Bypassing the Schwinn Tailwind LTO BMS

[Syonyk]

Canonical source here: http://syonyk.blogspot.com/2015/07/bypassing-schwinn-tailwind-battery.html

=================

You might recall one of my previous posts about the Schwinn Tailwind battery pack. In a nutshell, it's a 24v, 4AH pack built of Toshiba SCiB lithium titanate cells. It's really a weird pack - but that's fine, because it's on a really weird bike. I did some analysis of it, but wasn't willing to hack it up as it wasn't my pack. So I fixed the glitch and bought the bike. It turns out, ebikes with bad battery packs aren't worth particularly much. I knew this, though.

Bad BMS - Bummer!
The problem with the pack I now own is a bad BMS (or a bricked BMS that won't allow anything to happen, because the batteries drained too far - same result). The pack won't charge, and it won't discharge. The BMS controls the normal charge and discharge ports with a set of power semiconductors of some form or another, and it's stubbornly refusing to run them.

Except... it does charge. The standard charge port for home use (8A charger - 2C rate... *blinks*) doesn't work, but the bulk charger (40A... 10C charge rate!!!) does, and I can charge the pack through that with an external power supply.

An inspection of the BMS shows that if I were to add a few wires to jumper current around, I could get pack voltage to the output - and get the normal charger port to charge the pack as well.

So, of course, this being an absolutely terrible idea, I set out to do it.

Lithium Titanate Chemistry
The only reason I'm considering this otherwise mostly idiotic idea is because the cells are lithium titanate. An extensive amount of research indicates that they're a very well behaved, safe chemistry that can actually be discharged fully without damage (unlike a lot of the other lithium chemistries). They're a low energy density chemistry as well, and I'm not too concerned about running them in light duty without a proper BMS. I wouldn't do this with other chemistries, though. Consider this a special case for a particularly exotic pack.

Don't Do This. Seriously, don't do this. And if you do, I'm not responsible.
This is a terrible idea, bypassing a BMS like I'm about to. You shouldn't do this. You definitely shouldn't get creative with my steps and do it better than I did. And if you do, I'm not responsible.

Seriously. This is battery hacking at a level that simply doesn't need to be done. Buy a 24v battery of some other type and hook it up instead. Don't do this.

... so read on to find out how to do this.


Getting to the BMS
To bypass the BMS, first, you must reach the BMS. This bit of wisdom was received from a sage on top of a mountain, though why a desert plant felt the need to tell me this, I'll never be quite sure.

The first step is to remove the 8 screws (6 Torx T20, two phillips) from the bottom of the case and remove the lower case.

That done, marvel at the amount of wasted space in this shell, remove the 8 screws holding the actual battery in, disconnect the indicator light wire (it doesn't work anyway - bad BMS), and pull the battery pack out. The interesting stuff is, of course, on the bottom of it as you access things.

Once the pack is out, go ahead and remove the 7 screws on the side that protect the individual cell terminals. Once this is off, pack voltages are exposed, so be careful where you stick your tools. It's only a 24v pack, but it'll source some serious amps if you short it - don't do that.

Flip the pack over, and marvel at the more screws on the BMS cover. Four of them go into the plastic, and four mount the board. Remove all of them. Also remove the two screws on the front connecting the pack positive and negative terminals into the BMS. Be careful - they've got full pack voltage on them, so don't short them to anything! Or do, since you're not doing this anyway because it's a terrible idea.

Once the 10 screws are out, you can pull the BMS board. Be careful of the ribbon cable on the front - that's the balancing cable. Pull it out gently before you remove the board. Once the board is clear, flip it over and remove the two small screws on either side of the charging connector, and the BMS should come free from the metal plate.

With the metal plate clear, you can see the business side of the BMS. There's some serious current carrying capability here - but I'm about to mess with it.

BMS Layout
Here's what the problem is. I've annotated the below diagram with lines, red for positive, blue for negative (and also, there's a common ground plane on the board so the negative lines are connected to each other - I've got 99 problems, but a split ground plane isn't one of them).

Without the BMS happy and functioning, there are three different positive islands that are causing my problems, because the two I care about (charging and discharging) are not connected to the third, which is the battery pack.

The two connections in the top left of the picture are the output connectors. They press against the spring loaded contacts in the bike mounting bracket, and transfer power to the bike. If they've got about 24v on them (plus or minus a good bit), the bike works. If they don't, it doesn't.

The lower charging connector is somewhat interesting (you can see more detail on it below). It's two different charging connectors in one. The outside two pins (the large ones) are for the high amperage charger. The inside pins are for the home charger. And then there are a few other pins that I assume are for diagnostic communication, or communication with the high amperage charger, but I sure can't find any details on them.

In any case, the inner pin that I'm interested in is not connected to anything unless the BMS connects it. Neither is the output. So this is what I need to fix.

Here's a better shot of the connector side of the BMS. The left two plates are the output plates - negative on the left, positive on the right. The righthand charging connector is really interesting. The outer sockets are for the high amperage charger - it's about 40A. The inner socket has two pins, and is for use with the home 8A charger. And I have no idea at all what the four other pins are - possibly contacts for the commercial charger, but that seems delicate. Maybe just for diagnostics. Nobody will tell me a damned thing about these packs, so I'm left to guessing. And doing terrible things to them in the hopes that someone at Toshiba will be horrified enough to tell me what I want to know.

This is the other side of the board. This really is a "belt and suspenders" type board - the bolts go through the board... into where the traces (for some insane value of trace - these things are copper, coated with zinc or tin) are soldered onto the board. I really have no clue what went into this BMS, but I strongly suspect some engineer was very proud of it, figured it could comfortably drive a 50A load (it's got a 60A fuse), and was very sad when told it would drive a 250W bike (so about... 10A). Either that, or someone thought a safety factor of 6 was about right for this. I'd love the story sometime, if anyone happens to know. It's absurdly overbuilt, and I love it! It does make my hacking a bit harder, though...

The Goal
The end goal here is pretty simple: I'm going to fully bypass the BMS. I'm going to tie the charger connection, the battery terminals, and the discharge plates together, and they'll all be live, all the time, and this should just work.

Like this:

Things that do not work
I've found many ways to tie the rails together that do not work.

It turns out, these huge chunks of coated copper (WTF is coated copper doing on a 10A system???) are insanely good heatsinks, and my soldering irons are not capable of heating them very well. Certainly not to soldering temperatures.

The first attempt was to use a chunk of copper wire, bridge the connections, and be done with the output side.

This didn't work. I could not get the bars to take solder. They simply would not get hot enough for the wire to stick. Not even close.

I also attempted to create some large solder bridges. This looked successful initially (I could get the solder to sort of stick to the edge of the traces), but a bit of fiddling with a screwdriver afterwards popped them right off - they weren't well attached at all, which means they won't flow much current, and I'd rather not have balls of solder floating around inside my battery pack. So that's out as well. Also, it's very ugly.

Another option that does not work is using circular terminals and wire. They're too thick (at least the ones I have). I'd need something very thin to fit the profile that I have available.

And yet another idea that does not work is trying to spot weld some nickel strip onto those traces. My spot welder (788+) is simply not powerful enough to do it. Though, to be fair, I didn't turn it all the way up. I've blown it up once and would rather not do it again.

Something that actually works
After some experimentation, I finally came up with something that works. I removed some of the unneeded bolts, and created a jumper between the charging port and the output port. It's still difficult to solder, but I was able to make it work. I went with two wires on the output side for somewhat increased current carrying capacity. This is a low power bike, so I'm not carrying many amps. I'd prefer to have done something more robust, but this is the best I could do with the tools available to me.

For the charging side, it was a lot easier. I was able to solder a single jumper wire and get it hooked up nicely. This should handle the charging current from the home charger just fine - especially since the pack charges up quickly and stops soaking as much current.

If you're doing this, of course use the largest wires you have available. This is a proof of concept.

Does it work? Yup!
The pack voltage is available on the outputs, and the charger works! Excellent!

Does the bike work? Yes. It does.
"Runs & drives." For whatever limited value of running a Tailwind does.

What works? What doesn't work?
One of my biggest questions about this bypass is with regards to balancing. The cells should be balanced after charging, and I don't know if the BMS will do this or not in this weird bypassed state. So, data time. I rode it a bit, took some numbers, charged it, took some numbers, and then let it sit, and took some numbers.

Post-ride Pack Voltage
After a bit of a ride, this is what the pack looked like:

Total voltage: 23.40v
2.340v
2.340v
2.341v
2.341v
2.340v
2.340v
2.340v
2.340v
2.340v
2.340v

Well balanced, and nearly identical voltages. This is actually quite impressive for an older pack (the pack is 7 years old).

Let's see how it looks after a charge.

Charged Voltages
Pack voltage: 28.35
2.900v
2.845v
2.830v
2.800v
2.795v
2.797v
2.815v
2.805v
2.803v
2.806v

Oh. My. Wow. That's ugly. 0.1+v between the highest and lowest cells. This is not great, but maybe it works by bleeding off charge after topping off - I have no idea. The pack could be using magic for all I care. I do notice that the cells lose voltage quickly after a charge, so let's see how it looks after a few hours.

A few hours later...
Pack voltage:26.12
2.633
2.613
2.612
2.607
2.601
2.604
2.614
2.607
2.608
2.609

That's a lot better. Not perfect, but within reasonable ranges. So it's probably fine. I suspect the voltage/state-of-charge curve is nearly vertical at fully charged on these batteries, and I'm OK with a tiny bit of variance if it settles in nicely while discharging, which it does.

So: Should you do this if you have a Tailwind pack?
Well, this is pretty simple.

If the steps I outlined are somewhat confusing and sound a bit scary, absolutely not. This is not a trivial hack, and there's a lot that can go wrong.

If you're reading this, nodding, and occasionally saying things like, "Man, what an idiot. My 5kW soldering iron could totally solder onto those traces! And what's with that crappy solder job???" - yeah, you should probably consider doing this. It's not that difficult a job if you know what you're doing and have the right tools. Also, get some voltage into the cells before you plug the main charger in - I don't think it likes low voltages. I'd try to get about 2.5v per cell before using the main charger.

What's the best option for a Tailwind?
The best approach is probably to find another 24v battery, preferably one with a lot more capacity, and wire it in. I'll cover this in a future blog post at some point. This battery is just a weird bit of exotic oddness that makes no sense for an electric bike.

Will I do this for you?
Probably not. This is a quite ugly hack, and I'm not yet convinced it will hold for the long term. Sorry. But your local electronics geek might be willing to do it for you.

 

[slowhands]

Schwinn Tailwind Battery Failures

I have investigated this battery problem also. My conclusion is original battery is beyond economical repair, and since it has too little capacity (4.2 Ah) to be useful anyway, it would be better to gut the battery case and install higher capacity cells and a new BMS. I propose an alternative new battery layout for longer range.

Here's a little background. I got a like new Schwinn Tailwind with two dead batteries at a good price. There are some things about this bike I like. It has the Shimano 8 speed internal gear hub and Bafang front hub motor. The suspension seat and fork make the ride comfortable. The built in wheel lock is great. Build quality is excellent, though it's heavy.

There was a reason the bike was cheap: dead batteries. Turns out that most of these batteries are dead, and nobody knows how to fix them. So I tore into the batteries to see what's what. In both cases, the BMS seemed defective, not the battery cells. Based on the original posting, I bypassed the BMS, but still had challenges with one of the batteries.

Bootstrapping: Here's a Catch 22: The Schwinn charger would not start charging my dead batteries because their voltage was too low. When I tried it, the charger fan came on for about 20 seconds, indicating attempting to charge with a solid red indicator. But then a red blinking light showed on the charger, indicating fault, and the Schwinn charger shut down. The problem was getting some juice in the batteries when flat, and the Schwinn charger was too smart to help. It seemed I was dead in the water.

As it happened, I did have a dumb 24v charger, for my kid's Razor scooter. This cheapie Razor charger did work. I plugged a Razor scooter 24v battery charger into the high current receptacles, using some bent up 12 gauge solid wire to connect from the female Razor connector to the much larger fast charge receptacles on the battery. The bent wire was bent back on itself to make the big plug for the fast port, and 12 ga wire fits the Razor charger plug OK. The Razor charger is not fussy. If voltage is low, no problem. It puts out constant 1.5A current to bring the battery up (blinking red), then halts and indicates full charge (green). It took several hours, but all went well. It charged to 28v, with nearly equal cell voltages, and has settled to about 26v over time, no load. Seems like the cells are OK, so the problem is the BMS.

I followed the circuit a little with my DMM, and found that the output current is switched on/off by to P channel MOSFETS, labeled Q41 and Q42, which are controlled by hidden circuitry in the dead BMS. In the charging side, charging current first goes to two Schottky rectifiers for reverse polarity protection, D24 and D25. Then it passes through a P channel FET, Q40 which acts as a switch controlled by the BMS.


The BMS won't turn on the output nor charging input for reasons unknown, but I speculate that the pack had low voltage at some point, and that fault is stored in the BMS processor as a fatal error, hence shutdown. But that puts an expensive ebike out of commission, for possibly no good reason. The battery cells have held voltage for several days here, so I think they are still good. Some short rides have barely drained them. What to do next?

Can I Fix the BMS? I had bypassed the BMS, as suggested, but just as a stopgap.


It left me queasy because it's potentially not safe. Can the BMS be fixed? Not likely. The BMS circuitry is potted in black elastomer, so nothing can be tested nor replaced. Since there are no spare parts nor any documentation, repairing the BMS is not practical. Possibly the BMS could be substituted. But the cells have an unusual voltage, 2.39V, and no off the shelf BMS will replace it, although I'm sure you could modify a standard 10S BMS if you knew which sense resistors to change. For me, that's too much effort for too little return on this 4.2Ah pack.

Now this battery does work, but I regard it as suitable for use only in a laboratory, and not usable commercially. I would not advise anyone else to modify their battery or to use a modified battery either. Absolutely not.

To recap, this modified battery is not really safe, and its capacity is too low. But looking at the battery case, I think it's possible to gut the case, cutting out some of the internal screw bosses, and get enough room to put a quite respectable battery in it.

My idea is a 7S/6P layout for a 24v/13.2Ah Lion battery with 3X range, packaged in the original case. So that's in process. To be continued....

 

[Syonyk]

Very interesting! I'll be quite interested in seeing where you go with this - I considered doing much the same (installing a different battery configuration), but the Tailwind just isn't a good enough bike for my needs to justify spending the effort. A friend is riding it as a commuter, so I'll see how the pack holds up.

Some comments:

...and since it has too little capacity (4.2 Ah) to be useful anyway, it would be better to gut the battery case and install higher capacity cells and a new BMS.

I agree that 4.2AH is small, but it's 100WH, and with a 220W motor on the front, that's a good half hour of full assist. Given that it normally doesn't run flat out, and on the low/medium assist levels it's putting out less power, it's adequate for the bike. Adding more battery doesn't help the fact that the motor isn't very powerful and the bike isn't very light.

Based on the original posting, I bypassed the BMS, but still had challenges with one of the batteries.

Glad it was useful!

I suppose I should mention that I keep a lab power supply around for random battery charging crap. It's not nearly as smart as a dedicated charger. It just makes power.

Now this battery does work, but I regard it as suitable for use only in a laboratory, and not usable commercially. I would not advise anyone else to modify their battery or to use a modified battery either. Absolutely not.

Yup. I see you took my warnings to heart.

One thing to consider, at least for personal use, is that the battery is the lithium titanate chemistry. Do some reading on it if you're bored and love reading battery papers. The chemistry is very much unlike most other lithium chemistries. It just doesn't misbehave like the other chemistries do, mostly because it's dimensionally stable during charging/discharging. And an engineer I talked to commented that, while not recommended, they will recover from 0v just fine, repeatedly. It falls into the "Eh, you probably shouldn't do that too often" category, not the "do not do this ever" category like other cell types.

It also seems like the BMS, even in "turned off" mode, still balances the cells. Or they're moderately self balancing. One of the two.

To recap, this modified battery is not really safe, and its capacity is too low. But looking at the battery case, I think it's possible to gut the case, cutting out some of the internal screw bosses, and get enough room to put a quite respectable battery in it.

My idea is a 7S/6P layout for a 24v/13.2Ah Lion battery with 3X range, packaged in the original case. So that's in process. To be continued....

If you live in an area where the bike makes sense, this is certainly a reasonable solution. The bike isn't drawing much power (~10A), so you can build a pack out of fairly high energy density batteries and be entirely fine on the C-rate.

I'd strongly considered doing the same thing, but it just doesn't make sense, given that I have other bikes that are better suited to my needs. I wouldn't really want to sell the bike to someone with a one-off pack, and I don't think there's enough market for a replacement Tailwind pack to justify the engineering effort to build it as a commercial product.

If you do get the other pack built, would you be willing to either de-encapsulate your BMS, or send it to me so I can work on this? I'd really like to see what's under there, but I'm still using it for the contacts.

Also, if you put a charging port on the back of the pack, you should be able to just eliminate the spring contacts entirely and run a wire into the bike side of the mount point.

 

[slowhands]

I agree the bike is underpowered. I live in a flat area, and my trips are short, 14 miles round trip, often less. I barely notice the motor, but then I find I am not sweating when I get to my destination, so I know it was subtly helping. I still work, but not as much, which makes me like the ride more. Still, I want more power.

I would like to try this motor on 36v. Others are using this same motor on 36v happily. My 36v BMS would handle undervoltage shutoff, no battery risk. I don't know if the controller has the capability, but I just need the controller to work and not burn up. I will be looking at that. Everything on this bike is so over-designed, I would not be surprised if it works. I will probably do it to the second pack, keep the first as a baseline.

It is possible to fit a 10S/5P battery in that housing, so it's a temptation. On the other hand, I dislike the weight of the rack and battery up high behind me, making it handle poorly in turns. I am looking at putting new battery and controler in the frame triangle. Think I will try fiberglass covered Masonite or possibly Delrin for the housing. Easy to machine with a router, plenty strong and I can make arbitrary shapes and rounded corners easily for a cleaner look. Since I don't have a machine shop, but I do have plenty of woodworking tools, this is the way I have to go.

I hate to maul a classic ebike, but it's mine and it needs it. Then it's certainly not a Tailwind, but that's OK. Schwinn gave Tailwind a light aluminum frame, then hung too much dead weight up high. Need to strip it and redistribute the weight to let it handle better.

By the way, I decided to change the bypass wiring to incorporate the Schottky diodes in the charging path for reverse polarity prevention. They are already in the path. I don't think I will plug it in wrong, the connector is keyed, but I costs nothing to be safer.

 

[Syonyk]

I agree the bike is underpowered. I live in a flat area, and my trips are short, 14 miles round trip, often less. I barely notice the motor, but then I find I am not sweating when I get to my destination, so I know it was subtly helping. I still work, but not as much, which makes me like the ride more. Still, I want more power.

I'd be interested in your thoughts on the bike, compared to the review I made: http://syonyk.blogspot.com/2015/09/schwinn-tailwind-review-in-2015.html

I would like to try this motor on 36v. Others are using this same motor on 36v happily. My 36v BMS would handle undervoltage shutoff, no battery risk. I don't know if the controller has the capability, but I just need the controller to work and not burn up. I will be looking at that. Everything on this bike is so over-designed, I would not be surprised if it works. I will probably do it to the second pack, keep the first as a baseline.

I would be very surprised if the bike, as-is, will take 36v off the pack without burning something up. It's not particularly robust.

If you're going to do that, you may as well just get another controller, and stick a throttle on it. Bypass all the Schwinn stuff. The motor is, however, rated as a 24/220W unit, so... *shrug* It might work for a while. The controller, not so much. Brushless motors are decent enough with overvolting as long as they don't overheat.

I hate to maul a classic ebike, but it's mine and it needs it. Then it's certainly not a Tailwind, but that's OK. Schwinn gave Tailwind a light aluminum frame, then hung too much dead weight up high. Need to strip it and redistribute the weight to let it handle better.

Is the frame actually that light?

I mean, I like the sealed driveline and the rear geared hub, but the whole of the bike just isn't very good, and I'm not sure adding more power to the front wheel will improve things.