My first e-bike build

BikerG

Active Member
Region
USA
City
Echo Park, Ca
Hello, from Los Angeles!
I’ve been into cycling for close to 5 decades now, road, MTB, singlespeed, & fixed gear. Love to tinker, been building bikes from the frame up for a very long time…. I’ve also been doing a lot of 1 or 2 nite bikepacking these past couple of years on my Ritchey Ascent ATB. Anyway, to make my bikepacking journey a bit easier I’m seriously thinking about building up an e-bike using my Surly Ghost Grappler with the Bafang conversion kit. I’ll be looking for thoughts and advice regarding the build. It has a 1x12 drivetrain so I’ll have concerns about feasibility of the narrow chain, chain line issues, maximum chainring size, etc…
Thanks!!
 
The important thing to look at is the Max Chainring on "frame spec" tab for the ghost grappler at www.surlybikes.com.

I have put BBSHDs on a Surly Troll, Bridge Club and Ogre. The Troll and Ogre only required a minimal spacer (0.5mm to 1.0mm) to get the BBSHD to fit without hitting the chainstay. The Bridge club required a few more (maybe 1-2mm) but I did modify the chainstay a bit for more clearance (put block of wood on chainstay where I needed clearance and wacked it with a hammer. The chainline in all cases was fine.

See the following for a template you can use to check for clearance
https://intercom.help/bafangusadirect/en/articles/4649613-bbs02-bbshd-motor-fitment-templates

If you do need alot of spacers, you might have to just not use the biggest rear cog(s) in the cassette. I have done that before. I just bought a wide range cassette(11-50) and use the limit screws to lock out the biggest rear cog.

Another option is the CYC Photon. As a lifelong cyclist you would definately prefer the torque sensing. I recently installed one and it is so much nicer to ride than my BBSHD setups. You get more chain clearance as well (see attached pics for my Ogres, there is a small clearance on the BBSHD)

I got my CYC photon here:
https://www.electrifybike.com/

A 34t chainring would definately work, 38t might need a few spacers.

I have a thread in the DIY section with lots of details on all my DIY installs

There is also the torque sensing TSDZ2 and many here praise them. I installed one and didnt like the torque response so installed the OSF firmware which made it better. In the end, it developed so much radial bearing slop after 1k miles the best place I found for it was the trash.

The CYC Photon is the nicest DIY setup I have used yet, miles above the BBSHD and TSDZ2. Torque response is really good
 

Attachments

  • IMG_20230926_121231123.jpg
    IMG_20230926_121231123.jpg
    282.1 KB · Views: 185
  • IMG_20230926_121246778.jpg
    IMG_20230926_121246778.jpg
    199.5 KB · Views: 155
Last edited:
Hearing your description of yourself, I can't help but say it sounds like we have a similar background in cycling, in terms of longevity and bike building. Myself personally so far as Surly bikes go, mine is a Big Fat Dummy that I've done quite a lot to, in addition to the BBSHD.

What @linklemming is describing above is very important. I did a series on building an ebike, and in particular how to fit a BBSHD so the motor's secondary housing has proper clearance from the chainstay root. The relevant part starts here in the Tinkering chapter, which assumes you now have a frame donor and a motor and you need to check/perfect fitment. One place to get your own custom spacers made is McMaster-Carr, and you will see a couple of bikes where I used them. You can do a wider outside diameter for a firm grip, and a SLIGHTLY narrower inside diameter so it hugs the motor better than a BB spacer with no slop.


I would be wary of the Cyc Photon. I have been using a Cyc X1 Pro gen 1 since Cyc's early days, and I have seen the reliability issues those motors seem to inevitably face. To Cyc's credit they continuously improve, and generally do right by their buyers who are affected by what seems to be a never ending beta test cycle with motors that are in production. Join the Cyc Facebook users group and just lurk for a few months. You'll get the idea. Worth noting: the Photon is a sealed motor and it has a known duty cycle where you must send it back in to a qualified dealer to have them crack it open and perform a regular service to it. When the Photon came out, I said it sounded wonderful but I wanted to give it two years on the market to see what cropped up. We are about a year in and while its been a good rollout, its definitely not perfect.

This is typical of the sort of issues you see with their motors. Not dealbreakers but PITA stuff that jumps up and bites you after you've been at it a while.


They seem to have gotten past all of their controller problems that they had that took 3 generations of the X1 Pro to put behind them.

If you want something with literally bulletproof reliability, wide aftermarket support and zillions of how-to's out there, I don't think you can beat a BBSHD. Especially on a steel Surly frame with a straight down tube and an English bottom bracket. The only thing lacking at that point is taming the pedal assist system so it allows natural riding feel, which is also do-able, but few do it and thus the motors get a bad rap for running away from their riders.

 
It is a boost 148 frame. The chain line will be too far out in my opinion. Bafang motors are large and wide. It is also designed for a small chainring so the Bafang may not fit. I predict that it won't. You would need the blue print geo of the motor and then take some careful measurements. Nice bike. A 130 frameset for a starter bike would be better with a pie pan profile offset Bling Ring.
1695755986318.jpeg
 
I think it'll work, but you're right it won't be a given. I've built successfully with worse matches. The solid chainstay root on that bike's drive side will let you tuck the motor right up to the frame, which changes things quite a bit vs. a std chainstay setup. If you can get a coffee or tomato can around the same size as the secondary housing this goes a long way to help you visualize and is what I usually use. Center it, drill a hole and fit some allthread thru it. Big washer and nut on the other side affixes it.

The big bling rings have lesser offset than the 40T and 42T and will make alignment worse. I have them in I think 46 and 52T and they are suited to the small cogs, not the big inner ones @BikerG will want with bikepacking.

I'd use a 42T Luna Eclipse if possible because of the added 3mm or so of increased offset, and the wicked tooth profile, which never drops a chain ever. But I expect that ring will have too much offset for that frame.

A Pro ring gives extra offset and has teeth optimized for 10-12s, where the standard Lekkie/Luna rings won't. The 40T needs the smaller drive cover and a little filing on the motor casing. But the 42T version fits over the standard. Since with an HD the power of the motor can let the rider use fewer gears, he may have no need/desire for the biggest cog if there is an alignment issue. But I will say if it was me I would never use 12s and instead would limit myself to 11s.

9s is preferred for availability of durable, inexpensive components, but I've found a 10s setup I love that is the cheapest of all and works like a trooper (Microshift Advent X running an 11s ebike chain).

PXL_20221020_232054714.jpg
PXL_20220531_170624673.jpg
 
I recently built a boost 12. It cannot hit first without dropping. We will take 11 and place it behind 1st as a spacer then pull in the low limit screw. It will have the same range and the owner will still have the smallest and largest cogs.

'You gott'a Ride it Ride it, keep on, Ride it. You gott'a Ride Ride it. Ride it Ride it!"
 

Attachments

  • LA Woman.jpg
    LA Woman.jpg
    43.2 KB · Views: 123
While the BBSHD is a monster of a motor, well understood with parts availability its PAS is just soulless when compared to the CYC photon.

Some like the PAS more than torque sensing, I ride solely for exercise and prefer torque sensing. Since building up the CYC bike, its my preferred ride unless I just want to go really fast or just not work that hard in which the BBSHD excels. If I was interested in no sweat commuting, the BBSHD PAS would be perfect.

In defense of the CYC photon:
Stuck chainring - yes its a concern and has been reported by a few on endless sphere. IMHO, its more of an issue with the included tool which is really meant just for installation. I will likely fab a tool this winter.
Bad Bearing - I saw that video before ordering mine and I was concerned. I had no issues with my installation (no excessive side load and bearing ran smoothly). Its really hard to know exactly why the author of the video had issues despite his analysis. So far 500 miles and no issues.

Bearings and my DIY builds
All my DIY builds have had bearing issues

GMAC build Motor - motor seized up (would barely turn) after about 4k miles. Grin fixed under warranty but still had to unlace the wheel, ship it at my own cost and then relace the wheel myself. Major Pita The bearing replacement wasnt 100% done correctly and soon needed another bearing after 2k miles or so. I bought all the tools to replace bearings and bought the best bearings I could find and didnt have any issue after that.

GMAC build Erider Torque sensing bottom bracket - I have three of these now. The stock bearings are crap and are not shielded well on the drive side. The installation instructions also put excessive side loads on the bearings causing them to fail very quickly. I finally bought the best shielded bearings I could find and used loctite and just handtightened the part that removes axial play so there was no play. No issues since

TSDZ2 - My TSDZ2 had OSF software and a brass gear. Radial bearing slop was so bad after 1k miles I figured it was best donated to the landfill. Some on this forum has stated that is was totally my fault since it wasnt stock although no explanation was given as to how my mods would cause excessive radial play. This was 2019 and I have heard the TSDZ2 is better these days even without the OSF.

BBSHD - My first BBSHD developed enough radial play after about 2k miles that I decided to replace it. Nowhere as bad as the TSDZ2 and it didnt seem to get worse but the clunking sound drove me nuts That being said, I just need to break it open at some point and replace a bearing. My current BBSHD has about 6k miles with no issues

The torque sensing of the CYC photon is the best DIY solution I have experienced so far, By far my favorite DIY bike to ride, everything just seems to be the way I want. I beat the snot out of it (for science) and it seems to be holding up like a champ. The quality of the components in the kit are top notch, a step above anything I have gotten from bafang. Just holding the BB axle in your hand, its pretty obvious.

Replacing the bearing on the CYC photon looks like a 10 minute job for me using this bearing press kit
https://www.amazon.com/Zdovivot-Bicycle-Bearing-Universal-Installation/dp/B09YCY3Q7D/ref=sr_1_1_sspa?crid=G7TUTO9582VE&keywords=bike+bearing+press&qid=1695791355&sprefix=bike+bearing+press,aps,107&sr=8-1-spons&sp_csd=d2lkZ2V0TmFtZT1zcF9hdGY&th=1

Lets not forget that Bafang had controller issues with prior mid drive setups like the BBS02 awhile back so while history is good to know, its not always the end all answer for other current products.

I was so impressed with the CYC photon, I almost bought a second unit to install on a Surly Karate Monkey MTB but decided to try a Luna X2 with M600 ludicrous motor upgrade which will arrive on thursday
 
Last edited:
I was wrong about that Boost 12 speed Marin Team 1. I do not like proprietary stuff. It has a Deore Micro-spline Hyper Glide hub and free hub body. The 12th cog screws into the 10th! 11 cannot be moved on the 10-51t cassette. The 10-t cog is smaller than the free hub. It is just like @m@Robersons bike executive video. Proprietary. No options. By the way, I have not had bearing paly in any TS-B builds. Here is a closer view of that Team 1 bike.

1695834304878.jpeg
 
While the BBSHD is a monster of a motor, well understood with parts availability its PAS is just soulless when compared to the CYC photon.

... If I was interested in no sweat commuting, the BBSHD PAS would be perfect.
Yeah but thats if you leave the BBSHD in its stock form, or maybe if you don't change its PAS settings so it lets you work hard. The BBSHD out of the box is a beast that will run away from you on pedal assist. One of my two main configs, even on full blast PAS9 the motor peaks at only around 400w. My hill-climb version puts out around 900w and of course you can tinker with it to come out somewhere in the middle of that by varying one setting, which still keeps the current decay, keep current pullback etc.

That 400w peak is on a 1500w-capable motor.

The best way to appreciate a BBSHD as a cyclist's tool is to change its behavior so its nothing like it came from the factory. You end up with a motor that lets you work yourself like a sled dog thats what you want (which I do).

I think my original assessment of the Photon needing two years in the marketplace to shake out the bugs is going to prove accurate. I still see no real solution to it being a sealed motor that needs a factory overhaul at intervals. It holds great promise but for my money its not there yet. And if they finally release the larger Proton, thats going to be one to watch as well.

At the same time, we're seeing two new motors coming to market from ToSeven that - if they pan out - will give us still more options. Jury is still out on them.


I'm not crazy about Bafang being the single source for a bulletproof mid drive.
 
Last edited:
I put a BBSO2 on a bike almost a year ago, and thought I might try a BBSHD on a twenty year-old FS mtb that I acquired for $30. But the BBSO2 proved to be more than enough power for me. I don't really need an ebike motorcycle (though I think that maybe an emotorcycle is in my future.)
So instead, I put a DD rear wheel hub motor, with torque sensing on the mtb. The mtb is much more enjoyable to ride. Maybe it's the torque sensing, or regen braking, or the DD's reliability? Both bikes are great machines that are both super quiet, and function very well. I just enjoy riding the hub motor bike better than the mid-drive, but, then again, there's that inner bicyclist in me that doesnt mind shifting down and pedaling up the steep hills.
 
I've put together two TSDZ2 and two BBS02B, The donor bikes used 6 speed and 7 speed freewheels, so I've never had chainline issues.
.
Would recommend the BBS02's for someone that wants to ride fast/hard. I know my TSZD2's are half the power and pretty wimpy. Nonetheless, the torque sensor ride of the TSDZ2 is attractive, kind of fun, and good for my cardio.
.
I haven't stood on the pedals since I was a teen, but have been doing that with my TSDZ2B on hills. Mainly to show I can still do it, Then I read hat that kind of stress can break the torque sensor. Bummer.
 
Before reading linklemming, I thought 4500 miles wearout on the gearset of my ebikeling hub motor was wimpy. 4000 miles on gmac bearing, 2000 on the next bearing? 1000 miles on TSDZ2 before bearing failure? Yeah, those ****ese bearings are ****. If they are standard size, you might buy replacement SKF made in Argentina from mscdirect.com . Did so for a DD motor although turned out bearing wasn't the problem. Loose cover screws were. DD motors should be very reliable although a pig to ride with the power turned off. My DD motor was retired at ~500 miles.
Hope my Mac12t holds up better than the gmac. 2000 miles ridden before rain burnt the wire harness connector off in 2020. (pins on ASI controller too close together for 54 v). Harness replaced 8/23, having a ball now blasting up the 12% hills with 60 lb groceries. Took me 3 years to figure out how to get the armature out of the Mac case.
 
Last edited:
1695917002160.png


This is your bike, right? Nice! I'm jealous.
My only experience with a DIY mid drive is the TSDZ2 which I doubt would work well due to chainline issues even with an aftermarket 10mm dished chainring or the direct mount 30t or 34t chainrings, which to my understanding aren't compatible with the newer "B" version anyway. And it might not look in place, quality/appearance wise, on such a nice bike. If the Photon allows for a normal chainline that might be the only current good option. Even if the Photon is "sealed" and requires "factory overhaul" it's hard to imagine that motivated DIYers won't figure that out themselves. Good luck with your project.

addendum:

something that you've already considered but depending on your setup carrying a battery will cut into your available cargo space.
Here's one of my bikes when I had a TSDZ2 (with a 10mm dished chainring) on it along with a 48v 15ah battery in a top tube bag. The battery is heavy and takes up a lot of space.
1695919268441.png
 
Last edited:
Update on the 12 speed Microspline: Cogs 1-4 need to connect and 9-12, but the others are free floating in the stack. We can remove one in the middle and stack 5.25 - 5.5mm worth of spacers behind first. I am almost finished making a narrow/wide 38 for that bike that should better hold the chain.
 
Thank you so much for all the suggestions and ideas!! I’m as confused as heck now.. LOL…
Seems like the XT micro spline hub with its 12-speed drivetrain is the major sticking point. Perhaps if I swapped my drivetrain to an HG hub, running 9, 10, or 11 speed might do the trick, but i still have to deal with the 148mm spacing.
Thanks again, for all your expert input! This is going to be quite a learning process for me.
 
Thank you so much for all the suggestions and ideas!! I’m as confused as heck now.. LOL…
Seems like the XT micro spline hub with its 12-speed drivetrain is the major sticking point. Perhaps if I swapped my drivetrain to an HG hub, running 9, 10, or 11 speed might do the trick, but i still have to deal with the 148mm spacing.
Thanks again, for all your expert input! This is going to be quite a learning process for me.
We will let you know how it goes-- it's a learning process for us, too. I was considering a drivetrain swap as well, someday might do that, though now I'm not sure I will need to.

As we consider the chainline issue for the TSDZ2B, which is mechanically better than the previous Tongshen, but has the crank limitations @EMGX identifies (and that Pedal and I knew we'd encounter as well) the issues are really only evident in 1st, 2nd, and (only vaguely) 3rd gear, and that is mainly a problem for moderate-to-long steep climbs. In this one specific situation on the Marin, it's a particularly important problem to solve, because this particular motor feels like it's 60-75nm of torque, (meaning not that powerful) under those conditions, though I like the torque sensor generally. A lot.

I think we automatically think in terms of a 34t chainring or less, but consider a typical configuration:

34t chainring + 42t low gear @ cadence of 70 = 5.6 MPH, which is totally workable for extended moderate-to-steep climbs.

However, sticking with the stock Tongshen 42t chainring option , which keeps the chainring as close to the bottom bracket as possible:

42t chainring + 51t low gear @ cadence of 70 = 5.7 MPH, which is basically the same!

Hopefully, removing one (or two) gears from the center cluster, where the individual cogs are replaceable, will accomplish this (with the necessary offset.)

@PedalUma do you think that we could actually remove TWO (non-consecutive) sprockets? It all depends on how each of the three cluster elements are attached to each other, but we're pretty confident we can at least remove one, which should give us an 8+ mm offset. On paper, this should put first gear, that now-critical 51t, about where 2nd gear used to be: Totally usable, still some friction, hopefully *might* reduce friction and clatter a bit more.

The only other issue is this weird issue with perceived effort: 42t+51t and 34t+42t may be the same ratios on paper, but many cyclists report that faced with these two options, the combination with the smaller chainring just feels like it's easier or more efficient. Is this some hidden physics or engineering thing, or an issue with familiarity, where it's just a change in technique that's required? I think part of it is or might be familiarity, but that's the thing: You really don't know until you try it. There's no reliable way to game out all the variables-- including the torque sensor on the TSDZ2B, which I mostly really like because it's so efficient and it does have grunt for those brief nasty sections.

(Also, this is a 29'er, so YMMV in terms of the numbers.)
I haven't stood on the pedals since I was a teen, but have been doing that with my TSDZ2B on hills. Mainly to show I can still do it, Then I read hat that kind of stress can break the torque sensor. Bummer.

Doing this recreationally and indiscriminately for bravado, yeah, that's probably not great for the hardware, but-- doing it strategically for brief and brutal segments in granny gear, I expect, would be less stress on the motor, you're asking less of it. But the motor can really deliver in that situation, feels like an eMTB motor.

Anyway, the Marin Build (I'm still zeroing in on a name) kicks ass around town, and is very usable in the Hollywood Hills, even before changing anything. However, the changes will hopefully make it more comfortable, and quieter in the Hills, and getting that 51t online should mean we can go steeper than that, like the Verdugos.
 
I originally purchased my Surly Grappler last year as frameset only from my LBS. I built it up with a mix of some new, and mostly with parts I already had on hand. Currently, the bike is back to its bare frame so I can rebuild from scratch at this point.
The rear dropout is actually 145mm, meaning I can go either 142 or 148 boost. It’s also 12mm thru axle. I am assuming going with a 142 rear hub would be ideal for a mid drive setup. I also have spare 8sp and 11sp mtb shifters, cassettes, and derailleurs on hand already.
Alternately, is there a rear motor hub set up for 142x12 thru axle? I don’t mind going that route as well…
This will strictly be a bikepacking rig to help me up some of the steep fire road ascents in my local San Gabriel Mountains.
 
Why not build a throw-away bike first? I did my last TSDZ2B for under six hunded bucks, including a $50 beater Giant bike. The motor was $250 and I used an inexpensive UPP battery. That way, you don't carve up a good bike, although most conversions are reversible,
 
Why not build a throw-away bike first? I did my last TSDZ2B for under six hunded bucks, including a $50 beater Giant bike. The motor was $250 and I used an inexpensive UPP battery. That way, you don't carve up a good bike, although most conversions are reversible,
I agree.
My only ‘non-use’ mtb is a late 90’s Dean ti Colonel, hardtail, rigid, which would almost be a perfect candidate, except for that it is non-disc (I feel much safer with discs on a loaded rig) and that it has no rack braze-ons.
 
Why not build a throw-away bike first? I did my last TSDZ2B for under six hunded bucks, including a $50 beater Giant bike. The motor was $250 and I used an inexpensive UPP battery. That way, you don't carve up a good bike, although most conversions are reversible,
I've installed and removed TSDZ2 mid drives several times on several bikes. Other than cutting off a non-essential support bracket mounting point on the bottom bracket of a folding bike every bike was easily and quickly reverted to stock. For someone with obvious skills like the OP it takes maybe an hour or less for a complete installation or return to stock. Potentially more if cables have to be re-routed or fiddling with other components (like addressing chainline/chainring issues).
Elephant in the room (or on the bike) is 11+# of additional weight that the motor and (a small) battery add, more for heavier motors like the bafang and larger capacity batteries. Consider if you really want to change a 30# bike into a 41+# bike.
 
Back