[rishimaze]

Hi everyone,
I've been doing EV builds since 2015. This is my latest one. These scooters came with a 6 phase hub motor. Who made them is unknown, but they are well made and were designed in Japan. It's possibly the best hub motor ever made. These motors can run at well over 120v and don't suffer from significant iron losses at higher eRPM's. These are images of these motors on Zapino and RMartin scooters. They all came with this large controller that has dual 12 fet controllers inside it.

 

[rishimaze]

I knew these motors could handle a lot more wattage than they were being used at originally. The dual 12 fet controllers were good for 1500 watts each for a total of 3000 watts. The technology was from 2008 or 2009 and controllers have improved a LOT in the past 11 or 12 years! All of this combined with 5 heavy SLA's made for tepid levels of performance on a scooter of this size. My Zap topped out at 47 mph and only got there down a long hill with the wind at my back. On level ground 40 mph was about it.

First thing I did was make a 16S, 40Ah LIPO pack for it so I could remove the stupid heavy SLA's that came in it. In removing those SLA's I saved about 100 pounds and had no reduction in capacity or range. I rode around on my LIPO pack for about 100 miles on the otherwise factory setup. This was to get to know the Zap scooter and to see if there were any other issues to deal with. It soon became apparent that barely getting to 40 on level ground would bore me so it was time to upgrade...a LOT!

 

[rishimaze]

I wanted to move to 82v which would get me more speed and since I can build battery packs, this was not really a challenge. I took apart that 16S4P LIPO pack and rebuilt it into 20S4P with a few more Multistar packs. I also had a pile of A123 LIFE cells so I welded up a 24S5P pack (12.5Ah) from them that filled the space under the floor that used to hold 2 SLA 12v batteries and the original DC-DC. A third space behind the seat held another SLA so in there I put together an LTO pack in 28S1P for about 11Ah. All 3 packs since they are different lithium chemistries are built to the cell counts so that their output voltages match up. They each had a smart BMS on them.

This is the LIFE pack being test fitted in the space.

This is the LTO pack in its space behind the seat.

Here's the 20S4P LIPO pack on the bench getting all its cells balanced with an RC charger before use in the EV.

 

[rishimaze]

Later on the LIFE and LTO packs after about 100 miles of use got replaced with 2 LION packs. For the weight, size and capacities of these 2 packs LION was a much better choice and I had the cells to build with. I origianlly used these other chemistries just to see if they were adequate or not, but once I was riding around it didn't take long to realize that I needed a lot more battery capacity than they could provide!

This is the LION pack under the floor that replaces the LIFE pack. Capacity for the LIFE pack was around 12.5Ah and the LION pack is about 40Ah. The space used and weight are almost the same as the LIFE pack so this is a HUGE capacity gain. The pack is made up from 10S4P mobility scooter packs. The BMS's on them are defective so hundreds of brand new mobility scooter packs got recalled and replaced. I was able to buy 12 of them for $1.40 per LG MH1 cell. This is crazy cheap for what is one of the best 18650 cells out there! To make the scooter packs fit best, I stripped off all the extraneous stuff and cut them in half. It is 20S12P. The green coating is PET or Poly Ethylene Tape. It's quite resilient to abrasion, tearing and weather. I'm still waiting on the XiaoXiang smart BMS to arrive for this pack.

The LTO pack was replaced with a 100% built from scratch pack. I used 4800 mah 21700 cells I had purchased. They are supposedly the same cell that Tesla uses in the Model 3 car. They do test to 4800mah and for the price of $2.50 per cell, I couldn't pass them up even if they weren't legit Tesla cells. The pack is 20S10P for about 44Ah.

The space needed a floor in it and the cells needed to be optimally fit into the space. This is the final mock-up that used the rear space best.

I then made a wood frame that holds the cells tightly together in that shape so I could glue them together. The placement of the cells up or down probably looks somewhat random, but it is the optimal layout to group them together in 5 cell groups in each layer. The left sub-pack is about to get glue between the cells and the right one is mostly cured and then I wrapped it in PET to make sure it stayed together while the glue setup completely.

The bottom side of each 20S5P sub pack after spot welding was done. I forgot to take a picture of the top sides. I used .2mmx8mm nickel strip for this pack. My estimates based on total lengths and parallel sections of nickel are that across both layers of cells and across both sides of each layer I am making a total of 20 watts of heat at 100 amps current draw. This could have been reduced by adding more nickel, but this is also 5 watts per side of each sub pack and almost not worth bothering over.

Both layers got balance and power wires added and several layers of craft foam between them to create a padded and electrically isolated layer between the sub packs. The entire thing got taped together with PET and more layers of craft foam sheet added to the outside of the entire pack for protection and water proofing.

Each sub pack has it's own set of balance wires which terminate in 6 pin JST connectors.

The pack getting a test fit before it gets a final close up.

The completed pack with a XiaoXiang smart BMS on it. The only place water can get in is where the power wires exit and those spots have been glued closed around the wires so this can't happen.

 

[rishimaze]

Nothing got built in the order I present it. I'm trying to present logical sections of the build instead of a timeline of what was done when. For example the dual 24 fet Nucular controller set up happened before any of the 20S LION battery packs were built, but I'm presenting it after them. Oh well!

There is fairly new group of smart people in Moscow Russia that started making 100v FOC motor controllers. I have been aware of them since they first showed prototypes in 2018. As soon as they had a 12 fet controller ready for release, I bought 3 of them. Same for 24 fet controllers. I have since bought 5 more of the 24 fet controllers. Anyway Nucular is a good bunch of people and their controllers are pretty good too!

I knew I needed 2 controllers to run my 6 phase hub motor and I knew I wanted to see how far I could push the motor...so I dropped $1200 for 2 of the 24 fet Nucular controllers so I could find out. They are rated for 300 battery amps, 500 phase amps and 20kw. This seems like a lot of money until you consider that everything else in FOC at these power levels costs 2X more per controller. These controllers are a bargain!

They communicate via CANBUS and with a bit of wiring and setup up to 15 and one LCD can all "talk" to each other to share control information. I always use waterproof connectors and so I made up a Y cable so that the 2 controllers and 1 LCD could all communicate with each other.

Hall signals, CANBUS and control signals all get there own IP68 connector on each controller. Phase and battery wires get terminated in 6mm bullet connectors. There's many hours of work in connectors here!

Once that was all done, I could bench mount the hub motor and start testing the dual controllers and the motor together. I originally set up each controller to run the motor individually and then later connected them all together to run the motor from both controllers at the same time. Setup and making everything identical in both controllers was a lot of work!

In the end, both controllers were talking to the same LCD and working as a single unit with the motor. Whew!

Here's the dual controllers in place on the scooter. This isn't quite final as I have a large heat sink I want to mount to them to aid in cooling. I intend to run the controllers failry close to their limits fairly often so better cooling than the shells provide is a good idea. I see 250 battery amps on the controller LCD all the time. This is both controllers amp draw added together or about 20kw right now. Both controllers have reported peaks of 309 phase amps.

Closeup of the 2 sets of hall cables and all 6 phases.

 

[rishimaze]

The Zap scooters were designed to run dual voltage. The motor and controllers ran from 60v and the lighting from 12v. I prefer this to the much cheaper and less useful single voltage systems you commonly find in 48v electric scooters. However since I was moving to 82v, the 60v DC-DC converter needed to be replaced. Also, I was adding a more powerful LED headlight and I needed more amperage. I also swapped out all the incandescent bulbs used everywhere else for LEDs. I plan to add a few more LEDs to the sides for better visibility.

The new DC-DC is working great. The battery pack under it has changed a couple of times, but the DC-DC is still in the same location. From the factory, they are placed under the floor, but I used that space for a battery pack.

 

[brian_]

Later on the LIFE and LTO packs after about 100 miles of use got replaced with 2 LION packs.

I realize that you understand what you're talking about, but using incorrect terminology will just confuse other people.

Both LiFePO4 (or LFP, which you're calling "LIFE") and LTO are lithium-ion cell electrode chemistries; "Li-ion" (which you calling "LION) just means lithium-ion, so LFP and LTP are both Li-ion. You're saying that you replaced Li-ion with Li-ion, which of course doesn't mean much. Apparently you replaced the LFP and LTO with some other lithium-ion chemistry, such as NMC or NCA. I realize that the marketing information for commodity 18650-size cells doesn't normally include any information about the actual chemistry - it's just something with about 3.7 V nominal voltage.

 

[rishimaze]

I was not able to find DC breakers capable of handling 600 amps. Originally I used 2 Chinese E9 250A breakers...which are garbage. I was seeing regular peaks of 250 amps across both controllers and I really want to use them a lot closer to their limits. I found some Optifuse 300A DC breakers to match the controller current limits. They are much better built than the Chinese ones and Optifuse under rates their products. There is a 300A breaker for each of the 2 Nucular controllers. On each breaker is a precharge resistor so the controllers can come up to pack voltage gently. To the left is an XT90 connector for charging. I use them universally for charging ports since they are robust and cheap and I can charge at up to 90 amps with them. To the right is the breaker for the DC-DC and the 12v system. This is all inside the very front of the seat compartment.

 

[rishimaze]

The underside of the seat compartment has a few additions too.

Dual watt meters so I can monitor each controller independently. I chose meters that used an inductor instead of a shunt to measure current. THe wires to the controllers have changed since I took this picture. They used to be 8 awg and are now 6 awg to each controller and I used silicon wire instead.

The Dashboard: Dual watt meters, Nucular LCD, GPS bike speedo and magnetic cell phone mount.

This is car audio stuff, but it works fine at higher voltage and since it's just a solid block of brass with some set screws, does well at the amperage I'm putting through it. All battery power comes here before going to breakers for specific things. All wires that terminate here have their bare ends tinned so they hold there shape and make the best electrical contact possible inside the terminal block. All 3 battery packs terminate here via their own set of 8 awg wires.

 

[rishimaze]

That's the major electrical stuff. On to mechanical things I did.

The factory tires were installed in 2009 so the first thing I did was replace them both with new Kenda tires. This needed to be done, but they are rated for up to 60mph and right off the bat I was seeing 75mph. Also, the rear tire was wearing quickly thanks to the 20kw of motor torque on that patch of rubber. I really needed a better tire rated for more speed, power and I wanted to ride year round.

I found some Dunlop D604 tires that fit the bill and I wanted to convert to 12" since that got me better tires and better handling. They are a good bit taller and wider than the 10" Kenda K413's I put on the factory rims.

I also found a set of 12" cast aluminum wheels that I liked. Mounting a new tire, getting the wheel centered in the front forks was all pretty easy to do.

Factory front end...eww!

New front end...much nicer!

 

[rishimaze]

The front brake got upgraded too. Braking was not that strong and with the added power and speed, this was a significant weakness. With the 12" wheels and tires, this problem would only be worse yet.

The Zap has 160mm rotors front and back. I found a 207mm rotor to use in front. It didn't fit the bolt spacing on the wheel hub so I used a rotor that did fit it and that became an adapter to the larger rotor. The ID of the new rotor got machined off to keep weight down. I almost completely removed the red section. This is much larger and makes the same braking force from the caliper a lot better at stopping the EV.

Both rotors bolted together.

Since there was a larger rotor, the brake caliper also needed to be spaced further out. This is on the 10" wheel, but the same bolt pattern is on the 12" wheel so there was no rework here. There's a large gap to fill between the old mount points and where the caliper needs to fit now. I machined a block of 6061 aluminum to both spread the gap and to properly space the caliper for where the new rotor sits so they line up with each other pretty closely.

The final product on the new cast front wheel. Braking is greatly improved thanks to the larger rotor. I got to "test" it for a scary couple of seconds last week and it performed beautifully in preventing hitting a car that pulled out in front of me!

 

[rishimaze]

The majority of scooters from 2008 and 2009 that had this hub motor on them, used a 12" or 13" wheel except for Zap with a 10" wheel. I had bought a set of 12" cast wheels with the intention of replacing the 10" rear rim with the cast one. This took a bit of work to accomplish.

The center section needed to be machined out at the motor diameter. This was later enlarged another 5mm, but I needed to be able to slide the rim snuggly onto the motor so I could transfer the motor bolt pattern to the new rim and retain concentricity between the motor bell and the rim. The aluminum here is 10mm thick. I as glad to see no voids or air bubbles in the metal as I cut away material. Chinese stuff can be dodgy sometimes!

After I had drilled all 12 bolt holes in the rim, I did a test fit to make sure it was all perfect...and it was! It's so nice to NOT screw up! This was all done with a spare motor I have so I didn't have to take the scooter apart until the new rear rim was ready.

Another Dunlop D604 was mounted to the rim and filled. These pictures were taken moments after the tire bead was leak tested. One thing I found out when test mounting the rim was that standard vertical or even most 90 degree valve stems were too tall to fit around the motor bolt flange. In the first image you see a really short vertical valve stem. I couldn't get a fill nipple on the valve stem as it was too close the bolt flange and the bolt head that was right there at the valve stem. I had to take off the tire, pull the vertical valve stem and install another one. This is no easy thing to do! I found some really low profile 90 degree valve stems that worked perfectly with the space constraints. Remount the tubeless tire, fill, leak test...times 2...LOL!

 

[rishimaze]

The motor shell got some machining too. I had already seen what happens when I run these motors hard. They get hot if they can't breath. The motor I used to test fit the new cast rim has bad halls in it because I cooked them. I decided that would not happen twice and there was no way I'd get stranded on the side of the road due to an overly hot motor. A detail that became apparent was that the factory temp sensor didn't work like I expected. It was a temperature switch, not a sensor! Since I had a spare motor, I took it apart, replaced all 6 halls and installed dual PTC 10K sensors in it. This meant there were no assumptions about how good those 11 year old halls were and I had a temp sensor for each Nucular controller.

Don't let anybody tell you this is a hubmonster. That's a made up name from a single individual in Costa Rica. He didn't make or design or in any way help in their production. He's also reticent to say who really made these motors and the model number. He's deliberately withholding this information which IMHO is silly. I have searched and so far not found the manufacturer so I can use the proper/official name and model number for these motors...not some random made up name by a dude in Costa Rica that had nothing to do with their production.

In this picture you can see the white triangular temperature switch. I removed it! In the hall wiring were 6 wires per set with one wire doing nothing. I guess the factory design just wanted to shut down the scooter if the motor got too warm. I want each controller to be able to independently cut back motor power as the motor warms up if needed. I needed dual temp sensors in the stator windings.

I reused the original wires, replaced all 6 halls and installed 2 new PTC 10K temp sensors. This was all temporarily held down with electrical tape while the thermal glue cured.

Next step, since the motor was all taken apart was to vent the sides. The removable side plate is cast aluminum. The motor bell is cast steel of some kind.

The aluminum side plate machined very easily and quickly. That groove inside is not supposed to be there. A pin that retains the stator on the core had worked loose and was scraping away at the side plate. Eventually it would have worn through and that would have been BAD!

The motor bell was much harder to machine. The casting material was super grabby and it would bind up mills with no warning! I snapped off several 1/4" mills. It was a royal PITA to cut these slots! Much cussing later, they were done. While I had it all apart, I painted the magnet ring with electrical paint. The outer surfaces got a nice coat of blue paint to hide all the scuffs and scratches the original paint has in it.

BLDC and PMAC motors usually have neodymium magnets in them. They are quite strong and in larger motors like this, there is no way to controllably take apart or reassemble the armature and stator with muscle power alone. A 3 jaw puller is essential for control so things don't come crashing together and magnets get broken or windings damaged.

Here's the motor all assembled again. I did consider coating the windings in electrical paint, but chose not to since they are laquered. It got new bearings too. The old ones were fine, but hey it was apart...why not?

 

[rishimaze]

The motor got vented long before the new back rim was finished. I rode around on the 10" rim for a good while before all the back end parts started coming together.

A detail that quickly became an increasingly large concern was the motor axle nuts were always coming loose. Then there started to be a thumping sound coming from the rear wheel area. This was all due to the axle shifting back and forth in the drop outs and torque arms and wallowing out the steel.

This is one of the factory torque arms, It is clearly twisted by the rotational forces applied to it by the motor. At 3kw, it was fine. At 20kw...not so much! With strong regen, they failed really quickly.

The drop outs which are 1/4" thick steel were also getting deformed. There was no tight clamping of the motor shaft flats so any little amount of rotational movement was slowly deforming the steel. I noticed this problem after about 70 miles of riding around. Now the axle nuts coming loose all the time was obvious!

I made these torque arms out of 1/4" thick 4041 steel. They don't clamp onto the axle flats, but they were a huge improvement over the factory torque arms. Since the drop outs were damaged, these torque arms were carrying all the load and after another 70 miles or so started to deform despite being a better grade of steel than the factory parts. They just could not hold 20kw long term on the small portion of the motor shaft flats they could get at. I call them "torque arms v2".

 

[rishimaze]

So after much pondering and thinking about how to do it better and because the new back tire was larger than the factory one, I needed to shift the rear wheel back some and also hopefully permanently deal with the motor torque issue. Hence...Torque arm v3! They are made from 1/2" thick HRO steel and clamp onto the entire axle flat surface. They have dual M5 screws to close up any gap between the shaft flats and the torque arms so there is no chance of movement between them. They also got 4 M6 screws into the swing arms instead of 2 like the v2 arms had.

Rough cut from 1/2" steel plate with a hand held DeWalt band saw to make 2 "rectangular" pieces.

I super glued them together with 2 sides that matched up fairly closely the same and then machined them like they were a single 1" thick slab.

When they got to this shape and had the bolt hole pattern that matched up with the existing bolt holes in the swing arms, I broke apart the super glue. This was the last time they were glued together as the next operation required they be machined separate. The rectangular hole is the same dimensions as the motor shaft and the 2 flats on it. The long slot is there to make space to squeeze the rectangle together on the shaft flats. 1/4" steel does not flex easily. It needed to be fairly long above the rectangle hole to get it to flex inwards at the hinge hole.

Next step required machining off 1/4" thick areas of the steel so the new arms could be inset in place of the original drop outs and be able to clamp onto the entire motor shafts on both sides. I was impressed! I got the entire first torque arm thinned down with the same mill I started working the steel with. Just about 1 minute before completing the first torque arm, I finally snapped off that mill. By then, it had about 12 hours of use on it and was worn out anyway.

I eventually need to get down to the scribe line, but here I am rough cutting away waste material.

The completed torque arms. They look rough on the milled surfaces, but that's just milling marks and not really horrible machining. End mills don't exactly leave gorgeous surface finishes and especially not when used on a drill press!

The swing arms needed some adjusting to account for the new torque arms and the drop outs were trashed anyway. Cut cut cut! Since they were off the EV, I ground off a bunch of ugly welds and a couple of brackets I would no longer use. Then they got a new paint job. It was interesting trying to secure them to an XY table so I could machine away waste steel.

The new torque arms in place for a test fit.

Once I was happy with fitment, they got spray painted. Looks so much better now! I should have taken a wire wheel to them to see if I could brighten up that dull finish some, but oh well...I knew they were getting painted so I didn't really care too much.

Mounted on the EV. I sure hope these things last forever! They have maybe 20 miles on them so far...TBD.

 

[rishimaze]

Finally! The rear wheel reveal. I wanted to show all the stuff that went into the rear wheel before I posted pictures of the basically final result. I still need to make a rear fender for it. These pictures show the v2 torque arms in place. Other than needing to fix up some dings and scratches in the motor paint, I think it looks great!

A detail I didn't think about was clearance to the battery box with the larger rear tire and rim. The v2 torque arms were made way before the new rim and tire were finished. They don't shift the wheel backwards to account for this. This is the v2 torque arms in this image.

Clearance at the back tire is much better with the v3 torque arms shifting the wheel back 1/2". It looks like the tire will rub, but the rear shocks bottom out before that can happen. It's close, but not close enough to cause a problem. I deliberately rode over curbs to see if it ever rubbed.

 

[rishimaze]

The front shocks have always been much too soft for my tastes. This is pretty common on scooters that front shocks are not stiff enough. They are hydraulic, but not adjustable for damping or springiness. All you can do with them is add or subtract hydraulic fluid to get a "feel" that's going to suffice. With the larger tire, they also collapse too far and the tire can graze the plastic cowling with really hard braking or going over a curb. All of this needs to be improved!

In investigating how to improve them, I discovered I had a leaking seal so they needed work no matter what! I also learned that more hydraulic fluid makes them less springy and less "collapsible" since there is less compressible air inside. To that end, I took them apart, cleaned out all the old shock fluid, found new seals and got new fluid. They will hopefully get put back together tonight. I'm half tempted to add a schrader valve to each one so I can add or remove air pressure to adjust the compressible or springiness effect. The problem is there is no place to add it that won't interfere with their operation.

The scooter looks weird with the front end all taken apart. While it's all apart, I want to adjust the front brake set up a little so the caliper will come off and clear the rim without removing the rotor. I had to take the brake all apart anyway...might as well try to shift the rotor outwards a little with spacers behind it.

So for anyone that has never seen whats inside scooter shocks, here you go...

 

[rishimaze]

People like to read specs...
This is all old information prior to a bunch of upgrades:
0-60mph in sub 5 seconds
75mph top speed
250 battery amps peak
20kw peaks
620 phase amps peak
82v at about 64Ah of capacity
Range with constant aggressive riding...about 20 miles.

New information post upgrades will come soon. This is stuff done since the above numbers:
82v at 124Ah (2 new LION packs) of capacity
Larger tires and wheels that roll better and are matched to the motors designed "gearing".
v3 torque arms
Improved wiring to the controllers
Improved breakers
Improved and identical smart BMS's on all 3 packs
Dual watt meters
Shortened phase wires

 

[rishimaze]

I realize that you understand what you're talking about, but using incorrect terminology will just confuse other people.

You are technically correct and I'm not disputing that fact. However, I have to say that to date LTO, LIFE and LION has proven to cover these 3 lithium chemistries pretty well without confusion to what is being referred to. The specifics of anodes or cathodes aside, when I say LION, no one thinks I am referring to LTO or LIFE cells and the same for LTO or LIFE. These are commonly used acronyms that most people readily understand as specific to a certain lithium chemistry with a specific voltage range of operation. I'm going to keep using these generic terms since they are well understood and accepted for what they refer to. I hope that doesn't irritate you too much. It's not ignorance, just ease of terminologies...

BTW...I have seen LION cells that use a technology called LFP. I used to think this was specific to LIFE chemistries and it is not. A friend was telling me about an LFP cell and I immediately dismissed it as LIFE. He then sent me the manufacturers page on the cell and sure enough it was LION chemistry using an LFP technology in it.

Anyway, since this is a build thread, this topic of cell chemistries and correct terminologies really ought to be reserved for the battery section and NOT here.

 

[rishimaze]

The hydraulic fluid is in, new seals are in place and the shocks are stiffer now. They compress a lot less now. Since I've have had them all apart, adding more fluid is not terribly hard to do. All I need to do is take off the front cowling to get access to the caps on the tops of each shock. Unscrew them and add more shock fluid. Not exactly ideal, but doable. A syringe with a tube on it sure helps with this. I added a mark on the tube since I can't see inside the shocks. Extend the tube into the shock to the mark and pull on the syringe plunger. If I pull up fluid, then that shock has too much fluid in it. If it sucks air, add some fluid.

Shifting the rotor over did the trick. I used washers between the rotor and hub to move the rotor closer to the shock. This got me just enough spacing for the caliper to clear the rim so it can be slid off the rotor without loosening the rotor bolts. The rotor bolt heads clear the shock by 2-3mm. In this case a near miss is good enough. I now need longer rotor bolts since about 6mm of length is not in threads anymore. These are the original rotor bolts and the hex heads are pretty wallowed out. They need to be replaced anyway. Details, details, details!