Fishpaper and cell level fusing, USE IT!

I am currently working on a PCB for my next build that will have SMD fuses for each cell and also a passive temperature sensing which means that while the cell is not over heated the temperature “measuring” does not require any input power yet still is active. Are you guys interested in such a thing? Like this if you are so i can gauge the interest. :slight_smile:

Here is a thread with a simpler design, which I have already improved and I am waiting for a quote from a factory. That will be up this week hopefully.

8 Likes

It will not be fun when riding your power is gone and brakes do not work…

The temperature will be probably set to 70°C which is 10°C above the safe temperature. At that point it is about having the board explode/burst into flames (or other catastrophic scenarios) or to stop the temperature from rising.

It’s better to stop using burning board then facepalm at 40mph I guess…

Plus the temperature difference of inside the cell and the case + plastic wraping will be like 10-20ºC off

If you reach that temp even cutting power will not help you it will be prompt to catch fire by self igniting.

In a scenario where the cells are actually getting to such temperatures the difference between the temperature sensor and the inside of the cell or its metal body might be different but there would be a small time difference/delay between the cell reaching say 60°C and the sensor reaching this temperature, the delay may be a couple of seconds because of the minimal distance between cell and sensor.

I did not mean that this is supposed to prevent your board from being damaged (not in first place at least) but to prevent your board from destroying more things or hurting the user. Say you have your board charging in your garage and the cell is slowly climbing in temperature, when it reaches the temperature threshold of the sensor it will get disconnected and stop charging therefore stop heating up more since the power source is cut. Another scenario would be (in most cases) when you are climbing steep hills and your battery is providing high current to the system, in such a scenario you dont need to worry about brakes.

If your battery is on fire as you said you wont be able to stop because the battery would not work. Inside the battery is a plastic-like electrolyte in form of a film that separates the two other “films” being the opposite potentials. In this case when the film melts it shorts the battery which would mean that the battery is useless which results in the same thing, your board would turn off plus the fire. This is what I am trying to avoid.

When battery catches on fires it still works as a resistor so at least your brakes will work. Ask people how do they like then BMS cuts the power randomly…

1 Like

When the battery catches on fire/gets pierced etc. it becomes a 0 ohm resistor since the two terminals become shorted on the inside.

The temperature (70°C) is a temperature where the battery can still operate but it is getting to the dangerous state. I had a look at a 30Q specification and that is the temperature to which a 30Q gets after 20A discharge which is 5A over limit.

I should say that I am basing this on research and theoretical education, I have never measured a resistance of a battery on fire.

2 Likes

It your using a Vesc Brake will still work a bit event if you lost complete power, because of the regen. It happens to me quite a few time while trying prototype bms.

1 Like

Oh yeah I forgot about that, thank you! If I am not mistaken it is the same thing as was discussed in a thread about riding with battery disconnected. Where the motor acts as a generator and powers the VESC or other ESC for that matter on.

You should try it. It’s pretty easy. You can buy A4 size, 2mm thick copper sheets from hobby stores. They use them to scratch build model trains or WW2 scenery. It’s cheap and you can cut it to any size buss bar you want. Here in Australia, old houses still use fuse wire in the electrical mains box. So rated fuse wire is readily available.

Haven’t had any issues with my battery set up so far.

How much of an issue would it be if the p-groups are laid out long ways. Would that mean anything at all or would it be fine and dandy? I am thinking that it might be an issue with the closeness between the two groups potential still making the fused cell in series with the next group. Especially with the hauge strips welded to the positive terminal. I guess you could just make sure to isolate with fishpaper between groups if that makes any sense.

@chaka @Jinra

What happens when a fuse blows? I mean, do they blow one fuse at a time, or the whole P group of cells blows?

I worry about the following scenario: If someone rides hard so that a fuse blows, then the current through the remaining fuses in the same P group sharply increases, so that all these fuses are also likely to blow at the same moment. Now this guy ends up in a situation when he rides very fast and cannot control the board - not good.

I guess this is a lesser problem for a 6P setup. But I am planning to build 10S4P pack with 30Q cells. Should I worry about this?

http://www.electric-skateboard.builders/t/ollinboards-cell-level-fuse/7392

When mine blew it would only be 1 or 2 fuses each time for which I’d just resolder thicker wire to replace. It’s never failed completely for me.

@scrapheap Thanks, I’ve carefully read through that thread. But my concern still remains.

For my first build, I was thinking a lot about using a fuse on main battery terminals, but after I read that @Jinra had an accident because of a blown fuse, I decided not to use it:

Cell level fusing looks like a safer option because one blown fuse does not mean that your esk8 shuts of. But it may not be that much safer in this respect.

As many people wrote here, fuse wires often come loose (because of vibrations, bad soldering, etc.) If one or more fuses wires come loose, you many not immediately detect it unless you balance charge your board every time. (And even if you check balance every time, you can detect it only when you charge.) The cells will become unbalanced, and I am not sure that a BMS can help with this.

So the battery pack may be compromised, but the rider will be totally unaware about this.

Say, in a 4P setup, if one or two fuse wires in the same block are disconnected, then this P block is effectively 3P or 2P. If the rider puts it under stress (hard acceleration, going uphill for a long time, or both at the same time, etc.) it may easily reach the breaking point, where it was working fine last time. If this fuse blows, this immediately makes the current through the remaining fuses in that P block (which are already close to blowing) 2 or even 4 times higher than they should be. I am pretty sure that these fuses will immediately evaporate. So the board will shut off at the moment when the rider least needs it.

Tesla uses cell level fusing, but they have huge number of cells in P blocks. If a few fuses blow, it will not have a big effect on currents through remaining fuses. For esk8, where people typically use 3P or 4P setups, cell level fusing is more questionable.

I am not trying to say that it is bad. I am actually trying to figure out if and how to do it in a safe way for my next build. But there are lots of questions here.

With commercially available fuses, we know exactly when they are supposed to blow. With fuse wires we can only guess it. There are so many parameters here, not only the gauge of wires, but only the length, the amount of space and ventilation around them, heat dissipation, etc. Basically, there is no way to figure out the blowing amperage, unless one does a thorough testing of a given battery design under various conditions. Apparently @chaka does it for his packs. But I am not sure that I can do it (or even have means to do it).

If the amperage is too low than the battery is not safe (can shut of easily), if it is too high than cell level fusing defeats its purpose.

1 Like

@Jinra Thanks. While I was typing my post, I found that you already answered some of my concerns.

So what is your recommendation? No fuse on main terminals, but cell level fusing?

Do you recommend 20 awg wire? (In earlier posts I found that you advice against 22 awg.)

I am trying to figure out how to do it for 10S4P pack arranged in two layers.

Here is my idea:

For each 4P block cut a blank PCB to the rhombus shape that covers the positive ends of the cells. Drill 4 holes in the PCB for the cathodes of the cells. Solder a 12 awg flat wire to the PCB so that it goes between the holes. Glue this PCB to the positive ends using silicone. Solder fuse wires between the cathodes to the flat wire.

So fuses wires will be very short - probably a few mm. But I don’t know how to figure out the correct gauge without doing extensive field tests under various riding conditions.

Any suggestion? Do you think that this will work?

Seems like a bit of a design limitation in the way we use BMSes. In theory they should be able determine when a P group isn’t healthy (say a cell fuse blew), as the voltage sag would be higher. With the way we wire them up they have two choices; keep the current flowing, or shut down to protect the pack (and forget the rider).

Really what you want is for the BMS to tell the speed controller that the battery pack isn’t doing so well, and to back off the current. I’d imagine something like that happens in Boosted/Evolves.

It would be much better if, instead of shutting down, a BMS would make a warning, say, a loud beep if something is wrong. There is a simple LiPo voltage tester that does just this - gives a loud beep if voltage of some cell gets too low.