My first DIY electric motorcycle

[SafeDiscDancing]

The first batch is testing now and It's looking good so far. The first 20 cells have all been within 0.05v of each other and are charging nicely to 4.2v.

Don't forget:

Never let the cells go below about 3.0 volts.

Never let the cells go above about 4.2 volts.

When in storage always leave the cells at from 3.5 volts to 3.9 volts.

So are you thinking about that?

It's okay to cycle test these cells, but simply charging them to 4.2 volts "just to prove they can make it" then letting them sit fully charged is already your first mistake.

Lithium cells love the middle voltages and hate the extremes of high and low.

I never leave my cells full and even 4.1 volts (which many do for storage) is a bit high.

No real damage occurs with a middle voltage cell so for maximum life don't leave cells fully charged.

And if you are planning to build a pack I'd suggest getting them all down to 3.0 volts before you start spot welding.

 

[Millhouse_5]

Yup, I'm accounting for all that. Before I even got my first bike I got back into electric RC planes and cars so did a lot of reading on how to treat lithium cells as a lot of my battery knowledge was a little dated with NiCd and NiMh cells.

The cells will be charged on my programmable hobby charger rated for use with any kind of battery from lead acid to every lithium construction. It's got charge, discharge, balance, and storage modes. and once they are tested on the slower 18650 test charger, they will be cycled back onto the hobby charger and put at a "storage" voltage, which is around 30-40% (forget the exact voltage at the moment). The hobby charger can charge 8 at a time at 0.5c so I'll be able to feed the two testers I have constantly and keep the process going for at least 16 cells a day. If I can keep it up, that's 35 days of testing, which I'm ok with for now. I'll let you know if I begin to go insane after 2 weeks of constantly feeding chargers with more and more batteries.

Once a cell has been tested and passed, I'll apply the terminal ring paper for protection so I don't double test any cells. Once I have a good collection of cells tested and can start matching strings of 20 in groups then I'll start spot welding and get the first few parallel strings constructed.

I know it's rather ambitious for my first even battery build, but I'm willing to give it all the effort it deserves. I don't want a motorcycle with a crappy range because of one bad parallel string letting down the voltage curve.

 

[Millhouse_5]

Good news, bad news.

Good news:
I no longer have to test all 560 of my cells which would have taken over a month of non stop charging and testing.

Bad news:
My testers are so inconsistent that the same battery tested on the same tester, but in a different slot, returned two VERY different results making any and all tests I've done or will do, pointless.

As you can all guess, I'm somewhat upset. The testers aren't a complete loss as it can still tell me generally what kind of capacity I have in a cell, but not to the point of using the data to build a perfectly balanced cell by cell pack. Basically I noticed I was getting more "failed" cells on slots 2 and 3 and more "good" cells on slots 1 and 4. Moved a failed cell from slot 2 to slot 1 to test and it gained about 10% capacity. This test was repeated with all of the lower rated cells I had and confirmed that there is a repeatable and predictable pattern. As I don't know which is correct, the inner or the outer banks, or neither, all my data is invalid. I also tested a previously failed cell in the same charge bank to see if it's a cycle thing and it returned failed a second time.

All that out of the way, the data isn't complete garbage. It showed me that my first 60 or so cells are all testing within the manufacturer specified range. So odds are that this batch is a pretty safe bet.

I don't want to just start slapping a pack together and let the rule of large numbers take care of the balancing, but it's looking like that's what I'm going to have to do. And no, I don't plan on purchasing a professional tester or sending off all the cells to an independent contractor for testing. That would likely cost more than just building another pack if this one dies a little early.

So, may the random number gods be in my favor as I build this pack and let me get something at least mostly balanced. With a 20p pack, the odds of getting an entire bad string of cells is astronomically low so I'm going to let the dice land where the may on this one.

 

[SafeDiscDancing]

Bad news:
My testers are so inconsistent that the same battery tested on the same tester, but in a different slot, returned two VERY different results making any and all tests I've done or will do, pointless.

Welcome to the world of cheap Chinese electronics riddled with inaccuracies and outright errors.

Yeah... I had bought a simple battery cell voltage monitor and applied it to my battery and at first I thought "oh great, now I can read six cells at a time (per module) and it will do all the work of balancing FOR ME."

My childlike innocence was in a state of trusting bliss.

Then it started to seem a little weird that the balancer was moving energy around in crazy ways and always seem to favor one cell over the other.

I grab my voltmeter... oops... everything the balancer was saying was COMPLETE BULLSHIT.

It was murdering my battery after all.

So be very Thankful that you learned to be very, very skeptical now before your trust gets devastated entirely.

It's the trust in these badly made devices that cause most of the battery fires. (in my opinion)

The only way to be certain of repeatable performance is to test everything with the same tester.

Expect the testing to be completed by about 2035. :lol:

But seriously... I'd say randomness will be "good enough" to protect your investment.

Just be sure when you get to balancing your entire pack that you remember this nightmare... the BMS options do all this same stuff and this is why I've gone total caveman and balance and manage my own batteries. (bottom balancing)

One last thought... don't forget that if you plan to spot weld these cells together that the welding process itself can cause minor errors in your overall pack design. So even if you "had" perfected the capacity matching beforehand the assembly process would likely screw it up somewhat anyway. Just relax and do the best job you can.

 

[Millhouse_5]

I dont trust anything I buy, china overstock or otherwise, which is why I bought the testers in the first place. So I like to test most of the things I buy.

As for bottom balancing, I'm afraid I'm going to have a hard time once this pack is all assembled and covered in heat shrink. So ill have to do what I can with an at least we'll reviewed 28s bms. At this still early phase in electric vehicles it seems that anything that can handle over 100v is rare and there doesn't appear to be anything high end available at all.

 

[SafeDiscDancing]

As for bottom balancing, I'm afraid I'm going to have a hard time once this pack is all assembled and covered in heat shrink.

So you plan to completely bury your balance wires inside the pack making the whole thing 100% dependent on a BMS?

Why not design things so the BMS is connected outside the pack?

Then you can choose to manually check it or bottom balance if you felt like it.

Just bring all those balance wires outside.

 

[CONSIDERABLE SHOUTING]

I dont trust anything I buy, china overstock or otherwise, which is why I bought the testers in the first place. So I like to test most of the things I buy.

As for bottom balancing, I'm afraid I'm going to have a hard time once this pack is all assembled and covered in heat shrink. So ill have to do what I can with an at least we'll reviewed 28s bms. At this still early phase in electric vehicles it seems that anything that can handle over 100v is rare and there doesn't appear to be anything high end available at all.

It's also an issue of rarity and danger; 100 Volts can do some damage, and I know of only a couple motors offhand that support that voltage natively. I think it's also an issue of need too, like individual Tesla packs are only 25 volts; I think it's just once you hit 20 series connections the assumption is you would be better off having individual packs.

I don't understand the evangelizing of bottom-balancing. Keeping your charge to 80% SOC with BMS, randomly doing 95-100% when balancing and never charging when cold will literally do ALL that and more without additional work.

 

[SafeDiscDancing]

I don't understand the evangelizing of bottom-balancing. Keeping your charge to 80% SOC with BMS, randomly doing 95-100% when balancing and never charging when cold will literally do ALL that and more without additional work.

Bottom Balancing makes the most sense when you have a small pack and use all of it over and over.

The key strength is being able to hit the bottom of the pack and do no damage.

Like for me today... so far I think I rode four full 250 wh rides back to back with 20 minute fast charging.

Each time I reach the bottom region the voltage drops off very quickly and I am not tempted to abuse it.

When you have a big heavy oversized pack then you can not be so concerned because you just don't take risks on the low end.

My pack is 10 lbs. I have a 600 wh pack but it's a pig by comparison at 20 lbs.

So bottom balancing is best in the high performance scenario of hard use, then hard charging.

Nothing can break with bottom balancing, but we all know top balancing kills the weak cell and you might not even notice it.

 

[Millhouse_5]

The initial plan was/is to include the BMS within in the battery pack to keep everything as one clean unit with only a charge and discharge plug coming off it.
Lately I've been tossing around the idea of skipping the BMS in the discharge circut and having it wired as a charge only BMS. Heat is my main concern as, like I said, a high end 28s BMS is impossible to find. The best one I've found is still cheap and rated for 120a in a well ventilated area. Not ideal for when I inevitably find myself needing to use the highway.
Anyway, an external BMS wouldn't be that much more difficult to wire up especially if I don't need to keep it near the main discharge wires so perhaps I'll consider that when I get a little closer to finished the pack.

I've made 4 parallel strings so far and for those wondering, the Sequre SW-2 is not the right tool for this job. I am using 0.15 pure nickel strip and while Sequre says that it can do it, I can testify that it doesn't like it. I need to run it at 100% power for every weld and the leads get hot after about 10 welds so I need to take a break to let them cool.

While I love this welder and believe it was well worth the 89 USD I paid, It's not the best tool for this specific job. If I weren't using 0.15 pure nickel then it would be perfect. I'll finish the job with it, but if I ever want to make another pack, I'll be upgrading to the SW-1 for it's extra power.

 

[SafeDiscDancing]

I'm not actually suggesting you do this but you might find it interesting:

https://endless-sphere.com/forums/viewtopic.php?f=14&t=114594

What the Rotary Switch permits you to do is have very few connectors exiting your pack but you can "dial" up any cell and look at it or even add or remove charge from the cell being examined.

If you were like me and totally dedicated to bottom balancing then this gets you there easily.

And when charging you would get to know which is your weak cell and target it for the charge termination.

Anyway... I'm not suggesting you go down this route.

 

[Millhouse_5]

That is a very interesting concept,
I and also agree that it sounds like something i should not do. Probably fine if it's applied in something more industrial like and not at what I'll call "hobby grade" level that many of us build at.

For now though I'm out of cells and need to wait on my next shipment to arrive. I only ordered my batches about two weeks apart so hopefully they show up pretty soon and I'll be able ton continue working on my pack. At the moment my battery in is a 7s 20p configuration marking the 25% mark has been reached.

I'll probably buy my BMS today or tomorrow so I can avoid delays and get everything built in an appropriate amount of time.

I'm working via a 3d model paper template to ensure I don't weld a pack of cells backwards and using a 3d printed cell holder to ensure I get nice uniform parallel strings when welding.

 

[Millhouse_5]

Received my charger just the other day, but unfortunately I am still waiting for the remainder of my battery cells. AliExpress has 7 more days until I open up an order dispute to get a refund. Strange thing is that I received my other 140 cells from the exact same seller about a month ago and the orders were only placed 10 days apart.

Hopefully I'll get my batteries very soon and not have to deal with the hassle.
Worst case I have to order a full new set of batteries from a new seller and have unmatched cells.
Best case, I open a dispute and get my refund and shortly after my batteries show up taking the price of my bike down by 1400 dollars, which would be helpful that's for sure.

So far the cost of the project is just under 2000 dollars for all the parts I have received. So basically not counting those undelivered cells. Undelivered total cost is at 3511.32 exactly. The only large parts I still have to add to that total are the motor and controller so as long as nothing else huge comes up then I should be able to get away for just over 5000 Canadian dollars, which was my original budget.

 

[e-beach]

...but unfortunately I am still waiting for the remainder of my battery cells. ....

Sounds like you may have been nailed by the fabled Chinese New Year. February, give or take a few weeks, isn't the best time to order from China so that could be part of the delay.

 

[Millhouse_5]

Sounds like you may have been nailed by the fabled Chinese New Year. February, give or take a few weeks, isn't the best time to order from China so that could be part of the delay.

Sounds very possible. The batteries were only ordered 10 days apart so I expected them to show up 10 days later. With some room for error of course.
Thankfully the seller did respond to my message and was able to give me some updated tracking info showing that it is at least reached Canada. I know delays are crazy, but I also believe a 3 month shipping time is equally crazy. (76 days and counting at the moment)

Anyway. Not quite on topic, but I installed the CT-22 dash I ordered for my proton just today and it's working a treat except for the clock which is all sorts of horrible. I'm guessing that's because I don't have the "memory line" hooked up. It will likely be sorted next weekend when I have some free time.
I was waiting to order another one for the CBR build until I got this one figured out and could see if I liked it. Short answer is, yeah, I like it. So I'll be throwing one of those in the mail pretty soon.

 

[Millhouse_5]

84 days
It took 84 days for the second batch of batteries to come in the mail. I'm happy I got them when I did too, as the price has gone up by about 20% since I got mine.
Gas has also gone up a bunch lately too so my rush on this project also just went up a notch. its now 2.07 per liter in newfoundland (6.07 USD per Gallon for my American friends). So I'm now in a rush and need to get my bike registered and also get a bike license before I leave the province so I can stop driving my crazy inefficient car ASAP.

Anyway, the pack is half built now. I'm just trying to figure out a good way of mounting a connector. I want to go with a classic Anderson connector but I'm not quite sure how to interface the two. Obviously most of these connectors are designed to be crimped to classic stranded wire, and my 3 twisted 12 awg will be much smaller and unable to crimp the same way. So what can I do? Is there a special crimp or other joint to from solid to stranded wire that I've never heard of or will I have to make a terrible ugly bulging solder joint to get what I need accomplished.

 

[amberwolf]

I'm just trying to figure out a good way of mounting a connector. I want to go with a classic Anderson connector but I'm not quite sure how to interface the two. Obviously most of these connectors are designed to be crimped to classic stranded wire, and my 3 twisted 12 awg will be much smaller and unable to crimp the same way. So what can I do? Is there a special crimp or other joint to from solid to stranded wire that I've never heard of or will I have to make a terrible ugly bulging solder joint to get what I need accomplished.

You have a hydraulic crimper, so I suspect it'll make a good solid crimp regardless of the wire, as long as you're using the correct die size for the contact housing. If it doesn't keep the housing straight during the crimp, or makes "wings" that stick out, and keep the contact from floating in the anderson housing, they may not mate properly and cause connection issues.

I'd test using an untwisted segment of parallel wires, however many are going to go in the contact, (just a sample of the same wire, not the actual ones on your battery) and then cut the resulting crimp open to see if it is solid inside or if it just left a mess of open spaces (which is not a good crimp). If it's a solid crimp the wires are not going to just come out while cutting it open (lengthwise or crosswise across the crimp), so that's another indication. There are good pictures on a google image search comparing good crimps to bad ones on power-carrying wires for various contact types, if you're not sure what to look for.

When you do your actual crimp, I would use the end going into the contact as untwisted but parallel conductors. They'll fit more tightly into the contact that way. If you have spare wire and contacts you can experiment to see which works better for you.

If you are using SB50 or larger genuine andersons (whatever you use should be rated for more than you ever will put thru it, worst-case), the housings are strong as are their contact-mating springs. The PP75/etc shells are not as strong and I don't find their springs are either because of that. Non-genuine anderson shells seem often made of softer plastics that distort more easily and don't put as much force on the contacts to keep them properly mated.


If you have to mate two different wires to get what you want, use crimp lugs. There's plenty of different kinds; I use these for my simple needs:
https://www.amazon.com/gp/product/B076H3VHGN
you can certainly find better ones.

 

[Millhouse_5]

Thanks for the Input.
My background is in avionics so I know a lot about handling wire that's 16awg or smaller, but a little less when it comes to pushing actual power delivering current.

I'm looking at the Anderson SB175 as my battery is rated for 2C continuous (130A@13.5Kw) so I want to be sure that my wire and connectors aren't my limiting factor. Princess auto sells welding cable so last time I was in town I picked up a bunch of 2awg and that's what I'm planning on using for my main battery to controller wiring.
I'll try and order a couple spare connectors to do some experimenting with, and I literally just forgot that butt splices existed. One of those days I suppose.

 

[Millhouse_5]

Well i promised infrequent updates and I'm here to deliver on that promise.

Some progress has been made since I last shared anything. Biggest DIY thing is the battery is almost complete. Everything but the connector is ready to go as i simply haven't ordered them yet and I'm not in a rush as I've got a lot of house project to do since i moved this summer.

That battery took me quite some time to make and the little spot welder i picked up really punched above its price range, It was a bit too small for this project but I'm still really happy with it. The pack has been padded and secured with heat shrink since that picture, but a black rectangle doesn't really excite the senses.

Biggest news is that I've changed up my strategy for this project. Now that I'm in a big city and have access to more used things, I was able to find myself a new already electric sport bike chassis, which will be a lot easier to upgrade. With the CBR I would have had to do a lot of fabrication and of course that would involve buying a small pile of metal to make things. So this swap will save me time and money.
Anyway, here's the new bike. Its a 2019 Armada Predator that has only 18km on it. It's a touch bigger than the very tiny CBR and just as comfortable as it as well.
Hopefully I can get a little work done on it over the winter and spring to get it up and running for at least one ride on it this summer. The last time I did a full motor and controller upgrade it took me about a week, but I had a lot more free time then. Lets hope this process doesn't take too long this time.
Updates to follow...eventually.

 

[Shail]

Sounds very possible. The batteries were only ordered 10 days apart so I expected them to show up 10 days later. With some room for error of course.
Thankfully the seller did respond to my message and was able to give me some updated tracking info showing that it is at least reached Canada. I know delays are crazy, but I also believe a 3 month shipping time is equally crazy. (76 days and counting at the moment)

Anyway. Not quite on topic, but I installed the CT-22 dash I ordered for my proton just today and it's working a treat except for the clock which is all sorts of horrible. I'm guessing that's because I don't have the "memory line" hooked up. It will likely be sorted next weekend when I have some free time.
I was waiting to order another one for the CBR build until I got this one figured out and could see if I liked it. Short answer is, yeah, I like it. So I'll be throwing one of those in the mail pretty soon.

Hey, Can you please share instructions on the CT-22 dash connections? I am building an e-ATV and plan on using the CT-22. I have got CT-22 from aliexpress but couldn't find any instructions.