Saturday, August 13, 2016

BionX 26v 9.6Ah Teardown & 18Ah rebuild

It's been a while since I've done a pack teardown and rebuild of a new pack.  The most recent one was the Trek Valencia Ride+ pack teardown and rebuild back in January (Part 1, Part 2, Rebuild).  That one was a bit of a mess - corroded cells, held together with some sort of adhesive foam, and generally not at all fun.

This one is somewhat nicer to work with!  It's a very welcome improvement.

I have here in my hand a list of 205 a very well used BionX 26v LiMn battery pack.  It's rated at 9.6Ah, and is in the same form factor they seem to use for all their packs.  Interestingly, the capacity is the same as the 36v packs (9.6Ah), but the voltage is lower (so fewer cells).  It could make one wonder...

Read on for tons of photos, a full teardown, and a stunning rebuild to a whopping 18Ah!

If you're just here because you have one of these and it needs to be rebuilt, you probably want to go here for details on the rebuilds I offer.

But, first, the warnings.

The Warnings

Ah, yes.  The warnings that tell me not to do what I'm about to do.  The remainder of this blog post involves a lot of "Disassemble."  The warranty is already void, so I'm not worried about that.  And, really, "No disassemble!" would be just as adequate, and most people reading it would read it in the voice of Johnny Five.

Also, related to this label, have I mentioned I hate RoHS?  Lead-free solder is just junk compared to the lead stuff.  It's not good solder.  And don't get me started on lead free brass plumbing fittings.  If forced, I can say good things about them, like "They only leak profusely from the threaded parts."  I'm sure there's a trick to getting them sealed, but I sure can't figure it out, and neither can anyone else I talk to.  There's a reason that everything is PEX now, and I'm pretty certain lead free brass is a major part of that.

Yes, I get it.  Lead isn't great for humans.  But neither are tin whiskers causing failed boards, cracked plumbing, or the variety of other problems one has with lead free stuff.

The people who developed leaded solder weren't nasty, mustache twirling villains, cackling in their dank, dungeon labs as they tried to find the most vile substances they could possibly put in solder, because they were being evilly evil for evil's sake.  They used lead and tin because it worked the best.  Of all the elements one can put into solder, tin and lead combine together to make a really good solder.

And, besides, there are enough exemptions to RoHS that it's basically useless anyway.  Your car's circuit boards are lead free (so, they're probably growing tin whiskers), but don't ask about the huge lead acid battery under the hood.

On the other hand, I guess it's safe to repeatedly lick circuit boards now.

Anyway.  I'm not planning to incinerate this pack, at least...

The label at the other end is just fun.

"Do not allow this battery pack near a crippled droid with a broken axle"?

"Do not fire your large cannon into the ground"?

"Do not shoot this really complex bullseye"?

"I got rejected from the museum of modern art, so now I make random stickers"?

Something like that...

In any case, I don't plan to throw this pack away.  I plan to fix it, and make it better in the process.  We need more of that these days, by the way.

Diving In

The pack opens just like any other triangular BionX downtube pack, with four hex head bolts running through the pack, and two plastic latches along the top.

Inside, there's the usual stuff.  A battery, and the BionX BMS.  All of this is routed (with wires, imagine that!) to the connector in the lower left corner.

Interestingly, the newer BionX packs seem to be a bit better about holding the battery in place.  I'm not sure if there's a dimensional change somewhere, or why they've started adding the padding blocks and the silicone caulk, but there's nothing wrong with keeping the battery in place better.  And, this stuff is sticky enough to stay in place, but it doesn't glue the whole pack together - which is awesome!  I'm not a fan of the stuff used in the Valencia Ride+ pack, which really does glue the whole thing together.

A quick count of bulges indicates the pack is four cells tall on the right, and three on the left.  This case looks identical to one that can fit a five-tall pack all the way across, and that looks suspiciously like a big block of foam across the top...

Further detailed investigation reveals that, yes, that's just a block of open cell foam across the top, filling space and keeping the battery from rattling around.

I suspect this is why the caulk is used - otherwise, the battery could shift inside the pack.  It's not a bad idea, though, and I've started using some silicon caulk in my rebuilds.

With the big block removed, the smaller, denser foam blocks holding the pack in place are clearly visible, as is a thermal fuse that will trip if the pack gets too hot.  All the older BionX packs have this, and it's wired on the negative side.  It's worth noting that the body on it is electrically connected to the wiring - so be careful where you let it swing.  It's a dead short if it touches the exposed pack anywhere.  It's not a bad idea to wrap it in tape when you remove it from the pack.  I hear that it creates some impressive arcing if it swings onto a pack.

The BMS looks about like an old BionX BMS does - no surprises here.  There's nothing particularly exciting on this BMS - a beeper, a fuse, and a set of connections to the battery and to the external connector.

Observant readers, familiar with modern packs, will note the lack of balance leads going into this pack.  They're not needed due to the nature of the battery chemistry used.

It's a bit hard to see in the photo, but the "gunk" on top of the board and components is a conformal coating of some variety or another.  This is great to see!  A conformal coating helps protect the board and components from water and debris.  Some of the BionX packs have this, and some don't.  I'm a big fan of it, though.  Personally, I'd rather see a full potting, like Rad Power Bikes did with both the Rad Wagon and Rad Rover controllers, but I'll take what I can get, and a conformal coating is much, much better than nothing (which, sadly, is also found on some BionX packs).

The Battery

The battery pulls right out, and as usual for the older BionX batteries, there's a very helpful label on it.

This tells me it's a BMZ 7S6P pack (7 cells in series, 6 cells in parallel for each group), with a nominal voltage of 25.2v (3.6v/cell), and a fully charged voltage of 29.4v (4.2v/cell).  A 6P pack with 9.6Ah capacity means 1600mAh per cell - so it's almost certainly 42 of those cute green Sony US18650V cells.

I'm not sure exactly what the caulk they use is, but it feels like a standard silicone caulk, and peels off easily enough.  The combination of the foam spacers and the caulk keeps the battery solidly located, and gives it some shock absorption as well.

Inside the Pack

With the black plastic cut off, it's clear exactly what's going on here.  Groups of six cells in parallel, with seven of those groups in series.  I'd personally describe this pack as a 6P7S pack to make the layout explicit - this would be a 6P7S pack because the parallel groups are first, whereas a 7S6P pack would have strings of 7 cells in series, with 6 of those in parallel.  The standard BionX 36v pack layout is a 10S6P pack, and I tore one of those apart a year or so ago.  I've also seen a 6P10S 36v pack, so... I have no idea why they switch around.

A ruler and a micrometer indicates that the nickel strip is 10mm wide, 0.15mm thick - which I have on hand for the rebuild.

As I expected based on the fact that this is a BionX pack of 9.6Ah, it's built around the green Sony US18650V cells - 1600mAh per cell, and capable of quite a few amps.  I rebuild packs with Sony US18650V3 cells, which are the same chemistry but 2250mAh.  However, I think I can cram a few more in this pack...

Looking at the welds, it's clear this pack was hand built.  That's not a problem at all - I hear people will pay good money for Artisanal Small Batch Handcrafted Batteries (and if you're one of those people, the contact form is on the right - I build such things).  It's just interesting to see.  They use two welds per terminal for the most part (some have three), and I use three welds on all my terminals.

The cells are hot glued together, just like all the other BionX packs, and the battery as a whole is very solid.  I have no complaints with how these were made - and they're certainly easier to rebuild than some of the newer exotic layouts!

That's it for the teardown part.  There's no point in pulling the battery further apart, because there's nothing to be learned there - the layout is obvious.  Also, it still holds voltage, so I'd rather not rip it apart "just because."  I have a use for these worn out packs.  I'm not entirely sure what.  But I'm sure it's something.

Rebuilding to 18Ah

As I was pulling the pack apart, with all that empty space filled with foam, I started thinking.

After I got it entirely apart, I stopped thinking and started test fitting.  I confirmed that, yes, a 6x10 brick of cells will fit just fine in the enclosure - it's the same enclosure size as the 36v packs!  So, that means I can fit up to 60 cells.  The largest multiple of 7 that's less than 60 is 56 - 4 cells short, but still a big improvement from 42.  If I rebuild with 8P groups, my 2250mAh cells will give a whopping 18Ah of capacity on this pack.  That's a big improvement over stock, and I'm pretty sure I can do it!

A bit of playing around on a sheet of paper gave me a layout that should work - strings of 8, wrapped around as needed.  You'll see the details later.

The first step is building blocks of 4 cells.  This will let me build my "8 long" and my "4x2" clusters I need.  I'm using my standard 18650 jig,  and hot gluing cells together.

These groups of four then get glued into some groups of 8 as needed.

Once the groups of 8 are glued together, it's time to group them electrically.  Out comes the nickel strip and my spot welder.

To make the 12 long grid I need, I've got 8 cells in a line on the right, and a 2x4 grid of cells on the left.  It works out.  I have to build the pack and weld it up as I go, given the short arms on my welder.  I'd love a freeform spot welding pen, but I don't have one, so I'll work with what I have.  Figuring out how to build new pack layouts with the tools I have is a bit of a puzzle, and I enjoy it.

Importantly, one should insulate once one no longer needs access to voltage regions.  These cells are at shipping charge (3.61v), but they'll still source more than enough current to spot weld any short they get the opportunity to.

And, test fitting the completed pack.  It fits!  Why BionX didn't do this in the first place is beyond me.  I guess this was supposed to be the "discount" system.

For installation, the ends are covered with insulating paper.  This is a thick, adhesive paper that does a great job of keeping everything in place and preventing shorts.

And, back in it goes!

I used the small foam pads and the same type of silicone caulk as the pack came with.  It's properly secured in the case, the case is back together, and everything beeps as it's supposed to during testing.

Nearly twice the stock capacity.  Not bad for a rebuild!

Final Thoughts

This is very clearly a BionX pack - show me any of the older packs and I'll recognize them at this point.  I suspect they went with the parallel-then-series layout because it's easier than making a series-than-parallel arrangement work for this particular cell configuration, but they've experimented back and forth with different layouts over time, and (spoiler alert for future posts) it doesn't make a bit of difference in balancing behavior.

I have to assume this was supposed to be the "discount system."  Otherwise, I have no idea why it isn't full - it's no harder to fill it than it is to toss in foam, and going from 42 cells to 56 cells doesn't add that much cost.  Even in 2008 or whenever it was built.

These older packs are really just a joy to work with.  They're dead simple, obvious in layout, easy to rebuild, and cell technology has advanced enough that adding capacity is easy.

Despite the lack of a fancy BMS, the spinel LiMn chemistry does a great job of staying balanced.  It's really just not a problem.

I know they've moved to newer chemistries and much more complicated layouts, but I'm not sold on the advantages yet.  These old packs are really, really nice - and I'm happy to help keep them rolling around as long as possible.

How do you like your 26v BionX system?  Let me know in the comments.  If your pack is getting old and could use the same treatment, the contact form is in the right column, near the top.


  1. Hi, thanks for this very interesting article. I have an older 26V 250W system. I took the pack apart so I could mask it off for painting (hated the white on my black bike) and I think it looked the same inside as the one you worked on. As an older big guy, 275lbs (when I bought the bike/system) I was worried that it wouldn't be powerful enough to haul me around, but am very happy with it. I think it should be noted that quite a few of the cheaper systems list peak Watts for their motors and if Bionx did the same the 250 would be listed as a 350 so I think that is why the little motor does better than I expected. My rides were a mix of flats and small hills and, yes, I did have to build up my legs and stamina to handle some of the steeper grades, but it wasn't long before my little mule was able to take me where ever I wanted to go. I am interested in a pack rebuild/upgrade after I decided how to transfer my 250 to my cruiser so will bookmark this page.

  2. Any info on the BMS and balancing circuit here? how do you balance the 7s? I dont see any rainbow cable coming from the BMS?

    1. Nick -

      The BMS only does low voltage cutoff. You're correct in that there's no balancing circuit/ribbon, though the cells remain very well balanced.

      There will be a future post on this, but the older BionX packs are using a Spinel LiMn chemistry (Sony US18650V cells) that have enough of a self balancing behavior at high state of charge that the packs remain balanced if used normally. I've tested some of the different packs that come in for rebuilds still holding a charge (many of my rebuilds are stone dead), and the cells are balanced to within 0.01v (the tolerance of my current voltmeters). It's an interesting way to build a pack, and it works fairly well as long as the cells remain in good shape, but the highest capacity one can get on those cells right now is 2250mAh - so about 50% lower than the best energy density out there. You can't use the higher energy density cells due to the lack of self balancing.

      I have some ideas on ways to work around this, but I've not built any proof of concepts yet. I need to reverse engineer the I2C and CanBus protocols used first...

  3. I found your blog last month and have been reading them all. Thanks
    I can't believe that you still use lead solder, well i agree with you that lead solder works the best. I still have what is left of two 5 pound spools .032 and .050. The fact that they were given to me in the late 70's is a blessing. When my battery pack dies i'm sure to hope that you still will be rebuilding them. Again thanks for sharing your ideas and work.

    1. It's still quite trivial to buy leaded solder for repair use and the like. I'm not bound by RoHS. My primary spool right now, purchased a few years ago, is 60/40 Sn/Pb.

  4. Hi Russell, really like what you are doing with the Bionx battery rebuilds. I have a used Bionx system that was given to me but it needs a battery pack rebuild. I have done a teardown on the 36V 9.6Ah pack, it is the same as your posts of Aug 2015. This pack was previously rebuilt and the temperature sensor thermistor is busted, the bead is broken and the wire for connection to the neg side of battery pack is gone. By chance have you ever measured the resistance of one of these thermistors? Don't want to proceed with a rebuild until the thermisor is replaced. Brian in Edmonton, AB.

    1. Brian -

      To the best of my knowledge, the BionX packs use a standard "10k at 25C" thermistor. That's what I use if I have to replace a thermistor on a rebuild. I seem to recall testing one at some point and concluding that's what it was.

    2. Thanks for the info about the thermistor. I did some more circuit tracing for the thermistor circuit and found that it goes to the charging connector. However, the charger has no connection to the thermistor, only two wires go to the 42V 2A charger. The charger would need to provide current to the thermistor since on the battery board there is only a connection through a 510K ohm resistor to one of the analog inputs of the PIC18F252 processor. Looks like the thermistor is only there for temperature monitoring in case a high current charger is used. I put the battery pack back together for now with a 10K resistor soldered on the board and dispensed with the wires for the thermistor. Did a test ride and it works OK.

  5. Great article! I can see why you are such a fan of RoHS labels :)
    I find your blog very useful as I am currently in the process of rebuiling a bionx battery.
    Just wondering if there are any new spinel cells out there beside the US18650V3.
    ~3500mAh is becomming mainstream these days - would you consider using inr/nca for your future rebuilds (+bms) ?

    1. Alex -

      I haven't found anything compatible except the US18650V3s, as a direct drop in replacement. You could, potentially, build a higher capacity pack if you added an external balancing BMS, but there's not really space for that in the current packs. You might be able to fit a very small one in this pack as there's some extra space, but it would be tight.

      I'm working on some paths to get higher capacity cells in the packs, but it requires a non-trivial amount of reverse engineering and design work - so nothing in the immediate future, unfortunately.

  6. 10S5P will leave enough space for a BMS while almost doubling the original capacity.
    The only problem is that a BMS will not balance anything under 4.2V.
    The Bionx charger stops at 41.5V and the benefit of lower voltage will depend on the quality of cells.

    How the US18650V3 is handling the current compared to the old US18650V ?
    The old cell is rated 20A and despite the good remaining capacity of my 10S6P pack (about 8.8Ah), my battery sags a lot under load.

    The battery gauge is depleted in 15 minutes but the bike usually runs another 20-30 Km until the battery becomes flat.

    1. As far as I can tell, the US18650V3s are holding up to the current just fine - the BionX systems don't pull very many amps out.

      I have seen a number of the older BionX systems reporting an empty battery when things will continue many miles past empty. I think there may be some sort of aging built into the BMS that assumes batteries degrade at a certain rate, and this is wrong over the 8+ years a lot of the packs are working. Plenty of packs I've seen come in are in the 8+Ah range out of a 9.6Ah pack, yet report empty very, very quickly while the pack still has plenty of voltage left.

      Check the voltage under load - you can use pins 2 and 4 on the charge port to read the direct pack voltage. I don't think you'll see it sagging that far.

  7. Interesting - I am using the console to read the voltage. Usually, after about 10 km, the max assistance level is decreasing (<64), the error 25 is flashing and the voltage drops from 39/38V to 34/33V under max load. I've tested how much current is flowing during acceleration and is was in the 20-30A range which is almost 1kW. The battery has a 30A fuse and I assume that the max continuous current is limited to 30A.

    One more thing, during the rebuild, I am supposed to keep the BMS connected to a power supply to avoid resetting the kit ? What happens if I disconnect the BMS ? I am supposed to enter an activation code ? Thanks!

    1. Sagging under high load is fine, though I'm surprised the system is pulling >20A - this is a 26V system?

      On the older systems, I've never had to keep the BMS powered - I don't know about the newer systems, though.

  8. 36V/350W system - I have another 36V kit which is limited to 500W and despite being a mid drive, it has noticeably less power. I am currently measuring the efficiency of both systems. Until now, it looks like the mid drive is using about 1-2W less power per Km than BionX - with constant gear changing.

    As for the rebuild, I might experiment either a 11S5P arrangement or perhaps a mix of 2S VTC6 with 4S V3 - in this way perhaps high current and self-balance could be achieved.

    1. Ah, ok. I've wondered about how much BionX was playing fast and loose with power ratings. The 350 I rode felt like it had a lot more than 350W behind it when climbing hills...

      I would really caution against doing mixed battery chemistry/model builds. That's a good way to get some insane equalization currents flowing around the pack, which won't do anything good. Getting 30A out of a 6P pack is only 5A per cell, which is well within the limits of anything modern.

  9. You are right - With 11S I will no longer be able to regen and if I mix low/high resistance in parallel I will probably stress the cells. I would probably stick to the original configuration hoping that the voltage drop will be less important. Thanks a lot for your advice ! I will post some pictures once I complete my rebuild.

    1. I'm certainly interested in what you come up with!

  10. Omg, you are so bad at welding batteries. Only 2 strips per series connection? Are you kidding me? It is insane if you make packs like this and charge money for it.. Learn first!!! BMZ pack was also shit design but they couldnt probably care less. You should if you instruct people and offer services.

    1. I'm building a replacement pack for a reasonably low amperage system - the fuse is generally 20A on the older packs.

      At the current draws involved, two strips are just fine - the resistance on such a short interconnect is quite low, so the heating and power loss are minimal.

      What, in particular, would you suggest I change on this particular build to meet your standards? Additional strips on the series interconnects? Cooper wire soldered on top of the strip? And, given whatever changes you suggest, what would the additional power delivered to the motor be with those?

      I'm not building 50-100A packs here. BionX doesn't draw that many amps.

  11. Hello,
    I have a BIONX 26V 7S6P battery with BMS Smart Connect 5.2.
    I would like to replace the battery with a 37V (10S4P) battery
    Do you the think Smart Connect Board could manage this voltage ? is it compatible with any voltage 26/37v or must I reprogram something ?
    Best regards

    1. Crossman - I don't believe that you can change voltages on the boards. At least, it's not something I've been willing to try with how rare the original BMS boards are.

      However, I do have a substantial capacity boost on the 26V battery available as a rebuild (nearly twice the stock capacity). Get in touch with me via the contact form in the right column if you're interested in that.

    2. Ok Russell, I thought that the Smart Connect Board could manage different voltage. Is it programmed inside ? because when looking to a BMS 26v, and another one BMS 37v , I really don't see the difference ( I have the two batteries 26v and 37v ).
      What are the best cells at the moment ? NCR18650PF ?
      Best Regards

    3. I believe the same board can deal with different voltages. I don't have enough information to understand how to program it, and BionX doesn't exactly publish that stuff. If you can get the information out of them, I could certainly make the changes...

      I use Sony US18650V3s for my rebuilds as they're the same spinel LiMn chemistry with the self balancing behavior the LiMn cells exhibit. Since the packs don't have balancing wires (until you get to the 48V packs), you need self balancing cells or things will drift over time and become dangerous.

  12. how to contact you for buliding a 7S6P ?

    1. Right column, a scroll or so down from the top. "Contact Russell (based in the United States)"

  13. I did a rebuild on an 8 year old battery pack.
    10s6p, 42V, Sony Ericsson US18650V1, March 24/2008 is the original cells.
    All is good but for 1 issue: When starting the dash, the battery symbol appears, then disappears.
    Everything still works fine but no battery indicator so we do not know what is left in the pack.
    Charges fine, discharges fine.
    All the case to bike connections are connected to the board, no broken or loose wires.
    Battery pack only has 2 wires.
    Is this a reprogramming issue as the battery was disconnected from the BMS/Controller in the case?
    Any ideas?


Comments on older posts are moderated due to spam issues. If you don't see your comment immediately, and you weren't just spamming me with some irrelevant comment and a link to whatever site you're trying to SEO, your comment should show up relatively soon. If you're trying to use my blog for your SEO purposes, your comments will never show up, so don't waste your time.