Electric Scooters Before They Were The Rage

When I first started restoring my Isetta, I swore off any non-Isetta related projects.  I knew that project was so huge, and had so many different parts to it that it could act as its own heap of projects, and if I was ever going to get the whole thing done, I was going to have to focus.

Somewhere just before the end of the second year of working on the car, my project composure cracked. I wanted to do a big project.  I wanted specifically to do something unrelated to the Isetta.  I wanted to bust out. I did NOT want to be derusting or restoring something.  I wanted to make something new.  Preferably something a bit electronics heavy, something fun. Thus was born the Electric Scooter project.

Now let me just state that I knew I was biting off a fairly large project when I started this.  The electric bike project was going to pale by comparison.  First, with an electric bike you’re starting with a complete vehicle.  You have wheels, a frame, steering, brakes, a place for cargo, in short, a whole host of things that are already there, and all you have to do is not screw them up as you add motors, batteries, etc.

Secondly when starting off on the electric bike, I had specifically decided: no drive electronics. Just a two-speed  relay system that could be wired up in a few hours.  Doing a high current motor driver is a big deal.  There are a lot of tricky issues that come up that complicate the design.  In short, I have studiously avoided projects that require chopping more then a few amps through a motor.  The whole point of the electric scooter project was to face some of these tricky electronics issues (and have some fun doing it of course).

I drew some sketches and decided on an approximate size.  Then it was off to the surplus stores to find a suitable motor.  The motor would dictate most of the rest of the design.  I spent a day digging through various piles of grubby motors looking for a DC motor with enough torque.  I finally settled on a 50v motor which had started life as the spindle motor for one of those ancient reel-to-reel tape drives for computers in the 60’s.  I also picked up two wheels, some sprockets, and some matching bits of chain.  I decided I could run the motor at 48v from 4 12v lead acid batteries.

trombone_slideIt seemed like the first order of business was to determine what kind of reduction ratio to run on the motor to give me some decent torque, but also have a reasonable top speed.   I bent a piece of black pipe into a ‘U’ shape much like the slide of a trombone.  (Well a rusty thick-walled trombone slide.) I machined an axle, welded on some fitting so the axle could bolt across the two legs of the ‘U,’ and welded on a piece of angle iron so I could bolt on the motor.  Now my trombone slide had a motorized wheel at one end.  Still this was a long way from being a scooter, but I was itching to see what kind of power this thing was putting out.  What to do?

Motorized Hand-truck Terrorizes Neighborhood

hand_truck2Then it struck me.  I didn’t need steering or brakes or much of anything to give this puppy a road test.  All I really needed was a few more wheels.  So I bolted the trombone piece to the back of my hand truck, put a board on top to sit on and to hold the batteries.  I stuck the whole contraption into the street and climbed on.  I touched two alligator clips together, and zoom! I was shooting up the street.  Now let me tell you,  I have done a lot of wacky projects on this street.  The electric bike never even got a second glance.  The completed electric scooter didn’t cause a stir, but by gum a motorized hand truck really brings people out.  Within moments I was surrounded by neighbors asking what the heck that was.  I guess my other projects just end up looking like something I might have bought, but throw together a vehicle with clamps, a hand truck, and loads of loose wires, batteries, etc, and everyone’s impressed.  Go figure.

scooterRearDriveCrop2It was clear that the motor had zip, and I guess I lucked out on my guess about the gearing because it seemed to be just about right.  So it was back to the shop to build the host of things needed to turn a trombone slide into an actual electric scooter.  I welded the rest of the structure using 1/4 iron rod to form box sections.  It was strong, but wasn’t about to win any awards for being light weight.  Still it looked nice and was easy enough to do.  The next big issue was the front wheel.  It needed to be articulated so I could steer, and my design also called for another pivot so the whole scooter could be folded up.  What to do?

The Great Bicycle Caper

Now let me tell you I hate bicycle thieves.  When I was in high school someone stole my red Schwinn World Sport right off our porch.   It had been my first full-sized bike, had had many component upgrades, and had been my faithful touring bike on numerous bike tours.  I had gone thousands of miles on that bike, and I knew every scratch and ding.  When it was stolen, I wandered around town in a fog hoping against hope that I might find it parked somewhere. I knew that the thief didn’t appreciate that bike.  Not the way I did.  So you can understand why it took me some time to warm to the idea of The Great Bicycle Caper!

Now one thing I haven’t mentioned yet in the project is that I was not taking this on alone.  My friend and former co-worker Pioneer was also on the case.  He was there from the very first sketches on napkins over lunch.  Now maybe Pioneer wasn’t doing the welding, or surplussing, or electronics design, but it was his enthusiasm for the project that really saw it through to the end.  He was an excellent sounding board and made sure we maintained the appropriate goofy perspective on the project.  He was also the person who talked me into The Great Bicycle Caper.

We had been looking for a cheap donor bike at various Salvation Armies, Goodwills, etc. with no real luck.  For some reason those stores were either out of bikes, or wanted so much for them that it didn’t make sense since all I was really interested in was the front fork.  Now I knew that one would show up at a garage sale at some point, but we were looking for a bike, and the sooner we found one the better. Pioneer had noticed a bicycle frame locked to a tree near his apartment.  He pointed out that the wheels and components had been stripped, and that the frame itself was bent, but the front fork was still OK.  I balked at the idea of stealing even this abandoned wreck, but Pioneer kept after me, working the angle that we would actually be doing a public service by removing this abandoned hulk.   After a while he managed to bring me around, and I began to plot how best to steal this bike.

coneThe facts: The bike was U-locked to the tree.  The tree was at a very busy corner right next to the entrance to a Cost Plus.  The plan:  I knew my angle grinder would make short work of the U-lock, but with an obvious shower of sparks, and we’d need electrical power.  I also knew there was no way we were going to be able to do that unobserved, so going with the “public service” theme I outfitted us with face shields and orange safety vests.  We put out some safety cones, and Pioneer ran an extension cord into the Cost Plus.  We acted very businesslike. My grinder cut through the U lock in a mater of seconds.  We rolled up our extension cord and ambled off with the bike frame.

This would have all been much easier if we’d had some sort of vehicle, but since I only had a motorcycle, we weren’t going to look too official showing up on that.  Also riding a motorcycle while holding a bike frame would have been unsafe.  So we had walked over from Pioneer’s place, and we walked back with frame in tow.  I still feel a bit funny about the whole thing, but I do really think we were performing a public service, and that no one was victimized, so I guess I should just stop worrying about it.

The scooter takes shape

front_fork_cropI used a large lever arm to adjust the set of the bike fork, cut the fork much shorter, collapsed the ends, and cut slots in the flat areas to create a place for the front wheel’s axle to bolt.  I mounted the fork, extended the handlebar’s stem to be much longer, and suddenly the project started to look like a scooter.  As a quick initial test I cut a piece of plywood to act as a deck,
stuck some batteries in, and wired a switch to the handlebars.  No brakes, no speed control, but it was time for a test run.  I climbed aboard and I was off. After a little bit of use the power switch welded itself shut, and I found myself abord a runaway scooter with no brakes!

I leapt off and hauled the scooter into the air to keep it from running away. It was heavy, and the rear wheel kept brushing the ground. Now the question was how to disarm this howling squirming scooter.  I clawed as some wires and eventually managed to disable it.  Now we were having fun!  After that all tests were carried out with an exposed loop of wire which could easily be yanked free. A kind of “nearly dead man’s switch.”  The next order of business was to try a relay controller.  I wired up a big relay and hooked that to a control switch.  Pioneer and I tried this configuration out with a bit more trepidation, but things seemed to be going well.

Eerie Lights shine under the scooter

However, we did notice two things.  One was that when you released the switch, the motor did not cut out immediately, and two as darkness fell, we noticed the occasional eerie glow coming from under the scooter.  Was the scooter possessed?  Was it space aliens?  Nope.  It was the glow of an electrical arc shining out from the relay.  Because the motor was running on high current DC power when the relay opened, it was striking an arc, and electricity was still flowing through the relay even after it was fully open.  Needless to say the relay did not last super long under these conditions.  Scratch one relay.  I’ve often wondered if there is a simple passive capacitor circuit that would keep the arc from starting until the air gap was wide enough.  Who knows?  It was back to the motor control drawing boards.

Boring Technical Stuff About the Motor Controller

protoboard_cropNow I had sworn off high current chopper-based electronics projects for the same reason I don’t do high frequency digital electronics.  To many spooky issues involving high speed transients taking advantage of parasitic inductances and capacitances.   The circuit designs can get pretty finicky, and you end up putting a lot of ground planes around and just hoping for the best.  Also when you are doing high-current stuff, you end up having to deal with more expensive components, heat sinking, thermal runaway, great big gate capacitances, larger explosions, etc.  In short it’s a much bigger pain.  There are some very nice websites about these issues.  The Q4D folks have a very helpful site that talks about some of this, and SGS Thompson has a number of very interesting (well for motor control geeks) technical papers relating to this.

scooterSideView_cropI spent two months making various prototypes of my motor controller before building one that wouldn’t blow up when I actually tried it under full load on the scooter in the street.  (That’s two months when I didn’t have a day job, so that was a LOT of time.)  However the design I have in there now is pretty much exactly the design I started out with before I had decided I’d try and make all but the power MOSFETs be components that you could get at Radio Shack.  What a mistake.  I was building push pull stages to play tug of war with the MOSFET gates, and those stages always ended up having crazy noise issues, or suffering from thermal runaway, and just popping right off the board.  So I finally caved in and bought a SGS Thompson gate driver chip.  It pretty much worked right out of the gate.  I didn’t even blow one of them up, and I think they cost less than $2, so it was dumb not to be using them from the outset.  Still I learned a lot.

controller_box_cropI etched a number of different designs, but somehow even though they worked on the bench on protoboards under pretty heavy load (sticking a 2×4 against the rear wheel), they’d still blow up once I had them etched and was testing them under full load.  Finally I managed to build one that didn’t blow up.  It has some issues where when it goes to 100% on it has a lot more power then at less then 100% so it feels as if you have variable slow speed control, and then a kick of extra power when you go full throttle.  Which is fine although I’d like to know what is really causing this. I boxed the whole thing up in an aluminum box with a big heat sink sticking out one end.  The nice thing is that it’s all in one module, so I could pull the controller out and stick it into another project if need be.

Making The Throttle Handle

hand_grip2_smallAfter the motor controller was working, I needed to make a nice interface for controlling the variable resistor that was the throttle.  I decided a “motorcycle” style twisting hand grip would be swank.  So I cut off part of one of the handle bars just inside the hand grip. This part was to be the twistable throttle. I measured the amount of twist needed to go from idle to full on my motorcycle, and I made that be the amount of twist available on the hand grip.  Now Mark has been dabbling in clock restoration, and he was a great source for some spring steel that I used to act as the return spring for the handle.  Finally I used a toothed belt to hook the twisting grip to the potentiometer.  All in all it makes a nice motorcycle hand grip, but it was a fair amount of work to make.  If I were to do it again I might opt for a snowmobile style thumb lever or some such at least as a first pass.

So It Goes.  What About Stopping

break_super_closeup_cropThe next thing to consider was a brake.  You would think I’d have made one of these earlier, but where’s the fun in that?   I got a surplus stainless steel disk, and mounted that on the rear wheel.  I had been planning to use the spent brake-shoes from my motorcycle in a custom caliper made from some chunks of AL plate, but there were space issues, and it seemed like a complicated build.  So I eventually opted for an off-the-shelf caliper that I got at a go-cart shop.  It was fairly small, and was already fitted for cable drive.  So that made things a lot simpler.   I brazed various cable connections on the scooter just like a real bike, and ran the cable up to a brake lever on the handle bars.  Suddenly trying out the scooter was much less scary and entered the realm of something casual guests could try.

Dressing It Up

right_side2_cropSo then there were the countless cosmetic improvements.  I made a center stand for the scooter.  I made a clear plastic deck for the scooter with a nicely patterned grip tape surface (thanks to the local skate shop).  That and a nice paint job, and it was looking pretty good.   The scooter still sports a fairly silly plywood “kick tail” that I added for ergonomic reasons.  It turned out that otherwise the foot you had at the back would be at a somewhat uncomfortable angle. Also I still have never made a build-in locking mechanism for the folding action of the scooter.  I’ve always just used a C clamp.

Speed, Range, and Showing It Off

drew_riding_cropWell at this point, it was pretty much ready for folks to ride.  On several occasions my friends came over to check it out and ride it around.  Top Speed?  Well I’m not entirely sure; my best guess is something like 12mph.  Weight?  The bathroom scale claims it’s a chunky 56 lbs.  When I built it
weight was not the highest priority.  I’m sure it could shed quite a few pounds with a cast AL frame and lighter handlebar stem and tires.  As it stands, it’s quite rugged, but no featherweight.  Most of the weight comes from the 4 lead acid batteries and the motor.  Not much I could do about those without a full redesign.  So now comes the question of range.  How much range did it have?  Everyone wants to know about the range.  Honestly I just don’t have a clue.  In all the times it’s been ridden we either didn’t ride it enough to run down the charge, or it wasn’t fully charged to begin with.  So I guess the answer is that it had enough range to out last our attention span for riding it up and down the street or around the block.  It just isn’t really comfortable enough to want to take it for range trials around and around and around the block.





Building a Custom Waffle Iron 

Sometimes a project gets started with something as innocent as having a few waffles.

I was at my uncle’s house having breakfast, and I noticed that he was using a quaint electric waffle iron.  It looked like it dated from maybe the 30’s or 40’s, and it was pretty cute.  The thing that really caught my interest was that the waffle “iron” had aluminum grids on it.  I realized that I could probably buy an electric waffle iron, make some new grids for it, and turn it into a personalized waffle making machine.  That thought put quite a sequence of events into motion.

The first order of business was to select the right candidate for conversion.  I went to a number of thrift stores, but most of their waffle irons didn’t have much character.  However, I did discover a few things about electric waffle irons.  There seemed to be a wide variety of ways the grids were mounted and heated.  Some grids were clipped in and could be popped out for cleaning; others had the heating element and grid as a single integrated unit. Clearly I needed to steer clear of those.

So it was off to eBay for me.  What did I find?  154 waffle irons, almost all of them different.  Luckily I could ignore the cast iron ones which were designed for use with a coal stove, and that cut the list down to something like 50 electric irons.  After much indecision, I eventually bid on an iron which I thought looked nice and seemed to be in excellent shape.  I couldn’t tell from the pictures if the thing was going to be easy to retrofit with custom grids, but it seemed worth the risk.  No one seems to take pictures of the inner workings of the waffle irons they’re selling on eBay. Not surprising I guess.

I won the auction, and a few days later the iron was at my house undergoing a thorough cleaning and disassembly.  I was lucky that the iron had heating elements which simply screwed onto the grids.  Removing these elements was a bit ticklish because I could tell that the fragile heating element would be very easy to break, and there’d be no real hope of re-attaching it if it did. So I removed the heating elements from their nest of hook-up wires, which felt a bit like defusing a sticky 50-year-old time bomb.

After the heating elements were out, it was nice to give the rest of the  iron a very thorough cleaning.  WD-40 did a good job of cutting through the ancient baked-on oil and flour mixture that filled every nook and cranny.  I wouldn’t recommend using it on any section that you can’t later de-grease, or on parts which are going to be coming in contact with food, but for cleaning parts like the hinges, it really did the trick. I was thinking about giving my Mom a little fire extinguisher as one of the auxiliary gifts.  95% as a joke.

The plan was to cast new replacement grids out of aluminum so I needed to come up with a pattern. The original grids had been die cast and had various knobs and bumps that were drilled and tapped to take mounting screws.  I realized it would probably be a lot easier to use the back half of one of the actual grids as the pattern for the back of my new grids. I used auto body filler to fill all the waffle grid holes and remove some of the extraneous details which would make it harder to cast in a sand mold.


What about the actual grid design?

I thought it would be best to make a board-mounted pattern and design it so the deepest parts of the mold were less then halfway through the board. As long as I kept the outline of the design symmetric I could use the same pattern to cast both the top and bottom grids.


The plan was to give the waffle iron to my mom for Christmas. She and her husband Stan have been building a victorian home in Historic Rugby, Tennessee. Since Stan likes Calla Lilies, they named the house “The Calla Lilly Cottage”.  I thought the house would be a nice theme for the grid design, so I decided to make the Calla Lilly Cottage Waffle Iron.

I drew sketch after sketch trying to make a design that embodied Calla Lilly Cottage but was simple enough to act as a practical waffle iron.  The design had to have a balance of high and low regions so that the waffles wouldn’t be too fat and doughy or thin and crispy.

Eventually I settled on a design that produced a simple house silhouette-shaped waffle with a calla lilly embossed on it.

Carving begins

I decided to lay out the design on thin poplar, cut out the rough high and low areas, carve in the details, and then glue and fillet the whole thing back together. This was by far the most time consuming part of the construction.  I put more then the usual amount of draft on the pattern because not only would the waffle pattern have to come out of the sand, but waffles would have to come out of the iron, so I based my draft on the amount of draft on your average waffle griddle squares.

Casting begins, but time is running out

Finally the pattern was finished.  The iron was cleaned up and ready to go, and the stage was set to cast the new grids and get a move on.  I made a board-mounted pattern with alignment pins so I could use the original waffle iron grids in conjunction with the calla lilly pattern. This required making a flask with some very long alignment pins so the flask could maintain its alignment both with the board in place and after it was removed.  So I made a new flask with long dowels, but that was the cause of the first big failure of the project.

The Christmas holiday was fast approaching, and I still had a long way to go with the project.  The special flask made it so I could only cast one iron at a time, and that meant pretty much one casting per night. It was down to the last week, and I really needed to get two good castings right out of the gate in order to be able to have enough time to finish the project.

Then disaster struck.  I rammed up the mold as usual, separated the flask, and removed the pattern. So far so good.  The cope (the top part of the flask) was pretty darn heavy.  The grids are perhaps 7″ across, and that meant the cope was maybe a foot on a side. With it rammed full of sand it weighed as much as two cinderblocks.  As I was heaving it back into place and aligning the unfamiliar alignment pins, I (unbeknownst to me) brushed the face of the sand with one of the alignment pegs. This cut a channel from the mold cavity to the outside world.  And I didn’t know.

Fountain of molten aluminum!

So I fired up the furnace, and melted a large charge of aluminum.  I used cut up pressure cookers and other things that I felt were probably semi-food-safe flavors of aluminum. I got everything up to temp, skimmed the dross off as usual, and then went to pour.   As I poured, a stream of aluminum shot out the far side of the mold and began to pool on the ground.


I stopped pouring into the mold. I poured the remaining aluminum into the muffin tin ingot molds, but I could tell that the casting was shot for sure.

The interesting thing about the resulting casting was that you could see the exact path the the rouge dowel took as it cut that channel.  The aluminum frozen into the channel had the exact shape of the dowel’s path.


The other interesting outcome was that the form created by the spilled aluminum was rather beautiful.


The aluminum was on concrete, but the puddle actually spanned a gap in the concrete that was filled with a piece of wood.  The wooden strip burned and produced a nice colored line across the form. I’m thinking about making that into a lamp.

As far as casting screw ups go, no one was hurt and I got two interesting pieces out of it, but it didn’t exactly help my time table.  The next two nights of casting went without a hitch, but I still had the task of cleaning up and machining the castings.

I had designed the main casting with a sprue that was centered and perpendicular to the face of the iron.  I did this by taking advantage of a long mounting screw that was centered on the original waffle iron grid. This gave me a handy shank to mount in the lathe so I could clean up the casting and make it fit exactly into the waffle iron body. Now you know why I selected a round iron rather then a rectangular one.  This turned out to be a god send in terms of getting the whole thing to fit.  The cast plates had warped a little as they cooled, so it was nice to have some extra meat on the casting that I could face off on the lathe so they would fit together nicely.  (There should be a little space between the plates so that steam can escape from the waffles.)  If I had it to do again, I’d probably leave a good 1/8″ of extra material on there and just face off the whole thing.

In the end I took one extra day of vacation and worked from 9am to about 2am. Fitting, machining, sanding, filing, drilling, tapping, wiring. *phew* Finally the iron was finished.  A few hours later I was off to the airport with my waffle iron in a carry-on bag.  There was no way I was going to check that baby!  I was hoping to get to see an X-ray of the iron at the security gate, but no such luck.

Where there’s Smoke there’s Wire!

Of course, I hadn’t had much time to run a test of the iron.  I had checked for shorts and continuity across the coils, but that was about it.  On Christmas Day the iron was put to the test and FAILED. The iron would not heat up.  I quickly discovered that the cord that had come with the iron was a dud. All that testing of the iron’s wiring, and I never checked the cord.  Crazy.  So we were off to Drogan’s for a new cord.


Was that the happy ending?  No.  As the iron came up to temp, we noticed some smoke!  I had given my mom a kitchen fire extinguisher as a warm-up gift for the iron, but I hadn’t actually thought we might need it!  Actually it wasn’t that big a deal; it was just a thin plume of smoke rising from the top grid.  The ancient wiring was slowly smoking because it was still somewhat soaked in oil and pancake batter.

Argh!  This fix required another trip to Drogan’s for a few lengths of high temp appliance wire and some crimp connectors.  I spent an evening doing the ticklish bomb defusing dance trying not to break the heating elements while attaching replacement wires.  The iron was reassembled with fresh new wiring, and finally it was taken on its first fully successful waffle run.

It works!

We set about making a whole pile of waffles. The waffles came out easily and were lovely. For added fun, the pattern in the waffles fills with butter and syrup.  Mom, Stan, Cheryl, my sister Inge, and our friend Heather all tried a few waffles.  Finally the Calla Lilly Cottage Waffle Iron was fully operational.


It’s funny casting a waffle iron grid because the iron grid, in turn,  is used to cast the food batter.  Now mom can have fun casting waffles in the kitchen. I had originally thought it would be fun to make more custom irons, but I think I’ll put that off for a good long while.  It is rather a crazy amount of work, but can be pretty rewarding.


I was thinking it would be somewhat easier to make two matching “mirror image” grids by sandwiching two thin boards together, cutting them at the same time with a scroll saw, and mounting them on two different boards.  If the pattern were a simple two-level design without any carved bass relief, it wouldn’t be to hard to knock a pair out, glue them up on two boards, and then fillet them with filleting wax. That would be a lot faster then my approach.  Maybe in a few years I’ll have forgotten enough about how much work it was that I’ll be willing to embark on another one.

The original page in the Internet Archive.