Classis Vespa Smallframe to EV


100 mW
Sep 13, 2018
Hey all,

New to the board - looking for thoughts / advice / moral support as I start an EV conversion of a 1971 Vespa Primavera. I've read through a few of the various forums, with a lot of attention paid to Jim Delgado's conversion, posted here but mainly over on

The starting point for the conversion is a scoot in absolutely great shape - bought it for my wife to cruise around San Francisco this summer. But after driving a few of the Scoots around SF we decided that electric seemed more civilized than kickstarting and 2-stroke mixing.

I'd like to do the conversion using bolt-in options - meaning that I'd also like to be able to convert it back to ICE if I ever wanted, without having to do any welding. I'd also like all the things I do to be bought off the shelf. Not that I'm afraid of a little fabrication, but I've seen a number of people looking for ways to convert their kits, and a ton of projects what are now just awkward 404 pages (VespaEV, SoundSpeedScooters, to name a few).

To that end I'll try to keep a list of everything I buy and a random journal of the process here:

Parts list:
Random Log:

Here's the little Vespa as it stands today:
I've gone ahead and ordered the same motor and controller that JD used in his conversion ( 6kW should be more than enough to get around San Francisco, and the controller matches up perfectly.

I decided not to get the rear disc brake kit as I'm curious how the regen works. I know that it's not to be used as a brake, and won't even turn on if it detects that the battery is at full charge - but I've never used the foot pedal in the first place, so there's not much to lose, and so much in simplicity to gain.

My biggest question is the swing arm. The one that JD had built is wonderful, and if there's anyone here who knows of a good shop to have work like that done I'm all ears. Bonus points if someone's drafted a digital version (SketchUp?) that can be used for CNC milling or 3D printing :). If I'm serious about making this a kit, a printable swing arm would be pretty baller.
Also, thanks in advance for any help and advice. I'm a total newbie to the forum, and I expect to ask a bunch of stupid questions. You guys are a wealth of knowledge and I can only hope I don't embarrass myself too much :).
I don't have any machining advice, but it'll be cool to see that going. Does it look like the one Reki rides in Haibane Renmei? Or is it a different one?

At one time I had a rusty old '80s Honda (elite?) I wanted to convert, but I went the electric bicycle route instead. (I did use some of the scooter stuff on the bike, though, and a decade later I'm still using the taillight and turn signals, with LEDs in them instead of incandescents).

The hardest part is the battery, because you have to have one that can supply the current as well as give you the range while under the loads it'll see, which in SF could be quite a lot depending on where you go. And then still fit it in the scooter....

As for regen--it *is* a brake, because it will slow you down whenever it is engaged. How much it slows you depends on the controller design (and settigns if it has any), as well as how quickly your battery can be charged and it's present state of charge (the more full it is, the less it can take).

It might not be able to brake to zero speed, though, and most controllers only have on or off, no control over the amount of braking. Some of them do have "proportional" braking, that allows you to either have an analog braking lever (or other control), or you engage the brake lever switch and then use throttle to control the amount.

A few use only the throttle, and actively brake whenever the throttle is moved below the demanded speed/current/power (depending on throttle type of that controller).

There are some controllers that have versions of "regen" that aren't regenerative, or not wholly so. These can be called EABS or other things, and they actively brake the wheel by applying power to it against the rotation direction, so they work the same regardless of battery state (unless it's empty). They brake more strongly than regen, as well. I don't know what options Kelly has.

FWIW, your links don't seem to work; they go to here
and here:
If you instead enter the stuff directly into the thread it'll be better for people following (and trying to help) with the build.

If you need the info off the old sites, some of it is archived here*/*/
She's pretty similar! Same color , although the headlight is on the handlebars, not mounted to the front like that :). Specifically, she's one of these:

Yeah good thoughts on the regen. The controller seems to have several of the options you mention listed in the manual ( Brake switch regen, release throttle regen and 0-5V analog signal variable regen.

Supposedly it also uses "a novel ABS technique" that "provides powerful and smooth regen" which is probably along the lines of the EABS you're describing. My hope is that I'll be able to hook up the now-unused clutch cable to something that spits out 0-5v, and effectively use the variable regen as a variable rear brake. As far as braking to zero speed, I think you're right in that it probably won't go there. That's totally fine as the front brake is still there. And if it ever feels like too much of a weird brake hack, I can always add in the disc kit :).

Thanks for the info on my broken (now fixed) links! Turns out I didn't include the http://.
If it's active braking (EABS or whatever they call it) there's no reason it shouldn't be able to bring the wheel to a complete stop. Ideally it would regen as much as possible and then gracefully shift to active braking.
Yeah, that’s my hope. Fingers crossed it’s fairly intelligent.

I’m also looking for a way to use the existing throttle and clutch cables to modulate the throttle and rear brake (regen), both of which should work off a 0-5v variable input. I’ve seen a couple of solutions from Ballarat Ebikes, especially the Mobipus Moro throttle box:

Unfortunately their site lists this as sold out. Do you guys have any other suggestions?
aaron_mason said:
She's pretty similar! Same color , although the headlight is on the handlebars, not mounted to the front like that :). Specifically, she's one of these:
Unfortunately, like the first post's image, it just comes up as a minus sign in a circle. If you attach the image directly to the post, it'll show up for anyone that can see the post itself.

Similarly, while the links to the pages now work, the linked pages themselves come up blank for me except for a "toolbar" at the top that doesn't do anything. I've seen this problem with various google pages a lot, where they work sometimes for some people, all the time for others, and never for some others. For myself, they work once in a while, except for certain "shared docs" pages by certain users that work almost all the time. If you want to be sure that readers of your thread can read the linked content, I'd recommend inserting it into a post within the thread rather than linking it to elsewhere.

(additionally, external linked images and data can also disappear at any time whenever the site linked to goes away, or changes their TOS like photobucket did. Because members here often used PB for critical images, charts, data, etc, in image form, there's hundreds or even thousands of threads that are pretty well useless now because all that stuff is now inaccessible unless those members reupload it somewhere else and reedit their threads, or pay pb's ransom, neither of which is likely to ever happen).

My hope is that I'll be able to hook up the now-unused clutch cable to something that spits out 0-5v, and effectively use the variable regen as a variable rear brake.
You can setup a pot in a box with a lever arm/gears, and use the cable to move the lever. Of there are pot-throttle-boxes already existing if you want off-the-shelf. If the voltage range can be altered to use an analog hall signal (around 1v-4v) then there's several ways to use a hall throttle mounted inside a box for the cable to pull, or there's hall-based "trhottle boxes" that are even simpler than the throttle versions; easy enough to make if you don't need off the shelf.

For some DIY examples, you can look at Cboy's trike thread for a recent example. There's also a number of others but I don't have a quick easy search term set, because "cable throttle" also finds all the stuff about throttle cables. :(

BTW, make sure you get the "torque" throttle version of the Kelly, or whatever controller you get. They are also called pure-current throttle. These are much more controllable than "speed" or "PWM" throttle control.

As far as braking to zero speed, I think you're right in that it probably won't go there.
It might, if it's EABS.

Regular regen can't, because what it does is basically just take the voltage the spinning motor generates and put it across the battery, so when the voltages are equal, (motor spins down far enough) it doesn't generate any current.

More advanced regen (synchronous) "shorts" teh windings momentarily (flyback) to generate a higher voltage, once it's down to a certain speed, so it continues to generate a higher braking force, but it still has a lower limit to the braking speed.

"EABS" actively brakes the wheel by forcing current out of the battery into the motor against the rotation of the motor; it may also use regen methods at the higher speeds they're effective at, then switch to the toher method at lower speeds down to zero (like mine).

For an example of the capabilities of electric braking:

My SB Cruiser trike is a few hundred pounds with me on it, and much more than that with one of the dogs on it too, and even more with another dog (or more) in one of the trailers. I use "EABS" on the left rear wheel, and regular regen on the right, and between the two they drag the trike to a complete stop pretty quick from 20MPH (fastest allowed here), about two to three trike lengths depending on conditions/load, without using the either set of front rim brakes (there's two on the front wheel, none on the rear). Keep meaning to add them to the trailers, too, but never get around to it--has to change with the new trailer, especially if I haul anything like the piano I did a few months ago...).

Anyway, I'm just using generic cheap trapezoidal controllers, with pretty low current limits on both acceleration and deceleration. I'm sure you can get way better braking out of the kelly, depending on the motor itself. I"m using a pair of MXUS 450x motors, a 4504 on the left and a 4503 on the right; braking and acceleration are much more effective with these wider-magnet&stator motors than previous skinnier ones I've used. (FWIW, acceleration is under 4 seconds from 0-20MPH with just me on the trike).

But one thing you have to make sure of if using electric braking is that the motor's axle is very securely mounted, and that it is not going to subject the axle itself to forces higher than it can handle, repeatedly, or it can break the axle or the dropouts it's mounted in. Been there, done that. :/ Working on a radical way to ensure that never happens again. :)
Hmm. I’ll see if I can fix the photos and links when I’m not on my phone.

Thanks for the info on throttle boxes. Doing a quick search seems to bring up a lot of the same kinds of things, which should be fine. Mostly I’m looking for something that I can just drop a cable into, that doesn’t require a lot of bracket fabrication.

For the swing arm, i’m in total agreement that it needs to be strong. I think I’m going to go the CNC route, and have the arm machined from aluminum. That way anyone who wants to go through this in the future can just sent the file to a machine shop and get a shiny part in the mail. It will also allow for me to 3D print a few plastic prototypes to fit test.

I’m going to start in SketchUp and see where that leads me, but and advice on that front in general would be welcomed!
On the dropout part, it kinda depends on what kind of torque levels you're after, and what the specific motor you're going to use is going to have in combination with the controller and battery pack you're running it with. As well as how that specific motor is designed to do it's torque transfer.

The worst possible way is the wya that almost all "typical" hubmotors do it--via flatted axles, because only a very tiny amount of surface area transfers all that torque--both in accleration, and in braking. If the surface area is small enough, and there is *any* wiggle room, it'll rock back and forth and damage eitehr axle or dropout, until there's enough room to spin the axle in the dropout and twist up / rip the wiring, or break the axle or dropout.

If you do have to use this type of motor, I highly recommend making clamping (also called pinching) dropouts. There's a number of examples around ES, some with explanations of how they made theirs, some with just pics of the finished dropouts. You probably already know that it matters which version of which metal you're using, but I thought I'd bring it up that whatever you use has to be hard enough to resist being dented/etc by the axle as it twists. :)

There's a number of threads about the problems of this design, and some about working out alternate designs. Several of these have been started by or participated in by Justin_LE, of Grin Tech, and an analysis thread of one potential solution exists by Buk___ .

A very few motors (mostly Heinzmann) transfer torque via integrated arms kinda like IGHs (Internally geared hubs) on bicycles do. This is much safer, more secure, and not likely to fail.

Some may use other means.
That’s super useful. I think the motor comes with a torque arm, but if not I’ll be sure to make one. I hadn’t thought of clamping dropouts though - that’s a great idea.

The last thing I want is to have a beautiful aluminum swing arm arrive in the mail only to have it strip out on the first hill I try to climb...
aaron_mason said:
Yeah, that’s my hope. Fingers crossed it’s fairly intelligent.

I’m also looking for a way to use the existing throttle and clutch cables to modulate the throttle and rear brake (regen), both of which should work off a 0-5v variable input. I’ve seen a couple of solutions from Ballarat Ebikes, especially the Mobipus Moro throttle box:

Unfortunately their site lists this as sold out. Do you guys have any other suggestions?

I was midway through converting a honda throttle body and TPS for use on my motorcycle when I found this from Eastgem. Pretty sure it's the same as a Mobipus part.

I'm much rather run something like that (or a cut down throttle body) than mess around with my own DIY solution. I ordered a couple of them in case I decide that I want to set up variable regen on my clutch lever.
So nice to see another Vespa/Smallie conversion thread!

And straight off the bat, that throttle adapter you found is so much nice than the one I bought, and broke, and never got around to fix :D Just ordered one, thanks!

I am extremely curious about how the rear braking will work out on this. it says specifically in the user manual for Kelly's controllers that the regen braking is not to be used as the only brake. And from 15+ years of experience with smallframes, I have yet to try one with a front brake that you can really rely on, so....
Motor / controller arrived in the mail. I’m going to take it down to a local shop to have a tire put on it.

I also think I’ve figured out a neat solution for the swing arm. Instead of trying to fabricate something from scratch, I’m going to try and fit the motor and wheel into a swingarm from a Honda CT70. The dimensions are approximately correct, except for the main pivot, which I think I can widen with a sturdy piece of aluminum tubing. The swingarm are set up for dual rear shocks, which will also reduce any twisting of the wheel. If it works, it will be an easy swap. Best part is how inexpensive the parts are - cheapest swingarms start around $60.

My project for today is to get my workspace set up
And pull the old motor / tank out of the bike to get a look at what has to be done to accommodate an additional rear shock. Seems like there’s structural metal all the way across the upper inside of the wheel well, so maybe as little as drilling a hole and bolting in a second rubber isolator.
Nice to see that you are progressing!

The smallframe chassis has a reinforcement where the shock originally sits, so you will have to reinforce the opposite side too, to accommodate a 2nd shock.

Also, the original location has an indent/pocket where the shock sits, which doesn't exist on the opposite side, so you might have to be a little creative and use two different length shocks :)
Hey thanks for the info on the shock mount. I think I'm actually going to just use the single shock, and beef up the swing arm, as this seems like the simpler path. Also, the Honda swingarm was just a bit too narrow. Plus no matter what I did, it just wasn't going to look sexy. My wife chose this little Primavera because it was the cutest vespa we could find, and it would be a shame to not have a stunning end result.

I've been able to model a few different swingarms and upload them to 3Dhubs (the cheapest of the online quote places), and it looks like it'll cost ~$400 to have one CNC milled from 7075 aluminum. 6061 is about the same price. This is fun as I can integrate a rear brake caliper mount :). I'll most likely just add it above the arm, but hey, I'm having this thing milled, so I might integrate it like this:


A bit overkill, but if the machining costs about the same, why not?
Also, I'm starting to look at sourcing batteries. I'll need a 72v pack, with 15-25AH capacity to make it useful around town, and something that can burst up to around 100A for a little oomph on takeoff. Trying to find something that will fit inside the smallframe is a pain. I've seen a bunch of enticing options on aliexpress, but I simply don't trust them, especially after someone's response about their house burning down.

I've found seemingly more reputable packs at Lunacycle, LA eBike, EM3ev, Ebike MarketPlace and 2Twenty. I think LA eBike's 16" x 5.5" x 3.4" pack ( would fit nicely, straight down in the main space where the carb used to sit.

Do you guys have any recommendations for places to look?

I think I mentioned this before, but I'm trying to make this as much of a purchasable kit as possible for folks who want to replicate the conversion.
So for the past week or two I've been working on the swingarm, brakes, and the battery. My swingarm designs are going all over the place with the consideration of the rear caliper. Here's one of my latest sketches:

Screen Shot 2018-10-18 at 6.43.54 PM.png

However, the biggest thing that I'm grappling with now is the battery, in particular the BMS. I've just ordered a new fuel tank with the idea that I can cut off the top, and then build a custom battery that fits inside. I've bought a few Vruzend 2.0 kits and 80 20A 18650s, and now I need a BMS to run the thing.

Cost is much less of an issue for me - I would like something that is tidy and works well. Looking for 20s (72v) 80A. I love the look and presentation of the Speedict Neptune, but it will only handle 15s.

Do you guys have any thoughts on high-quality BMS?
Made another update to my swingarm design, and looks like I'm very close to having this thing made. I've sent it out to the shop (3DHubs) to make sure everything looks good.

I designed the swingarm to work with the Kelly 10" 6kW disc-brake motor, and included a few holes to bolt on a Brembo P30C caliper that should be properly positioned (with a few small spacers) to use a 180mm disc. I had to move the suspension point a few cm forward from the original geometry, and I might have to do some ride-height adjustment, but otherwise it's pretty much good to go.

I'll most likely add two more holes for brake and power cable brackets, but otherwise to have this made out of 7075 Al is US$450. Here's the current design:

Screen Shot 2018-10-22 at 4.31.59 PM.png
View attachment 1

I'm also excited about building myself a battery. I've decided to run with a BMS, but only for monitoring / charging. I want the controller connected directly to the battery to make sure that regenerative braking is never interrupted. I very much like the look of the Neptune BMS, and the guys seem responsive, so perhaps I'll use one of their 20s devices that are due out soon.

Anyway my Vruzend kit and 18650 cells are shipping soon, and I'll be able to actually get the motor to turn :).

Also, if anyone else is interested in one of these swingarms (I can send you digital files to make sure you're OK with the design), after the first one they're only $200 each to make more.
My swingarm came in! I had to do a little bit of filing on the axle to make it fit, and I need to add a few spacers to get the brake caliper and rotor aligned, but at first glance everything works! I love how hidden the little brembo caliper is.

View attachment 1IMG_1500.jpgIMG_1501.jpgIMG_1504.jpgIMG_1491.jpg
Like I've said on PM, that's an absolutely stunning swingarm, and I can't wait to see how you get on from here!

Have you test fitted it into the chassis yet?
Did they thread the holes for you also? I’d have 3d printed many tests before sending in my final cad file. I’d be nervous it wasn’t right.