KMX Franken Trike 12-18kw Build

Hi everyone,

This is a bit early yet, but I thought I'd start the thread and see what people think. I won't be starting any actual construction for a few months yet, but never the less, I am planning for this build-out.

My goals are this:
1. 0-60 in less than 5 seconds
2. 3 wheel drive (3 C80100 outrunner motors and 3 controllers, one per wheel)
3. Top out at 70mph+++
4. Incorporate custom software in a nano to detect steering angle so that all 3 wheels can pull proportionally based on their arc in the turn.
5. Incorporate some form of traction control so that wheel spin and loss of traction is controlled in the nano.
6. Incorporate a clutch mechanism so both front motors can be mechanically engaged or disengaged at the wheel sprocket with the flip of a lever.
7. Disable both front controllers with the flip of a switch or key turn for lower power, back wheel only riding.
8. 100 mile range.
9. Custom build a 24S pack or packs out of 21700 LION cells.
10. Build this monster on a reinforced KMX recumbent trike frame.
11. Full regen on all wheels.
12. Each controller will have it's own display (android phone) for status and configuration.
13. Each motor will drive from 219 chain and sprockets.
14. Reduce losses via ceramic bearings, cowlings and covers for improved aerodynamics.
15. Custom LED lights for crazy bright lighting.

I think I can accomplish all of this for around $5000 in parts. This project will take 6 months or more to build and perfect.

I have 40 Samsung INR21700 cells coming next week. I've been working with Power Velocity for a good while now on his controllers and have modded quite a few of the 12 fet version for 4kw continuous. I think with better mosfets than the AOT290 which are OK, but spotty on consistency and quality that I can achieve 5kw in a 12 fet controller. I might also use 3 Lebowsky controllers since they are FOC and have a better than 100,000 eRPM limit and I can custom build my own power stage using TO-247 mosfets. I have the boards and MCU's for 3 of the through hole controllers.

I'm expecting to rewind the C80100's or have Alien Power do it for me. There's lots of room on those stator teeth for more copper. When I took this picture I swapped out the shielded bearings for sealed ceramics. All 3 motors will have at least sealed ceramic balled bearings.



I will also be reinforcing the shaft to bell union like this. I've talked to Alien Power about doing this for me and the cost is reasonable. It puts 6 M5 set screws into the shaft onto flats or shallow holes. This picture is from my own work on my Currie scooters motor.

At that point I think I would just buy a shifterkart.

I did implement a proportional rear drive on the Terex MT6300 Mining Truck. We had twin 3 megawatt triple reduction gearmotors with a Skoda AC asynchronous motor. General Atomics built the invertor, basically the same unit on a nuclear sub.

Used a pair of offset rotary encoders on the steering pinion output from the hydraulic steering box. Encoder wheels had custom spacing between offsets so we would have a multitude feedback of position "resets" to avoid any errors from missed signals. Turns out we didn't need it, but in the event we converted to autonomous drive it was the safe thing to do.

Once you know steering angle its a piece of cake to write the velocity offset formula.
Also had offset encoder pairs (equal spacing) on each motor with 180 signals per revolution (42:1 gearbox, 7560 pulses per tire revolution). We needed this level of resolution to avoid backlash between the gears.

Worked really well. Initial testing was in the Ft. Mac Tar Sands at the Syncrude mine. On sheer ice with an empty dump body this truck was more nimble than an AWD Ford pickup.

Proportional drive is simply an incredible machine to drive. Best of luck.

Triketech said:

At that point I think I would just buy a shifterkart.

I did implement a proportional rear drive on the Terex MT6300 Mining Truck. We had twin 3 megawatt triple reduction gearmotors with a Skoda AC asynchronous motor. General Atomics built the invertor, basically the same unit on a nuclear sub.

Used a pair of offset rotary encoders on the steering pinion output from the hydraulic steering box. Encoder wheels had custom spacing between offsets so we would have a multitude feedback of position "resets" to avoid any errors from missed signals. Turns out we didn't need it, but in the event we converted to autonomous drive it was the safe thing to do.

Once you know steering angle its a piece of cake to write the velocity offset formula.
Also had offset encoder pairs (equal spacing) on each motor with 180 signals per revolution (42:1 gearbox, 7560 pulses per tire revolution). We needed this level of resolution to avoid backlash between the gears.

Worked really well. Initial testing was in the Ft. Mac Tar Sands at the Syncrude mine. On sheer ice with an empty dump body this truck was more nimble than an AWD Ford pickup.

Proportional drive is simply an incredible machine to drive. Best of luck.

Click to expand...

Buy a shift kart? What fun would that be?

ElectricGod said:

Buy a shift kart? What fun would that be?

Click to expand...

About as much fun as one can have with their clothes on, that's all.

Over the years I started on a trike, then onto bikes, karts, sports car racing, formula cars then back to sports cars, shifterkarts (with my son), bikes and now trikes. Ok its a boosted trike with a MAC 10T. Not for speed but to climb some of the hills in the mountains I live in with a bum knee.

No doubt shifterkarts were more exciting and demanding to drive than the Prototypes I've driven in endurance racing. But the older you get, the less physically capable you are to pilot them.

Electric motors for many years was more of a way to earn a living with the machines I designed, but over the last decade battery technology has finally become practical for propulsion. What you're doing harkens back to what many were doing with cars 50 years ago. That's a wonderful activity of creative passion.

Building a trike with 3 RC motors is a project for one who does prefer building than riding. You are going to make yourself a neverending maintenance schedule.

MadRhino said:

Building a trike with 3 RC motors is a project for one who does prefer building than riding. You are going to make yourself a neverending maintenance schedule.

Click to expand...

Why do you think that?

The C80100 is hardly an RC grade motor. 6-7kw, 12mm shaft and includes halls. THe only thing it lacks is improved bell to shaft support and sealed bearings. If you said that about a CA80-160kv Turnigy motor, I would agree with you.

I've built several EV's that use a single motor and they need occasional chain lube and that's about it. Otherwise I go many months of daily riding between needing to do anything beyond charging.

Triketech said:

ElectricGod said:

Buy a shift kart? What fun would that be?

Click to expand...

About as much fun as one can have with their clothes on, that's all.

Over the years I started on a trike, then onto bikes, karts, sports car racing, formula cars then back to sports cars, shifterkarts (with my son), bikes and now trikes. Ok its a boosted trike with a MAC 10T. Not for speed but to climb some of the hills in the mountains I live in with a bum knee.

No doubt shifterkarts were more exciting and demanding to drive than the Prototypes I've driven in endurance racing. But the older you get, the less physically capable you are to pilot them.

Electric motors for many years was more of a way to earn a living with the machines I designed, but over the last decade battery technology has finally become practical for propulsion. What you're doing harkens back to what many were doing with cars 50 years ago. That's a wonderful activity of creative passion.

Click to expand...

I meant that buying something would be no fun. I want to build it myself. Are shift karts fun to drive...well sure, but that's not what I meant.

There's lots of derived joy and pleasure and pride in making something with your own hands. And then when you use it to destroy everything else on the road, well there's joy and pleasure in that too. ;P

I'm about to finish up my third 2 wheel EV. I can either go lots more power on 2 wheels or lots more power on 3 wheels. The 3 wheel machine has the advantage that it's virtually impossible to drop it on slippery surfaces and you can drive all 3 wheels. How many times have you laid down a 2 wheel machine? I've done it quite a few times myself. I'm always pushing the limits of grip and 3 wheels has 30% more grip than does 2 wheels. And if course since it's still a recumbent trike, I can ride in bike lanes or bike trails or on the road. Good luck doing that with a kart!

Looks like an ambitious project. I'm curious why you plan to go with a Nano instead of something like a Raspberry Pi Zero. Nano can handle one program at a time, but the Pi can run multiple, making it easier to have traction control, and steering angle based proportional torque work together.


@Triketech: Dumptrucks On Ice is something I would stand in line to buy tickets to go watch.

Drunkskunk said:

Looks like an ambitious project. I'm curious why you plan to go with a Nano instead of something like a Raspberry Pi Zero. Nano can handle one program at a time, but the Pi can run multiple, making it easier to have traction control, and steering angle based proportional torque work together.


@Triketech: Dumptrucks On Ice is something I would stand in line to buy tickets to go watch.

Click to expand...

I don't know much about the PI zero. Does it have multiple DACs and ADCs built in? It does have a more powerful CPU. The nano has most of the hardware already on board so it seemed like a good option. I'm not married to it and probably I'll want better DAC and ADC resolution than the nano can deliver. That would mean external IC's. Who cares what talks to the ADCs and DACs as long as it does the job. I really hadn't considered a PI zero.

Drunkskunk said:

@Triketech: Dumptrucks On Ice is something I would stand in line to buy tickets to go watch.

Click to expand...

Not sure if I still have any of the old videotapes. I mounted a camera in one the first time I drove a prototype without proportional drive.

Our tech took me out to a flat sheet of ice about a square mile and turned me loose. Before rolling he said "turn it all the way to the right" and 4 turns of the wheel later he said "now give it a little throttle".

The 20 cylinder MTU engine revved up to about 1300 RPM, the Kato alternator exciter diodes kicked in, and the General Atomics invertor began switching the 220 lb IGBT's. The truck went straight, about 10 MPH.

Then he said "floor it". I did. The engine picked up to its peak of 1700 RPM. And we started doing donuts. In a 330,000 lb truck with a viewpoint of Santa Claus about to slide down a 2 story chimney. Unforgettable experience.

I shot a video about two days later after installing a Motec data acquisition unit, its an old wmv format but gives a pretty good idea of how large the beast is.
http://www.kartweb.com/Trailers/MoTec_Performance_1.wmv

ElectricGod said:

I meant that buying something would be no fun. I want to build it myself. Are shift karts fun to drive...well sure, but that's not what I meant.

Click to expand...

I can relate with that. Shifterkarts are pretty intensive on constructor skills.

I admire what you're doing, just wasn't clear that you're ambitions were more of an innovator than rider. I'm a bit of both in my own way.

Here's a possible architecture that may be a bit advanced but worth considering as an eventual target. Not kid gloves stuff, and a bit complex but capable of providing incredibly accurate sine wave signals to all 3 motors with no aberrations at very high commutation speeds. Maybe overkill.....

TI TMS320F28069 PIC. Can be set up to interface with 3 TI 8312 driver chips. Probably a bit beyond your current needs but I thought it was worth tossing out. Worth a little investigation maybe for taking a proportional driver system all the way to bright.

No matter'whatever you come up with will be pretty cool I suspect.

Triketech said:

ElectricGod said:

I meant that buying something would be no fun. I want to build it myself. Are shift karts fun to drive...well sure, but that's not what I meant.

Click to expand...

I can relate with that. Shifterkarts are pretty intensive on constructor skills.

I admire what you're doing, just wasn't clear that you're ambitions were more of an innovator than rider. I'm a bit of both in my own way.

Here's a possible architecture that may be a bit advanced but worth considering as an eventual target. Not kid gloves stuff, and a bit complex but capable of providing incredibly accurate sine wave signals to all 3 motors with no aberrations at very high commutation speeds. Maybe overkill.....

TI TMS320F28069 PIC. Can be set up to interface with 3 TI 8312 driver chips. Probably a bit beyond your current needs but I thought it was worth tossing out. Worth a little investigation maybe for taking a proportional driver system all the way to bright.

No matter'whatever you come up with will be pretty cool I suspect.

Click to expand...

I like to ride, but I like to build as well and to ride what I build is the best ever.

Are you suggesting I use the Motec monitoring device for this build?

Donuts in a 330,000 pound machine...awesome! Just don't hit anything, you'd destroy anything you got near. I guess that the wheels on one side spin forward and the other side spin backwards to do those donuts. At 10mph max there's no way the back end could break loose so you could do proper wheel spinning donuts.

I rode a KMX powered by a pair of big RC motors driving only the rear wheel. It was insanely fast. I thought I was going to die several times while riding it on the streets of San Francisco. The steering is pretty twitchy at high speeds. On dirt it could easily break loose the rear tire and drift, which was fun but rough on the tires.
Being so low makes it harder to be seen by traffic.

fechter said:

I rode a KMX powered by a pair of big RC motors driving only the rear wheel. It was insanely fast. I thought I was going to die several times while riding it on the streets of San Francisco. The steering is pretty twitchy at high speeds. On dirt it could easily break loose the rear tire and drift, which was fun but rough on the tires.
Being so low makes it harder to be seen by traffic.

Click to expand...

I forget who did it, but it looked like a SpinningMagnets project with 3 3220 motors with Divinci Drives stickers on them on a recumbent trike. All 3 motors drove the back wheel so of course breaking loose the back end was ridiculously easy even on pavement. My theory is that 3 wheel drive will level the playing field a fair bit so that at full power more energy is transferred to the pavement and much less into burning rubber.

Did the KMX you rode have front shocks? Talking to a friend of mine who has rear wheel drive on his KMX said that he spent the money for the suspension kit after he got his bike up to 60+ and it get pretty twitchy in the front. He said that the suspension kit which costs about $1000 was worth it for the added control it gave him at speed.

There are flags and whip LEDs for recumbent trikes and other uses that will greatly increase visibility such as this. Mount one or 2 of them behind the seat and of course several very bright LED lamps should do the trick. Beyond that I'm already a 2 wheeled EV rider. I'm always watching what the cars are doing so they don't run into me. I assume I'm the one watching out for them, not the other way around.

https://www.ebay.com/itm/5ft-White-LED-Whip-lights-with-Red-Safety-Flag-for-Sand-ATV-Buggy-Dunes-UTV-/182923172313?_trksid=p2349526.m4383.l4275.c10

When I had the MAC 12T on the back of my Scorpion FS26 at 40 amps it would spin the tire about 1/2 a turn on launch. Didn't like the gear groan it made though.

With the 10T on motor only it peaks out about 28 MPH @ 52V. I ride in the mountains and with the bodylink seat lean far enough on turns to drift. Downhills with Big Apple tires at speeds over about 45 MPH the long wheelbase and full suspension is pretty stable. Indirect steering is also very nice. Direct steering tends to be far less forgiving at speeds over about 30 MPH. In part due to the larger caster angles found on Catrikes & KMX's. I've had a few of both. Catrike headsets are too wobbly at speed, KMX isn't too bad in that respect.

Whips and flags help with visibility and in the daytime nothing beats adding a bright yellow or green "tailbox" above a rack on the back. We avoid busy streets and just run whips at night. A simple driveway marker wrapped with X-mas light LED's and a battery. Those won't be on the Home Depot shelves much longer.

Triketech said:

When I had the MAC 12T on the back of my Scorpion FS26 at 40 amps it would spin the tire about 1/2 a turn on launch. Didn't like the gear groan it made though.

With the 10T on motor only it peaks out about 28 MPH @ 52V. I ride in the mountains and with the bodylink seat lean far enough on turns to drift. Downhills with Big Apple tires at speeds over about 45 MPH the long wheelbase and full suspension is pretty stable. Indirect steering is also very nice. Direct steering tends to be far less forgiving at speeds over about 30 MPH. In part due to the larger caster angles found on Catrikes & KMX's. I've had a few of both. Catrike headsets are too wobbly at speed, KMX isn't too bad in that respect.

Whips and flags help with visibility and in the daytime nothing beats adding a bright yellow or green "tailbox" above a rack on the back. We avoid busy streets and just run whips at night. A simple driveway marker wrapped with X-mas light LED's and a battery. Those won't be on the Home Depot shelves much longer.

Click to expand...

It looks like direct or indirect steering methods are both workable and both have their lovers and haters. My friends KMX tops out at 70mph. He's never mentioned any issues with high speed steering. Of course at those speeds, you are NOT going to be weaving about at all! I didn't see if the KMX implements ackerman steering. Do you know? I assume it does.

Hubs in bikes and trikes is super common. I'm not a fan of all that unsprung weight inside the wheel. This is a big reason why I want to use 3 C80100 motors. They can be mounted to the frame and power is transferred to the wheel via 219 chain.

I have an idea for this build that would simplify things a LOT. I ran across a 4WD quad bike made by a Chinese company called Camel Trikes. The drive mechanism is a bit complex so that pedal power goes to all 4 wheels. I was thinking possibly of adapting something like it to the trike and running form a single large outrunner such as an Alien Power 12090. The back wheel would have constant drive and would not implement a freewheel so that regen can be used. The front wheels would both incorporate free wheels so that in turns the wheels can spin at their separate speeds, but engage in straight line riding. This is how all quad bikes drive both back wheels. It's not a true differential, but it's light weight and trouble free. If I could find a real diff that was small and light weight, that might be a better option. Freewheels also drive the inner wheel instead of the outer wheel which hurts turning under power. With this drive solution, I would not need to worry about adjusting throttles or traction control or turning angles at all. It would be a simple mechanical solution that would take away a lot of the complexities of my other option. The Camel Trike quad uses a flexible shaft that drives the front wheels. I could use a single chain from the motor to a central sprocket between the front wheels. Either side of the front sprocket would have a freewheel to drive the front wheels separately.

ElectricGod said:

Did the KMX you rode have front shocks? Talking to a friend of mine who has rear wheel drive on his KMX said that he spent the money for the suspension kit after he got his bike up to 60+ and it get pretty twitchy in the front. He said that the suspension kit which costs about $1000 was worth it for the added control it gave him at speed.

Click to expand...

No, it was stock in the front. My camera fell out of my pocket and got lost on that ride, so I don't have a picture of it.

To me, it seems like just driving the rear wheel was good enough. There's a fun factor to being able to break loose the tire. If something happened to the drive system for one of the front wheels while going fast, it could be bad.

fechter said:

ElectricGod said:

Did the KMX you rode have front shocks? Talking to a friend of mine who has rear wheel drive on his KMX said that he spent the money for the suspension kit after he got his bike up to 60+ and it get pretty twitchy in the front. He said that the suspension kit which costs about $1000 was worth it for the added control it gave him at speed.

Click to expand...

No, it was stock in the front. My camera fell out of my pocket and got lost on that ride, so I don't have a picture of it.

To me, it seems like just driving the rear wheel was good enough. There's a fun factor to being able to break loose the tire. If something happened to the drive system for one of the front wheels while going fast, it could be bad.

Click to expand...

With the freewheels to both front wheels, that will stop wheel spin. Both front wheels can only spin as fast as the slowest wheel and be driven. IF a freewheel broke internally and just spun, the other wheel would be the top speed. I bet there would be some pulling to the side away from the driven wheel. With the addition of a clutch mechanism to disengage the front drive, that quickly gets the front end disengaged. In turns, the outer wheel in a turn won't see drive with dual free wheels in the front.

The clutch mechanism that disengages the front drive system would put all of the 18kw to the back wheel for spinning donuts and burning rubber. I'd want to break the power connection as close to the beginning of the front drive system as possible so that none of those parts are still being driven. The 2 freewheels in the front section would disengage the front wheels from the front drive system. I think a clutch like what typical car and truck AC compressors use could do the job. I may need a bigger one since I'm looking to transfer something like 12-15kw through it, but there are much larger compressors such as are used in busses that have large clutches that should serve well for this use. Selectable drive vehicles also use a similar clutch at the wheels or in the transfer case. It's just an electromagnet that drives 2 spinning friction surfaces together to transfer torque. A simple toggle switch can turn this on or off.

Single wheel drive is probably sufficient for road riding, but I want 3 wheel drive. I want to be able to maximize drive traction and to do off-road riding and to power out of mud and bogs. This is things a single wheel drive won't do. I'm really thinking of going with the 12090 18kw motor and drive all 3 wheels from the single motor instead of the 3 motor idea I originally proposed. It simplifies a lot of things and I get similar power levels and similar functionality. I'm still looking for a small differential or posi unit. I'd like to also incorporate a 2 or 3 speed transmission. I want the motor to be dead in the middle of it's power band and yet be able to creep around at 3-20mph for slogging through deep mud or up a steep hill. Then I want a second gear that will be used mostly for typical street riding where I top out at around 50mph and then a high gear for those speed runs. If I can find a lawnmower transmission that uses bearings and doesn't incorporate a worm gear, that will get me a small transmission that will serve my purposes quite well.

I've seen nice differential units for pedicabs but they tend to be pretty expensive.

I found this:
https://www.searspartsdirect.com/part-number/794292A/0071/142.html?utm_source=google&utm_medium=cse&utm_term=794292A&sid=SPDxGPROD&k_clickID=8eb2878d-0259-42d2-a560-f5d8fe40ad09&gclid=CjwKCAiApo3SBRA4EiwAty8i-n23Hc5a5jwnMeUOEIn4IlqNdrfyaqJNO-EK8AR5Cx8yPbdKUvR8axoCNyAQAvD_BwE

"Lawn Tractor Differential" looks very similar to the ones on pedicabs but much less expensive.

Another version:
https://www.northerntool.com/shop/tools/NTESearch?storeId=6970&ipp=24&Ntt=go-kart+differential

They use them on some go karts as well, so may be available from the go kart parts places.

fechter said:

I've seen nice differential units for pedicabs but they tend to be pretty expensive.

I found this:
https://www.searspartsdirect.com/part-number/794292A/0071/142.html?utm_source=google&utm_medium=cse&utm_term=794292A&sid=SPDxGPROD&k_clickID=8eb2878d-0259-42d2-a560-f5d8fe40ad09&gclid=CjwKCAiApo3SBRA4EiwAty8i-n23Hc5a5jwnMeUOEIn4IlqNdrfyaqJNO-EK8AR5Cx8yPbdKUvR8axoCNyAQAvD_BwE

"Lawn Tractor Differential" looks very similar to the ones on pedicabs but much less expensive.

Another version:
https://www.northerntool.com/shop/tools/NTESearch?storeId=6970&ipp=24&Ntt=go-kart+differential

They use them on some go karts as well, so may be available from the go kart parts places.

Click to expand...

Those might do what I need. Thanks for finding them. They would eliminate the need for freewheels and would drive the outer front wheel in turns. Mounting the diff rigid to the frame via a support bearing at the ends of the shaft would probably work well. A sprocket on the end of each shaft and then chain to another sprocket on the wheel would allow for full suspension travel. I doubt I would need anything to take up chain slack either.

LightningRods makes a wheel hub that takes 2 sets of 6 bolt pattern mounts on the same side of the hub. One for a sprocket and one for a brake rotor and free wheel threads on the other side. This hub would serve well for my uses.

ElectricGod said:

Those might do what I need. Thanks for finding them. They would eliminate the need for freewheels and would drive the outer front wheel in turns. Mounting the diff rigid to the frame via a support bearing at the ends of the shaft would probably work well. A sprocket on the end of each shaft and then chain to another sprocket on the wheel would allow for full suspension travel. I doubt I would need anything to take up chain slack either.

Click to expand...

What happens to the chain when steering?

fechter said:

ElectricGod said:

Those might do what I need. Thanks for finding them. They would eliminate the need for freewheels and would drive the outer front wheel in turns. Mounting the diff rigid to the frame via a support bearing at the ends of the shaft would probably work well. A sprocket on the end of each shaft and then chain to another sprocket on the wheel would allow for full suspension travel. I doubt I would need anything to take up chain slack either.

Click to expand...

What happens to the chain when steering?

Click to expand...

Steering? Who needs to steer? Yeah...good point. I wonder if the sprockets on the ends of the diff shafts could be mounted on some kind of drivable pivoting head...sort of an mini CV joint.

This guy did something like that:
https://www.pinterest.com/pin/293085888227802911/


Those U-joints look pretty small for anything more than human power.

The suspension parts from a 4wd ATV would be much stronger but very heavy.

More ideas here from our members:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=49547

David Bruce built a handful of FWD trikes.
http://www.dbtrikes.com/

Nobody makes small enough Rzeppa CV joints. Polaris has the smallest I've seen they're made for quads.

I've used some small Tripodes on a machine I designed about 30 years ago that came from Stock Drive Products but haven't seen them in the catalog for many years now.

U-Joints tend to get "wobbly" with much more than 4° angular displacement as the shaft speed cycles fast-slow in relation to the stub axle speed. "Unconstant Velocity" due to the pivot axis's being in the same plane. Automotive steering shafts are a good budget source FWIW.

Proportional drive will be difficult to implement without independent drives to each wheel. Three high resolution encoders (>300 pulse/rev) running at wheel speed and an offset pair of encoder inputs for steering angle (>1200 pulse/rev) could be used to independently control throttle signals to each motor driver chip. The actual throttle would be connected to a PIC programmed to compare left and right speeds to a slew factor applied to the rear wheel speed. Rear wheel is driven directly proportional to the throttle with the slew factor gaining speed on the outside wheel and losing speed on the inside wheel.

Of course you'll want some safety features in the program for overspeed and current draw differentials too.

I've looked at the Samagaga products before...like 6 months ago. There site is not responding ATM. I wonder if they went out of business?

David Bruce has more access than I do to be able to make custom parts. He essentially made everything himself. The cost for a one-off project is quite a lot. I'm going to have to use commercially made components and piece them together to make this work.

A samagaga diff is an option, but small diffs for lawn mowers are pretty common too.

I spent some time looking at transmissions last night and found the Peerless products are decently made. I'd still want to mod it somewhat to use bearings on all the shafts instead of brass bushings. It appears that there are bearings on the input shaft only and everything else is typically bushings. I will be running the tranny at more than 10 RPM at the output shaft and I'll want to minimize losses as much as I can. Bearings will have to get into the mix on everything. Also, they typically fill them with grease. I'll replace that with a synthetic gear oil instead. Bushings will leak, but sealed bearings will keep it inside the gear box.

I thought I preserved it in a txt file I have for EV stuff, but now I can't find it anywhere. I found a web site for a UK manufacturer some time back. They made small CV joint like units, flexible shafts, flexible drive components and other things commonly needed for driving something that has to move. I'm limited by the suspension so I can't drive through that. This means going around the KMX suspension components for drive. Some kind of pivoting mount on the ends of the diff shaft would allow for wheel turning and keep the chain line straight...hopefully. The pivoting end on the diff shafts would need to slide in and out too so that the chain line was always straight. It may be that there would be so much custom machining on this that I'd do better to go back to 3 separate drives instead. LOL!

I remember running across these companies for flexible shafts.
http://www.flexishaft.com/
http://www.suhner-transmission-expert.com/en/home.html

Using a universal joint or gimbal on the diff shaft ends might work. Of course, any kind of universal joint won't take large angle changes at higher RPMs very well. I'm guessing the front wheels will make a 20-30 degree arc stop to stop.

Really small CV joints exist commercially. All the way down to RC car and truck stuff. A CV joint may be a viable option. It also means a fairly complex front end set up between the diff and the wheels and CV joints typically need to drive through the axle. That would mean getting below the KMX suspension components. I suppose an add on block that brings the wheel axle down below the frame might work. I could probably get away with driving from a flexible shaft instead of a CV joint in that case. Some custom one-off work can't really be avoided. This may not be simpler after all...just different complexities.

Maybe 3 motors and 3 controllers is a better option after all? Running the front motors would need to drive through a flexible shaft in order to mount them on the frame. I'd still have to drop the front wheel axles below the suspension travel range. Either single motor and mechanical or multiple electronic setups, both ideas have their own set of complexities and common issues to overcome.

Just found these guys in Oregon. Their site is really old and who knows if they even exist anymore. They make CV's and diffs for bikes.
Contes Engineering: http://www.contesengineering.com/