DIY 12V LFP starter battery for my motorcycle.

[rebelrider.mike]

A fun thought experiment for now, that if it goes well, I may actually build.
Recently I discovered that my old store-bought LFP battery kicked the bucket, and I'll probably replace it with a lead acid battery. But I have almost all the parts sitting around to build an LFP battery, and I figured designing it would be fun. I have 16 LFP cells sitting around from the failed cargo camper project, and I'd like to make some of them useful. So here's a few considerations I've been thinking about:

- Will the battery get too hot? No. The motorcycle has no engine bay to collect heat. It will be sealed in its own container away from the engine.
- Will the battery get too cold? Kind of. I generally don't ride when the temperature is below 40F, so it's not usually an issue. I may have some flat 12V heating elements that could go in between the cells. That might be more efficient than using the current to the headlight to warm up the battery before starting. (Which is the conventional way to do it I'm told.)
- Will the battery turn the engine over? Maybe. The maximum momentary current for each cell is 300A. That should be enough to turn the engine over. But I'll need to give it a try.
- Will the alternator over Volt the battery? No. The maximum voltage of the battery would be 14.6V. The alternator puts out less than that. I don't remember the exact value.
- Will the alternator over Amp the battery? No. The alternator maxes out at 40A. The cells are rated for 60A maximum charge current.
- Will the battery over Amp the alternator? Not sure. If the battery were dead and had to be jumped... I'd have to measure how much current the battery would draw. Worst case: the inconvenience of not using the motorcycle until the battery is fully charged.
- Is a BMS feasible for this application? Not sure. Theoretically, I'd need 300A to turn the starter motor, and a few more Amps to run the headlight and ECUs. So I guess I'd be looking at a 400A BMS. But 99% of the time, the battery would see less than 40A in or out. Also, a 400A BMS is going to be way too expensive.

The old battery didn't have a BMS. It did have a tiny cell balance board. I'd want at least that, plus a way to check each series, and a fuse on the positive wire feeding all but the starter motor. Something like this:



Maybe if I could find a 40A BMS that used the positive side instead of the negative side. It would at least protect the battery when not starting. Anyone have any creative ideas there?

 

[Korishan]

and I'll probably replace it with a lead acid battery.

 

[floydR]

What are the Lifepo4 cells you have on hand?
Many starter LiFePO4 batteries bypass the BMS for discharge using it for charging only.
My DIY LiFePO4 jumper box 4s4p(headway cells) 32 Ah uses a 100A non smart Daly BMS with all the BMS functions except overcurrent protection.

my 12v battery box

the main purpose of this battery box is to get a 6 cylinder diesel tractor to turn over, at least the get the starter unstuck. The tractor is an old Oliver row crop model. there are otherthings wrong with it but getting starter unstuck. is first on the list. The tractor still has 12v positive...

secondlifestorage.com

Later floyd

 

[rebelrider.mike]

Interesting. I hadn't thought of a bypass relay. I see you had a positive ground. I'll have to think of how to adapt that to a negative ground.
The cells I have are Liito Kala. Lots of bad reviews on You Tube, but they were what I could afford at the time.

 

[floydR]

The postive ground was on the Oliver Tractor.
The battery box is set up with a nonsmart daly LiFePO4 bms.
There are BMS's just for charging.
The specs on the litto kala look good but as you say reviews haven't been that good.
Later floyd

 

[rebelrider.mike]

I've been testing my Liito Kala cells, and so far, they've lost very little, or no capacity. I'm still on cell 10 of 16 for the discharge testes, so still a ways to go.

In the meantime, I've been looking around for possible BMSs and haven't found anything impressive. Closest I can see that would work is a BMS with Both P/B-, and P/B+. I could wire that up and just bypass the starter. Power never comes back from the starter, and low voltage would be immediately evident when the engine doesn't turn over. I'm not sure how those types of BMSs work though. Do they cut off current through both the negative and the positive? Cutting off the negative won't work. Current would still go through the big negative cable needed for the starter. Can't route that through the BMS unless the BMS is rated for 300+ Amps. Maybe I could use an extra solenoid to cut off the big negative when the starter isn't active. I'll have to think on that.

Anyway, the best thing I think a BMS could do that a fuse and active balancer can't is shut off output from the battery if one of the cells "dies". Or shut off input to the cell for the same reason. Here's my latest idea:


I have a few depleted cells available now that I could use to test if the alternator would overwork itself trying to charge a dead battery. I need to move the bike into a sheltered area so I can do some experiments.

 

[Korishan]

I would just go with a decent JK BMS. They have come a long way on their capabilities and you can easily monitor with your phone via BT. They are designed for a specified current and can surge for a brief moments.

JK-BD6A2412P - JK BMS JIKONG BMS JKBMS JK-BMS

Overview A lithium battery smart BMS, designed for big capacity series lithium battery packs, is a management system that provides a range of features, including voltage collection, active large current balancing, overcharge, overcurrent, overtemperature protection, Coulombmeter, Bluetooth...

www.jkbms.com

jk-bd6a20s20p can do 350A surge

https://www.aliexpress.us/item/3256805517850287.html

Actually, several listing on that store shows they can do 350A, 450A, etc.

Mentioning that because you can still do starter through the BMS this way. If you absolutely still want to bypass the BMS for just the starter (nothing wrong with that tbh) you can go with some of the lighter weight JK units. They are still in the $50 range for the good ones. You might be able to find some in the $30 range for ones that are previously used or were in old battery packs that were broken down.

Most of the BMS only cut power through the Negative rail. The unit is powered by a single Positive thin wire; at least with the JK BMS's.
You don't have to cut power when starting. Just bypass it. Altho I'm not sure how well they would handle the Negative side being connected to both sides, the Load and Battery sides of the BMS. I wouldn't disconnect, though. That can upset the BMS and possibly cause issues with the balance leads as they would then be trying to carry current.

 

[rebelrider.mike]

I've heard way too many bad things about JKs to ever trust one. But here's my idea on how a negative only BMS might work:


Both the regular and negative solenoids would activate with the key switch. If the negative one were to fail, I suppose the current would try to go through the BMS. Maybe too much would anyway? Current takes all paths to ground, not just the easiest one. So this setup may trip the BMS anyway. I'm kind of thinking that a BMS isn't going to work for this. Unless it's 350A. But if it costs more than $100 it's not worth it. I can get a very reliable lead acid for that much.

 

[Korishan]

I've heard way too many bad things about JKs to ever trust one

such as ???? I know in the past several years they've been extremely reliable.

Are you sure the starter pulls that many amps?
Yeah, everything is grounded to the Negative anyways, so yup, going through the BMS is almost a certainty.

 

[rebelrider.mike]

Well, I'd rather not start bad-mouthing a company that I've never dealt with. I probably shouldn't have said anything.

The bike has an old style starter motor that is very inefficient compared to modern ones. It also has an EFI computer that won't fire the spark plugs if the voltage gets too low. Urals are weird combinations of old and new technology. There is a redesigned starter motor available, but I can't afford it. I actually rebuilt it last year. Even made a 3 part video series. It was fun and I learned things.

I'm actually not sure how many Amps the starter needs. I've read 300A but have never tested it. I did manage to get the bike under the shelter yesterday, so soon I can actually test it. A lot of other Ural owners have ditched their motorcycle batteries and have installed much larger car batteries in the trunk of the sidecar. I didn't want one taking up my trunk space, so I opted for the much lighter LFP battery that I put in a box near the starter. I also put larger than stock wires in to feed the starter motor and ground back to the battery negative.

Anyway, I have some car batteries and a clamp meter that can read up to 600A, so I plan to do some experimenting.

 

[rebelrider.mike]

I finally have some real numbers to work with!

First off, the bike uses 10A just to run the lights, fuel pump, ECU, and whatever else it might be using. I measured current through the negative cable to be sure to get the total current of everything at once.

All the batteries I used sent around 270-280A to the bike while running the starter.

So the truck battery was not so great. Since I don't drive the truck much, the battery has been sitting on my bench on a maintainer. I hooked it up to the bike and got the engine to turn over, but it went a bit slower than I would have expected. Turns out my load simulator tester only read 200CA out of the 700CCA rating. Looks like it's time to replace it.

The lawnmower battery did much better. Turning the engine over was a bit slow, but the battery is only rated at 285CA. So it was being used at its maximum rating. It tested at 600CA on the load simulator. I guess that means the simulator is good for gauging the health of the battery, but not necessarily it's actual CA ability?

Both of these batteries are ordinary lead acid by the way.

I put together two batteries from my pile of LFP cells. One that was fully charged, and one made of cells that were fully depleted after discharge capacity testing. The resting voltage of the charged battery was 14.2V, and the resting voltage of the discharged battery was 12.1V.

First I tried the fully charged LFP. It turned the engine over faster than I've ever heard it. Even with a brand-new 300CA motorcycle battery. It sounded like turning over my old Honda Rebel engine. Anyway, after starting the engine, I looked to see how many Amps were being sent back to the battery. At idle, a couple of Amps were still being drawn from the battery. But at some RPM I forgot to check, I was able to get 27A back into the battery. I'm guessing it might have been around 3,000RPM, but I can't be sure. Just for fun, I put the charged LFP on the load simulator and it was off the scale. Well over 1,000CA. I'm estimating at least 1,200CA.

Next I installed the discharged LFP. I wasn't expecting it to turn the engine over, so I wasn't looking for starting Amps. But it turned the engine over just as well as the charged LFP. Very surprising! What I really wanted from the depleted battery was to see if it would draw too many Amps from the alternator. It's only rated for 40A after all. Turns out as long as I keep the RPMs below 3,500 it should be ok. Might still be a little high as the bike is using 10A to run itself, so the 36.7A I got at 3,500RPM would be in addition to that I think.

I had assumed that a "dead" battery would not start the bike and so I'd be alerted to an issue. Turns out this is not the case. I'll need a Volt and Amp meter on the dash to keep me from overtaxing the alternator if I build this for real. I'll have to think about all this data and decide what to do with it.

 

[rebelrider.mike]

Well, I've been thinking on it, and I've decided to forgo the BMS. But I do want to add one more safety feature: a current limiting resistor. Also, it turns out I don't have any 12V heating plates. I thought I did. So I'm leaving them out of the design too. As powerful as those cells are, I wonder if heaters would even be needed in 40F weather like the old battery. An experiment for next winter.

So for the limiting resistor, I've had to renew my understanding of resistor math again. For some reason it just won't stay in my head. The condition where it's needed is if the battery is low voltage, and the engine is running high RPMs. Like on the freeway. The rest of the time it's not needed and I'll have a bypass switch.

In my earlier experiment with the low voltage battery (12V), higher RPMs were sending about 40A into the battery. I don't remember the exact voltage the alternator puts out, but I'm guessing not more than 14.6V. The voltage difference, 14.6V-12V=2.6V and the current, 40A suggests that the circuit might be something like 2.6V/40A=0.065Ω. I don't know if that's enough to be relevant, and I also it probably changes depending on the temperature and state of charge of the battery, but it's somewhere to start.

The electrics on the bike need about 10A to run, and the alternator is rated for 40A, so more that 30A going to charge the battery is going to be bad for the alternator. Worse, the maximum DoD of this battery would be closer to 10V (but that's under load) so I'm thinking a resistor to slow down the current could be useful.

I'm thinking if the battery is really low, like 10V and the alternator is making 14.6V, then the difference would be 4.6V. A resistor capable of reducing that 40+ Amps of current down to a more reasonable 20A would be 4.6V/20A=0.23Ω. With the circuit already having a resistance of 0.065Ω, I'd only need a resistor with 0.165Ω. I plugged all that into LTspice and sure enough, it worked out.

New problem: This resistor would have to dissipate some heat. The voltage drop across the resistor would be 20Ax0.165Ω=3.3V. And so the heat dissipated would be 3.3Vx20A=66W. Looking for such a resistor, I found one for over $900. Screw that! I played around on my spreadsheet for a while and came up with putting six resistors in parallel. Each one would be 1Ω, and would only need to dissipate about 12W. I can buy a 10-pack of those for less than $10. A bit more reasonable. I figured 1/(1/1Ω)x6=0.166Ω. Close enough right?

With all that plugged into LTspice, I got this:


Some good news I hope, is that this could all be off as much as 10A in either direction and still be fine. I spent a while on the simulator trying out different conditions and it all seems to work. I've also updated my diagram:

 

[rebelrider.mike]

I've started building the box that will hold the battery. Many years ago I relocated the battery, and it has had its own custom box ever since. Here's version 1:

Turns out I didn't like the new location, and the box rotted out from weather exposure anyway. So I built version 2:

That one went on the nose of the sidecar, and worked very well for several years. Both boxes were made to resemble an ammo box. In fact, the second box has the same outer dimensions as a real ammo box, except its a bit taller. It is still in ok condition, but it flooded once and the volt meter was ruined. I could refurbish it, but since I'm building my own custom battery, I've decided to make the box out of a real ammo can.

The old battery was mostly empty space, with a small 4s2p LFP pack inside. Here's a comparison of the old cells vs. the new cells:

The new box has enough room for the four cells, but it's a bit tight height-wise. I've transferred the mounting bolts, and added a couple brackets to support the box once it is on the bike. I also removed the main wires for the battery and cleaned them up. The ends were re-crimped, and the failing vinyl tape was replaced with shrink tube.

I'm experimenting with cell placement and adding EVA foam, Styrofoam, and hard plastic sheets to protect the cells electrically and physically.

 

[OffGridInTheCity]

You do the most interesting projects! Appreciate the sharing

 

[rebelrider.mike]

Thanks!

I got the old hardware cleaned up and ready for paint.

The old bolts had shoulders which was great for the plywood, but not for the sheet metal box. So they got replaced. Also discovered the ammeter fits into a pipe coupler. So you know I've got to do something with that.

 

[rebelrider.mike]

Experimenting with insulation/protection placement again. Hard plastic on the lid, and probably Styrofoam. Spacers between cells to keep them from moving. Mounting bolts and brackets are in place. I've soldered wires to two bronze washers to make one of the bolts and the ammo can grounded to the negative of the battery. Holes drilled for the big negative cable and the starter positive cable. Found some gromets for the big wires too.

 

[rebelrider.mike]

Got the current limiter designed and built.

Also fit the ammeter into some PVC bits. Not sure how I'll mount it yet.

Still experimenting with bus bars, but I've got the balance wires for the active balancer and the cell checker done.

Instead of going by an Amps/Gauge chart to size wires and buses, I've decided to go by voltage drop and Watts lost to heat. I'm not sure how many Watts are ok in a wire or bus bar, but I'm guessing if its pretty low then it should be ok. The important thing is how the Watts lost translates to heat.

 

[rebelrider.mike]

Finally figured out how I want to mount the ammeter. Got all the pieces shaped, painted, and assembled.


With that and the battery box pretty much done, its now time to install it on the bike and see how it all works. Unfortunately, that may have to wait for a week or two for nice weather.

 

[rebelrider.mike]

I've got the bike back together and ridable!
The ammeter went on with more difficulty than I thought, but it's on.


Got the new battery box installed and wired.


My son and I went for a ride yesterday for a couple hours. For the most part everything worked great. I need to crimp the small positive cable connection better. And at about 4k RPMs, the ammeter goes all spastic and I can't get a reading from it. Not sure if it's vibrations or some kind of electrical thing. I may need to re-think having a mechanical gauge. Which is too bad because digital stuff doesn't really go with the look of the bike.

The DIY bus bars I made work perfect. Didn't even get warm after starting the bike. No issues with the battery at all. Now I get to ride around and test it out long term. What a chore.