[MattsAwesomeStuff]

The Bike: A 1985 Honda Nighthawk 750S. I chose it because it was the first cheap bike I found (paid $20 total, already had a failed rebuild of an engine and the guy wanted it gone). I would have bought literally any motorbike carcass for $20, but by coincidence I happen to actually really like this one.

It was a 750cc bike with an 80hp engine capable of going 128MPH (212 km/hr). It's the original Widowmaker, the bike that term "superbike" came from. Original bike specs: http://paulsnighthawkpages.blogspot.ca/p/nh-700s.html

The style is a "UJM" or "Universal Japanese Motorcycle", very beginner friendly, you just sit on it comfortably like you would sit on anything.

An added challenge is that this particular bike is shaft drive (versus chain). More on that later. However, again, I would have bought literally any bike for the $20 I paid for this, beggars can't be choosers. The driveshaft ends in a U-Joint that couples the driveshaft to the transmission.


I've been looking to build a bike for 6 years already, and asked a friend to keep an eye on cheap bikes. He told me: "If you pay less than $500 for a bike, you're going to end up spending over $500 for it anyway. Even a bike with a blown engine is still worth more."

Well, the seller had wrecked the engine on his rebuild attempt and given up on it and stripped it down to part it out. Some of the parts were gone before I bought it.

The headlight, front brake lines, and the front fender/windshield piece were already gone from it when I bought it, sadly, and I'll have to find a way to make a facimile from sheet metal (it's $250 on Ebay, nope).

I got a rolling frame, a half-disassembled engine, and a rubbermaid full of miscellaneous parts and all the hardware.

I know nothing about motorbikes. I don't even know how to ride one. I know nothing about engines or mechanics. I am coming from a position of complete novice.

I started laying out the parts and trying to guess and group them to what they might probably be (I know some of these are labeled wrong, they were my first guess):

And, after a bunch of heaving and pondering and figuring out exactly, barely, by an 1/8" the one possible way to tumble, twist and pivot the engine out through the frame, I got it out.

Great. Onto the next thing, I need a motor.

 

[MattsAwesomeStuff]

Motor:

I first mocked it up with a treadmill motor to see if it would turn and how hard it was to get the orientation right. It was very difficult. That in itself almost made me give up, there seemed to be nowhere I could hold the motor that wasn't clunking the driveshaft around it's enclosure.

Yeah, that's a toilet plunger. It happened to be the right shaft size for the driveshaft's U-joint input, so I hammered it on to "machine" the matching splines.

A treadmill motor is good for 1500 watts, probably 5000 if you feed it higher voltage. So that would be decent for a moped, and for my junk-built recumbent electric bicycle from many years ago, but not for a motorbike.

BIGGER!

I phoned a local forklift repair place and told them my goal was to build an electric motorbike from garbage, and asked if they had anything they were about to throw away that I could strip parts from. The nice folks there said sure, and let me strip a small 1969 Yale lift. They bought it for its batteries and the rest wasn't worth selling. It had sat in a flood.

It had 3 motors. Because it was small, the biggest motor was the pump motor, and it had a pair of shorter drive motors (one for each tire).

Big Motor Specs:
- 7.25" (18cm) diameter and 12" (30cm) in length
- Weighs 91 lbs (41 kg).
- Series wound DC (good).
- 4 brushes. Appears to have neutral position.
- Armature feels good, no detectable wear between the brush path and the edges.
- Built-in blower fan (great).
- No HP/Wattage ratings. No spec sheets turn up on google.
- Not sure what voltage it was designed for. Was probably on a 24v system.
- Female output shaft, slot cut (awful).

The two slightly smaller motors have a threaded output. They were my backup plan, but looks like I'll be okay, more on that later.

(guess which direction it was laying in the flood?)

I clean it up and, it runs!

First Sparks!

Great, now the motor needs to mate to the driveshaft somehow.

 

[MattsAwesomeStuff]

Transmission:

TL;DR - I decided I don't need one so it's gone. Ta da. Skip this if you don't care for details.

The transmission shared oil with the engine, they're the the same assembly. I could have cut the engine off, but then I would also need to:
- Plug all the oil ports.
- Adapt the electrical motor to... whatever part of the engine spins the transmission.
- Add 60 lbs of weight for the transmission.
- Take up a transmission worth of volume inside the frame I'd rather use for batteries.
- Keep the gearbox and clutch and clutch lever.
- Actually learn how to ride a motorbike (I've never even been a passenger).

So, clearly, if I can get rid of the transmission, I should.

I figure I would need to keep the transmission for 3 reasons if they're issues:
1 - If my motor speeds don't match the correct range to spin the driveshaft for the speeds I want to go.
2 - If my motor isn't powerful enough to accelerate reasonably from a stop.
3 - If my motor needs more power than it can handle at low speed hill climbing.

Let's go through those:

1 - Napkin calculations. It's a 16" tire, so, 16x3.14 = 50" tire circumference or 4.2 feet. So every tire rotation moves the bike forward 4.2 feet. 5280 feet in a mile, so, 1257 tire rotations before it covers one mile. 60 minutes in an hour, so 1257/60 = 21 RPM for each mile per hour I want to go (13 RPM per km/h).

So if I want to go 60mph (100km/h), I need to spin the back wheel at ~1260 RPM. The driveshaft to rear wheel has a 3:1 gear, which means 3780 RPM on the shaft goes highway speed.

Gut check time. Most motors this size are fine spinning up to about 5000 RPM. The limits are bearings and centrifugal force ripping the rotor apart. So it can physically handle spinning that fast.

Series motors can spin that fast no problem, I don't know how fast it used to spin, but that seems the right ballpark.

Let's doublecheck that and make sure it matches the ballpark for the original engine.

Most transmissions only exist to slow the engine speed, where top gear is just straight linked with no reduction. The bike engine originally redlined north of 11,000 RPM, but then it also had a top speed of 125 mph. 11,000/125 = 88 driveshaft revolutions per MPH. Which does not match the tire RPM to make the bike that fast (21) so, something somewhere is still gearing down by 4:1? The driveshaft:tire gear is 3:1 or so, so, sure, it's in the right ballpark. The numbers don't matter, just a gut check to make sure I'm not worlds apart and made a mistake somewhere. If it was 7000 RPM I'd be in trouble, if it was 1000 RPM I'd be in trouble. I'm right in the sweet spot, as most motors generally will be compared to their gas counterparts.


2 - Kinetic energy = Mass * Speed * Speed / 2

It's the first, oh, 40 miles per hour (70km/h) that are going to piss people off in traffic, what would be the low gears. Above that the effects of wind resistance are dominating and you're having to consider that you only have the left over power not being used just to maintain speed. So, 40mph is 18 meters per second.

Mass is probably around 500lbs of bike plus 200lbs of rider, 700-ish pounds, call it 350kg.

E = 350kg * 18 * 18 /2
56700 joules.

Joules are useless to think about, 3600 watt-hours in a joule, so about 16 watt-hours. About as much as 4 cell phone batteries holds, to bring a 350kg bike + rider up to 40mph, ignoring wind resistance.

Anything above 0.3g of acceleration is considered aggressive, anything above 0.5g is considered dangerous. 0.3g would be 6 seconds for any vehicle to reach 40mph. That's a good target, a 27hp motor should be good enough for acceleration.

3 - Unlike with wind resistance, the power needed to climb a hill is linear. If you want to climb it twice as fast, it only takes twice as much power. The total energy used to climb a hill is the same regardless of whether you do it slow or fast.

The only thing that matters for the hill climbing part of power requirements, like with acceleration, is weight. In this case, rather than making weight go faster, you're lifting a weight.

700 lbs, 350 kg.

Potential Energy = Mass * Gravity * Height

If a hill is 330 feet tall (100 meters), regardless of how steep it is or how fast I climb it, the energy to climb it will be:

350 * 9.81 * 100 = 343,350 joules. Or, 95 watt hours.

That's only 330 feet tall, as tall as a 33 story building.

If it took my an hour to climb it, it would add 95 watts to my motor requirements.
If it took 15 minutes to climb it, it would add 380 watts to my motor requirements.
If it took 1 minute to climb it, it would add 5,700 watts to my motor requirements.
If I climbed a straight vertical cliff, and wanted to do it at highway speeds, 330 feet is 1/16th of a mile, highway speeds are a mile a minute, so I'd need 91200 watts for a cliff, ignoring traction.

My motor can't put out 100 horsepower, but luckily a 20% grade is about the steepest hills there are to climb, which makes it 20% of that, so, ~20 horsepower (plus ~12 hp to maintain highway speed itself from air/rolling resisistance), on a super steep hill that they probably don't ever make highways... 32 hp total. Very close to the 27hp needed to accelerate "aggressively" from a stop.

Good enough.


Boring post, no pics.

No transmission is the way to go, so I just have to couple the motor to the driveshaft now.

 

[brian_]

Some of the parts you are uncertain about don't matter to the build, because they're part of the engine, but in case you're selling them... and others you might need:

• yes, that's the engine oil filter
• the "rear something" looks like it might be the tail lamp bracket
• yes, that must be the rear fender
• the top "heads" item is the head; the bottom "heads" item is the valve cover
• yes, that's the throttle control

 

[MattsAwesomeStuff]

Motor Coupler:

The motor has to couple to the driveshaft, without a transmission.

The advantages to shaft drive are:
- Quieter than chain. (Nice for electric which is already quiet)
- They don't tend to break or fall apart. (Nice for series wound DC motors, which, if a chain ever breaks, will accelerate off into low Earth orbit in a seconds without a load).
- Nothing to tangle or grab your cuffs or shoelaces or pick up gravel/grass offroad.

The disadvantage to shaft drive are:
- Coupling to the shaft is difficult. You can't just throw a sprocket on and set the chain tight.
- Machining odd couplers is hard.
- Layout is restricted, you can't put the motor in the top triangle and leave the main frame for battery compartment. The motor shaft has to go in line with the driveshaft, on the left edge of the bike.
- Balance is an issue. The motor is on the left edge of the bike.
- Alignment is difficult. The motor shaft has to be lined up in both position and angle, nearly perfectly, or the u-Joint rattles and clanks as it maxes out. A chain drive you only have to get the 2 sprockets into the same plane, the chain does the rest.

I don't have a choice, so, I have to work with the disadvantages.

This is the old tranny output:

This is the driveshaft:

The U-joint couples them together.

Since the motor was the hydraulic pump motor, it is mated to a hydraulic pump. So I have 4 possible pieces.

Here's the struggle:

1 - Female splines on U-joint.
2 - Male splines on Tranny output.
3 - Female slot on Motor output.
4 - Small male splines on hydraulic pump shaft that drives the vane carrier (pump internals).

1 and 2 naturally mate. 3 and 4 naturally mate. I need to mate 1 and 3 (both female), using the chopped off shafts of either 2 and 4 (males). Life umm, finds a way.

Ideally, I would just machine down the splines on the hydraulic pump shaft to match the u-joint... but while Male #4 was made for Female #3, he isn't girthy enough to satisfy Female #1.

If I use Male #2 from the transmission output, that fits the u-joint perfectly, but then I have to figure out how to mount it into the motor. It's much easier to just cut a giant screwdriver bit out of a shaft, but it's harder to mount it.

Clear as mud? Here's the plan. Add the vane carrier to the pump shaft, then machine it down to match the old tranny output, ignore the tranny entirely:

Still don't understand? HERE'S A MONTAGE:

I "machined" that coupler while visiting family, without my tools, in a back alley, using a masonry wheel in a wood-chopsaw, a grinder, and sander to roughly shape it (any would have been fine on its own), and then a dremel or a hacksaw to cut the teeth.

I used the last bit of hot glue still inside the glue gun I took from the sewing room (shove it forward with a pencil) and some packing tape to "fixture" the coupler to the tranny so I could mark to mimic the splines. I had 1 dremel wheel I found under the plastic in the case, and no wheel arbor (I'm using the tip of the sanding drum, the wheel broke right after I was done marking).

When people complain about not having the right tools, or not knowing how to machine things, or how expensive it is... I drown your complaints with this splined shaft coupler I made on my first attempt with a rusty hacksaw in a back alley. Just try harder.

I cut the pump in half and mounted it all up:

Lent my welding mask away? Haven't welded more than an hour in my life? Want to tack weld it just enough so that it holds in place and I can test the spin?

Let's go.

How about some speed control?

 

[brian_]

Most transmissions only exist to slow the engine speed, where top gear is just straight linked with no reduction.

That hasn't been true of automotive transmissions for many years (they all have overdrive ratios). That's how traditional automotive transmissions did work, with the top "gear" being just a direct connection of input to output, and some transmissions still have one direct gear (it's just not the top gear anymore); however, motorcycle transmissions are typically "all indirect" designs (like transverse car transmissions) so there is no reason for any ratio to be 1:1. I have no idea if motorcycle transmissions tended to have only reduction gears.

You might have noticed that the clutch is on the transmission input shaft, not the engine's crankshaft. The engine drives the transmission through a pair of gears, which likely provided a stage of reduction gearing. That's one reason not to use the bike transmission - it could be difficult to mount the motor in a way which fits and using the right gear on the motor shaft.

 

[MattsAwesomeStuff]

Speed Control:

This seems to be one of those places you can't cut corners, you gotta spend some monies.

Supercheap Plan A:

Size the battery voltage to whatever was roughly the right amount to travel highway speeds, so, just lock the throttle on with a contactor and good enough. Then use water-cooled resistors to do speed control below that. Maybe a series/parallel switch to toggle to half-voltage.

If you say "That won't work" I direct you to my hacksawed spline-coupler and wave two middle fingers, and also to past vehicle projects where it works just fine. You pump 20x the wattage through a resistor it was designed for (to keep the size down), just long enough to give you a slow start and bring you up to speed.

Plan B:

Morning after pill. It works.

Plan C:

How was the forklift controlled? You pulled all those guts out, and labelled them before you did so you can put it all back right?

Umm...

Well... I was in a hurry. It was dark out.

What I did do was find a schematic glued to the frame, printed on wax paper, from 1969, that sat submerged in hydraulic fluid and flood water, and was still more legible than a week old newpaper. They don't print 'em like they used to.

How hard can it be?

Two weeks later

GE SCR control panel using 1960s electronics.

I don't know how it works. I know how it doesn't work.

Modern controllers you use MOSFETS or IGBTs, and you provide them a small signal and they turn on, and you take that small signal away, and they turn off. You pulse it the amounts that make sense, and the average amount of voltage going to the motor spins it whatever portion of full speed you're ON for.

This uses antiquated SCRs. An SCR is like a MOSFET, except you can't turn it off. You turn it on with a signal pulse, and then it just latches on forever.

To snuff it out, you have to kick its feet out from under it with a back-voltage that overwhelms it. Every single pulse. Where do you get that big of a spike of reverse-voltage? Another SCR! How does it charge backwards? Umm... coil... capacitor.. transformer... umm... well whatever, that's the theory.

It's like swearing in a foreign language. You know enough to get the impact you want, doesn't matter if you're not fluent.

From the wiring diagram I started trimming out all the pieces I didn't need that were forklift related:
- Pump motors? Cut.
- Hydraulic controls? Cut.
- Power steering? So cut.
- Horn? Cut.
- Forward/reversing relays? Mostly cut (this is called "Plug" braking, you slam it into reverse, it bleeds energy off as heat, it's impossible to recover, standard forklift operation).
- Safety switches? Cut and/or shorted so it was always on.
- Max speed contactor (shorts out speed control, just gives max speed)? Temporarily cut.

I short or open 90% of the circuit down to the bare bones:

After cutting out all excess stuff out, you're left with some pretty minimal connections to make:
- The motor connects to positive.
- The motor connects to the speed controller.
- The speed controller connects to negative.
- The foot pedal potentiometer connects to the controller.
- The foot pedal "power on" switch connects to a contactor, or a dummy load to fool it to thinking there's a contactor there.

GET TO THE POINT! DOES IT WORK OR NOT?

This Video Sucks, I know, don't bother complaining

The speed controller is only good for around 5000 watts though. So it's enough to test, but it's not going to work on the bike. Also it's the size of a goddamn toaster and weighs 30 lbs.

If anyone has a better controller they'd like to donate, this branch of the build has bottlenecked, it'd let me continue. I'm happy to pay shipping (to Canada), the whole "built from unwanted things" is somewhat of a mission beyond that though.

...

Meanwhile, I guess it's time to put the bike together, this might actually be a thing.

 

[Duncan]

Hi
How about a Paul & Sabrina controller - it's about $600 for a 500 amp and 150 volt kit

 

[MattsAwesomeStuff]

Actual Mechanics/Motorbike Assembly:

I don't know what I'm doing. I never took this bike apart, someone else did, I got a bin full of parts and hardware.

Front stuff at the front, back stuff at the back, maybe?

Sort bolts, numbers might help me figure out how stuff goes together?

There are at least 13 wrong ways, wrong orders, wrong places, and nearly-but-not-perfectly interchangeable sets of bolts that you can use to put this bike together wrong, that require nearly-complete re-disassembly again. Some places it doesn't matter if you use the short or medium ones, until you get to the place the needed the medium ones and you used them elsewhere.

The theory with this goes: "Just keep making mistakes until the only way left to do it is the right way."

I don't even know basic mechanics. I've never touched a motorbike. I don't know what side the throttle goes on. I don't know a brake from a clutch. I don't know what bolts belong with the engine I'm not using, and which hold all the cosmetics on.

If you ever try this, I recommend starting on a vehicle that is already assembled or at least with all the pieces.

Ta da? Close enough?

Well most of those wires might be in the right place.


Issues:

1 - Most valuable an iconic part of the bike, shown in the photo when I bought it, is not there. It's $250 to replace it. I trade the engine/tranny Honda covers for one that's broken in half and glue it back together, sort of, whatever.

2 - Previous owner cut and threw away the front brake lines. I order some for $20 from a scrapper. Paid using real Earth dollars. Local places want $80/line. China wants $15 but I don't want to wait.

Replacement lines don't reach the manifold though...

Because...

3 - Handlebars are bent. No problem.

4 - Headlight and license plate light are out. $20 for the pair. I don't know how to find compatible ones at a junkyard, so I just buy them new. More real Earth dollars gone.

5 - Brake lights out. It's because he cut the wires. There's more.

6 - Horn is wrecked. So, don't unplug the spade terminals, cut the wires off flush with the harness so I can't ever goddamn find them or know they're supposed to exist there. Thanks previous owner.

7 - All cosmetic panels are broken, cracked, shattered, ripped off. I now see why the previous owner who bought it fix it up, gave up. He wrecked everything he touched. Superglue and epoxy to the rescue.

8 - No windshield, I'll have to make one.

9 - No left grip. Ordered new ones. $3, China.

10 - Speedo cable mounting clip is shattered. JB Weld to the rescue.

11 - Master Cylinder (I learned a word!) does nothing. Brake fluid and 3000 frustrating pumps later while bleeding, fluid suddenly appears for no reason. Hurray. Brakes hold.

12 - Honda fuse cover at the handlebars is missing. Fuses are gone. Cosmetic pieces are missing. Speedo fender mount is gone. Rear middle grab handle is gone. I'm sure I'll keep finding more. I hate the previous owner now. Enough random crap that this would be a $500 bike if I had to order it, but the international Nighthawk community is very nice and everyone has parts bikes they raided the valueless bits for me.

13 - The "still good" tires are badly cracked, I was too ignorant to know what to look for. They'll need replacing.

14 - Mirrors? Said he had them, I don't have them. $15 locally.

15 - Turn signals are rubber clamps that grip the front forks. Rubber is old. It split when I opened the C-shape up. 5 different types of glue/epoxy/cement so far, nadda. Not sure what to do next with those.

That said... I paid $20 for most of a motorbike. There are worse tragedies in the world.

I buy 5 different used helmets for $10-30 each from different old timers. I don't know what style I'll want, might as well try a bunch and then buy a new one that's what I want. I don't think I'll buy full gear, but, without wasting money, helmet and gloves aren't places to skip out on on principle.

...

I temporarily rig up the throttle cable to a bracket on the forklift's potbox just because I want to be reminded this thing spins after a few weeks of mechanical musical chairs.

She's mostly mechanically reassembled. Let's get some batteries!

 

[MattsAwesomeStuff]

Batteries:

Of the major expense:

1 - Bike.
2 - Motor.
3 - Controller.
4 - Batteries.

... I got the first 2 nearly free, the 3rd... ehn, half assed. So just batteries now.

This project sort of started when someone told me you can buy used starter batteries at the junkyard for $10/each and I thought "What could I make with bottomless starter batteries?"

I know they're not meant to be deep-cycled. I know they will have severely short life. Blah blah. I don't care. Replace 'em every other month for that price, they're worth $10 in core charge. Revenue neutral! You still could build an electric motorbike from junk.

Half-way through the build I found out, no, $40/each. Well crap, now it's not worth swapping out.

I was going to use 4 or 5 of them in the frame somewhere. Probably 15 mile range at highway speeds.

Spoiler: Ain't nobody got time fo that.

Bending the rules here...

You're now caught up to the present day of the build. I've spent the last 3 months recycling garbage tool packs. I have around 1500 good cells. ~12kwh so far.

At 200wh/mile at highway speeds, that's 60 miles.

It's about 2 milk crate's worth of cells.

Disassembled, charged, capacity tested and marked, recharged, and sorted.

I don't know if I have room (bulk-wise) for that much lithium on the bike. They're about 1lb for 10 cells, so, 1500 would be 150 lbs... about the weight difference between engine and motor so far.

So it's pushing it for weight, and pushing it for space.

I got lots of cable from an estate sale for free (old farmer guy had 80 lbs of batter cabling his kids just wanted out of the garage, gave it away). So that's handled.

Oh, and the batteries...

For free from a local recycler. I picked up 2 each of 4x 18650 chargers and have been running them (on top of a charging bank) for literally 24/7 and 3 months straight. Because I'm persuasive. That's the bending the rules part. A random person can't get the sweetheart deal I worked out, this isn't something the average person can find from unwanted items.

The caveat is that, after I finish my rig, and figure out how to do this safely, I'm going to keep doing it and help support local community projects at the local makerspace and such. Kids (and adults) wanting to make E-bikes, kick scooters, skateboards, mopeds, go-karts, power-wheels, etc. Taking them out of the waste stream. Which is part of why I'm doing this anyway, 'cause I wish someone would've shown me how to make stuff when I was a teenager with no spending money.

...

So that's where I'm at.

To Do:
- Fix signal lights.
- Fab windshield from lexan.
- Put together some 18650 packs (I have the 4x5 holders, $20, China).
- Fab speedo fender mount.
- Fab battery cases.
- Build a charger (variac would do for now).
- And get a controller that works.

No more blogspam. Updates as they happen.

Did you make it this far? Here's a video of it "running" 6 months ago: Yes this video also sucks, I wasn't intending to show it to anyone

 

[Duncan]

I LIKE your spline adapter!

Keep an eye on the UJ - check to make sure that it does not start being able to move up and down or sideways

As the suspension moves up and down the driveshaft normally has to move in and out of that spline - you need to ensure that it does not crank the two splines apart

The tower and forks on that old forklift are saleable - the farmers here bolt them to the rear of their tractors - instant forklift

Great Job!

 

[MattsAwesomeStuff]

Some of the parts you are uncertain about don't matter to the build, because they're part of the engine, but in case you're selling them...

Yep, had those figured. That was from my first attempt. Eventually the bin got smaller and smaller as I went to do things, things wouldn't attach, and I'd go "Aha!" and find something that I'd use.

I gave the coils away to a gentleman who helped me out. Very little gas-stuff left. I did eventually notice there were no exhaust pipes anywhere, another thing that got sold before I bought the bike.

That hasn't been true of automotive transmissions for many years (they all have overdrive ratios).

It's true of this bike, I looked up the gear ratios, 6th gear is 1:1.

It redlines at 11,000 rpm, it doesn't need an overdrive

How about a Paul & Sabrina controller - it's about $600 for a 500 amp and 150 volt kit

Sounds like about $550 more than I'd spend

But, decently cheaper. I've never heard of this though, so I'll do some research.

The OpenRevolt project seems abandoned or, at least when I look through the big thread it's all in a state of maybe. I'm not good enough with electronics to reverse engineer things or troubleshoot anything that goes wrong. I could etch my own board and order components fine, but, if it doesn't work perfectly... I hate being that guy that has to ask 1000 questions and have someone do tech support over text.

I LIKE your spline adapter!

There's always one in a crowd...

Keep an eye on the UJ - check to make sure that it does not start being able to move up and down or sideways

That was the part that almost made me quit on the bike. I couldn't figure out how to center the motor exactly where it needs to be. I ended up taping a few straight edges to the rear tire's axle, through the swing-arm pivot, and then try to guess where that lines up along the front of the frame so I knew both the position and angle to mount the motor.

The driveshaft isn't constrained on motor end, the only thing that stops it from flopping all over the place is the position of the motor, which is somewhat fixed in place (welded to brackets and bolted in). I could easily be off by 1/8" in any direction, hopefully not 1/4". But howevermuch it's off by... it's not going anywhere.

As the suspension moves up and down the driveshaft normally has to move in and out of that spline - you need to ensure that it does not crank the two splines apart

Ahhh... that might be the most prescient piece of advice I've gotten so far. I'd considered if that mattered. It wouldn't be anything I would've noticed until I had it on the road and hit a bigger bump.

The U-Joint is in line with the pivot hinge of the swingarm, so at least there's no... what would on a bicycle be called "Chain stretch" or "Chain snap" when you use the suspension.

As the angle changes away from straight, obviously the U-joint bends (or it would've been made fixed) But it hadn't clicked that it might need to lengthen or shorten the shaft too, by sliding down the splines.

I might be bottomed out on the old pump-shaft's splines... will the U-joint ever need to push in, or just pull away?

The only other shaft-drive bike build I saw, the guy welded the motor shaft to the U-Joint... glad I didn't do that.

The tower and forks on that old forklift are saleable - the farmers here bolt them to the rear of their tractors - instant forklift

Lots of farmers around here, he might've saved the forks, but, the mast and the rest of the machine just went straight into the steel bin when I was done with it. If he could've sold it, he would've I'm sure, that's what he does, buys junk lifts, repairs or steals parts, sells them. This one he said he pulled the batteries out and the rest was junk to him.

 

[MattsAwesomeStuff]

It's going to be less concise updates from here out, since we're caught up to the present.

I had someone offer me an E-bike controller that isn't suitable to them anymore.

It's 36-95v, 60A. Meant for BLDCs. It's an 18 fet controller.

Now, that's only ~5400 watts, which is enough for moped speeds perhaps, but not much beyond that.

So I was wondering:

1 - Does anyone know if a BLDC controller can be used on a normal series-wound DC motor? I presume I would just gang up the 3 pairs of outputs?

2 - Does anyone know whether this controller could be upgraded by me? Can I just add more fets in parallel, and more caps on the input rail, or are there likely limitations on the fet driver as to how many fets it can drive?

I'm somewhat comfortable with electronics, but I don't have a great grasp of what's actually going on in there.

If it's the kind of thing I'll just spend more time and money to cludge the wrong controller to work, then I guess it's a poor choice, but if I can get away with it, that might solve my controller problems.

The good thing is that it's in the right voltage range (I'm thinking, maybe around 80v would be a good target?), whereas all the forklift or golf cart controllers I've got are only 36v.

If anyone has advice, I'm all ears.

 

[MattsAwesomeStuff]

1 - Does anyone know if a BLDC controller can be used on a normal series-wound DC motor? I presume I would just gang up the 3 pairs of outputs?

I had someone point out to me that, obviously no, this will not work.

Ganging up 3 rectified pairs of outputs would net me smooth DC yes... but without any actual speed control. So it would be a very fancy on/off switch.

Back to the drawing board.

I had a guy who works at a local golf course give me a questionable golf cart controller from a 36v golf cart.

It's a Curtis PMC 2586 (or a 2586, 4, G09?). I can't find any info on them what-so-ever. I looked up the MOSfets, they're 60v, 60amp fets and there's 9 of them, so, that means it could technically handle 60v and 540amps, max, no? That's be 33kw (44hp) which would be pretty much perfect (a little low on the voltage, high on the amps, but, beggars and choosers and all that...).

Or do you need lots of overhead?

In either case, I don't even have a pinout for it, let alone specs. It's not listed on Curtis' website either. The usual googling has failed me to find out further specs.

Anyone have an old Curtis catalog that knows better?

 

[Electroddy]

Try thishttp://products.jacobsen.com/img/manuals/28646.pdf Page E-17 might be what you're looking for. The 25864G09 is EZ-GO's part# for the
Curtis 1206. As to the FETs, if they are only 60V rated, they wont be happy at much over 48-52V.

 

[MattsAwesomeStuff]

Aha. It's actually an "EZ-Go" 25864G09 golf cart controller, which is a Curtis 1206-4301 (which is not written anywhere on it, you just have to know that).

And for future people who may come across this thread: http://products.jacobsen.com/img/manuals/28646.pdf <-- Manual for the golf cart, which has some controllery bits in it on page E-17, but not enough for me to figure the pinout.

[Edited to add .... Was typing this up yesterday, just saw Electroddy's post saying exactly the same thing here today. Good to confirm ]

...

Revisiting a couple rejected options I didn't discuss before...

I also have a controller from the forklift yard's junk bin that was marked: "Needs repair, low output". It's a Curtis PMC model: 1-187-067. 24-36v, 275 amps.

Again nothing comes up on google, and I couldn't get into it, it's potted shut. Drilling out the potting to find 6 screws underneath and some hammering later, it says on the board it's a "1204x-42". Hard finding that exact model, but at least it's a model that exists and I can find pinouts for.

So that's promising. You know, most of the time, in any place, when stuff is marked as damaged there's nothing wrong with anything but the technician, so, let me just see why the output might have been may be marked "low"...

Err... oh. Not the technician.

Looks like everything that's steel (mounting hardware) is rusted, everything lead (all the solder joints) are crumbly white powder.

I can't get inside the thing because the side boards are all soldered together. It's at least dozens of solder joints just to examine it further. FETS are I+R9207, (+- symbol?), can't find a datasheet on them.

I do like a challenge, and I do like repurposing junk, but:

1 - I don't know if it's in working condition.
2 - I have to make dozens of desoldering connections just to access the FETs and clean them.
3 - It's hundreds of soldering joints that are corroded, any one wrong and it probably won't work.
4 - Is the circuit board corroded internally?

Sounds to me like the 1204 is a shelf-it project, maybe turn it into a charger later.

I also pulled a GE Clark EV-T5 controller out of the junk bin last year which is 3x the size of the Curtis 1204 (as big as a double-slice toaster), but, it's only 24v and 75 amp (250amp at 50% duty cycle). The current is probably okay, but the 24v is just hot wet garbage (and I don't know if it's working).

So, I guess a bit more sleuthing on the 1206 as that's my best bet at something that might hit highway speed.

Normally I'd be happy with trial and error. Pain in the ass is, once I arrange the 18650s into their plastic holders, (head or tail in the correct directions)the entire pack has to be disassembled to get them apart again, let alone all the soldering. And soldering the batteries damages them a little bit each time I'm sure. A lot of my projects get abandoned when I hit a point where I have to backtrack because of how discouraging that is, so, trying to prevent that this time.

So, I'm kinda stalling out at this stage to see if I can get a higher voltage controller that I'm more sure will work before I double-down.

 

[MattsAwesomeStuff]

I kept digging for pinouts, found this: http://ev.evdl.narkive.com/0masMVv9/ezgo-controller-connection-diagram

"On the 5 pin style 1206 Curtis controller pin one and two are your 0 to 5 k/ohm throttle input. Pin three is half speed reverse and pin four is your key switch input to power the brains. The fifth pin is not used."

Sounds pretty straightforward. 1-2 pot. 3 ignore. 4 short to... positive when on. 5 ignore.

So... I try hooking it up on a power supply...

Confusion.

1 - If the power supply is on B+ and B-, B+ and M- have the same voltage on my multimeter. Is that because there's no load, or, does that mean blown-short mosfets and back to square 1?

2 - I have 3 "E-bike throttles", of different styles. None of wires on them do what a potentiometer should do. Nothing changes when I twist, resistance doesn't is either infinite or isn't variable. I (foolishly?) presumed all throttles were just pots.

Doing more research on pot boxes and such... these twist grips are all hall-effect sensor based. They provide 0.8v-3.8v, not 0-5kohm. They appear not to make actual potentiometer throttles.

The controller expects a 0-5,000 ohm potentiometer, but, it uses that to lower the voltage on some pin probably. It's probably not coincidentally the same range.

So...

- I have to find some kind of pot that will work inside the mechanism.
- I have to order a different throttle entirely.
- I have to convert my signal to the right voltage using a microcontroller or something, which I've never used before.

This is the first actual outright mistake I've made on the bike.

http://www.evwest.com/catalog/product_info.php?products_id=294 <-- I could spend $58 to remedy this. That's 3x what I paid for the bike, so, no.

I could half-ass a variety of solutions, epoxy a pot into the handlebar and have the end of the throttle twist it (presents a hazard if the bike falls over), etc. Don't really want to half ass it.

Not exactly sure what to do next. It's a minor problem, I currently have the original cable throttle hooked up to a pot box, but it's the size of baby's skull, cumbersome and annoying. It's solvable but not in an elegant and cheap way.

... More pressing is the fact that unless I'm mistaken, the controller itself appears blown and I'm back to square 1 anyway.

 

[MattsAwesomeStuff]

Controller Success!!

I basically sat with my thumb up my butt for 3 weeks, figuring the controller was broken because it showed full voltage on the motor output without the keyswitch or the throttle even connected, indicating blown mosfets. And then I was bummed about my throttles all being the wrong type and not being able to use them to test either.

So today I decided to actually just hook a motor up to it and a pot (had a 50,000 ohm rather than a 5,000 ohm, but, no matter).

Well lookie that that. As soon as there's really any load at all on the output, there's no voltage there, so, must have some high impedance path through the controller.

I used an 1800w treadmill motor (easier to carry onto my office desk), threw on the pot, shorted pin 4 to positive, and... nadda.

About to give up and then I tested the pot and even after checking it thrice, discovered I'd wired it backwards (so I was adjusting between 45k and 50k, rather than 0k-5k, and just never bothered to twist it all the way around to the far side).

And the damned thing spins. And it changes how fast it spins when I twist the pot.

A few interesting tidbits:

1 - The max supply voltage is 46.6v. Above that it shuts down. Since it was designed for a 36v kart, and 36v batteries are at absolute most 45v when boiling their electrodes, that's sensible. But it means if I want to use a higher voltage than that I'll have to reverse engineer it to find how it determines that.

2 - The min supply voltage is 28.4v. Below that it shuts down. Again since it was designed for a 36v kart, when your lead acids are less than 9.5 volts on a 12v battery they're plenty dead (heck they're dead by the time they actually get to 12v). Shouldn't be a problem for me, an 11series lithium pack is well dead below 33v (3v/cell) anyway. 28.4v is 2.58v per cell. That's already into damaging territory, but my low voltage cutoff is human-decision-based anyway, so I'm fine making the choice whether I want to stop or continue driving. I don't need it idiotproofed except perhaps to actual idiot level (I leave the lights on, ideally it won't murder my batteries).

3 - High resistance is low speed. The documentation didn't actually say which, controller might be built either way.

4 - Below some resistance, [Edited to add: the manual for the Curtis 1204 says 300-4400ohms is the operating range, I presume this is true of the 1206] the motor cuts out. So it goes faster, faster, faster, dead, if you bring the resistance too low. I'll need to have my stop switch sometime before this, as, you'd come to a complete stop while at wide open throttle, ease off a bit and then backflip as it slams max power again.

....

Decision Time:

Do I design the pack for 46.2v and start soldering cells, or, wait for a better controller?

It will be a huge pain in the ass to change my mind later, it involves desoldering every single cell, and likely rebuilding the entire battery compartment since it has to be built around the frame.

On one hand, this isn't the voltage I want to run at. It might not even hit highway speeds and there's no way to find out until I try. That would be a huge disappointment.

On the other hand, I'm getting pretty tired of seeing other riders on the road enjoying their summer, while I have a pile of parts. And I don't have another controller to use anyway.

Maybe I'll split my battery pack into two full sets of 11series so I can rewire them for double-voltage later. If I upgrade the controller down the road, I can design it for whatever voltage and, 46x2=92v is an okay target.

Controller claims 275 amps, 46v = 12,650 watts = 17hp when completely topped up. Should be enough for highway speed if the motor will be spinning fast enough with that load. No way to tell. Not sure if the controller has overcurrent protection either, so maybe I can manually demand 500 amps and as long as it doesn't overheat it'll be fine.

Next up, start making cardboard battery boxes and seeing how/where I can stuff them into the bike.

 

[MattsAwesomeStuff]

Battery Bulk:

My big goal was to get 60 miles (100km) worth of range at highway speeds. I need ~140-200 watt-hours to travel a mile at those speeds.

I have these 4x5 battery holders, an 18650's average voltage through it's discharge curve is ~3.7, maybe 3.75v, cells are roughly 2000mah each (good ones are 2600, I'll stuff in as many of those as possible), so each 4x5 brick of 20 cells is roughly 150 watt-hours and weighs 2 lbs.


Battery Weight:

At a brick a mile, I need to find room to fit 60 of the bricks in the frame (9000 watt-hours). Since I'm doing strings of 11, might as well round it up or down to 55 or 66 bricks. It'll weigh about 120 lbs to do that.

Motor is 92lbs, so motor+battery is 212 lbs, plus the weight of wiring and the battery enclosure.

The original engine according to spec was 181 lbs dry. Plus fluids. Plus exhaust. Plus fuel.

I should weigh the bike now and compare it to the 467lbs it was originally.

In any case, I'm in the right ballpark I'd say.


Battery Sizing Layout:

My first big happy surprise. I had more room in the frame than I thought.

66 bricks is roughly 2 milk crates worth, bulk-wise.

Just looking at the bike, I figured there was no way I'd find room to fit them. I've been putting it off forever, but tonight I dummied up a bunch of empty battery trays and started seeing where they could go.

1 - Below radiator, in front of frame: 4x4 bricks, +4 if I remove radiator. 16-20 total.
2 - Below/between frame, below motor (as in pic): 3x4 bricks. If I double-stack (each stack is 2.5" tall), double that. Brake disc is 6" above ground, so I figure I'm okay with that clearance for city riding? 12 or 24 total.
3 - Beside motor: Sloppily, sticking out of frame, 5 first tier, 8 on second tier, 8 more on third tier. 21.
4 - Above Motor Right: 5x2 bricks. Three tiers. The mounting for the motor sticks out 2" so I have to separate right from left. 30 total.
5 - Above Motor Left: 4x2 bricks. Three tiers, but it's getting ugly to not narrow at the top of the frame. 24 total.
6 - Under Tank: 2x2. Not much space, hard to fit, but room for 4 total.
7 - Above Swingarm Triange: 3x2. 3 tiers. Nice and narrowly tucked, won't interfere with my thighs. 18 total.

Grand total: 125-141 bricks.

Jeez, I only needed room for 66.

Heck, I only have enough weight available for 66 (not 250-280 lbs).

I don't even have 141x20 = 2820 cells = 21kwh of cells. That's almost as much as a Nissan Leaf.


This isn't the layout I'll be using, or even the orientation, it was just the easiest way to slab up cell holders and ballpark the spacing available. If I'm over what I need by at least double... I can afford to make some choices based on cosmetics, not "What used to look like a motorbike now with a bunch of bricks".

I can skip the whole row in front of the frame. The 2nd tier below the frame. I can slim a whole row of bricks off of the left and right sides each. Still have room for 66. Easy.

Expecting my hubris to bite me later, but, for now I have some breathing room.