A bike mechanics cost-effective DIY utility build

hexadrome

10 mW
Joined
Jan 8, 2023
Messages
24
Hello folks,
I'll document my build here. I'll also show possible ways to hack together a working bicycle. What I don't want is a bike mechanic pissing contest on proper or improper ways of building a bike, I'll just show what has worked for me in the past or what I witnessed myself as an inspiration to someone not willing to invest in specialty tools.

The starting point is this '99 Giant Box One freeride frame:
b1.jpg

The Frame
This one is a little bit of mountain biking history and it's precious to me because of it. You can find all information on the frame here: http://junglebiscuit.com/articlesgiantboxone/
I haven't been able to find more.
It is a Unified Rear Triangle design. In short, if you pedal out-of-saddle the suspension bobs. Interestingly, the original URT design featured a much higher pivot point which didn't have that problem. There's a good article on PB about it: https://www.pinkbike.com/news/the-short-turbulent-life-of-urt-suspension-mtb-history.html
It has 4" of rear travel. The frame has to be rated for triple clamp forks since it originally came with one.
I don't like to pedal out-of-saddle anyway, so the limitations of URT are not a concern to me.
Currently it runs on a Manitou Radium R air shock. I also got a decent enough coil shock at my workshop if the little Manitou won't handle as high pressures as I'll be needing.

Background
I work as a bicycle mechanic and live in a big city in Europe. Over the years most of my friends moved to the countryside and it's getting harder and harder to visit them. Using public transport to travel to their backwater villages takes a lot of time and is quite expensive. It's not practical for me to have a car. The city I live in is historical and not planned, traffic realities are such that half of your driving time is spent looking for a parking place and you need to take three right turns if you want to go left. I'm not exaggerating. In short, I have a strong usecase for a utility e-bike.

Build Goals
In no particular order:
  • Cost-effectiveness and durability - I want to invest once upfront and only have to replace wearing parts. On my bicycles I'm okay with breakage because Every Broken Part Is A Weakness Leaving The Bike, but this needs to be strong and reasonably cheap.
  • Using what I got - I already have 36V batteries, so I need to use either this system voltage or 72V to keep cost down.
  • Range - my farthest friend lives 80 kilometres away. Charging en route is not an option.
  • Stealth - without going into too much detail, even using a ready-made conversion kit to DIY an Electrically Power Assisted Cycle is a legal grey area in my country. I'm in need of transport, not trouble. I also don't want a law discussion here, I've made my choices and I'm willing to live with the consequences.
  • Cruising speed of 30-35km/h, top speed of 40km/h. I simply don't want to go faster in traffic.
  • Final weight of around 40kg or less would be nice. I need to be able to get that thing up and down two sets of stairs.

So, without further ado, let's install the forks today!
 
The bike looks ok to me. The air shock. if it holds air, will be fine, you are not going off-road right? You will need flat-resistant tires and tubes.

Build goals;
1) NA
2) 48V or 52V are the voltages you want. 72V will raise to cost as you will have to use heavier wire, connectors, worry about pre-sparking, etc. Besides, 72 Volts will have a higher top speed than you want for the motor speed ranges you want to work with. If you use a mini rear geared hub motor, which I recommend, you will have 3 speed range choices, low-speed = 201 rpm @ 36V, mid-speed = 250 to 260 rpm @ 36 Volts and high-speed, 328 rpm @ 36V. Forget the high speed as it's inefficient in a 26" whl. Speeds are
Low-speed (201) 36 V = 18 mph, 48 V = 20 to 21 mph, 52 V = 22 to 23 mph.
mid-speed (260) 36 V = 20 mph, 48 V =22 to 23 mph, 52 V - 23 to 24 mph.
So you are not going to get much speed from a sm. hub motor on 36 Volts. Maybe a direct drive or mid-drive is what you need to run on 36 Volts, but a DD on 36 Volts will not be as efficient as a geared notor on 52 Volts.
3)80 Km's, that's like 50 miles, a long way. I'm thinking you would need at least 30 Ah's of battery. I used to carry 25 Ah's of LiPoly and you can really feel it, and Lipoly is a third less heavy/bulky than the batt.s you intend to use. If you used LiPoly, which comes in bricks, it would adapt nicely to mounting on that square head-tube. I guess you could build a big box to house 30 Ah's of round-cells, but that wouldn't be very stealthy
3)Speaking of stealth, the best thing one can do is start off with a black bike so all the cables blend in.
I know you want to use some of the stuff you have, that only makes cents, but that doesn't make them the right parts for building a bike that will come close to your goals.
It would be a difficult task even if one was buying everything new. My bike, a bike much like the bike above and listed below always gets 1 1/2 miles per Ah (I know, that's not really the right way to measure usage, but it's convinent). Given how far to want to ride one way, you certainly have your work cut out for you.
 
How to install a headset with a brick

I'll show fitting a threadless headset, adapt for threaded as is needed.

The old headset would be fine and plenty strong. I like to overbuild. I got a cup each of two oldschool downhill headsets so that's what I'll be installing. For shits and giggles I'll try to avoid using the proper tools.

new_headset.jpg
Bottom half of a FSA The Pig and top half of a Tioga Alchemy.
Notice the size difference to the old headset. Most importantly the cup flanges are longer which makes for a larger area of force transfer from fork to frame. In a bike headset, the lower cup gets all of the shock loads from riding. This can result in pitting of the lower bearing races over time. If you have a headset with ball cages, you can strengthen it a bit by using loose balls in the lower cup and putting in as many as fit. Sometimes you can extend the lifespan of an already pitted headset this way and make the steering index less noticably. The ball cage leaves regular indentations but some of the loose balls will run in-between on the not-indented part of the bearing race. Tribological concerns on using loose balls in a ball bearing do not apply here because a bike headset rarely sees a full rotation, it's more about transmitting loads to the frame.
There's a downside however:
loose_ball_headset.jpg
I forgot I used loose balls on the lower cup. As you can see, they are a mess to service. Keep a magnet ready.

Now we'll need to remove the old headset. This entails driving out the cup flanges from the frame. There are some possibilities:
cup_removal_tool-handlebar-screwdriver.jpg
On the left is the professional option. It's called a headset cup removal tool, they come in different sizes depending on the headset diameter. You can also use any suitable pipe with a square edge, old straight alu handlebars work well enough and leave no marks on steel cups. The worst tool would be a large screwdriver. With a pipe, the force gets distributed over a wider area. The screwdriver tends to leave indentations easily.
Regardless of the tool, it's easiest to drive out the cups if you brace the frame against a suitable surface.
brace_for_cup_removal.jpg

With the professional tool just a few whacks on top would be needed. Without, it takes longer. Work your way around the flange using the biggest hammer you got. Sharp precise blows. Beware if you've got a headset with thin flanges. They can get damaged using DIY tooling. On high-end headsets the flanges can be so thin that even the professional tool struggles to get purchase.

I whacked them out using the handlebar.

Now we need to install the new cups. These all can be used:
headset_press_options.jpg
The professional tool is called a headset press. It is designed to guide and press in both cups at once. It is a fancy piece of allthread but the guiding part is the key. Without that it's easy to get the cup in crooked. Good tools have long press plungers.

What we'll now be doing is called an pressfit or interference fit. Basically the cups are held in the frame with friction. Some mechanics grease the cups before installation, some don't. Realistically speaking, greasing is irrelevant from a technical standpoint due to the comparatively low loads the cups experience. Usually I don't grease, my reasoning being "what goes in easier also comes out easier", but this time I found grease on the cup flanges and head tube so I'm not even consistent myself. Do what you prefer.

Now here's the important part that decides how cheaply you can get away with:
From a machinists perspective, bicycle tolerances are shit.

Your combination of head tube and headset might have a tight fit or a loose fit randomly. I've seen everything, from cups falling out of frames after install to cups bursting head tubes.
What that means is, ideally it will take some force to push the cups in. The amount of force needed depends on your tolerances.

To install the cups without damaging them we'll need to consider the force transfer. First of all, the bearings need to be removed since putting a large static load on the balls tends to result in pitting. Headset presses can operate somewhere in the low one-digit ton range. That's a lot of force for the small contact patch of the bearing balls.

The FSA uses a large industrial bearing. It runs smoothly but it's stuck inside the cup. I can't get it out and in this case, I don't need to. I can use either the sidewalls of the beefy cup or the outer bearing race to press it in. Only press on the sidewalls if your headset is sturdy enough.

There are two main ways to go if you don't want to buy the professional press tool. You can either press or hammer in the cups. Pressing it in is theoretically better, it's called a pressfit after all and the force gets transmitted gradually that way. In practice, it's a bicycle not a centrifuge.

The key is to be patient and do one cup at a time. I'll start with the lower one and try the cheap 8mm allthread. For comparison, I've got a selection of allthread next to the press tool.
The 8mm one is the thinnest I've successfully used to get cups in. Smaller threads than this don't work in my experience. It is the size to try for a one-time use with aluminium cups. Steel cups take more force since they deform less. It takes less force to install cups into an aluminium head tube than a steel one.
Next is 12mm. This is a good size to get if you want to build a reasonable DIY version on the cheap. The professional tool has 16mm trapezoidal threads.
On the right is 20mm trapezoidal allthread I used for my DIY press. Total overkill but I had it laying around. That thing is in the last corner of my workshop currently so I can't be bothered to make a picture.

Large washers can get relatively expensive. Let's see if I can forego them using leftover soft wood and small washers. I do use soft wood for 1.5" cup installation in the shop. We get that size so rarely it doesn't make sense to get the proper press plungers.

8mm_allthread_woodblocks.jpg

The allthread is barely long enough. I got the cup started carefully. It will want to go in crooked so be patient and go slow. Because there's no guide, you'll need to reposition the allthread as is necessary. If the cup starts to get crooked, you can use a piece of wood to tap the cup out a little. If it gets too crooked, start over.

In my case, I got a very tight fit with a long flange. The force it takes to press that in is enormous. The soft wood was... too soft. The small washers spread the forces well enough, but upon reaching a certain point the wood started to compress. I escalated to hard wood. This got the cup in a bit deeper but also ultimately started to compress the wood.

allthread_large_washers.jpg

Next I tried the large washers. They will bend under the pressure, even if you use multiple.
To build a lasting DIY headset press using washers you need to make them work as a single unit. For mine, I welded four thick ones together. That did the trick.

To press on the outer bearing race I found a BBR60/BB9000 bottom bracket tool to be a fitting adapter.

The 8mm allthread was not up to the task this time. On most other head tube/cup combinations it would have worked fine for at least a single use provided the fit isn't too tight. I've no 12mm of suitable length, so I used the proper press instead.

press_on_outer_race.jpg

Even the professional tools have limitations. The Cyclus Tools one is too short to press both cups in simultaneously on >60cm frame head tubes for example. This time the end piece had a better suited diameter than the press plunger.

I wasn't able to get this very cup in using the cheapest 8mm allthread. 12mm would've worked.
Next up is the upper cup installation. And for that, let's try the other method of hammering it in.

Using a brick. Yup.

Now the key to this method is proper force transfer. You need your blows to be precise and on target. You also want the force to not dissipate, so prop the frame up well until it rest planely on the other end.

proper_prop.jpg
flat_lower_cup.jpg

Now we need to start the cup. It is foolish to use the brick now as that would give us little control. Using something lighter, carefully start the cup with light taps until you got the cup square and deep enough.

tappy_tap_tap.jpg

Now it's brick time. Or essentially any suitable weight such as a hammer starting with 1kg or upwards. The smaller ones do work but not as well. We need to deliver a lot of force precisely. Use both hands to deliver hard and square blows. If you got the cup started well, it won't skew and seating fully should take two to three hits.

brick_installation.jpg

Now it's time to prepare the fork. I could skip the next step as I don't need the old crown race, but I'll show you regardless. Some forks have a screwdriver slot. This is easymode for getting off the crown race.
View attachment crown_race_slot.jpg

Let's pretend we're not so lucky and the fit is so tight we can't get even our smallest screwdriver in. What you need now is a knife with a thin blade and a sturdy enough back. It will suffer from the procedure. Make sure the blade is in the gap between crown race and fork, this is hard to feel sometimes.

sturdy_knife.jpg

Tapping lightly with a small hammer work your way around the gap. Place the blade, give it a tap, pull off square and repeat. Do not lever, the blade will break off. Rather use it as a thin wedge to open up a slot wide enough for a small screw driver and go from there. Given enough patience, this method never failed me.

But now that we got the crown race off, we'll need to put it on the new fork. This is straightforward enough, given that the proper tool is a glorified pipe called a crown race setting tool. I have these options:
crown_race_setters.jpg

Any suitable diameter pipe with a clean cutoff will do the job. You can chamfer the inside if you need to. It needs to have a tight fit around the forks steerer to not damage the bearing surface of the crown race. Instead we'll be pressing on the inside edge of the race. For 1" steerers, an old 30.9mm seat post is a good fit. For 1 1/8" you'll need to look around. Yes, I did use the vaccuum cleaner's tube successfully to fit a crown race. I didn't expect it but it even survived with little damage.

But again, tolerances are the issue.

First of all, crown races do come in different diameters. Buy the right size suitable for your headset and you'll have little trouble. The crown race basically only needs to have the right bearing race chamfer for the given cup used and fit the fork steerer well enough. You'd be surprised with what you can get away with depending on your tolerances. I've seen crown races so loose you could turn them on the steerer by hand(!) work just fine long-term. I've seen precisely machined headsets bind on high-end frames because of mating surfaces with crap tolerances. As long as it doesn't rattle it'll probably somehow work. Just don't expect your bearings to hold up well with a badly fitting crown race.

good_fork_bracing.jpg

Let's first take a look at a good crown race installation setup. This entails the single step of propping up the fork crown on something suitable so that the force won't be transmitted through the dropouts. If you set the fork on the dropouts instead they can bend. I've seen it happen. On suspension forks either your hammering won't be as effective, you could damage the fork internals with that high a shock load or... you can bend the dropouts as well :) Did it myself.
It's more effective to hit the setting tool with something heavy than to hammer it up and down the fork steerer regardless of how heavy your setting tool may be.

But sometimes you have no choice but to adapt the given race to an existing fork and there are a number of ways to do that if your fit is just that smidge off.

First case: crown race too tight. If it gets stuck on the top third of the slightly conical steerer seat, you'll be better off looking for a new one in my experience. Else, slowly and concentrically filing the steerer seat away starting at the base with frequent checks in-between will get you to a point where you'll be able to use the crown race setter successfully.

You can try differential heating, but only to a point. With too much heat you'll anneal the race, so red-hot is out of the question. With too little heat the seat won't expand enough. You should stop heating when the grease on the race starts smoking. Regardless, should you want to go this route be aware that the race will cool rapidly once in contact with the steerer. You have a couple of seconds to land your blows at best - bigger hammer helps - but if the race gets stuck halfway before seating it'll be a bitch to get it off again to restart.

Second case: crown race too loose. Much easier to deal with and a wider range of options. If it's only a little too loose, use a center punch to hammer a row of marks around the race seat. The punch drives up the material around it just a little. Ideally, it'll fit tight enough, whack it on with the setter and you'll be done.
If the fit is too loose for that, I've successfully used coke can shims. It's very fiddly. My technique is cutting so wide a shim that it does not overlap when wrapped around the steerer tube. It needs to be at least double the height of the crown race. Wrap the shim around the crown race seat and put the race over it. The steerer is conical at the base so it should start to seat. Now you'll need to hammer down the crown race carefully so that the shim stays in place and doesn't crumple up at the crown base. Once successful the overlap can be cut off using a sharp knife.
There is also the irreversible way... which is to epoxy that sucker on and start working on the cockpit. Been there, done that to a fork of mine as a last resort. Works brilliantly as long as the race is not worn :D

HCR_tool_vs_socket.jpg

If you own a socket set, you may be tempted to use a fitting one as either a press plunger or a punch. 12mm allthread fits through a 1/2" socket just fine and the socket shape lends itself well enough... provided you don't try to use the conical shoulder. It's a bitch to get that out once it wedged itself and all it takes is a bit of inattention. Guess how I know.
Be aware that the sockets faces don't have as large an area compared to a washer or the press plunger. That translates to leaving marks more easily or indentating the cups in the worst case.

Regardless of the methods used I ended up here:
finished_installation.jpg

Using anything other than proper tooling on carbon parts is bad taste in my opinion.
 
motomech said:
The bike looks ok to me. The air shock. if it holds air, will be fine, you are not going off-road right? You will need flat-resistant tires and tubes.

I prefer to set up suspension on my bikes through trial and error. I run the shock at 110 psi and it feels perfect. I finally found the right year manual. The maximum pressure is 175 psi and it recommends sag settings for downhill use. The shock does hold air since the last time I set it up was one or two years ago. It will be fine.

Many years ago I salvaged a 24V EPAC. It had a front hub motor, a 250W controller and a defective battery. I tried hooking it up to car batteries instead. Acceleration from stop was a bit slow. It had a top speed of only 18km/h (11mph) on 28" tyres. I was sure I did everything correctly, so I was really disappointed by the performance.

More serious EV experiences started only because I was able to get a couple broken e-scooters by chance. While I may be able to grasp mechanical concepts well, I can't get my head around electricity. It's frocking magic to me. I could risk damaging the parts since I didn't depend on them for anything and I knew the system did run previously. The parts inside didn't have labels, but were hooked up already so I was able to work with that. I would not have been able to wire something from scratch.
I had access to a welder at the time and managed to fry together a beer crate trike.
kösti.jpg
Now the scooter controllers were tiny. I also wasn't able to get them to work and had no clue as to why. I managed to afford a 50€ generic 36V controller and tried mimicking the existing connections. It was hard to wrap my head around wiring at first. I didn't bother to label the wires and make an electrical diagram.
xinaoma.jpg
I didn't have any sense of magnitudes. I'm still really struggling with that. The little trike actually ran without overheating, top speed felt like 25km/h (15mph) on 10" tyres. It was fun to drive so close to the ground. It was exciting driving a motor vehicle. I got away so far with not having a drivers license by living in a big city so it was the first time owning anything motorized. I rode it around in the park and on bicycle lanes. Peoples reactions were mostly positive but I felt like I may annoy other bicyclists. I was slower than most riders but took up much of the lane.

The frame was made from mild steel scrap and weighed about 40 kilograms. It wasn't able to manage steeper hills, it simply bogged down due to low torque immediately. That saved it from overheating. I wanted more torque and the other rear wheel wasn't hooked up yet.

I looked into it and realized running the two rear wheels with one controller wouldn't be possible due to electrical interference. I didn't try it since I wanted to avoid frying the controller and wasn't able to afford another. But it could work using two identical controllers on the same throttle. I wanted to try that to have double the torque.

I scrounged up the money a couple of months later and bought the same item from the same vendor. They sent me a controller with the same listed specs, but it was another manufacturer's model. It didn't have the same color or amount of wires. The wiring diagram on the vendor side was the old one from the XINAOMA. I contacted the vendor who told me they ran out of XINAOMAs and promised to update the diagram. They didn't.
I tried hooking it up using careful guesswork. Wire colors were off but I made assumptions based on wire gauge and connectors. I had trouble understanding the simple wiring of the trike I had built a few months back by then. Labelling the wires at least would have been a good idea. I managed to reverse-engineer the connections and hooked up the new controller to the other rear motor.

Unfortunately I ran into the old adage of You Get What You Pay For. On powerup the new controller let out the magic smoke. I disassembled it and I found put the PCB was put in crooked from the factory and the traces had shorted on the case. I had cut off the plugs to wire directly using lineman's splices and solder so I wasn't able to get a refund. I lost interest in EV after that due to monetary constrains.

Now I still had the components laying around. A couple of years later I got the idea to lace the motor to a 26" rim. I tried understanding the theoretical interrelations beforehand but it got over my head quickly. I didn't want to put in the effort only to have the e-bike run as slow as that first 24V EPAC. Somehow I was able to conclude that it would work fine enough in practice.

So I built it reaching a final weight of around 25kg/50 pounds and was pleasantly surprised at the capabilities. 34km/h (21mph) top speed seemed plenty fast and I offset the shit acceleration with patience. Over time I got more and more dissatisfied with the performance. I realized that in order to participate in city traffic safely, I needed to be able to stop and accelerate quickly. Rim brakes were strong enough for my use and the e-bike handled like a normal bicycle.

But the acceleration again depended on frocking magic and I wasn't able to understand key concepts well enough on my own. I wanted to avoid opening the hub shell at all so I didn't have any real specs to go on.
I was able to figure out the missing links with the help of ES.
https://endless-sphere.com/forums/viewtopic.php?f=2&t=118817

motomech said:
3)Speaking of stealth, the best thing one can do is start off with a black bike so all the cables blend in.

I will need to further clarify what my goals are regarding stealth. At the moment I'm not sure what that will entail. Basically, I don't want to run into trouble with the police because I'm riding an e-bike. These are tightly regulated here.
ebike.jpg
I've been riding this thing for a year. I didn't get any police attention even when stopping besides them at intersections. They key seems to be 1. riding in a safe and civil manner. I always put on a full face and a reflective jacket. The top speed is moderate and I'm terrified of crashing. And 2. having a clean looking build. A mess of cables make you look like a tinkerer to the public, but police is interested in keeping public safety. I had accidental shorts on other projects where I didn't keep the wiring clean, so I'll try and avoid that.

motomech said:
2) 48V or 52V are the voltages you want. 72V will raise to cost as you will have to use heavier wire, connectors, worry about pre-sparking, etc. Besides, 72 Volts will have a higher top speed than you want for the motor speed ranges you want to work with. If you use a mini rear geared hub motor, which I recommend, you will have 3 speed range choices, low-speed = 201 rpm @ 36V, mid-speed = 250 to 260 rpm @ 36 Volts and high-speed, 328 rpm @ 36V. Forget the high speed as it's inefficient in a 26" whl. Speeds are
Low-speed (201) 36 V = 18 mph, 48 V = 20 to 21 mph, 52 V = 22 to 23 mph.
mid-speed (260) 36 V = 20 mph, 48 V =22 to 23 mph, 52 V - 23 to 24 mph.
So you are not going to get much speed from a sm. hub motor on 36 Volts. Maybe a direct drive or mid-drive is what you need to run on 36 Volts, but a DD on 36 Volts will not be as efficient as a geared notor on 52 Volts.

I think you may have overlooked I was talking about km/h. In mph I'd like a top speed of 20-21. I'm happy our calculations match.

motomech said:
3)80 Km's, that's like 50 miles, a long way. I'm thinking you would need at least 30 Ah's of battery.
I got the same using the Grin simulator. Awesome!

motomech said:
I know you want to use some of the stuff you have, that only makes cents, but that doesn't make them the right parts for building a bike that will come close to your goals.

Exactly my point. Through the previous builds I was able to gauge my goals accurately.
Modifying the existing e-bike further felt like trying to get a cow to ice-skate.
I ran into a wall trying to avoid opening hub motors. I was afraid of either hurting myself due to the strong magnets or damaging a motor by using basic tools.

My budget is extremely limited but floating. I will spent the needed amount to reach my goal but try to get away under 1000€. A good quality three arm puller still is a big investment to me. I'm buying parts over time so I know exactly what to get beforehand. Even so, goals may change suddenly.

Since I started with 36V batteries, it made sense to either stay on that nominal voltage or go to 72V. I was too hasty and tried using Grins simulator without a solid enough foundation of theoretical knowledge. Somehow I ended up convinced that I'd need a 72V system voltage for my usecase.
Later I understood that at the speeds I plan to cruise a 36/48V system voltage is best suited.

This mistake resulted in spending needlessly for a 48V-72V controller that will not work out-of-the-box.
Luckily my intuition is right and I should be able to modify it to run at 36V system voltage. Again, thanks to ES for confirmation:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=118795

So now I'm confident the build will work.
electricals.jpg

Batteries:
36V - 15,9 Ah - 572W - Max. charging current 6A
Motor:
Bafang G040: Rear hub geared drive - 48V - 500W - 60Nm
Controllers:
The bigger one is generic chinese.
48-72V - 60A - 3000W
The smaller one is a Kelly KLS4812S.
24-48V - 50A continuous - 120A peak

Funnily enough I bought the generic one second-hand from a guy who made the same mistake. He also realized he wanted a different system voltage after buying it. Having a back-up controller makes me more confident so even though I don't need it that's money well spent in my opinion.

The size differences took me by surprise. From my experience the wattage ratings of cheap chinese electronics are almost always on the high side. For a good safety margin, I divide all ratings by 2.
The generic one came only with an image of the wiring.
The branded one has a manual, is programmable and you get the fitting connectors as well. I also can contact their support. From this I conclude they have a working QA and care about their product.

I made sure to find good offers. I don't have Paypal or Amazon accounts nor a credit card. Thus I won't be able to buy the cheapest offer worldwide, so I'll try to find European vendors accepting bank transfers.

I paid 150€ for the Kelly, 90€ for the generic one and 200€ for the motor. I have a backup controller (needs modding) now. But if I damage the motor I won't be able to afford another one. I will disassemble the generic controller before using it to make sure the PCB isn't shorted from factory. I plan on successfully wiring and programming the Kelly correctly on the first attempt. I will need to be careful.
 
hexadrome said:
Basically, I don't want to run into trouble with the police because I'm riding an e-bike. These are tightly regulated here.
ebike.jpg

At least here in California that lack of pedals would get you a lot of attention from police. Throttle e-bikes have lower speed limits compared to PAS e-bikes, so it's always good to have pedals here, even if your gearing and speed means you are just ghost pedaling to keep the popo off your back.
 
G040 is a nice motor, I have a G020 (SWX02) and it's the best Sm geared hub motor I've used (the MXUS was good too).
It's important to get the right wind (speed range) motor. If they say it's 48 V rated motor and does 40 Kh, it's probably a 13 Turn, especially if it's rated @ 325 rpm @ 48V. That would be about 280 to 290 rpm @ 36V and that's the highest you would want to go. I had one on 52 V and it would do 26 - 27 mph, too fast for me. I swapped the motor core for an 11 turn and now it does tops out @ 24 mph (but climbs well for a mini).
You want a 13 turn so you will see 20 - 21 mph on your 36 Volts.
But be careful with the Amps., with a higher speed/smaller motor, if you let it lug down (easy to do), the system will try and feed it all the Amps it can and that's when wires start melting. I don't think you will want to make any more than 25 Amp.s (max) available.
19 Ah's of battery will be around 25 Miles (40 Km's) of range ( w/ little to no reserve).
I use two torque arms,m even on sm. hub motor installs, cheap insurance.
 
hexadrome said:
Build Goals
In no particular order:
  • Cost-effectiveness and durability - I want to invest once upfront and only have to replace wearing parts. On my bicycles I'm okay with breakage because Every Broken Part Is A Weakness Leaving The Bike, but this needs to be strong and reasonably cheap.


  • Why would you want to build up an aluminum frame that's already decades old?

    Your own link shows what will happen very soon. :mrgreen:

    articlesgiantboxonefault2.jpg
 
I can gauge the scale of this build a bit better now. I'm no photographer and my smartphone camera is not the best. Taking decent enough pictures takes a long time. Plus I need to scale them better, 3MB/image is too much. I'll upload the image data to my own server, as well as organize and resize it. It'll take months/years to document, and I need to build it first, then take my time with proper documentation. I need the transport as soon as I can get it. I'm in for the long haul.

motomech said:
It's important to get the right wind (speed range) motor.
I agree. While it would be better to get all the part specs beforehand to put into Grins simulator, it wasn't possible for me. I simulated with the nearest presets and expect the final build to work "well enough" at around 21 mph.

https://ebikes.ca/tools/simulator.h...0&mass=120&motor=MG60_500&black=load&axis=mph

motomech said:
I don't think you will want to make any more than 25 Amp.s (max) available.
I will limit the controller. To what values I don't yet know. I simulated 100A max throughput and do scale the wiring/fuses/switches accordingly.

motomech said:
19 Ah's of battery will be around 25 Miles (40 Km's) of range ( w/ little to no reserve).
I got 31.8Ah's at hand and simulator says 45 miles at 26.5Ah's. That's motor-only. I don't think range will be an issue. I plan on accelerating quickly to optimum cuise speed and will build my own cassette to better reflect the e-bike's needs.

I like this gearing calculator. Used it to set up a 7-speed system that'll probably work well enough in real life:
http://ritzelrechner.de/?GR=DERS&KB=38&RZ=11,13,13,13,15,22,34&UF=2050&TF=90&SL=2.6&UN=MPH&DV=speed
It looks like a 5-speed because I'll assemble multiple 13-tooth cogs or the like. Together with an old (but not worn) bushed roller chain I know I have somewhere, the drive train will be near-indestructible for a couple of years.

Comrade said:
Why would you want to build up an aluminum frame that's already decades old?

Your own link shows what will happen very soon. :mrgreen:

In my experience, old frames with known faults like this one (or the 70's German Mars bicycles or all those Stevens frames I saw with a ripped-off downtube near the headtube etc.) that still didn't break are safe to trust.
Because logically, all the bad frames went extinct over the years and only the strongest & most well-built survived in spite of those faults. This one I got second-hand with an unknown previous history and the welding looks proper on those seatstays. Also note it's still a Freeride frame. I trust it will be fine.

Furthermore - because cost is a really limiting factor to me. And the frame is boxy like the batteries I got. Mounting will be really easy - I saw in other build threads that this is can be a real hurdle with big batteries.

The final BOM is this:
Code:
Kelly KLS4812S controller                                156
Generic MQ7260 controller                                90
Bafang G040 rear hub motor                               195
Bafang motor cable                                       16
2x Mechanical brake sensor                               13
2x KTY 84-130 temperature sensor                         6
Scooter handlebar switch unit                            14
Scooter rear view mirror                                 14
Scooter ignition switch w/ keys                          11
Scooter horn 12v universal                               5
3x Car fuse 50A 58V                                      3
USB<->RS232 adapter cable                                12
3x DC/DC stepdown converter module LM2596HV 1.2-50V 3A   6
Magura 203mm rotor                                       25
Mini audio mixer (as prefab junction box)                10
Coulometer 100V 100A with shunt                          25
Solar battery cut-off switch 48V 50A                     9
Solar 1-pole 60A circuit breaker                         10
Azonic Flow riser bar gold                               30
PAS sensor 10-pole                                       20
Kelly RS232<->SM-4A adapter cable                        40
2x Avid BB7 mechanical disk brake                        76
2x Quaxar semi-metallic pads                             14
2x Continental tubeless sealant                          13
Continental tubeless valve set                           12
2x Schwalbe tubeless rim tape strips                     10
Busch+Müller Brex e-bike rear light w/brake light sensor 34
Busch+Müller Lumotec IQ Cyo T e-bike front light         48
Matte black spray can                                    12
Metal primer spray can                                   4
Total Cost 933€/$1014

Plus shipping of around 40-50€ I guess. The rest of the parts I either have outright or will modify to my needs.

I work at a specialized vintage road bike shop. Thus I also need to build the bike in a way my colleagues would approve of design- & style-wise. So that's another consideration on top of everything told yet. Luckily I got some really nice parts already, for example:
lever_color.jpg

The dropouts will need minimal filing work and I'll also use two torque plates on the inside of the dropouts:
axle_fit_ds.jpg
axle_fit_nds.jpg

It'll be a lot of work. I'll update whenever I can.
 
hexadrome said:
https://ebikes.ca/tools/simulator.h...0&mass=120&motor=MG60_500&black=load&axis=mph

motomech said:
I don't think you will want to make any more than 25 Amp.s (max) available.
I will limit the controller. To what values I don't yet know. I simulated 100A max throughput and do scale the wiring/fuses/switches accordingly.

motomech said:
19 Ah's of battery will be around 25 Miles (40 Km's) of range ( w/ little to no reserve).
I got 31.8Ah's at hand and simulator says 45 miles at 26.5Ah's. That's motor-only. I don't think range will be an issue. I plan on accelerating quickly to optimum cuise speed and will build my own cassette to better reflect the e-bike's needs.

It's good to have the flexibility for input voltages, because you could change your mind later after a real world test.

What is your planned cruising speed? If you want quick acceleration up to your cruising speed, then you have to build a bike that can go a fair amount faster than that speed. For efficiency, you want to build so your cruising speed is close to your max, but that will kill the performance getting to that speed, since the torque curve takes nose dive at 16mph and is so low that 18mph to 21mph will feel like an eternity. You really need to decide whether you only want the efficiency, or whether to compromise in order to get the performance you're looking for.

You can drag the slider (dotted line) to the left to see what kind of acceleration you have at the various speeds.

The same setup at 48V accelerates steadily to 21mph, and then drops off to get to the 27mpg top speed. At 18mph it's accelerating twice as fast as when on 36V, with twice the torque available at that speed.

https://ebikes.ca/tools/simulator.html?batt=B3626_GA&cont=PR&hp=0&mass=120&motor=MG60_500&black=load&axis=mph&bopen=true&cont_b=PR&motor_b=MG60_500&batt_b=B4823_AC&mass_b=120&hp_b=0

Experimenting with different voltages can provide you with a lot of info to make decisions from. I tried voltages between 52V (14s) and 96V (22s) and felt like 18S was perfect for my riding. However, 20S is more standard, so that's what I went for. It comes with it's own issues (touchy throttle, etc), but 60V or 16S, which is the next standard voltage, didn't give the acceleration I wanted. Most of my riding is between 17mph and 21mph, but the bike will accelerate hard up to 45mph before dropping off and getting to 20mph take literally 2 seconds. You may not need that, but I'm willing to bet that if you build for 21mph that you'll find the acceleration lacking. It's a pretty noticeable dropoff during riding when you hit that inflection point and the torque curve takes a dive so you want it to happen after you hit your cruising speed.
 
That's really good information to have here. My hard compromise for cost is weight (I've a motor :lol: ) and the 36V battery system.

E-HP said:
What is your planned cruising speed? If you want quick acceleration up to your cruising speed, then you have to build a bike that can go a fair amount faster than that speed. For efficiency, you want to build so your cruising speed is close to your max, but that will kill the performance getting to that speed, since the torque curve takes nose dive at 16mph and is so low that 18mph to 21mph will feel like an eternity. You really need to decide whether you only want the efficiency, or whether to compromise in order to get the performance you're looking for.

Planned cruising speed is 21mph. I rode the black e-bike without pedals long enough to learn motor limitations. Let me rephrase a bit for clarity:
I plan on accelerating quickly to optimum cuise speed using pedals and I want to pedal along at speed.
Regardless of the final outcome I feel I did a good enough job planning to be satisfied with the end result.

99t4 said:
If there are any guarantees, its this: When you "finish" your build :wink: , you will want to either rework it, or start another one, to incorporate lessons learned, or better ideas that you have generated along the way. :D

I agree with that also. I can swap to 48V later when the 36V batteries age. Right now this should work for me and I'll be happy with the build.

E-HP said:
Most of my riding is between 17mph and 21mph, but the bike will accelerate hard up to 45mph before dropping off and getting to 20mph take literally 2 seconds. You may not need that, but I'm willing to bet that if you build for 21mph that you'll find the acceleration lacking.
But considering the only real e-bike I used long-term was the black one w/o pedals, rather I expect to find the acceleration just fine for my needs. Everything is better than that crappy 10"->26" acceleration torque loss.
I rode a 48V out-of-the-box conversion kit build once, it accelerated way too hard for my taste. Then again, I don't know how the owner set up the throttle response (he didn't probably).

Parts I got so far:
parts_1.jpg
parts_2.jpg
parts_3.jpg
parts_4.jpg
(Sorry for the watermark)
The rest is in shipment transit currently. In German there's that old saying "the cobbler always wears the worst shoes". It's quite true. I don't want to tinker around much after the initial build. This is a one-attempt project to me at this point in time.
 
hexadrome said:
In my experience, old frames with known faults like this one (or the 70's German Mars bicycles or all those Stevens frames I saw with a ripped-off downtube near the headtube etc.) that still didn't break are safe to trust.
Because logically, all the bad frames went extinct over the years and only the strongest & most well-built survived in spite of those faults.

That is absolutely not how aluminum fatigue works. :shock: Especially since now you will be using the frame at higher speeds than it ever experienced.

Here's how a properly designed URT frame looks. The seat tube rests on the shock. Nothing acts like a giant lever, that is welded at the end, out of aluminum. :shock:

urt.jpg

From a durability and strength standpoint, that frame is very high on the list of worst design I have ever seen. It might be a cheap build of around $1000, but what's it worth without a seat? It's ultimately your money and your build, so good luck. :mrgreen:
 
Comrade said:
hexadrome said:
In my experience, old frames with known faults like this one (or the 70's German Mars bicycles or all those Stevens frames I saw with a ripped-off downtube near the headtube etc.) that still didn't break are safe to trust.
Because logically, all the bad frames went extinct over the years and only the strongest & most well-built survived in spite of those faults.

That is absolutely not how aluminum fatigue works. :shock: Especially since now you will be using the frame at higher speeds than it ever experienced.

From a durability and strength standpoint, that frame is very high on the list of worst design I have ever seen. It might be a cheap build of around $1000, but what's it worth without a seat? It's ultimately your money and your build, so good luck. :mrgreen:

In italics, I disagree a bit. Riding down mountains in '99 at 21mph doesn't seem feasible to you?
The new riding style will be flat tarmac mostly. A foreseeable difference would be that I'll ride at a higher average speed.
It even had Suntour DC90 forks originally, that totally means it's a real big boy bike :mrgreen:
Look: https://www.retrobike.co.uk/data/attachments/46/46145-aa5d343efc6692148018b8c345562589.jpg

In bold, I completely agree. There's a reason most manufacturers limit aluminium frame warranty to 5 years - it's material fatigue. Ultimately, there's no difference in strength of materials - lab testing data is on youtube, sure. This is also the reason we sell steel bikes at the shop - we need to warranty the bikes for a year and since we build bikes from vintage frames, this is the way to go for us.

I'd like to discuss my options with you. I asked about that issue in this thread:
https://endless-sphere.com/forums/viewtopic.php?f=51&t=118776
Given the answers there, would you still advise against using the Giant frame? I'm not really limited by it, I have the Spank also.
I thought the high BB would lend itself to a future mid-motor conversion, but I can switch the build at this stage no problem.

I'd have to get a new front hub though (~120-150€). I made a mistake ordering, so the forks are 110mm spacing non-Boost and I got a 110mm Boost hub. As a result, the front rotor didn't fit on the inside of the forks.
Also, it's a bigger investment. For a hardtail frame I'd be looking for a suspension post (in 30.4mm diameter no less, so I'll have fun :confused: ). And the good ones are expensive. So another 200€ probably.
What's your assessment ultimately?

Edit: Just found a DT Swiss 350 20x110mm non-Boost hub on sale for 45 Euro. I'll get that.
For the rear I'd need a Mavic EX729 36 hole 26" or 24" rim. This'd be hard to find.
Regarding the suspension post, the proper way would be turning down a 30.9mm post to final insert depth at 30.4mm.
Realistically, I'd need to use a 27.2mm and shim it to fit. That's a 1.6mm sidewall shim. Not that elegant at that steep seatpost angle of the hardtail frame. Hmmm.... :?

Edit2: Found DT Swiss 535 26" rims with 32/36 holes. 50 Euro. Sounds reasonable to me.
 
hexadrome said:
The dropouts will need minimal filing work...
Yes, looks good there. :thumb:

hexadrome said:
...and I'll also use two torque plates on the inside of the dropouts
Better make sure there is sufficient space for that. What is the OLD (Over Locknut Dimension) of the hubmotor and how does it compare to the bike's dropout width?

hexadrome said:
The new riding style will be flat tarmac mostly. A foreseeable difference would be that I'll ride at a higher average speed.
(Talking about minimizing stress on the seat tube cantilever weld points) If by "flat tarmac" you mean smooth pavement, no bad surface imperfections, large potholes, curb jumps, railroad crossings, etc. then I would agree with you, that would tend to not overly stress the cantilever mounted seat tube there. Of course, you can further minimize the stress by unweighting when riding over the rough surfaces, as long as something bad doesn't surprise you. :shock:

Your photos look almost as if the camera lens is fogged or greasy. Have you tried cleaning it?
 
hexadrome said:
What's your assessment ultimately?
Since ultimately it's my choice and I'm confident the frame won't break under my riding long-term, I'll proceed.
The Spank is a streetetz build and it'll stay that way. A rear hub motor isn't right for it. https://www.pinkbike.com/u/bipolarexpress/blog/the-streetetz-era-staring-at-the-sun.html

Fitting the axle to the frame dropouts was easy enough. I used a combination of cheap milling drills and hand filing. It didn't take too long to achieve a good seat.

dropout_right_before.jpg
There wasn't much material to be removed in the first place. Only 0.03"/1mm. Mostly paint. I've seen dropouts in far worse condition on frames, even ones where the axle of Shimano internally geared hubs rotated fully with their comparatively massive tabbed washers.

dropout_right_after.jpg
A 9mm QR axle will still fit these dropouts just fine.


overkill.jpg
Compared to the manufacturer supplied axle washers, the old torque plate I made is far too much. I've some 3/8" (4mm) mild steel, so I'll use that.

metal_cad.jpg
I've started on the torque plates and I'm quite unhappy. Of course there's work involved in fabricating these by hand, but I'm not willing to spend that much filing time. I hate filing. There's a machine for that called a die filer and I don't have one...

big_drill.jpg
That drill is from the '60s and of full metal construction. It only turns right. 450 old-school watts. It will rip my wrists off and keep working if the drill bit ever gets stuck. With that much torque, the long additional handle is a must.

I tried using the angle grinder table I fryed together a couple of years ago, but the mild steel scraps I used were too thin to hold securely. The table makes a good enough mounting point for the vise though.
angle_grinder_table.jpg
angle_grinder_table_detail.jpg


99t4 said:
Your photos look almost as if the camera lens is fogged or greasy. Have you tried cleaning it?
I bought my smartphone pre-broken so I don't have to worry about dropping it. I've decided to invest in a second-hand digital camera since I've got unused tripods and I'm tired of holding the phone and trying to focus the terrible camera simultaneously.

Anyhow, the current torque plates form is too much unneeded work. I would realistically be fine torquing the axle nuts with the manufacturer supplied washers. But I think I'll rather try to fabricate this kind of clamping-style plates which will be done in a couple of angle grinder cuts instead:
torque plate.png
Much more manageable with as little filing as possible :)
 
How does the clamping feature work? It looks like it has a slotted hole maybe?
 
99t4 said:
How does the clamping feature work? It looks like it has a slotted hole maybe?

It's supposed to, yes.
Well, on paper it was a brilliant design. I intend to screw on some fenders so I'd use extra long bolts I thought.

Tools and tooling

I own only the chinesiest of drill bits. $10 for a set of 150 bits is hard to beat for scrap metal value. They served my needs so well I didn't bother to upgrade to real drill bits yet. But there are some considerations for successful use:

What matters more is the tool. Upgrading the chuck on cheap drills works out well cost-wise. On the other hand their rotors run untrue out-of-the-box, so it's a moot point ultimately. Even good quality drills like the 750W green Bosch I use in a drill press, more often than not come with a keyless chuck installed. In my experience, don't bother with keyless chucks unless it is a quality one in a bolted-down machine. Tools held by humans have little stiffness by default. A keyed chuck grips the drill bit harder, so the shank doesn't spin out as easily. Once the bit locks up, the chuck leaves marks on the shank. As a result, the drill bit runs out-of-true unless properly dressed again.

upat_plus_(metabo).jpg
I really recommend getting a very old professional tool on the second-hand market. I got two 60 year old drills for 30 Euro. Both chucks run completely true.

On old power tools the motor capacitor usually dies from age. Metabo had a distributors agreement for hammer drills with Upat at the time, so the drill I use is a Metabo, but Upat branded. The capacitor blew on the other drill of course, so I tried to find replacements for both. Beware that you may not find capacitors with the same values nowadays. It is possible to run the power tool without a motor capacitor.

upat_coincidence.jpg
I used to desolder electronics in my spare time, so I have an assortment of random components. Luckily I found used capacitors with the right values, so both drills run well now after overhaul. The one for the Upat is circled.

I mentioned the drill only turning right. That's because it doesn't need to run left. Thinking about it there is simply no use case, apart from needing to use a left-cutting drill bit. More often than not, the left-turning on a hand drill is used to free stuck drill bits.

The 750W modern Bosch bogs down easily on more serious drilling jobs. The 450W Metabo never does that. It drills with a constant speed under load. It will either break off the drill bit or my wrists. Maybe both.
In all cases, let the tool do the work for you. The Metabo weighs >8 pounds (4kg). It drills like a drill press under its own weight.

chinese_drill_bit.JPG
Point being, with the right power tool and some oil for coolant, >1.5 inches (4cm) depth-of-cut is no problem freehand even with the cheapest of drill bits. I went slow and used a lot of oil.

a_bad_implementation.JPG
Ultimately I'm not a good milling machine after all. It... didn't fit. :lol:

A couple of days earlier, I stumbled on a chinese steel track frame I had in the basement. That finally gave me the right idea.
track_dropout_fit.JPG
The track dropouts are about as hard as mild steel and have conveniently placed holes already. They also fit the motor axle perfectly as they were. Lucky me. Thickness is 6mm (just under 1/4") and they need to resist 60Nm of motor torque plus whatever I put out. I read the dropout failure thread
https://endless-sphere.com/forums/viewtopic.php?f=2&t=14195
and do recommend it.
Mine are sized quite appropriately. As soon as I tighten the axle nuts to >=60Nm, the system as a whole will resist the axle torque with quite a bit of safety margin.

track_dropout_ds.JPG
I cut one short to hopefully clear the rear derailleur. The other I crossdrilled to make a clamping dropout. Together with the tabbed washers I'm sure of the result.
track_dropout_nds.JPG
They're painted and drying right now. I wanted to paint the controller black and that gets hot, so I got some radiator spray paint. Now I use it for everything. I'm not much of a painter but that seems to work alright.

How to build a cassette from single cogs

Sheldon Brown's site has the relevant measurement information. To keep it short, cog spacing is a nightmare if you use different speed cassettes. It is bad without a milling machine regardless of manual effort. For higher speed cassettes, it gets worse. In order to fit more cogs into the limited space, the tolerances get that much tighter. I chose 7 speed as my optimal point in bike industry history where the components were most durable and not that heavy. Remember your local supply may be different.

Maybe buy a regular cassette instead?

Also most of the cassette is pinned together so it leaves less of a mark on the freehub under heavy load. I am not able to pin this cassette, so I need to consider greasing the freehub (so that I'll get it off more easily later) and torquing it down harder than usual.

Regular steel is most durable. Take a look at low-end Shimano components. They got the durability formula down to a point.
This time I puzzled together an 11-12-13-13-15-21-34 7-speed cassette using some 10-speed spacers. It looks funny.
cassette.JPG
Then there is the issue of shifting. All newer cassettes have special tooth profiles milled into the cogs. They make for a better shifting performance and need to be matched between cogs of course. That goes out the window immediately unless you're willing to read Sheldons site carefully and invest in the right tooth profile cassettes, but cost-wise that'd mean...

So shifting performance will be crap.

I don't mind. I saw many single-speed ebike builds on ES, made by more experienced people. I expect not having to shift all that often.

If you really want to go down that route, here we go:
Consider the difference between freewheel and freehub. Here's an image I found
https://images.squarespace-cdn.com/...MNFI485NV82WMHL6JN/freewheel+vs+cassette+.jpg
For maximum stability the bearings need to be spaced as far apart as possible on the axle, so a freehub design makes more sense for our application. For a long time, Shimano held the patent on the freehub design. Thus even an expensive Campagnolo Super Record rear hub had a freehub body on the outside, yet was build based on a freewheel design internally. I've seen many broken axles due to that general design flaw. Material-wise, steel Shimano freehubs are very durable.

Yes, it is actually possible to disassemble a freewheel and reassemble it with different cogs. But if you disassemble a freehub, many small bearing balls fall out (magnet!). If you disassemble a frewheel, even more small bearing balls fall out and at least one will drop to the floor and get lost forever...
So that's not practical in the first place.

As for freehubs and cassettes, the optimal route would be using Shimano UniGlide. This is the predecessor to the current HyperGlide system. UG had no special tooth profiling but twisted teeth instead, so shifting works well in my experience. The UG freehub body has outer threads in the first iteration. The cassette is pinned together but has a threaded smallest cog. The pins can be unscrewed, but the threads are only on the ends. Which is fine if you leave the cog with the threaded pin holes in place. I didn't, so my cassette is loose cogs only now. Which is not ideal in theory for ebike application, but the cogs are wide on a 7 speed system. Cheaper cassettes often come riveted together.
Then for a time freehub bodies with outer threads and inner threads got produced. They offer the widest flexibility, but only went up to 8-speed.
The downside is that these components are often hard to find these days. You can turn UG cogs around once they're worn and wear them out for a second time, except for that threaded last cog. It's an incredibly long-lasting system.

The difference between the highest-end components from a manufacturer to the second-highest is materials, finish and weight. Less weight is costlier to achieve. It also means the component is less durable. I always forego XTR for XT and Dura Ace for Ultegra components. The weight penalty is not that important to me. I weigh a lot myself.

Since on a chained gear bicycle system, the derailleurs are 'dumb' and the shifters are responsible for moving them the right amount to change gears, it makes a lot of sense to have a higher-end shifter and a lower-end derailleur.

Come to your own conclusions why the manufacturers do it the other way round. It is intentional.

trigger_disassembly_start.JPG
I wanted to have the shifter on the left. Logically, it seemed possible to swap the ratchet inside from a right to a left shifter and have it work. Many shifter units are mirrored to save cost on parts and tooling. Sadly, my favourite ST-M050A shifters had different width ratchets inside so I now have yet another couple of disassembled shifters for rebuilding later.
Those M050s are completely overbuilt and can run without the covers long-term just fine. The left one is trimmable, so the front derailleur can be adjusted properly across the whole cassette range for optimal chain guidance. They are the original Shimano RapidFire design. The good one with thumb only operation. I never understood how you're supposed to use the index finger for shifting on the later RapidFire Plus shifters under simultaneous braking. Attempting that always felt biomechanically wrong to me, so I stuck to thumb-operated shifters as a result.
But the only useful thing was ultimately the handlebar bracket.
trigger_disassembly.JPG

I chose to go the other path and installed a left '89 XT thumb shifter I've found at my personal workshop. The right one was missing. Has to be destiny.

Unindexed thumb shifters are the simplest, thus most robust shifters available. They almost never break down mechanically, given a proper design. Broken shift cables can simply be knotted together as a makeshift repair.
cockpit_left.JPG
I salvaged the metal box I built for housing switches from the old ebike. I decided I wanted to have the amp-/voltmeter as well as a bike computer on the handlebars.
hose_clamps.jpg
I used hose clamps for everything previously. They come in different sizes fitting bicycle tubes well and they got a nut welded on. If that nut fails, a heavy pipe may damage someones expensive property. Manufacturers don't want that. They weld the nuts properly. On the other hand, these are professional-grade clamps I got from a plumber friend of mine. The simpler ones may not have such stringent QA procedures implemented.
front_derailleur_clamp.jpg
A clamp may be fabricated by disassembling a left over front derailleur as well. Once upon a time, Shimano made banded MTB derailleurs. This is a Deore one.

The switch box needed a cutout for the 100V 100A coulometer. I used that sawblade thing made especially for drills as I've a cheap set. They are horrible. Not from a quality standpoint, rather it's the teeth that get caught easily on the material. A human is not as stout as a milling machine and that's where these saw blades would work best.
drill_sawblade.JPG
Still, I managed to cut a well-dimensioned opening with the drill attachement from hell. Using the leftover shifter bracket the end result looks fine to me. I don't intend to waterproof. The bike won't see torrential downpour and to me, avoiding condensation is more important on electronic components. Those imperfect cuts don't bother me in the grand scheme of things.
cockpit_right.JPG

Oh yeah, I also built the cockpit. How silly not to mention that sooner. It looks like this for now:
cockpit_so_far.JPG

For more height I used an aluminium stem extension. They usually look ugly and bulky, but with the forks it somehow works stylistically:
front_view.JPG

It's not as clean a look as I wanted originally, but its more important to me to have good ergonomics. The old setup was too low for long range commutes, so this will do nicely instead.
 
I really didn't want to open the motor but a temperature sensor sounds too nice to have.
Also it came pre-broken like my smartphone. The metal spring was damaged from what appears to be factory handling. It cut into the motor cable at one point and damaged two sensor leads. So I had to repair it anyway.
motor_cable_damage.JPG
To get a bit of experience I decided to first investigate the old ebike's motor. I overheated it multiple times, but it was working until disassembly. I tapped a knife around the hub cap until I could get some screwdrivers in there.
old_motor_opening.JPG



Interestingly enough, I couldn't last longer than two minutes.




old_motor_overview.JPG





The stench from inside was so overpoweringly burnt cheap chinese electronics death cancersmell I had to put on my gas mask.





old_motor_detail.JPG



Not kidding. I threw the parts outside immediately after dissection.




old_motor_stator.JPG
old_motor_rotor.JPG

The white powder on the inside of the hub shell looks a lot like aluminium oxide or hydroxide. But how would it form on the inside? No idea what the goop on the anchor could be. The windings speak for themselves. This thing doesn't look healthy for sure.



Opening a Bafang RM G040.500.DC

First of all, do not mess with the freehub or the screws around it. They are not needed to disassemble the Bafang at all. Guess how I know. :lol:

After removing the hub shell screws, I got this part layout. Take note of the thin spacers on the axle.
g040_open_hub_shell.JPG
The clutch with the three planetary gears comes off easily and has no alignment markings on the cogs.
g040_sun_gear.JPG
It is indexed to the motor axle via a key. There is an aluminium sleeve under the key. Would be nice not to damage it.
First I tapped a screwdriver around the key with a small hammer to loosen it. I then applied vice grips and simply pulled it out.
The sleeve could be taken off the axle after that.
g040_key_and_sleeve.JPG
The sun gears screws are quite horrible. Phillips screws weren't ever meant to be disassembled by design.
You need a good fitting screwdriver. However a ratchet is even better as you can press it down into the screwhead while turning the handle. I was able to remove all screws using the ratchet and a small PH2 bit. The bigger screwdriver fit worse from a torque transmitting standpoint.
Nevertheless one of the screws only came out after repeated localized heating.
g040_sun_gear_bloody_screw.JPG
After removing the sun gear there is a snap ring to be taken off. The simplest tool to use would be a flat mid sized screwdriver and some small ones. Insert the mid sized one between the ears of the snap ring, lever one side open and help the ring out the groove with the small screwdrivers.
I happen to own snap ring pliers and do highly recommend them for this job. Essentially, you pay for the tips on those. Good pliers have cleanly machined and hardened tips.
g040_snap_ring.JPG
Now the rotor can be taken off. It has the magnets, so I went with some tight-fitting leather work gloves. It takes a bit of controlled effort. I found putting the axle end down on some hardwood and pressing it off downwards works really well.
g040_rotor.JPG
Finally the stator is freed and modifications can start.
g040_open_stator.JPG
I wasn't able to count the turns. I quickly discovered I had no idea how to attempt that without de-wiring the motor at least partway.

stator_before_decabling.JPG

After cutting some of the goop around the cable, I cut off the wires leaving a length of them for identification. Then I ripped the cable out the axle. Pulling downwards and using your body weight works.

I installed a length of shifter housing into the drill and cleaned some goop out of the axle.
axle_goop_removal.JPG

The motor cable is damaged in a frustrating spot. I could pull it through until I can cut off the damaged part and re-solder, but the length of cable leaving the axle would be too short. I have the other motor cable with the female plug, so I figured I'd replace the broken cable completely leaving me with a longer new motor cable.

The worst part was feeding the new cable through. Grin Tech made a video about it:
https://www.youtube.com/watch?v=spgDaEw0QEA
Trust the video. It is the easiest way.
I stuck the wires I intented to wire up the thermistor with through the axle and fished the tape through. Feeding the tape through by itself wasn't possible. There is a cheaply machined sharp turn in the way.
feeding_motor_cable_through.JPG
Feeding additional wires through is impossible. Too little space. I sacrificed the Bafang's speed sensor wire (white) for the job.

The Kelly controller I'll use needs a KTY84-130 thermistor. I was able to get a hold of two. It has got a black mark on the one side and the datasheet states this is the cathode. Another datasheet stated the cathode is supposed to go to negative, so I wired an additional length of black cable to ground and sacrificed the white speed sensor cable for positive. I then put transparent heatshrink on the sensor leads. Didn't shrink any of it though, I figure it will shrink by itself once the motor gets hot under use.
I stuck the sensor under a phase wire lead and gooped it in place using MS Polymer glue. It is a kind of flexible sealant glue much like silicone, but is more resistant and sticks far better. I do recommend this goop.
cabling_underway.JPG
Then I prettied the wires up best I could and put everything back together in reverse.

And then I discovered an interesting issue:
critical_error.JPG
Mwahahahaha... :mrgreen:
 
Hello from Texas, I am a novice ebike builder with 4 DIY builds currently. Two of my builds have 20" rear DD hubs with a 26" front wheel. The smaller rim feels a touch faster and accelerates quicker with my 36v batteries. In addition I used BMX bars to raise the front further! I recommend this, it feels awesome! Also I have an ebike with a 26" rear DDhub and I swapped the front forks for a 29" wheel, it has BMX bars as well.When I get an a bike without these modifications, I cant stand the feel. These are full suspension frames similar to yours, and sitting in the back seat is where its at.
 
In italics, I disagree a bit. Riding down mountains in '99 at 21mph doesn't seem feasible to you?
The new riding style will be flat tarmac mostly. A foreseeable difference would be that I'll ride at a higher average speed.
It even had Suntour DC90 forks originally, that totally means it's a real big boy bike :mrgreen:
Look: https://www.retrobike.co.uk/data/attachments/46/46145-aa5d343efc6692148018b8c345562589.jpg

In bold, I completely agree. There's a reason most manufacturers limit aluminium frame warranty to 5 years - it's material fatigue. Ultimately, there's no difference in strength of materials - lab testing data is on youtube, sure. This is also the reason we sell steel bikes at the shop - we need to warranty the bikes for a year and since we build bikes from vintage frames, this is the way to go for us.

I'd like to discuss my options with you. I asked about that issue in this thread:
Old freeride frame suitable for 1kw?
Given the answers there, would you still advise against using the Giant frame? I'm not really limited by it, I have the Spank also.
I thought the high BB would lend itself to a future mid-motor conversion, but I can switch the build at this stage no problem.

I'd have to get a new front hub though (~120-150€). I made a mistake ordering, so the forks are 110mm spacing non-Boost and I got a 110mm Boost hub. As a result, the front rotor didn't fit on the inside of the forks.
Also, it's a bigger investment. For a hardtail frame I'd be looking for a suspension post (in 30.4mm diameter no less, so I'll have fun :confused: ). And the good ones are expensive. So another 200€ probably.
What's your assessment ultimately?

Edit: Just found a DT Swiss 350 20x110mm non-Boost hub on sale for 45 Euro. I'll get that.
For the rear I'd need a Mavic EX729 36 hole 26" or 24" rim. This'd be hard to find.
Regarding the suspension post, the proper way would be turning down a 30.9mm post to final insert depth at 30.4mm.
Realistically, I'd need to use a 27.2mm and shim it to fit. That's a 1.6mm sidewall shim. Not that elegant at that steep seatpost angle of the hardtail frame. Hmmm.... :?

Edit2: Found DT Swiss 535 26" rims with 32/36 holes. 50 Euro. Sounds reasonable to me.
I have this bike. The exact one I bought two years ago. They haven't even changed it.


I've customized it a bit. I want a new fork to raise the handlebars So i sit more upright. Should I not be wasting my money though on a fork for this frame? I think it uses an older headset technology.

I am thinking this frame could fail any day, specifically around the drop outs. I have punished this bike for two years straight and put over 10k miles on it.

Should I just start a new build? I am worried about alumnium fatique. I don't see any cracks, but that doesn't mean there could micro stress fractures.

I don't want to invest more into this build and not be able to cannibalize the parts off it.
 
Back
Top