in a cook-off.

 

[A.Justice]

Same guy... florst had two fires. There are three other fires I know of. One is a vague memory and the other fire totaled the guys van. The cause of that fire was never determined.

If a battery is properly constructed I don't believe there is anything to worry about. I always suggest using insulation spacers between the cells and around the cells if using a metal fixture, and braided busbars. Some battery manufacturers don't do either as can be seen in Will's tear down videos.

On a scale of 1 to 10 I would rate faulty connections being a 10 as the most likely cause of fires. Of course there "could" be other things involved. We have seen cells severely over charged and bloated and they did not cause a fire, Same with severely overly discharged cells.

I am not convinced any of these fires was caused from a cell suddenly shorting internally. Also keep in mind the old timers are referring to cells made a decade or more ago. While I value their input, chemistry and construction has changed since. There is a lot of data missing as well.

LiFePO4 battery fires are pretty rare, and when one person has two fires, in a very short succession, I tend to think user error.

Maybe my assumption will make me an ass, but I think the probability of two fires from random equipment failure is slim to none, even with not-so-nice equipment.

If I had to guess the cause of the majority of issues / fires, I would agree with you and say poor connections. Making a good connection was the most time consuming part of my build, every terminal and bus bar was sanded, twice, rough polished, coated in dielectric grease, and grub screws thread-locked. It's a pain in the butt, and requires things that most people don't have sitting around, so I understand why it's skipped.

 

[bdserv]

Seems like combustion is impossible without generating heat first, so if their are BMSes or other multi-probe temperature sensors they allow individual cell temps to be monitored, that would go a long way towards adding an almost perfect additional layer of security.

The problem I see with this approach is that while you can sense the temperature rise of a cell due to an internal short, there is nothing you can do to stop the process of self-destruction by then.

It seems to me that looking for a change in self discharge characteristics would provide a better early warning sign that a cell is on its way to becoming dangerous.

 

[bdserv]

If I had to guess the cause of the majority of issues / fires, I would agree with you and say poor connections. Making a good connection was the most time consuming part of my build, every terminal and bus bar was sanded, twice, rough polished, coated in dielectric grease, and grub screws thread-locked. It's a pain in the butt, and requires things that most people don't have sitting around, so I understand why it's skipped.

I did some basic cleaning of the bases of my battery terminals [M8 threaded studs, not tapped aluminum block] with scotchbrite and then assembled the pack using the supplied nickle-plated bus bars. My application is a golf cart which can draw from 50 amps running along on level ground to nearly 300 amps during initial acceleration. My terminals and bus bars did not even get warm to the touch when monitoring the pack during a normal run around the course.

As I'm sure you know, people need to be careful with dielectric grease - its purpose is to prevent oxidation and it is an insulator. It can actually be detrimental when used in excess.

 

[toms]

That is the wrong application for dielectric grease, you should use a product such as Noalox that improves conductivity and prevents corrosion.

Dielectric grease is for applications like inside a terminal block - to repel water and prevent fretting corrosion.

 

[A.Justice]

As I'm sure you know, people need to be careful with dielectric grease - its purpose is to prevent oxidation and it is an insulator. It can actually be detrimental when used in excess.

Absolutely, I should have clarified for anyone reading, I know there's a ton of strange information about dielectric grease floating around.

I use a cotton swab to apply a very small amount to the outer surfaces, after all the connections have been made. It is to prevent moisture from accumulating on the surface, or inside of the stud / bus area.

Everything is working well so far...?

 

[A.Justice]

That is the wrong application for dielectric grease, you should use a product such as Noalox that improves conductivity and prevents corrosion.

Dielectric grease is for applications like inside a terminal block - to repel water and prevent fretting corrosion.

See my post above, I added a little more detail to that statement. You are absolutely correct.

I've heard so many differing opinions on NoAlox, I really don't know how, or if it should be used on anything outside of aluminum grid feeder cables. I'm not saying it should or shouldn't, I'm saying that I have absolutely no idea.

I do, however, know how dielectric grease should be used, there's not nearly as much technical debate, in my opinion. It's more of a surface coating to prevent corrosion, and in my personal experience of putting it on the outside of already connected connections, it does a second to none job at keeping salty, moist air out.

 

[Daddy Tanuki]

First, I totally agree that the stress associated with typical solid bundled busbars is totally overlooked / minimized. That’s the reason I went to the trouble of assembling 23 6” battery cables from 2/0 welder’s cable. Even in a fixture, the expansion/contraction from a change/discharge cycle and the stress that would translate to on the soft aluminum terminals that were never designed for that stress made me uncomfortable.

But as a counterpoint, I don’t believe EV batteries composed from these LiFePO4 cells are in a true 300Kgf fixture. Assembling into a properly-designed rigid metal case seems to be just as safe (though cycle lifetime may suffer).

I agree. I’m not using my battery in a non-stationary application and movement/vibration no doubt brings along it’s own set of additional risk factors.

I wish everyone using their DIY LiFePO4 battery’s in non-stationary applications all of the best but will also repeat that the cost/impact of burning down your boat or RV is not the same as the cost/impact of burning down your primary residence (so the risk/reward profile is different and may be sensible for many).

After having spent a year building my own 560Ah 24V LiFePO4 battery, I totally agree that going DIY to save money is a false economy. Only go that route if you love the hobby, have an insatiable need to understand new technology, and want to learn…

Just over the past year, we’re seeing better and better and more and more cost-competitive turn-key LiFePO4 battery offerings here in the US (story still evolving).

To be fair, it’s a DIY Forum and it’s still early days. Have a look at a distiller’s forum if you want to get a sense of how the messaging shifts with maturity and once safety becomes a top concern.

sorry but you seem like a player selling premade packs. My assessment of you. Thats 35 years of dealing with the US government and how people manipulate each other. people have been using Lithium batteries for over two decades. while its not as well documented as we would like but they have been safely used for about two decades...

 

[MattiFin]

That is the wrong application for dielectric grease, you should use a product such as Noalox that improves conductivity and prevents corrosion.

Dielectric grease is for applications like inside a terminal block - to repel water and prevent fretting corrosion.

Wish it was that easy:

Phoenix contact(connector manufacturer) says:

"... terminal blocks from Phoenix Contact, the following guidelines on conductor pre-treatment must be followed: • The oxide layer must be removed from the stripped end of the aluminum conductor using a blade (suitable knife, e.g., WIREFOX-D 13 stripping tool). • It must then be immediately dipped in non-acid and non-alkali, i.e., neutral Vaseline."

And Tyco:
"Non-gritted sealants are recommended for flat connections and as a groove sealant in bolted connectors such as parallel groove clamps. Our gritted sealant is primarily used in compression connectors."

https://www.gvk.com.au/pdf/electrical_jointing.pdf

 

[MattiFin]

And since we got greased up lets continue with this

Choosing an electrical connector grease - A product selection guide

Choosing an electrical connector grease can be difficult but our guide will help you find the solution to your fretting corrosion, water ingress or contact wear problem

support.newgatesimms.com

All of them non-conductive mineral/synthetic oil based grease. Also mentions NOT to use silicone grease.

 

[Gazoo]

The most common end of life failure i’ve seen is dendritic growth failure, it results in the cell internal shorting and in the hundreds of LiFePO4 cells i have seen with this failure mode it has resulted in LVD disconnect of the pack, and one dead cell.

Is this due to aging or was cell capacity below 80%......or both? Manufacturers recommend replacing cells when they reach 80% capacity.

 

[fafrd]

The problem I see with this approach is that while you can sense the temperature rise of a cell due to an internal short, there is nothing you can do to stop the process of self-destruction by then.

It seems to me that looking for a change in self discharge characteristics would provide a better early warning sign that a cell is on its way to becoming dangerous.

Early warning / fire alarm was all I was going after, since there is literally no way to ‘stop’ self-discharge of a cell that has shorted internally.

But you bring up a good point - any shorted cell should experience dramatic voltage drop compared to neighboring cells.

Rigging up an alarm that gets triggered whenever the BMS goes to LVD might already provide one level of ‘early warning’ but rigging up a parallel monitoring system that triggers an alarm whenever one cell behaves ‘differently’ than all of the others and specifically when it’s voltage drops far faster than the others would be even better.

This would take a parallel monitoring board and writing some code to establish the alarm conditions / thresholds but I it’s probably the most reliable way to go about this. Heck, in it’s simplest form, the system could monitor the cells on an ongoing basis and send out an alert whenever anything ‘different’ / unusual starts happening.

Such a battery monitoring system would benefit from as many ‘senses’ as it can get, so I’d still be interested in monitoring cell temps and pack pressure.

But this seems like very cheap insurance (at least for use within a primary residence) so my long-term plan will be to add a monitor of this type to my DIY battery.

Thanks for the idea.

 

[upnorthandpersonal]

Making a good connection was the most time consuming part of my build, every terminal and bus bar was sanded, twice, rough polished, coated in dielectric grease, and grub screws thread-locked.

Agreed (in my case, add making flexible bus bars). However using something like MG Chemicals 847 makes for a good and stable connection compared to dielectric grease, and helps even out the contact with the terminals.

 

[fafrd]

sorry but you seem like a player selling premade packs. My assessment of you. Thats 35 years of dealing with the US government and how people manipulate each other. people have been using Lithium batteries for over two decades. while its not as well documented as we would like but they have been safely used for about two decades...

Wow, just when you think you’ve seen everything…

I didn’t say DIY LFP battery’s were ‘not safe’.

I said we really don’t know what risk we are taking purchasing these cheap aftermarket cells from China and connecting to them in ways in which they were never designed to be connected (and that unknown risk is almost certainly greater than 0%).

If I do decide to ditch my hard-worked-on 16-cell battery, it will be because I’ve decided I want welded connections like I’m seeing more and more of the higher-end turnkey batteries using (and which all of the EVs use), and welding battery terminals is a bridge too far for my DIY modest skill set.

The back-and-forth has convinced me that a bit of effort put into a parallel cell-level monitoring system, rather that relying on a cheap Chinese BMS as the only line of defense, is a viable way to address the (small) remaining risk (and within my DIY skill set).

I’ve only seen 2 turnkey batteries I’d have any confidence at all were more safe than the battery I’ve built myself (both welded). Half of the turnkey battteries are nothing more than the same DIY batteries we’re putting together (and using the same components) encased in a shiny plastic case with a nice label…