18650pack cell protection idea. PTC + fuse & diode).

Redeque

1 µW
Joined
Sep 16, 2017
Messages
1
After seeing different types of 18650 pack builds I have been thinking about protection.

I am going to build a 14S6P pack for my E-bike using Panasonic NCR18650PF 2900mAh - 10A.
This will be used with a 60amp BMS and 30amp controller.

After looking at the current/temperature curve for this cell I sure don't want to ever pull more then 10amps per cell.
To avoid this happening (in some case I can't think of yet) using a PTC fuse would be great.
Problem with a PTC fuse is that, if a problem does accure, you can't find it!
So the Tesla style fusing comes to mind but that makes a blown fuse very permanent.
That is when I thought of combining the two options.
The PTC fuse as cell overcurrent protection and a 1A tesla style fuse with diode as inrush protection, in paralel.
With this setup you can never pull more then 10amps per cell due to the PTC and if a cell dies the fuse will blow at 11Amps.

Price wise I am not bothered since it is around 10,- extra for the entire pack.
(100x PTC = 9,- / 100x Fuse = 1,50,- /100x Diode = 1,50)

I this "overkill crazy", "does't work crazy" or just "brilliantly crazy"....?!
 

Attachments

  • 18650_pack_protection_detail.jpg
    18650_pack_protection_detail.jpg
    202 KB · Views: 1,788
Read into how a PTC works. More power means more heat wich means more resistance. It is not a "switch". You will physically burn out the pack due to heat generation.

Buy a proper bms, its cheaper once you add up the cost of building such a rube goldberg sysyem.
 
What about using PTC fuses on the signal wires/balance leads to the bms?

I would like to implement this on a custom pack build for safety. The small wires that go to each cell seem easy to damage and could be a direct short if the battery got hit in a crash. I know several companies use them on all the BMS wires. I built one pack with 18gauge signal/balance leads because it was all I had on hand. I sometimes worry about that one because I feel like as a fuseable link the 18ga. wire could definitely start a fire in the battery bag before shorting itself open.

Do PTC fuses alter the voltage under normal use temperatures such that the BMS cutoff points need to compensate for them? Id like to be able to still use a balance charger with 100-200mA of balancing current. Would that get 22-24 gauge wires with a ptc fuse warm enough to throw off the reading drastically?

Is it as easy as buying a few $ worth of a common size of PTC fuse and just adding them in or is it more complicated than that?

What would be the thought process when selecting the propper PTC fuse?
 
Please be more accurate, a PTC inside a cell is called Pressiure Temperature Current. A regular PTC in electrical terms a polymeric positive temperature coefficient, aka: a self resetting temp fuse. They both work as a fuse but one is permanent and the other is just a fancy resistor
 
I am talking about the auto resetting fuse. Im guessing based on your description that the resistance goes up with temp until its basically "tripped" to stop the current.

I guess I'm wondering if there is such a ptc that would have little perceptable effect on the voltage signal under most normal temps/use of the pack so I dont need to recalibrate the BMS Voltage limits.
 
If you want to go that route then you need a regular fuse, never ever put passive or active parts between the bms and the cells.

Please note that it is still academic, the risk of a cell popping does not cause a immediate problem, you just lose capacity and the bms will keep trying to balanc out the voltage difference created by the popped cell.
 
flippy said:
If you want to go that route then you need a regular fuse, never ever put passive or active parts between the bms and the cells.

Sorry I don't get it. A regular fuse (i.e. non PTC) is a passive element so he cannot go that route either then. Do you mean he should only use a regular fuse after the BMS (as probably everybody does) and not protect the individual balance leads?

I'm also interested in the idea of fusing each balance wire but as the OP, i'm concerned it could alter the balance function if the fuse creates a voltage drop. Yet many people are using balance wires of considerable variable length for each parallel group depending on the design (since some are further appart from the BMS than others) and this is not a subject of concern although the voltage of the farthest group (longest balance wire) should be seen as slightly lower than the voltage of the closest group (shorter balance wire). So in practice would a 2A fuse placed on each balance lead cause substantial voltage drop as to alter the balance function and overall security of a BMS or balance charger?
 
The fuse will only create a voltage drop based on it's resistance and the current flow at the time.

If a fuse has say, 1ohms resistance, and there is balancing going on at max current, perhaps 50mA, then the voltage drop across the fuse would be 1ohms * 0.05a = 0.05v.

That might cause issues with balancing the last bit at the top of the range (causing the BMS to switch the balancer on and off repetititvely), but if there's no detectable current flow (just sensing, no balance), then there is no voltage drop, so it won't cause problems except in balancing (if even then).
 
kilou said:
flippy said:
If you want to go that route then you need a regular fuse, never ever put passive or active parts between the bms and the cells.

Sorry I don't get it. A regular fuse (i.e. non PTC) is a passive element so he cannot go that route either then. Do you mean he should only use a regular fuse after the BMS (as probably everybody does) and not protect the individual balance leads?

I'm also interested in the idea of fusing each balance wire but as the OP, i'm concerned it could alter the balance function if the fuse creates a voltage drop. Yet many people are using balance wires of considerable variable length for each parallel group depending on the design (since some are further appart from the BMS than others) and this is not a subject of concern although the voltage of the farthest group (longest balance wire) should be seen as slightly lower than the voltage of the closest group (shorter balance wire). So in practice would a 2A fuse placed on each balance lead cause substantial voltage drop as to alter the balance function and overall security of a BMS or balance charger?

no, the problem is a popping fuse. once that happens the voltage difference on the bms is suddenly doubled. the balancing and measurement circuit is usually not rated beyond 5V max. so if a board develops a problemn that would pop the fuse in the lines it would suddenly also increase the voltage potential on the board in the problem area and you would get a cascade effect and light the entire board up. so in practice it does nothing and it it has to work its a problem more then a benefit. if something bad happens on the board you want it to happen with enough force to blow the enitre problem into another dimension and make the problem part of the board inert. if you have slow blow fuses you will not get that instant "bang" and you will get smoldering instead of vaporising.

dont forget, the bms does not know there is a fuse. it just sees weird voltages and starts compensating or trowing errors. if you dont know the board is in error because the fuse prevented a big bang you will still end up with a dead battery as there is no way for you to know there is a problem. but if there is a proper bang/pop you will smell it from a good distance you have a problem.

TLDR: you want a catastropic failliure on a bms be as impactful as possible and blow the A/D's to kingdom come, a soft fail just results in a low voltage reading.
 
flippy said:
no, the problem is a popping fuse. once that happens the voltage difference on the bms is suddenly doubled. the balancing and measurement circuit is usually not rated beyond 5V max. so if a board develops a problemn that would pop the fuse in the lines it would suddenly also increase the voltage potential on the board in the problem area and you would get a cascade effect and light the entire board up.

If this were the case, then all those packs where a balance wire breaks off would fry the whole BMS every time that happened. I don't recall any cases of that happening posted here on ES; I believe the balance wire was just fixed and then the pack worked right again.

It would also cause failures of BMSs whenever people unplug the BMS balance connector to store a pack for a while, or for troulbeshooting, if during the plug or unplug process the connector made some contacts before others, as the plastic connector slightly warps under finger pressures, and the contacts slide around in the shell. I don't know of any cases of that, either, though they could be masked by other issues.
 
a balance wire that breaks off is not helped by a fuse. as its considerably more likly it will fail before the fuse then after it making the whole enterprise useless. just use wire with the proper diameter, real thin stranded copper (sillicone wire!) and make sure the connections cant move or shake. and dont have solder joints in the wire. just full runs without extentions or extra connectors.

putting fuses on balance wires is looking for a solution when there is no problem in the first place, and by using fuses you are creating more problems then solutions. KISS still applies.

if you need to remove the bms to prevent it from draining you need to replace the bms for a better one. a bms should NEVER be removed once the battery is built.

not even tesla uses fuses on their balance wires.
 
In my case, the battery would be used without BMS and would be charged with a balance RC charger. Since the balance wires must be connected/disconnected on every charge, withthe balance plug sitting outside the pack, I'd have seen some potential benefits if fuses placed on the balance leads (the fuse would be connected directly to the nickel strips so as close as possible from the battery). But I see your point with the balancing. However I thought balance charger (and maybe some BMS too) shortly cut the charge when measuring voltage, specifically to avoid issues with voltage drop while balancing the pack... Not sure whether this is common or if all BMS/charger do that.
 
john61ct said:
flippy said:
a bms should NEVER be removed once the battery is built.
That is ridiculous, what possible reasoning would support such a statement?

Because a major cause of BMS failure is reconnecting it to the pack. Most of the cheap ones can only handle around 12v on each channels. When connecting, it is sometimes possible to make a battery connection across many channels and a slight imbalance can result in one channel's balancing shunt to turn on, which places more voltage on the remaining channels, and you get a cascading turn on where the last cell to turn on goes way over 12v and fries the channel.

As long as you are careful to avoid this condition, you can safely disconnect and reconnect them.
 
flippy said:
a balance wire that breaks off is not helped by a fuse. as its considerably more likly it will fail before the fuse then after it making the whole enterprise useless.
I didn't say anything about a balance wire breaking off being helped by a fuse.

I said that balance wires breaking doesn't blow up a BMS, based on examples previously encountered here on ES.

That means a fuse in a balance wire wouldn't blow up a BMS either, as it would cause exactly the same condition.

Please show how this condition occurs, if you have seen it, with examples, etc.


putting fuses on balance wires is looking for a solution when there is no problem in the first place, and by using fuses you are creating more problems then solutions.
One possible need for a fuse on a balance wire is if the balance wires are shorted together or to other parts of the pack, etc., by a crash or other unpreventable physical damage to the pack or vehicle.

if you need to remove the bms to prevent it from draining you need to replace the bms for a better one. a bms should NEVER be removed once the battery is built.
Has nothing to do with the point I was making, and you know it.

Please respond to my previous post with examples of exactly how a BMS will be destroyed by disconnecting a balance wire (which is exactly what the fuse blowing would do).

You have repeatedly, in various threads around the forum, made incorrect and/or absurd statements, that may not even relate to the thread or post in question, then when called on it, you simply never bother to respond and explain.


If you're not willing to read posts and respond to the points within them, and explain how your conclusions were made, please don't post anything at all, as incorrect information is simply confusing people and redirecting threads into unnecessary convolutions to correct you.
 
goatman said:
check this link by Chalo

https://endless-sphere.com/forums/viewtopic.php?f=3&t=91925&p=1340872&hilit=hailong+battery+contacts#p1340872
What does that have to do with disconnecting or reconnecting a BMS, or fuses on balance wires?

All I see in the linked post and immediate subsequent ones is a discussion about main pack connectors to the cradle.
 
amberwolf said:
goatman said:
check this link by Chalo

https://endless-sphere.com/forums/viewtopic.php?f=3&t=91925&p=1340872&hilit=hailong+battery+contacts#p1340872
What does that have to do with disconnecting or reconnecting a BMS, or fuses on balance wires?

All I see in the linked post and immediate subsequent ones is a discussion about main pack connectors to the cradle.

don't know why its here, I thought I put that in, power cuts out thread.
 
amberwolf said:
flippy said:
a balance wire that breaks off is not helped by a fuse. as its considerably more likly it will fail before the fuse then after it making the whole enterprise useless.
I didn't say anything about a balance wire breaking off being helped by a fuse.
I said that balance wires breaking doesn't blow up a BMS, based on examples previously encountered here on ES.
That means a fuse in a balance wire wouldn't blow up a BMS either, as it would cause exactly the same condition.
Please show how this condition occurs, if you have seen it, with examples, etc.
putting fuses on balance wires is looking for a solution when there is no problem in the first place, and by using fuses you are creating more problems then solutions.
One possible need for a fuse on a balance wire is if the balance wires are shorted together or to other parts of the pack, etc., by a crash or other unpreventable physical damage to the pack or vehicle.

if you need to remove the bms to prevent it from draining you need to replace the bms for a better one. a bms should NEVER be removed once the battery is built.
Has nothing to do with the point I was making, and you know it.

Please respond to my previous post with examples of exactly how a BMS will be destroyed by disconnecting a balance wire (which is exactly what the fuse blowing would do).

*removed the personal shit*

i will start using smaller words, your inabillity to understand does not mean you need to make it personal.

lets start at the beginning: a the reason a balance wire disconnects or breaks is 90% at the solder joints as they are by far the most fagile connection. in a PROPER pack you only have 1 solder joint (coverd in hot glue/sillicone), on the pack itself. in a pack with fuses you suddenly have a minium of 3, 2 of them are before the fuse. so a breaking connection that shorts out is most likly to happen BEFORE the fuse, not after. its also WAY more likly to happen because solder joints are weaker then the crimped connector.

so solder joints = bad.

if you have a crimped connector wire that is iffy you need to fix it, use proper connectors, change the bms, use better crimping tool and/or better quality wire. get sillicone insulated wire whenever possible and use hot glue on the wire/connector part to stop viberations from degrading the connection. that way its basically impossible to get ANY physical problems barring shitty pack construction and not glued/taped down the wires properly.

a fuse is NOT a solution to solve shitty pack construction.

most of the bms that are out there (chinese crap) use A/D converters to measure voltages, try removing/breaking a single input. the bms generally does NOT fault out when that happens, the voltage is just lower on 1 or more groups depending on the implementation. so a fuse does not do anything to help in the most common issue. and even worse: you dont notice it because the bms does not fault out because it still measures voltages.

fun fact, a fuse does not work because the bms is not isolated, current can and will flow FROM the bms to the short so popping a fuse one one end does nothing because the current can still flow from the other end.

high end partially isolated A/D converters like the ones you see on a tinybms for example dont have this issue and will correctly report a fault BUT most of the time in the litteraly dozens of those i have used its usually the A/D converter itself that shits the bed when something shorts out the wires and instantly blows up the A/D converter bricking the bms because these things generally pop at anything above 5v on a single input. and yes, i have had serveral tinybms that failed at the converter (nice chuncks blown out of the chip) that had to be warrantied. and no, i dont keep collages of broken boards, i have a business to run.

you want a bms to fault out hard, and always trigger when something breaks. the cheap chinese ones often do not do this, they will happely continue working with missing balance wires and still register voltages on those missing lines.

if you are in a accident and you have the problem that the bms if physically separated from the rest of the pack and balance wires are ripped apart i will bet you have bigger problems at that point. a couple a fuses will not help you in any way at a point that the pack is physically destroyed. especially if the pack is inside the frame of the vechicle....

i will bet a crate of beer that anyone that is folded around a lamppost at 35mph+ is not thinking about the balance wire fusing of the bms.

in short (little pun), there is no reason to add fuses and add even more points of falliure then you already have. if you build shitty packs that tend to blow up you need to fix that instead of trying to make up solution from a problem that should never be possible in the first place. using dozens of fuses to fix bad construction habits is like the twisted logic of saying its ok to drive 150mph because the car has airbags.
 
Panasonic and other cells already have a thermal fuse in them, and the datasheet has details of destructive test results (usually they test a sample by subjecting them to dead short, over charge, over temp etc)

The reality is there are millions (billions?) of lithium ion cells in service all over the world and they are incredibly safe. If you design the pack to run within discharge and temp specs, keep an eye on cell balance; you don’t really need any extra protection.

If you can get it nicely spot welded and use a BMS rated for double the current you plan to draw; you’ll probably get years of good service out of it.
 
flippy said:
lets start at the beginning: a the reason a balance wire disconnects or breaks is 90% at the solder joints as they are by far the most fagile connection.
While every point you make may be applicable to the stuff you yourself are talking about now, almsot none of it answers any of the questions asked.

I wasn't being personal--I was trying to get you to answer specific questions, but instead you were (and are) going off on tangents that don't do so.

I don't need small words--I understand everything you said. But your first reply to me did not answer the questions implied in my post here:
https://endless-sphere.com/forums/viewtopic.php?f=14&t=90581&p=1563612#p1563447
and the actual questions in the post you just responded to.

This time, you finally did answer part of the question:
flippy said:
high end partially isolated A/D converters like the ones you see on a tinybms for example dont have this issue and will correctly report a fault BUT most of the time in the litteraly dozens of those i have used its usually the A/D converter itself that shits the bed when something shorts out the wires and instantly blows up the A/D converter bricking the bms because these things generally pop at anything above 5v on a single input. and yes, i have had serveral tinybms that failed at the converter (nice chuncks blown out of the chip) that had to be warrantied. and no, i dont keep collages of broken boards, i have a business to run.


However, most of the cheap BMSs I've seen (similar to the ancient Signalab) do not use A/D converters at all, nor MCUs, smart BMS chips, or any of that advanced stuff. Most of the cheap ones I've seen (including those in threads with broken-wire problems) misuse the individual voltage detector chips.

How does a disconnected balance wire (regardless of reason, method, etc), blow those up?

And if it does, always, every time, do so, then why doesn't every person with a broken balance wire in a pack (regardless of reason, method, etc), or that disconnects a BMS balance plug (regardless of reason, method, etc), end up with a destroyed BMS every single time?


I'm not trying to be a dick; I'm trying to get a real answer to the question, so that anyone that comes along will understand exactly what is going on, so that they will know why whatever idea they have will either work or not work. Or why when they have a specific problem, a specific thing happened, and thus what they should do to fix it.
 
amberwolf said:
flippy said:
lets start at the beginning: a the reason a balance wire disconnects or breaks is 90% at the solder joints as they are by far the most fagile connection.
While every point you make may be applicable to the stuff you yourself are talking about now, none of it answers any of the questions asked.

I"m not being personal--I'm trying to get you to answer specific questions, but instead you're going off on tangents that don't do so.

I don't need small words--I understand everything you said. But it does not answer the questions implied in my post here:
https://endless-sphere.com/forums/viewtopic.php?f=14&t=90581&p=1563612#p1563447
and the actual questions in the post you just responded to.


do i really have to point out that a wire can short on BOTH ends?

a bms is NOT insulated. so fusing wires does NOTHING as the current simply flows from the main negative to all the other balance wires that the fuse broke from.

grab a bms, connect a 20S battery, add fuses on the pack balance wires (at the pack) and short out the S3 and S15 line for example at the bms and tell me what the bms did. i will bet it involves LOTS of popping and burning pcb smoke, all the fuses will pop and the short will not be stopped because the bms is directly connected to the main neg so you get main pack amps going right into the balance board as the A/D converters are not isolated. so no, a fuse will do exactly shit when its needed.

ps: voltage detectors are A/D converters. they convert the voltage into something the main chip can use. its a family of devices. you have cheaper and more expensive iterations.
 
I was in the middle of writing and rewriting the post, and accidentally hit submit instead of preview, so what you quote there isn't the version intended, and is not complete. :oops:

If you can, please read the version I just finished (the same post ID), and respond to that one instead.



EDIT: Nevermind. I'm obviously just too braindead, sick, and exhausted to grasp any concepts here. Maybe someday I'll get better.
 
amberwolf said:
I was in the middle of writing and rewriting the post, and accidentally hit submit instead of preview, so what you quote there isn't the version intended, and is not complete. :oops:
If you can, please read the version I just finished (the same post ID), and respond to that one instead.
EDIT: Nevermind. I'm obviously just too braindead, sick, and exhausted to grasp any concepts here. Maybe someday I'll get better.

there is a difference between a disconnected/broken wire and a dead short. one is fine-ish, the other is not.
right now i am just considering dead short situations.

that is probably a disadvantge on forums, assumptions dont transfer well.
 
Back
Top