How To Build a Mid Drive Ebike That Doesn’t Break

No wearing-out of the drivetrain early. No broken chains. No motor shifting…
No whining. Read this so you don’t become That Guy on the internet.

This article is the follow-up to How To Ride a Mid Drive Ebike Without Breaking It.  That article points out how most of the online tantrums about unreliable mid drive ebikes are bad riding, not bad equipment.  

This article’s main points are all found in different posts here on this site, and together, here and there, they all cover the ground I am re-covering here. However, this subject comes up so often I decided to try and consolidate things into one place.

I am not trying to list every single thing you need to do to build a bike (I don’t mention tightening all the bolts or putting air in the tires, for example). I’m trying to shine a light on the more common mistakes. Don’t make them and you stand a good chance of having a trouble-free bike. Some of this stuff is pricey, so maybe you’ll want to work it in bit by bit as budget permits.


Remember: a successful DIY mid drive is about both building and riding optimally. Mid drives – particularly the ones made for USA-legal and adventurous, off road DIY builders – up the ante on the required competence of both builder and rider. There is no way around this. If you want idiot-proof and simple do a hub kit. If you want the versatility that comes with a mid drive, though, you need to put in the extra time and effort. There is no way around this.

I’ll focus solely on the mechanical bits this time, and break the process down into key component areas. We’ll start with:

Pick a Frame…

To get a mid drive to work properly, you need to pick a frame that can handle it, and this is not a given. Lots of frames are a bad choice. So what are we looking for?

… That Handles The Torque

I can still remember looking down at my very first mid drive build, a 4kw Cyclone, and saw the motor flex when I hit the throttle. With that flex, the whole bottom bracket flexed with it.

Thats a bad thing. Pick a frame of very sturdy construction. You are going to have an electric motor giving one hell of a pull on a chain that is connected to your back hub. That pull can flex the entire bike frame.

Can your typical mountain bike do it? Yes. Can your road bike with Columbus tubing from the 1980’s do it? Ehhhh lets say no on that one. Whats the problem? Designed for light weight and strength keyed to relatively smooth roads and human power, the stays are too spindly. All that power can pretzel the poor, innocent chainstays and seatstays when the power of ten pro riders yank on the chain.

How do you fix that? A lighter-duty BBS02 with the amp output dialed down is one solution I have seen done successfully several times. The lower amps do not yank on the chain hard. Some effort on the builder’s part to change the settings so pedal assist is kind and gentle is also important (and also preserves an authentic cycling experience).

… That Fits The Motor

Modern downtubes on mountain bikes tend to be curved and swoopy, coming into the bottom bracket at an arc that equates to a roughly 3-o’clock position. That arc means an installed external motor like a BBSxx has no choice but to hang straight down. Kiss goodbye your ground clearance. Here’s a picture of my 2018 Guerrilla Gravity Smash:

Take a look at how the down tube is curved, and how, on the right image where the drivetrain has been removed, its clear the only way to put a BBSxx-style motor on a bike like this will result in that motor hanging straight down.

You can see, sitting on the floor in that right picture, a Cyc X1 Pro motor, which has long arms that mount the motor as far forward as possible to avoid this ‘angle of the dangle’ problem.

How did I do in terms of mounting the motor and preserving ground clearance? Well, look above at the final installation. Look how low the chainring is. Draw an imaginary horizontal line parallel to the ground from that chainring… the motor is above that line.

The higher the better, but bottom line here is the motor is above the drivetrain so we haven’t lost a lot of ground clearance on this frame by using it (with the right mounting kit a Tangent Ascent motor will fit inside the triangle, just under the shock). So, this frame that is totally unsuited to one kind of motor can be pretty well suited to another.

Here’s another example frame. This is the type of frame typically recommended for a BBSHD, BBS02 or similar motor:

The down tube of this frame is straight as an arrow and attaches to the bottom bracket at a high angle. This allows you to rotate a dangly motor like a BBSHD up as far as possible so you lose as little ground clearance as possible. How did this shake out once this bike was built?

It came out pretty good. Bearing in mind this bike used a smaller-than-usual 40T chainring, the motor is roughly at the same height above the ground as the chainring. Maybe just a bit below. This is as good of a fit as you can expect on a bike for this kind of motor.

So… the lesson here is to think through your motor choice if you already have a frame to work with. Or the reverse if you have a motor on the shelf that needs a frame.

… That offers Good Chain / Crankarm Alignment

This is a tough one to nail down in advance. On chain alignment, you can come close during frame selection but you’ll never know for sure until you actually fit a motor into the frame, along with an assembled rear wheel so you can drape the chain and figure out how it lays.

As to crankarm alignment… thats one you will have to work out once you have a bike on a stand during the build. The key is to remember that your desired final result is to align the pedals directly underneath you… not the crankarms. Focusing on the pedals gives you a couple of extra options over and above finding an offset pair of arms to even things up. I have used uneven-width pedal spacers for some big changes, and different washer counts on one side or the other (from zero up to two) to move the pedals an additional 1.5mm to 3mm in either direction.

uh oh… lots of extra space on the drive side. Can’t use standard crankarms on this bike

I covered these issues in excruciating detail in How to Build an Ebike From Scratch: Tinkering. So I’m going to link you to that page and stop, rather than going through it all over again.

Buy The Right Drivetrain Parts

So very many builds fall on their face because the builder left cheap parts on the bike rather than replacing with strong ones.

Chainring

Here is another topic already done to death elsewhere. The options laid out here result in significantly different chain alignments. This is a pretty good general resource on the subject:

Generally speaking for BBSxx style motors with a large secondary housing located where the chainring ordinarily resides, you want a chainring with an inward offset that helps with chain alignment.

You also don’t want to overdo the size of the ring. Generally, “smaller is better” thanks to the rules that go along with riding a mid drive ebike (See below). You don’t want to pick a big ring unless you know exactly what you are doing with regard to gear selection and chainline. 42T is typically considered ideal unless you are riding singletrack. You go as small as you can get away with in that case.

Do a little shopping and you’ll find quality offset chainrings from the usual big name players aren’t cheap. Especially post-COVID. Here’s a new one I found recently on Amazon that offers about 15mm of offset and is well constructed. The offset looks to be more than that simply because the alloy used in construction is so thick.

I bought this ring and it appears a solid alternative – with a great tooth profile – for about half the cost of the other high end rings.

Chain

A beefed up chain is often overlooked and just as often results in an Epic Fail. You can’t run a powerful mid drive and expect to use the cheapo chain you already have on your bike. Whatever you do, don’t use a cosmetic (painted) pretty chain, or one of those pricey skeletonized weight-weenie chains. Instead, spring the bucks and buy a proper strong chain.

I run 11-speed systems with a KMC e11 chain – the KMC ‘e’ line is specifically designed to take a mid drive’s punishment. These 11s chains are brutally expensive (today’s price is over US$47), but this is part of the cost of admission if you want to run 11s and a high powered motor that doesn’t snap chains (or wear them out really fast). 11-speed is a wonderful thing to have on an ebike – particularly on a bike you pedal instead of throttling – but the cost of durable 11s drivetrain parts is a serious deterrent.

The story gets a lot better if you are using an 8, 9 or 10-speed setup. The SRAM EX1 ebike chain is used for their (hideously expensive) 8-speed EX1 mid drive-stressed drivetrain system. Here’s the thing though: That chain has an MSRP of US$28, and is usually sold for about US$25. The link above to Amazon has it on sale right now for US$18.95.

It gets better still: The EX1 system is 8-speed, but the chain is sized for a 10-speed system. That means you can use this chain on 8-, 9- and 10-speed drivetrains. Since 9s is probably the ideal sweet spot for mid drives, this chain can be considered almost everyone’s inexpensive default. To sweeten the deal, the chain comes with 144 links. So it will fit everything but a longtail without having to buy and section two chains together.

Rear Derailleur

This one is quasi-optional. A quality derailleur is always a good thing. Especially if your alternative is something like a cheapie Tourney or similar.

11 Speed

I use a SRAM GX 2.1 long cage rear derailleur. This is not meant for a 1x drivetrain but if you just use a narrow-wide front chainring, you’re fine. At over US$100 a pop this is not a low cost option, and one more reason why you need to know you want more gears and more finely-diced cadence options to bother with 11-speed.

9 Speed

I have found two I consider to be stars:

Box Components Prime 9
I have two bikes with these, including my most recent build: The Apostate. My preferred version is the Box Two Prime 9 Extra Wide, coupled to their Box One single shifter, which is tailored for ebike use. The Extra Wide part refers to extra wide range on the rear cluster, which in English means Big. You can go up to I believe a 50T rear cluster with one of these. Combined, shifter-plus-derailleur is about a US$185 solution. I like it better as a premium option because… it shifts spot-on, was super easy to initially adjust and runs silky smooth.

Microshift Advent
I like the Long Cage version, which runs US$60 and is meant for 2x systems but works fine on 1x. The Pro shifter is single-shift like the Box, and costs a whopping US$29. This setup is significantly cheaper than the Box. It works just a little less slick, and the fit/finish is more… workmanlike… but its nothing to complain about given the price range its in.

I have to say I like the adjustable clutch on the Advent better than anything else I have ever used.

Shifters

I have mentioned but not explained why single-shift is needed for a mid-drive so here goes: Shifting under power is a Very Bad Thing. If you shift a bunch of gears at once, as is easily possible with a standard trigger shifter, its Way More Bad. If you are trying to snap your chain as quickly as possible (or taco a chainring or cog), thats how you do it. Single-shift shifters only allow one gear shift per trigger pull, so you can’t screw up. So either learn to carefully shift one gear at a time (certainly you can do that, as I do on my older bikes) or buy something that eliminates the issue.

This looks like a good time to link to the discussion of whether or not you want to install a gear sensor. Spoiler alert: I don’t use them and I still stay safe.

Rear Cluster

Put simply, you want durability. Forget about light weight. You want all-steel cogs, preferably pinned together into a monolithic block. That block distributes the force across the cassette body/freehub underneath so you don’t dig a trench into the poor thing when you hit it with a shipload of watts.

11-speed
The Sunrace CSMS7 is an all-steel 11-42T cluster, pinned together. I haven’t found anything else that is all-steel in an 11-speed. These clusters are often tough to find in stock and there’s no telling how long the link above will remain valid.

The CSMS7. Steel spiders, steel cogs and permanently pinned together. Wunderbar.

9-speed
For bikes that don’t require a lot of range, you can’t beat the Shimano HG400-9. They are found in a range of tooth counts – my favorite is the coated 12-36T. They are dirt cheap and functional. All steel except the smallest cogs, and pinned together. Nearly indestructible and easy on the cassette body.

The Microshift Advent 11-46T Hardened Steel cluster is my absolute favorite. A nice wide range up top. No spiders (each cog is 1-piece). Pins are all over the place to hold it together and prevent any torque tacos.

And its cheap with a US$35.99 price at the time of this writing. Unbeatable for a 9-speed mid drive build.

Rear Cassette Body (Freehub)

Whatever it is, it needs to be steel. If not, with an unrestricted DIY mid drive you’ll dig into it like its made of cheese. This is a big deal. I’m only devoting a couple of sentences to it, but thats because there’s just not that much to say. If you want to avoid tearing up this rather expensive and possibly difficult to replace part, use one made of steel. Period.

Left: Alloy at 1300 miles (and an AWD bike so it had an easy life). Right: Steel replacement. A year later I checked the steelie and it wasn’t even scratched.

Rear Hub

With all the talk above of strong chains, steel cogs and steel clusters, you can see where this is going I bet. With all that durable hardware we are reinforcing the drivetrain further and further down the line until we find the next failure point. After we get to the surface of the cassette, the next thing that breaks are the pawls inside the freehub. Whats a pawl? Here are a couple of pictures where they are visible:

The little things sticking out on the end are ‘pawls’: ratcheting couplers that hold the hub firmly while you and your motor apply force to the rear thru the chain. If they give out, the chain ‘freewheels forward’ and you lose the ability to apply force to the rear wheel to make it go. Mid drives putting out big power put big stress on these poor little pawls, and they can die as a result.

Its not hard to see why. All torque is transmitted through those pawls. What you need is more points of engagement. You can find 4, 5 and even 6-pawl freehubs. You can also buy quality rear hubs that can take this level of abuse (I hear Salsa makes some, as do Sun Ringle). You can also change the game and go to a different kind of mechanism that is effectively indestructible to a mid drive…

Hooray for a ratchet engagement mechanism

Bike Radar did a good piece on freehub construction so I’ll just link to that article here and let you read the details, and the comparison of pawl vs. ratchet. So far as I can tell across several wheelsets and several thousand miles, the DT 350 hubs with standard 18T and 24T ‘Hybrid’ ratchet mechanisms are effectively indestructible.

Rear Wheel

So we’re walking the points of failure back still. ‘Under’ the inside of the cassette mechanism is the hub, spokes and the rim itself.

This is not something anyone wants to hear, but if you want a trouble-free mid drive you put on a quality wheel built for it, by someone who knows what they are doing. This is not cheap, hence the reason nobody wants to hear it. And you may be able to do without. Heck… a lot of Mongoose Dolomite conversions are out there and running just fine. But if you expect to pour on the miles and keep doing so for years, trouble-free. Well, your money will be well-spent on a beefy wheel.

Lets take a typical 26″ example. I really like the SunRingle MTX39. It comes in 32- and 36-hole versions. At US$60-90 each they are not cheap, but also not crazy-expensive, either. Its ridiculously strong, but isn’t light weight.

My MTX39 on the rear of my Bullitt. I also use them on my Envoy.

If I want to blow off the budget and go for strong, lightweight (and tubeless) a DT Swiss FR-560 is a winner. I have them on 26″ and 29″ wheels, and they are also available in 27.5. But at about US$150 a pop you have to want em pretty bad.

Next, what about the spokes? Choose good, strong ones. Sapim Strongs are an excellent choice. So are DT Swiss Champions, or DT Competitions. If you are made of money, DT Alpines are pretty awesome, too. These spokes as a body are not the great big 12-gauge spokes you see on some imported Asian rims. the European manufacturers substitute quality materials and smarter engineering and don’t need that massive construction, but they still provide superior strength.

And lets not forget the nipples! Spoke nipples that is. Once again, choose strong ones, not light ones. That means no alloy. Use the brass ones.

Don’t Ride Like a Dumbass

Build it as smart as you want, but if you do bad things you will get bad results.

How To Ride a Mid Drive Ebike Without Breaking It

So last but not least, follow that link. And don’t be a …

Larry vs Harry Bullitt – The Cargo Box

This is the heart and soul of your Bullitt. The reason it exists. There are many brilliant cargo area customizations out there. This is not one of them, but I did do some stuff worth at least mentioning.

The Bullitt Build
1. Battery and Battery Box
2. Cargo Box (You Are Here)
3. Brakes
4. Front Motor & Wheel
5. Rear Motor & Drivetrain
6. Bits & Pieces

The Cargo Box

Truth be told, I didn’t do all that much to the cargo area. At least compared to some of the engineering marvels I have seen some owners put together. But I did do some things a bit differently than I have seen done before, so I decided to write them up.

Mounting Changes

I used the stock honeycomb deck sold by Larry vs Harry. As detailed in the Battery Box episode, the downward-facing bolts are now upward-facing studs. That means I need to use a nut up top – sticking out in the cargo area. So I need to manage that. Additionally, because under the honeycomb floor is a very, very expensive battery, I wanted to introduce as much security-through-obscurity as I could to the installation.

Security Nuts

As seen in the pictures below, I used some oddball security nuts. Now, anyone who has ever owned a pair of vise grips, or maybe some good channel locking pliers, knows you can get these nuts off of the stud without needing the special keyed security socket. But it will be a pain in the ass, and probably a 10-minute process (you’ll need to know in advance you must put a wrench to the socket cap underneath or the whole bolt assembly will just spin and not loosen) versus … what? 30 seconds? Its just another layer of security to make things more difficult for a thief thinking about stealing the battery out of the bike.

The ‘studs’ are in fact M6 socket cap bolts, sized in advance for this job. When I used a nice, thick, wide Grade 8 washer and screwed down one of these security nuts on top, the height of the washer and nut equals the length of exposed bolt. I did not use any sort of thread locker. Instead I made a visible registration mark – a dot – on the washer and a matching dot on the nut. So long as the dots line up, I can tell at a glance they are still tight.

Early on I played a few games to cap the nuts so they didn’t intrude unduly on the cargo area, but none of the ideas really worked well until I came up with my final solution, described below in the Padding section.

Wiring and the tubes

As mentioned in the episodes covering the front motor, the battery box and the brakes, I decided to run my brake and front motor cables through the cargo area. Partly because a battery box occupied the space where the brake cable was to be routed, and partly because I wanted the cables to run internally to keep them out of harm’s way.

As seen in the pictures above, I ran one cable down each side of the cargo bay. To keep the cable from getting in the way and snagging on stuff during daily life, I used furniture grade (thin but sturdy) PVC tubes as cable guides. Each tube is mounted to the side panel brackets very simply with criss-crossed zip ties at each bracket. The tubes are solidly mounted and are not going to budge.

In a couple of pictures above, some kind of black glop is visible at the point where the cable exits to the front of the bike. That is 3M 2228 moisture-sealing (mastic) tape. Slight variations of the same product are sold as ‘electrical insulation putty’ and thats what I used the tape for here. I wadded it up into a blob and stuffed it around the cable ingress point. It not only forms a waterproof seal, it holds the cable firmly in place at this point so it can’t rub back and forth on the (sharp-ish) edges of the honeycomb board.

Padding

As noted above, I was looking for a way to deal with the mounting studs/bolts sticking up into the cargo compartment. At the same time, I was also looking for a way to pad the cargo bay interior. As I go bouncing down the road, my cargo can bounce along with me in time with the potholes. Something to deaden and reduce that bouncing around was in order. I also wanted to use padding that was thick enough to support a person. Something that would not squish down paper-thin if someone’s backside were plunked down in the bay.

After some poking around, I settled on Minicel-T600 closed-cell EVA foam in a 1/2″ thickness. 1/2″ doesn’t sound like a lot but this is some serious foam. It is very tough, and incredibly fine-celled. It feels – and somewhat looks like – a fine-grained suede.

Its a perfect foam for a cargo bay that is going to have stuff tossed into it and dragged around. Because it is so strong, I can get a lot of mileage out of a foam that doesn’t cost me much in the way of lost cargo volume. I can sit on it and the foam doesn’t bottom out. Better still, it forms around and over top of the security nuts so you can’t tell they are there, unless you sit on one, but even so its not especially uncomfortable (I give a passenger another layer of that same foam in a square to make sure nobody has a problem).

Better still, its another layer in the way of getting at the battery box. If you are looking to help yourself to my battery, peeking into the cargo area (I do leave the tonneau on the bike when I go into a store) won’t give an immediate clue as to how to get underneath. You’ll have a minute’s extra work pulling up the floor, at which point you will learn about the security nuts. Again, not perfect… but layers.

I bought a large sheet of the padding and, after a lot of careful measuring, cut out a pattern that lets me lay a single, fitted piece into the cargo bay that covers the floor, the sides and the front. I drilled a couple of holes in just the right spots at the back panel and ran extra long bolts thru the padding and into the panel, so the layer of padding is actually bolted down at the back. I could have done two more bolts in the front but its not necessary. Its not going anywhere as it is.

CAD? Schmad. My finely detailed technical drawing used to cut out my pattern. Dotted lines are folding points.

I have toyed with the idea of cutting down the right and left sides of the padding so they are edge-fit-flush with the side panels (maybe needs about a 1″ lip on each side), rather than running up them and padding the sides. The idea would be to recover that 1/2″ of lost storage space on each side, at the expense of side padding. But, once I cut it there’s no uncutting it, and I have yet to need that space more than I want the clean appearance and full coverage I have now.

Front Compartment

I found I have enough garbahhge carried with me that its better to wall it off into its own semi-permanent compartment. If I need the whole bucket I can always mount panniers on my rack and move all this stuff to the back of the bike.

Thats a cut-down knee pad from Harbor Freight doing the job of compartment divider. And yes, thats a folding stool under the green bag (which is my 5a traveling charger)

The Divider

I really like these things. This is nothing more than a super-dense knee pad that you can buy on Amazon for about US$18-$20, Except, if I go to Harbor Freight I can get them for about US$6.50. I literally have a half-dozen of the things doing various jobs as a knee pad, a sun shield for my SoGen, or as is the case here a compartment divider. I had to cut it down some to fit my cargo bay tightly, and a bit off the edges to fit under the tonneau. It can also be pulled out for use as a knee and body pad if I am unlucky enough to have to work on the bike, roadside.

The Ridiculous Lock

Actually its a chain and two locks. A Pragmasis 2-meter boron steel noose chain, with a long Pragmasis motorcycle-grade U lock, and a medium version of the same lock. Since Godzilla is a utility bike I need the locks to always be there for me should I stop at a store. There’s nothing worse than going by the supermarket and needing something, but having to come back after I go and get my lock. There’s a mistake I made once or twice before deciding the locks stay on the bike always.

I’m not fooling around when it comes to locking Godzilla. This method is quick to deploy and requires multiple cuts with an angle grinder to free up the bike.

I keep the locks in a well-fitted, oblong MOLLE pouch that also holds their keys. I never have an issue of forgetting those keys as a result. Which means I don’t have to pack back up and leave before I even get in the door at the store (once again, thats a lesson learned the hard way).

The Chair

Well, really its a stool. I found back in the bad old days of the pandemic shutdown that riding a bike to a store and waiting for curbside delivery to show up was MUCH easier if I had something besides the bike to sit on. A collapsible stool fits the bill and so I keep one with the bike. Godzilla’s is an 18″ Walkstool which, despite its cost, I highly recommend.

The Charger

For many of my bike builds, I make my own weatherproof chargers. I often mount them permanently on the ebike. Its the ultimate in range-anxiety relief. But this time Godzilla has a 32ah battery. A monster battery for a monster bike so I figured I didn’t need a charger this time.

After I got a few hundred miles under my belt, I found I could still screw up, forget to charge and remember this in the middle of a ride clear across town. If I’d had a charger on board, I could have stopped at a local city park, kicked back and taken advantage of the electrical outlets present on the canopies, for the use of picnickers. But noooo I didn’t want to lug around a charger… so lets turn the pedal assist all the way down and hope we make it.

I didn’t need to do that more than a couple of times before I put one of my 320w, 5a chargers into my kit. I think the one you see above, in pictures I took in 2019, is the same one I am using in Godzilla now.


Thats it for the cargo bay. Who’d a thunk I’d come up with almost 1900 words to describe a big empty box? Lets see how much space I can take up talking about something simple like…

The Brakes

Larry vs Harry Bullitt – Project Odds & Ends

The Bullitt Build
1. Battery and Battery Box
2. Cargo Box
3. Brakes
4. Front Motor & Wheel
5. Rear Motor & Drivetrain
6. Bits & Pieces (You are Here)

Whats Left?

I’ve hit all the big ticket items. Now for the bits and pieces that may be of interest.

Fahrer Panel Bags and Tools

The Fahrer bags were a late addition. Partly because they cost so damn much. I was not willing to spend the money until I finally realized my idea for a MOLLE panel, trimmed to size, with bags attached was just more effort than I felt like going to get the same result in DIY fashion.

For me, the point of these bags was to move my tools out of the cargo area. I seldom need to mess with/rummage through them so I wanted them in a place that kept them out of the way. At the same time, there is a fair bit of money wrapped up in this toolkit, not the least of which are the battery powered air pump and the ridiculously expensive but wildly useful Knipex pliers/wrench. For a bike that gets left outside of a store a lot, I wanted to be able to easily, routinely remove these bags for carry inside with me.

I settled on putting the tools in a bag that fits inside of the bag. Specifically I used one green and one brown Condor Field Pouch, which have handy top-grab handles for pulling the bags easily and quickly up and out.

Two of these bags, in two contrasting colors – green and brown – so I know which is which on sight.

Ordinarily these Condor bags are used for something like a 1L water bottle, but they are a near-perfect shape for form-fitting directly inside the Fahrer bag. The different colors help me remember straightaway whats in each pouch.

The contents of my tool kit fit along my usual lines: I want to be able to use what I am carrying to do just about any typical field repair. I like to work with full-sized tools, so I am not at a disadvantage on the side of the road where life sucks bad enough already if I’m broken down.

Worth mentioning are those two black disks in the pictures above. Those are not emergency field rations (unless its REALLY an emergency). Those are two regulation hockey pucks. Solid, dense rubber blocks. What the hell for? Set one each under the Bullitt’s 2-leg kickstand. Now either the bike’s front or back wheel is up and off the ground by several inches, depending on which way you tilt the bike. Instant quickee service stand.

Rear Rack

I’ve heard its tough to fit a rear rack on the back of a Bullitt. There are M5 bosses built into the dropouts, but it can be a challenge to utilize them as the brakes get in the way on the drive side.

I found an Axiom Streamliner 29er DLX rack fits perfectly.

UPDATE: Mar 22 2022. I actually used this slightly different Axiom rack that was sold as the Streamliner DLX. Nowadays this same rack is marketed as a 29er rack and now includes the extra-strong center mount (I used a spare from one of my Fatliners) as well as the stay mounts. The current Streamliner is narrower and sold strictly as being compatible with 700C tires (although I have one of them on the front of my Envoy that fits 26×2.8″ tires just fine).

Its lower mounting arms move the rack 4cm out and further back to ensure there will never be any heel clearance issues. Its strong – rated to a whopping 50 kg, although I don’t know about that. I’d say 25 kg is still pushing it, but as racks go its very strong and can stand up to any reasonable level of use. I use it to hold the more delicate items that come from the store. Bread, chips and such. Or if I am just expecting a large or heavy load I may attach the panniers so I can clean out my cargo bay so its one big empty box.

But the rack needs just a bit of surgery to fit. Note in the picture below I did just a bit of filing to give clearance for the brake adapter bolt. Without this small adjustment it won’t clear.

A bit of hand filing with a fine half-round file took just enough off of the mounting arm so it fits – without compromising carry capacity.

I also used the center mount from an Axiom Fatliner rack (I have several of them). It has about – literally – four or five times the material in it as the Streamliner center support. I did have to drill new engagement holes for it. Maybe I can get away with that 50 kg rating after all.

Kicktail / Reinforcement / reflector

I have done a reflector/kicktail before on many bikes as sort of a signature mod. I usually take a 6″ x 18″ street sign, stand on it, grab an edge and bend up until I have a lazy L shape of roughly 35-40 degrees. I drill some attachment holes and the result is a 1-piece rear rack deck plate. The kicktail ensures no mud comes up at the rider, ever. Coupled to other measures I like this approach better than a traditional fender.

To add utility, I also cover the underside of this kicktail in 3M 3432 Red Micro Prismatic Reflective Tape… Exactly the stuff used in manufacturing municipal street signs in the USA. It so happens that a few years ago I scored a 12″ wide roll of this stuff at a small fraction of its regular retail price, so I can just unroll a bit of it and slap it onto the back side of the kicktail and voila… the world’s largest red bike reflector. I guarantee you I am visible to an overtaking automobile even if I didn’t have my steady and blinkie taillights fired up.

But my usual racks are fat bike racks where a 6″ deck plate works in conjunction with mounting panniers. Not so with the much more narrow ‘normal sized’ tires of the Bullitt. I needed just a narrow strip on the top of the rack with channels on each side to allow the pannier hooks to be engaged.

So I used a narrow 2″ strip of aluminum flat bar in the full length that I needed, and then bent that bar up to the desired angle. Following that, I bolted a second, full-rack-width 4″ x 12″ piece of aluminum flat bar onto the strip to make the kicktail. Line the back side with tape and functionally the job is done. However, since I had discovered a green duct tape that is a near perfect color match for the Larry vs. Harry ‘Lizzard King’ green, I wrapped the upper surfaces of the rack with it for what I think is a nice look.

The ‘Bobtail’

So… the 2-piece kicktail… in the end I decided it wasn’t working for me. The giant red reflector was a big winner as they have always been, but the need to do a 2-piece thanks to the narrower rack caused issues that just grated on my OCD (just kidding I am not diagnosed with this condition). Being 2-piece, it jiggled around some while riding simply because it was heavier than its predecessors. Also, the metal plate rested directly on the rack and rubbed, scraping off the protective powder coating. Trying to pad that with 3M 2228 mastic was only marginally effective as the stuff inevitably wore down and mooshed away under the force of the metal edge bearing down upon it.

To solve this issue I re-thought my approach. I discarded the wide flat plate. Next I cut down the rack deck plate so it was just long enough to intercept all spray coming up off the tire. Then I lined the now-exposed rear with the same prismatic red reflective street sign tape used on the full kicktail. In the front, I affixed a swatch of orange prismatic tape to provide another reflective surface. To finish it off, after filing down the upswept end so there are no sharp edges, I took the added precaution of making a bumper with a layer of that same rubber 3M mastic.

Job done.

With this change, I hit the sweet spot. The bobtail still keeps water off of the rider, it still has a large reflective surface and I quickly grew to appreciate the shorter profile. Godzilla is already about 8 feet long before that tail comes into play, and I don’t miss bumping into the kicktail as I walk around the bike.

Rear Thru Axle

Surprisingly, sourcing a proper rear axle was a major pain in the ass. The rear axle is a 12×166 (12mm in diameter by 166mm in length). The ’12’ part is standard stuff. The ‘166’ is a little unusual but still not a dealbreaker. Mix in the required P1.5 threading and thats where Google and Amazon searches go off the rails.

After a fashion, I ended up finding two of them. I ended up using the cheap one and have kept the fancy one in reserve.

The Cheap One:
Dymoece brand 175mm with 19mm thread length. Its exact specs are written right on it in the picture below. I was surprised at how well this was machined. Its 175mm length is too long, but I had a 3mm thru axle washer in my parts bins. Putting that on sizes the axle so there are 2-3 rows of threads visible coming out the derailleur side. Perfect. Everything lined up on this one from the thread pitch to the slightly longer thread length to allow the use of the washer. I like this axle because it is tool-less. I am already carrying hex keys and they’ll work just fine if I need to remove this axle roadside.

The Fancy One:
DT Swiss model 2160062210. This is a model that has a snap-on lever – that is just a glorified hex key – for removal. It looks nice on the bike, and you can remove it so nobody walks off with the thing. For me, since I am already carrying hex keys, I don’t need another one hanging off the end of the axle. I also like the fact that – on a bike that will sit out alone in public – no lever means no invitation to passersby to remove said axle. Still… its a DT Swiss part so its top quality.

Good luck trying to find details on this thing without looking deep across multiple web sites. The Amazon product listing gets it all wrong, calling it a 142mm length (which is correct for the internal dimension) and saying absolutely nothing about thread length or pitch.

Here’s what it really measures out to: 12x171L, thread pitch is 1.5 (it had better be) and thread length is 13mm.

With the shorter length and shorter thread pitch, it fits the Bullitt frame and dropouts perfectly without a washer, and the thread overhang out the drive side is minimal.

Rear Fender

As usual, I cobbled together something from parts and raw materials. Here’s what the finished product looks like from a few different angles.

This is a pretty non-traditional fender set up that does the same job as full coverage fenders without the full coverage. We’ll first cover the materials, and then we’ll discuss the assembly as a whole.

Gather Your Materials

Every bike builder should own some of this stuff: Flexible black cutting board. Unfortunately the black stuff seems to be off the market for the moment, or I am just not looking in the right places. Here is a link to something made of similar material to give you an idea of what I am talking about. What you want to do is replicate the same base material that is used for flexible MTB mud fenders. Once you have that you can cut that raw material to shape as needed.

Next, I used two of these front fenders, that I interwove together, facing each other, so they made a single longer fender. total cost for the pair – which are the core of the fender project – is US$12.84 delivered.

I created the rack deck plate and bobtail with a 1/8″ thick, 2″ x 24″ piece of 6061 aluminum flat bar stock – sold as a remnant – that cost me $8.00 on Ebay. That is way cheaper than normal. Here is a 48″ long piece that is 1/8″thick x 2″ wide. You can cut down to your needed length. Save the rest for your next project. You can go much wider than this for a rack deck plate, but I decided I wanted to go narrow for reasons that will be clear when you see the next item.

I continued to use the paint-matching duct tape that I found by accident across all parts of the rack. If your Bullitt is not the older, brighter Lizzard King color you will want to go with something different. If black works for you then Gorilla Tape is probably my favorite overall material for this kind of work, and in fact I use it here, as described below.

Put the ‘fender’ together

Its not really a fender, but rather a series of pieces that do the job of a fender.

First of all, there’s the rack deck – the bent-up piece of aluminum flat bar. Its length is determined by how long it needs to be to intercept all water spray coming up off the back wheel. I also want enough of a ‘tail’ so I can add enough reflective tape so it is a sufficiently big rear reflector. I used a relatively narrow piece that is 2″ wide. Why?

I wanted enough free space on each side so I could hook panniers to the rack. I *could* have done a full coverage deck – which would have allowed me to use a wider bobtail, and then used a router or hand files to cut out slots for pannier hooks. Frankly… it was easier to do a narrow piece and then follow up with …

Tape. The color-matching tape is nothing more than a wide series of strips folded over onto themselves to make a 2-sided sheet fully as wide as the entire rack deck. Then I bolted the aluminum flat bar on top of it. To allow for the hooks to be popped onto the top of the rack tubes, I simply took scissors and snipped the tape where there is a crossbar, to allow it to fold down when I clip panniers on. This is a bit cheesy, but it got the job done. If some time down the road I have some time on my hands and about US$30 I will cut a full width bar and slot it so its prettier than what is there now.

The rack and bobtail thus do most of the work keeping water coming up off the wheel. Now we have to deal with water coming forward. That is accomplished with the cheapie, doubled front mud fender. It is attached with small green zip ties to the seatstays, and its length in the back lifts up and contacts the rack deck. So… no water is coming forward past that. The front half of this fender also comes down and butts up against the seat tube fairly low, so there is very little area uncovered. But lets cover that too.

To deal with that last bit, I used the cutting board material and snipped out a rectangular length that is as wide as possible while still staying out of the way of my legs and the crankarms (look for it, installed, in the pics). I poked a couple of holes in the right spots and simply screwed this length of ‘fender’ directly to the frame using the two M5 bosses that are already there for fender mounting, on the back of the seatpost.

Front Fender

On the front, I cheated and found something I could just buy and install.

Godzilla uses a Greenguard-belted Schwalbe Super Moto X tire in 20″x2.40″ size. This is larger than most tires – in fact its a hair under being too large to fit the frame. Most fender solutions will not fit.

The SKS Rowdy 20-24″ fender set provides the solution. I shelved the rear fender and used only the front half of the front fender. The front mount without the fender attached is shown below.

The spacer under the mount keeps the fender from interfering with the tonneau.

Since the fender slides onto the mount bayonet via a firm interference fit, I keep the fender in my toolbag. If I need it I just slide it on. I like to be able to see the tire when riding, as part of my general need to be over-informed of every aspect of the bike as I go down the road.

In addition to it not fitting the bike, there’s no need to use the rear half of the front fender. The front of the cargo deck shields the rider from any splash coming up off the back of the tire. I did add some waterproof rubber tape to the frame at strategic locations to ensure it is a fully sealed face to splashing water.

Steering Dampener

I am using the generic Larry Vs. Harry steering dampener. I found the bike had a horrific speed wobble prior to adding the side panels, which almost completely eliminated said wobble. It only reoccurred on certain downhill road segments where there were repeated road undulations… that set up just the right harmonics for it to happen again. Installing the dampener and setting it to its lowest level cured the problem forever.

Jones Handlebars

I use Jones bars to alleviate pain from pressure on my wrists that came from getting hit by an inattentive motorist in 2017. Their 45-degree angle works wonders. They also give me a long grip area that lets me change positions, and lots of real estate to attach all manner of gadgets. An AWD ebike with one throttle for each thumb, dual PAS panels and dual displays, headlights, a dashcam, missile launchers etc. benefits from the wealth of mounting points available.

I also use two sets of Wolf Tooth Fat Paw grips. 1 and 1/2 grips per side. The extra diameter works well with my hands and thanks to those wrists, I need the extra padding. I cover the 2-piece-per-side grips with a wrap of silicone plumbers’ tape that keeps them grippy but not sticky. If I tear up the grip tape somehow (this happens rarely) I just wrap on another layer.

Lights

Long ago I gave up on trying to find a light that suited my needs, that I could wire into my main ebike battery. Instead all of my lights have their own internal, rechargeable batteries. Once a week at the office I charge them all on the same day, via a USB hub that is a part of my office garage gear.

I use my lights as daybright, daytime running lights and night use. The need to have the lights day-visible means I am going brighter than most folks’ lighting setups.

My lighting choices are informed by the knowledge that blinkie lights are the best to improve the ‘conspicuity’ (yes thats a word) of a rider, but they inhibit a motorist’s ability to track motion. So I use a combination of steady lights and blinkies, front and rear.

Front lights

Out front I use a pair of Niterider Lumina Micros for my main headlight beams. I split my headlights into two widely separated beams for the same reason cars do it. It makes for a wider beam pattern than you can get even with a dual-bulb single headlamp.

In between the two steady headlights is the front version of the Knog Big Cobber. This is a pricey light, but its also probably the brightest, smartest front blinkie you can get your hands on. It actually has an app that lets you program its modes, and I have mine set to ‘eyesaver’ mode coupled to their econ mode (short but intensely bright) blink. The Cobber lights from Knog shine in a 330-degree arc. So you want the eyesaver mode to keep from having that blink coming back at you.

Rear Lights

Out back, I have another Knog Big Cobber as my blinkie. Again it is set to ‘econ’ mode which is an intense, sharp, short blink. This time I let it ride for its full 330-degree illumination. That blinks light to the ground, to the motorist and up into the red reflector of the bobtail. Hopefully they’ll see me.

To aid said motorist in tracking my motion, I have two Knog Blinder Square lights, set to a steady illumination of the outer half of the light.

Running the light half-strength like this lets it last the full week as I like to charge everything together on Friday.

Clip-In Pedals

After literally decades of riding first toe clips and straps, and later clipless when they were introduced into the market, I have been riding flat pedals and Five Ten Free Rider sticky platform pedal shoes for the last couple of years.

I decided to go back to clipless not too long ago, and I have found the Funn Ripper pedals to be amazing. Their spring loaded cleat mechanism – the cleat sticks up and makes cleating in ridiculously easy – is a wonder. Highly recommended. I’m using Giro Rumble shoes which let me cleat in, or get off the bike and walk into a shop or wherever comfortably without clackclackclacking on the pavement.

As you can imagine, being locked into the pedals again made for an exciting transition period, and I managed to fall over at an intersection within 24 hours of making the switch. But once that little reminder was hammered home, I haven’t done it again.


So… Thats That

It seems I’ve come to the end of this somewhat tedious breakdown of all things Godzilla – The Big Green Bullitt. If I come up with any more I’ll amend this post. But, honestly I’m happy this series is done as it lets me get on with other projects I have been looking forward to, but can’t begin because this monstrosity remained unfinished.

Larry vs Harry Bullitt – Rear (BBSHD) Motor and Drivetrain

The Bullitt Build
1. Battery and Battery Box
2. Cargo Box
3. Brakes
4. Front Motor & Wheel
5. Rear Motor & Drivetrain (You Are Here)
6. Bits & Pieces

The Mid Drive

Godzilla is a 2wd/AWD ebike, with a Bafang geared hub motor in the front, set up as a helper to the Bafang BBSHD mid drive powering the back wheel. This is not the first AWD ebike I have built and we’re now beyond what I call my Gen3 configuration. I don’t call it Gen4 as it is still effectively Gen3, and incorporates all of the lessons I learned (and mistakes I did not repeat). Its different in that the two motors blend together in a kinder, gentler fashion I am calling Drama Free AWD.

Fitment to a late model Bullitt

Lets not sugarcoat this: The BBSHD does not fit a Bullitt. BUT… its pretty easy to rectify that. Thankfully there is very little needed to make it fit.

You will figure out for yourself what the problem is real fast. Just try and slide the motor axle into the bottom bracket. Oof… one of the threaded ‘ears’ that holds the bottom bracket clamp overlaps the frame. There are three ways to fix this. I will warn you right now: If you try Method #3 you will be brought up on charges and imprisoned.

Method 1: File down the ear a little

Thats all there is to it. The ear is quite a substantial piece of alloy sticking out of the motor casing. Just file enough of the corner off so it clears the frame. Do a little at a time with a hand file. File a bit, check to see if it clears. File some more. Check again. Repeat until it fits thru. Do this right and the motor will clear the frame and there will still be plenty of material left so there is no concern of any kind as to the structural soundness of that ear.

Figure 1 – Red arrows: Thats it? Thats it! (ignore the heat sinks. We’ll cover them later).

Looking at the picture above, its kind of hard to see exactly where I filed it, because I took a black Sharpie to the freshly exposed alloy and just blacked it out. A few months later, its worn off a touch but you still need to know where the work was done to be able to see it at all.

Method 2: Partially disassemble the motor

This method is for the Felix Ungers of the world. If you cannot bear to take a file to the motor casing… take it apart, then position it and reassemble the motor in situ. There are only a few screws involved to make this happen… but Jesus H. Christ this is waaaay more effort than it is worth in my own personal view. If you ever want to pull the motor off you have to take the damn thing apart again to get it off the frame. For me… give me the file and 5 minutes for Method 1 and I’m good.

Even so, if someone goes this route I can understand and respect the insane attention to detail such a course exemplifies. HOWEVER…

Method 3: Take a file to the frame

Yes really. I’ve seen it done. Rather than filing off metal on the motor, some folks decide they want to file down the frame instead. The thing that is in the way on the Bullitt frame is a weld seam, and I suppose that structurally, the stuff in the way is not critical to the frame’s structural integrity. But… as far as I am concerned a bicycle frame is a sacred temple and anyone who violates the sanctity of that temple… well they should be taken out and shot. Don’t do that.

How Well Does It Fit?

Once you get the motor fit into the bottom bracket, Its a great fit for the frame. One of the Lekkie (42-52T) or Luna Eclipse 42 or 48T chainrings will give you great chain line. The motor does hang almost straight down, and this looks a bit disconcerting, but the reality is you have other things on the Bullitt that hang lower. The motor is not in any way a hindrance to ground clearance unless you decide to start rolling up and/or off of curbs and such. I have never hit anything and I have never heard of anyone on a Bullitt having such a problem.

Crankarms

In your typical BBSHD or BBS02 installation, the secondary gear housing – just behind the chainring – is present where, on an ordinary bicycle, nothing exists. Consequently the chainring is pushed perhaps as much as 2cm outboard from where it would be. To counteract this, its common to use special chainrings that offset themselves inward to undo what would otherwise be a disastrous chain line. That fixes the chain line, but it does nothing for the alignment of the pedals underneath the rider.

If you move the chainring mounting 2cm out that means – even if you undo the misaligned chainring – your pedals are also 2cm outboard from where they would be only on the right hand side. Yeah thats right. Your pedals are not centered underneath you.

On many builds, this is dealt with by using crankarms that have a left arm that is offset outward. Typically by 18mm (like I said… “about” 2 cm). That centers your pedals underneath you. Thats the good news. The bad news is you can either use the Bafang stock crankarms – which are cast alloy and not particularly robust – or you buy quality forged crankarms – typically from Lekkie who pretty much owns the aftermarket for strong, high quality Bafang-compatible crankarms. So you get either very cheap kaka crankarms that may or may not survive a proper pedelec rider honking on them, or you spend a bundle (and get top quality stuff).

Figure 2 – Note the secondary gear housing tucked in behind the chainring on the left.

So, Forget Everything I Just Said

The Bullitt is one of the very rare frames that does not need any of this offset crankarm business. You will buy a Bafang BBSHD whose axle is compatible with a 68-73mm bottom bracket. Install that and there is enough axle sticking out the non drive side that there is no need to do anything further. Not only is there no need… you don’t want offset arms since the goal is to center the pedals underneath you. Straight arms with minimal Quack Factor will do that.

I used forged 175mm Shimano FC-E6000 crankarms meant for use with a Shimano Steps drive. They got quite a bit more expensive before the Pandemic ratcheted everything up. I paid less for both arms than you are going to pay for just one of them. In fact they are only sold individually these days it seems. Here is the right 175mm arm and here is the left 175mm arm. If 170mm is more your bag, I know they are made because here is a 170mm left arm. I can’t find a right one. Maybe you can if you want the shorter arms.

Still, the good news is you can use straight crankarms from any vendor… so long as they are square-taper (don’t shoot the messenger on the square taper part).

The Wire Tunnel

I have seen numerous BBSHD’d Bullitts where the wiring from the motor is run forward underneath the main … girder … or whatever that part of the frame is called, and then run up once it splits into the deck support. I don’t like this because it creates visible wiring. Also, wires directly under the bike are potentially subject to ground impacts. A risk I’d prefer to avoid.

Figure 3 – the wire tunnel. Note the velcro wrapping at the point the wire bundle enters the tube. This snugs it to the inside diameter of the tube.

I looped the wires up and over the drive side. The secondary housing naturally protects and hides them. The wires then cross over the top of that girder, hidden by the wire tunnel. My original idea was to use a bit of green or black furniture-grade PVC. However, early on in the build process I stumbled across ‘Duck’ Brand neon green duct tape. This is commonly found in hardware stores in the USA. Its a hobby-grade household product and not construction-site worthy, but its fine for this job.

By some miracle the Neon Lime Green color is nearly an identical match to the LarryVsHarry Lizzard King paint. You literally have to be looking at it just right to see any difference between the two. It was perfect to cover the dark green PVC pipe used for the wire tunnel. I did use two very large dark green zip ties to hold it in place. These were the closest color match I could find and, while they are not ideal, I have yet to find a better solution (including taping the thing onto the girder with the matching tape. I tried it and it looked awful).

Figure 4 – The wires look a lot more visible here than in person. Wires include both front and rear motor wiring as well as the battery charge plug (green cap).

The inside edges of the plastic PVC pipe are not what I would call ‘sharp’, but they are edged and a little more unforgiving than I would like when rubbing every day on flexible, soft wire casing. I beveled/chamfered the inside and outside edges using a quick pass with a pipe reamer. Problem solved before it becomes a problem.

As it stands, the wire tunnel protects the wires coming out of the motor that have to be run forward and up the steering tube, where they rise to the handlebars. You don’t even see any wires now the Fahrer bags are installed.

Figure 5 – With the Fahrer bags in place, what little visible electrical wiring that exists … disappears

Heat Sinks

Godzilla presently lives and works in Fresno California USA. Whats the weather like there, Ollie?

And once it starts getting hot, it stays hot…

When faced with this, you have to take steps on a variety of fronts to at least mitigate the issues that come up. Insofar as the BBSHD is concerned, I have already covered this subject in a fair amount of detail in a separate article that predates the Bullitt build and, yes I pretty much covered the motor in heat sinks just as you see in that article.

Since the crankarms are not offset on this bike, I was not able to use an endcap, but as shown in the linked article above I covered every bit of the motor I could with the things, using both the silver center sinks and the little black squares around the edges.

BBSHD Settings

The settings for the BBSHD are another subject already covered in detail. First in this article that introduces my approach to the subject, and following on with this one that covers some very minor refinements.

To jump specifically to what I am doing on Godzilla, you want to jump straight to Version 2 and look at the right-side image.

Here are the three screens below, but all the explanation for what they mean is in the linked articles above.

Figure 6 – The BBSHD settings used on Godzilla

Drivetrain

If we’re discussing the mid drive, we have to at least touch on the rest of the drivetrain, which is integral to making the bike go places.

SRAM GX 11-speed shifter

I prefer SRAM shifters because the way they mount onto the handlebars, they take up much less real estate than their Shimano or Microshift counterparts.

SRAM GX long cage rear derailleur

SRAM derailleurs in general, once adjusted properly, seem to stay that way. There’s not much to say about this derailleur other than it just works, precisely, smoothly and reliably, does a great job wrapping chain and has no issues with the 42T big cluster I have on the back. I don’t use its clutch feature. It is able to handle a 46T rear cluster just fine, although I don’t have one on Godzilla. I do have one on another bike set up that way with the same drivetrain bits.

Figure 7 – The long cage SRAM GX is sweet!

KMC e11 chain (mid drive-specific)

When I build a mid drive bike I use a mid-drive specific chain. That is I think part of the reason why, in all of the thousands of miles I have been riding powerful mid drive builds, I have never once broken a chain. Unfortunately the e11 is frightfully expensive. Bide your time, keep your eyes open for deals anywhere in the world (particularly in the EU) and you can find an online deal.

Sunrace CSMS7 cluster: 11-42T

For all mid drive bikes I build, I prefer steel, welded clusters that are 1-piece. Such a thing is not really available in the 11s world, but the Sunrace CSMS7 is as close as it gets. It uses steel spiders and high tensile steel cogs, which is what you want for durability with a mid drive.

Figure 8 – The Sunrace CSMS7 showing off its steel underwear

NOTE: The CSMS7 has turned into a unicorn these days. I have managed to score two NOS in auctions on Ebay. One for use on this Bullitt build, and a second recently as a backup for the inevitable day when one of mine fail (I have another on my Surly Big Fat Dummy’s street wheelset). But thats it, Nobody in traditional bicycle retail has had them in stock for months.

Given global bike supply issues, there is no telling how long it will take to fix this. I have found that a Sunrace CSMX8 appears to be an excellent performer despite its alloy spiders (one is pictured below). I use this on my mountain wheelset on my Surly Big Fat Dummy and so far it has given me no trouble. Its range is 11-46T which has proven to be no issue for the (lower end) SRAM NX derailleur that is on that bike.

Figure 9 – The Sunrace CSMX8: Lighter weight scaffolding under the hood… but still steel cogs.

Rear Wheel

The core of the rear wheel is a DT Swiss 350 Hybrid rear hub. The Hybrid line of DT’s 350 is, compared to the standard 350, beefed up in just about every way. The flanges are beefier, the cassette body is steel, the internal ratchet mechanism is a solid piece rather than being relieved for light weight, and it is a 24T engagement vs the stock 18T, which gives better response while still maintaining the super strong ratchet mechanism that makes this hub almost indestructible when paired with a powerful mid drive.

Figure 10 – The DT 350 Hybrid. The Sherman tank of hubs

The rim is a SunRingle MTX39, which has a 30mm internal width – the ’39’ in ‘MTX39’ comes from its external bead width. The review of this rim over at MTBR described it as “monumentally strong” and “impossible to bend” while noting the penalty for this strength is weight. Since weight is not a factor on a twin-motor cargo bike, I’ll take that strength any day of the week. The rim has proven itself un-dingable both here on Godzilla and on the Mongoose Envoy build I first used them on.

Figure 11 – That deep, triangular, thick-walled structure makes this puppy plenty strong.

The MTX39 is commonly available in both 32H and 36H configurations. So is the DT350 Hybrid. However, at the time of my builds I was never able to find 36H hubs, so both of my builds are 32H… and they are none the worse for this choice. In particular, a Bullitt’s construction does not put anywhere near as much weight squarely over the back wheel as do for example your garden variety mid- or longtail. That makes it easier to live without 36H. 1100 miles so far in a few months and the rear rim is still perfectly true.

In between the rim and the hub are Sapim Strong spokes with brass nipples. Once again I used the expert services of Stoic Wheels to provide cut spokes to my specs.

Front Chainring

Ordinarily, I advise a builder that you have to keep the motor spinning on a mid drive, and a 52T front chainring is way too big. You’ll bog the motor as its trying to lug itself up from a stop. This is the worst thing you can do to a mid drive, and among other things is a great way to snap your chain.

Figure 12 – 52T front Lekkie Bling Ring. As big as they come.

But Godzilla proved to be an exception. The choice of a 52T chainring (in country that is flat as a pancake) was still not initially obvious. As noted in Musical Chainrings, I have a lot of chainrings available from a variety of builds. With any new bike build, I do not expect to get it right on the first try. Unfortunately I came close here (it would have been nice not to buy another one) but in the end, no cigar.

Plan A

At first, I tried a 46T Lekkie ring pulled from the parts pile. The offset was good and planted the chain in a straight line … about four cogs inboard on the rear cluster. So with that 46T ring I was spinning a little too much with good chain line, and the bike was outrunning me. Time for …

Plan B

I switched to a 130 BCD adapter and a 48T chainring. If the 46T with strong offset was too hot. The 48T with minimal offset was too cold. Now I had a big chainring; good on the flats, but with no offset, good chainline was only on the smallest cogs in the rear. That was great once I hit cruise but bogged the motor when starting from a stop… and thats very bad. Lets try …

Plan C

My third try is a Luna Eclipse body with a 48T (proprietary) ring. So, same chainring size as Plan B but with the most offset possible in any BBSHD aftermarket chainring. This felt pretty good. I was running up in a larger rear cog so my chainline was great. That kept the motor from bogging from a standing start.

Figure 13 – Left to Right: Plan B, Plan A and Plan C

But (!) … I was not able to get down into small enough cogs when I was up to cruising speed. The bike was running away from me. The only solution was to let the bike run at slow speeds as it peaked with usable cadence and effort at about 18 mph. I lived with this for awhile – and it was ok (just ok) – until I decided to finally buy a chainring in a bid to have my cake and eat it too.

So now its come to Plan D

A 52T Lekkie chainring that has the same goodly offset as was found with Plan A. I would be biasing my chainline into the upper half of my cluster. But not so much that I would be unable to comfortably use the smallest cog (as was the case with Plan C). I figured the increased tooth count would cause me to run one gear up the cluster from where I was running with Plan C, and slow my cadence enough at that bigger rear cog to still run the bike faster, while at the same time keeping a straight-ish chainline, preserve my ability to gear down to bigger cogs and up to smaller ones.

And this time, finally, I was out of parts and had to buy the thing, using my usual source for Lekkie rings here in the USA. Ouch that one smarted a bit.

My gear calculations were done on the Speed at Cadence screen at bikecalc.com. Its an invaluable resource for the thoughtful builder working out the right gears for a build.

But the money was worth it as it completed the bike. Plan D worked perfectly. This is, as far as I can see, the only time a 52T chainring works well on a mid drive build. Worth noting is the fact that this bike lives on table-flat ground, and if there were hills in its way, I would have likely had to go with a 42T front Lekkie ring.


Thats it for the rear motor and drivetrain. Lets wrap up discussion of this build with

Bits and Pieces

BBSHD Performance Settings (slightly refined)

In BBSHD Programming for the Pedaling Cyclist, I laid out what my preferred settings were for my ebikes. I have settled upon them after a fair bit of tinkering over time. Since that was published, I have made some refinements.

I’m going to skip all the background I went into before. None of those details have changed. I’ll just note what I have done, and what it seems to accomplish.

Remember, There is no one perfect suite of settings for the Bafang BBSxx series of motors. Most likely, your perfect setup will take a bit from here and a piece from there to give you exactly what you want. Use what you see here as a basis for your own experimentation.

In fact, I don’t even have a single flavor I stick to as you will see below.

The Basic Screen

From then to now, I changed nothing on this screen, and I’m only showing it here to give you a complete settings reference in one place.

Figure 1: There actually is one change from the past but its only visual: the 30a current limit. The screen said 25a last time and that was not something I ever stuck with.

The Pedal Assist Screen

There are several changes here, all geared to toning things down and making behavior more gentle. I have two versions of this screen in play on different bikes. The first is the more general-use, but the things I get from Version 2 could be very desirable on any bike.

I can’t say enough that there is no one best choice of settings for your BBSxx motor.

Version 1

Figure 2: Before on the Left, After on the right

Pedal Sensor Type

Ignore this one unless you are in the mood to play around and see whats what. Some motors come programmed with BB-Sensor-32, others with DoubleSignal-24. This is something I just leave as-is depending on what the motor originally came with. Frankly its not a setting I know much about so I don’t fool with it.

Start Current

This has been reduced to 4 from 6. Start Current dictates how much juice goes to the PAS system when it initiates. A lower number means a lower ‘jerk’ on the drivetrain as the PAS system engages. This is a percentage number, so we are going from 6% to 4%. Restated, that is from ‘very gentle’ to ‘mouse-fart’! Still, as small of a change as is indicated by the numbers, its a noticeable decrease. Bear in mind I am riding some pretty big fat and cargo bikes and when starting a bike with a heavy load, a little less is much more.

Also don’t forget no setting is an island unto itself. These things taken together make up a whole, and its that whole that matters.

Start Degree

This has been increased from 4 all the way up to 10. Start Degree is the number of signal points that are crossed (as you turn the crankarms) before pedal assist engages. A full rotation of the crankarms equals 24 signals, so I have kicked up the pedal assist engagement from a 1/6 turn to almost a half-turn.

This comes in handy when I am at a stop light, but I stopped smart, several car lengths back from the actual edge of the intersection. I can then stay in the saddle, balanced with feet on the pedals. I slowly turn the crankarms a bit at a time to maintain my balance and crawl forward. If I do it right, the light turns green before I reach the edge of the intersection, whereupon I can hit the throttle and move forward without having to dismount (and then re-mount) the bike. If I decide not to use throttle and pedal my way out, I’ll need about a half turn to re-engage assist, and when it comes on it will be a very gentle ramp-up in part thanks to the new lesser Start Current above.

Current Decay

I have set this to maximum now, meaning the system pretty much doesn’t want to cut back power as my cadence increases; virtually disabling the cut-back feature from the Basic screen, which conserves power (if you can spin the crankarms superfast, you don’t need assist). This is a setting I will very likely continue to play with. If I set it to the minimum of 1, then the motor will aggressively reduce the provided power as cadence increases. A lower setting is much more in tune with the pedelec ‘philosophy’ which I usually stick to. So don’t be too surprised if you come back in a month and this text has been supplemented with talk of a different, much lower setting.

Why did I change it? Just playing around. I’m leaving it here to illustrate that these settings often interact with one another in some big ways.

Version 2

Version 2 is a riff on Version 1 above. Its what I use on my 2wd/awd Bullitt cargo bike.

‘Before’ on the left is the Version 1 ‘After’ screen just above in Figure 2. On the right is Version 2

Slow Start Mode

The slow start mode – the strength of the initial punch the motor gives out when it fires up – has been turned down a notch to ‘2’ from ‘3’, which was already a low number. I have noticed the bike this is used on has a noticeable, audible clunk when the chain engages the upgraded, 36T internal ratchet mechanism of the DT Swiss 350 steel cassette body. Every time I hear that clunk is a time the hub gets beat on. Even if it was already a reduced amount, I wanted to notch it down further. This is a good thing on a heavy, often-laden cargo bike, and its probably not such a terrible thing for any ebike.

Start Degree

In Version 1 I have this set so that there needs to be just under a half rotation of the cranks before the slow-start of pedal assist engages. When all of the settings are taken in together, it typically means pedal assist doesn’t begin to be felt until the bike crosses a 5 mph speed (If I want assist earlier there is always throttle).

In Version 2, I have reduced this to a 1/6 turn of the crankarms. The idea with all of these settings put together is that assist from the rear motor starts faster – Remember, on this bike the BBSHD is supplemented by a front motor that starts almost instantly from a stop, so the BBSHD is not trying to haul a bike up from zero on its own. On such a big, heavy bike as this cargo bike, some assist from both wheels asap is a good thing. Net result is that with two motors, I can get a good result from pure pedal assist without having to resort to the battery-sucking thing that is hitting the throttle. I start pedaling and the bike sedately glides forward even if its weighing 400+ lbs.

Current Decay

This is a full 180 from Version 1, where Current Decay is set to do as little as possible, meaning the motor is the least likely as it can be to pull power as your cadence increases. Here, it is at its next-most aggressive setting. In part this is informed by the fact that there is a front hub motor (running off the same common battery) that – if faced with a sudden power drop from the rear motor – will still provide forward assist, and in practice that means it will increase wattage as its workload increases. This is part 1 of 2 of this story, with Part 2 being …

Keep Current %

The amount of assist that is retained when Current Decay pulls power back has been reduced to 30%. So the two settings together make it more likely that, as cadence increases, power will be reduced, and the amount of power that will continue to flow has further been reduced by 25% from its previous (fairly low) value.

Got all that? Here’s what the rider feels: Not much. Despite my characterization of these changes as ‘aggressive’, which they are if you just look at the numbers onscreen, reality is the result is muted. We’re using a little less power, which we don’t miss much, and that is going to give us a bump in range. At 20-25 mph, pedaling strong at 60-70 rpm cadence, the BBSHD on flat ground is eating about 200-250w on the typical Assist Level 4 and 225-275w at Level 5. Thats pretty light (power consumption will be more on a bike with just the BBSHD for a motor!)

If you want to have the motor back off when you demonstrate you don’t need it (if you did you couldn’t spin your legs and the crankarms) the Keep Current and Current Decay settings are your go-to’s to making that happen and reducing your power consumption.

The Throttle Screen

Before on the left, After on the right

There is just one change here, and its along the same lines as what I tinkered with on the Pedal Assist screen

Start Current

This is the same kind of effect as Start Current on the Pedal Assist screen, only it works when applying throttle. I have halved the initial response the throttle gives down to a 5% delivery.

What’s the actual effect of that? Well, in conjunction with the broadened Start Voltage and End Voltage already in place, setting such a low value lets me feed in a constant 50-100w to the motor. Its important to understand this is not an initial value that increases: This is the initial value that will output continuously at the lowest level of throttle engagement.

Lots of times, I want to engage throttle just a little bit to slowly navigate some delicate or narrow pathway of some kind; especially where pedaling might throw my balance off (remember I could be riding a loaded 500+ lb cargo bike). This lets me do that. It also means when I engage throttle, the thunk of the cassette body engaging itself is no longer a thunk at all.

So not only is the bike more controllable at a low level, no thunk means stuff lives longer!

I guess I won’t be selling this alloy cassette body on Ebay after all (400 miles on it and its easy to see what cogs I like to use)

Thats It?

Yup thats it. At this point in the development of what I like and dislike on a BBSHD setup, I’m down to the last of the fine-tuning.

Exit mobile version
%%footer%%