2WD (dual motor, two wheel drive) Trike build.

[llile]

Welcome to the saga of my Tadpole recumbent Trike build, aka Frankenbike 8.0.

I've forked this off of a really excellent thread about 2WD bikes. http://endless-sphere.com/forums/viewtopic.php?f=2&t=33429

Most are focused on standard diamond frame bikes, which us 'Bent (recumbent) folks affentionately call Buttbusters, since your butt and your nether parts will invariably hurt after a long time no matter how cushy the seat is and how many springs you have in the forks. Maybe that's just me, call me Tenderbutt.

I've got a really successful SWB 'Bent (Short Wheelbase recumbent, for those of you not steeped in the lingo) that I have been using for almost a year - I ride more, farther and faster than I ever did on a standard 'Bent and I am in better shape than a year ago. I am addicted. So how could I change things so that I use my E-Bike more often?

Weather is a big factor. Sure, I ride in the rain and in the cold, I have ridden in rain that would fill your shoes after a minute, and ridden in cold that would have frozen Sam McGee. But those aren't conditions I want to endure every day, and not both at the same time. Speed is a factor. I live 15-20 miles from town on good roads, sometimes I don't have an hour to get there and an hour to get back. In the winter, darkness hems in both ends of the ride. I have lights, ride at night in town where traffic is slow and streetlights are common, however I'd rather be a more visible target at night on these lonely backroads. I have a 50 mile round trip - a formidable distance. I'd like to have a 40 MPH capable vehicle, a formidable speed.

I am fixated on building a tadpole trike with two front motors. A challenge, no doubt. I'd like it to have a faring, for better wind resistance and to keep some of the wind and rain (not all - I'll still have my balding helmeted head poking out the top) off of me. Whether or not this is a good idea remains to be seen, but I am convinced it will be.
Missouri, being a liberal state (NOT) allows motorized bikes up to 30 MPH and 3 HP (2200 watts!). Such a bike is allowed on ANY road, even the interstate shoulder, as long as it is on the shoulder where a minimum speed is posted. I actually ride a four lane road a lot - you have your own 8 foot lane with a rumble strip between you and the big trucks! Call me insane, I am not alone.

Playing with the wonderful calculator is the next step.

[llile]

Playing with the calculatorhttp://ebikes.ca/simulator/ I find it correctly estimates my current bike. Although it is a stokemonkey, and will perform significantly differently on steep hills because of gearing, at high speeds it looks pretty accurate:

Clyte 406 motor, 20A controller, PING 15AH battery, full recumbent wind resistance, gives 0% grade speed on the flats of 27 MPH, 15.3 WH/M, 414 watts, and 37 mi range. That's pretty close to what I can do now. With the Cycle Analyst I can dial down the power a little and get 40-50 mile range at 20MPH, which matches up with the calculator if I dial the throttle back to maybe 90%.

So how to model the new bike? A basic rule is that 2X the speed takes 4X the power. The simulator also tells us that hill climbing is *linear* with speed - air friction isn't the main drag, it is oomphing up the hill. That I didn't know. Try modeling a steep hill if you don't believe it.

So I am looking at 4X the battery to go twice as fast the same distance. About 60 Amp Hours. Ping says that is 50 Lbs of batteries. Wow! My current battery is 12 lbs.

So hauling my tub-O-lard butt and my relatively heavy trike and battery up a hill will be a 120 kilogram affair. Say 150 KG so I can haul a few groceries.
I have a lot of 6% grades and a few very steep grades. 15%-20%. They are short, however. Let's say I stay with the Full Recumbent air resistance, although my air resistance is likely to be lower with a faring.

With a 150KG bike, full recumbent, Clyte 5303 motor, 52C 0.2 ohm 30 AH battery (48V nominal) this vehicle should go 39 MPH on the flats and have a 46 mile range. Hmm. How to model dual motors and hilly terrain?

I am going to say that the black "load line" doesn't change, however the motor power will be 50% of the load since there are two motors. THe load line crosses 40MPH at 1000 watts. This is simulated well with a drag CdA of 0.25 and a Cr of .0005. If I change the CdA to .012, that seems to simulate the load on one motor of 2 pretty well, making the load line go to about half or a little over 500 watts.

With a 72V 0.2 ohm 30 AH battery (half the expected battery pack), a 25A controller, 62% throttle, 0% grade I am at about 40 MPH and 550 watts on one motor. Range estimate is nuts at 121 miles.

I will cut the wieght to half for estimating grades since there are two motors. On a nice easy 6% grade as encountered on 4 lanes, I am at a respectable 34 MPH with no overheating. On a 15% grade, I am at 20 MPH and overheat in 25 minutes, which is fine since the steep places are short, maybe 1/4 mile. THis is looking like a really respectable machine, sumpin' that'll get me to the store and back.

How to estimate range? Modeling my old bike at a 1% grade gives a pretty good estimate of the range I am getting. Some hills are short and steep, sometimes I am on the flat or downhill. So if I estimate the above parameters at a 1% grade, throttling back to 40 MPH, I get 51 miles range with a 15 AH battery per motor. Hmm - this is hard to swallow, since I had earlier estimated a 60 AH (30 AH per motor) battery which is twice that size.

I could certainly set it up with two 72V 15 AH batteries with provisions for two more, then ride the darn thing and find out for sure what the range is in my terrain. Might be a plan!

Efficiency - comparing a single motor (150KGs, .25 CdA, 40 MPH) with two motors (half the wieght and CdA) I find that I am at 75% efficiency on the flats, 75% on a 12% grade and 65% on a 15% grade with the dual motor setup limited to 40 MPH with the throttle. THe single motor is 81% on the flats, 31% on a 12% grade won't go up a 15% grade at all. Here is where the dual motor shines, hauling my lard butt and a big battery up a hill with high efficiency.

[corkscrew]

I've considered doing this to my trike also, at least since I found the single side axle models on ebikes.ca.

Good luck with the project!

[llile]

SIngle side axles are the classic way to build a trike, my friend had some aluminum blocks custom machined into a little axle-holder that holds teh side axle and mounts it on a bearing, really successful homebrew recumbent trike, maybe ten years now. He used hub brakes on the front - uncommon, but obtainable. These are cable-actuated brakes that look and work kinda like the old coaster brakes.

I am figuring that I'll want something sturdier for higher speeds. I am going to skip the side axle front wheels and use a couple of disk-brake-capable mountain bike forks for the front end. These will be stronger, I believe, and also have some spring as they are the suspension forks. You NEED springs if you are going over 30 MPH believe me. I am not sure how you'd put brakes on a side axle front hub motor otherwise.

There are some tricks about the steering geometry of a trike. If you draw a line from the front steering axle, through the tie rod axle, that line should meet in a triangle at the point where the back wheel kits the ground. THis will cause your wheels to steer correctly around corners, the outside wheel turning at a wider radius than the inside wheel. THis avoids the outside wheel skidding along at the wrong angle. Very important, something we found out the hard way by doing it wrong before we found the answer in some texts. Dover press has a book called "Bicycle Physics" that is very instructive for DIY frame builders who are not building a classic diamond frame two wheeler.

[llile]

Playing around with the calculator more, I get more efficiency, maybe 20% more, out of a 9C 2807 vs 2806. Still climbs hills OK and hits similar top speeds.
I will be using 26" front tires. Kevlar with thornproof tubes filled with slime.
Faring will start with a clear bubble faring from the pros that make such things, with the rest possibly hacked out of coroplast or perhaps a body stocking, not sure yet. THere have been some really nice coroplast farings made, quite cheaply and pretty good looking. Coroplast makes a decent living hinge, simplifying entry and exit.
Here is a preliminary parts list, the rest being made out of junk bikes:

(2) 9C 2807 on 26" rim @ $270 = $540
(2) PING 36V 30AH batteries @750 = $1500
(2) infineon 25A controllers @ 125 = $250
Cycle Analyst @ 150 $150
Throttle @ 20 with ebrake button $20
Halogen Lights a@ 100 $100
Have homemade red LED flashing lights for rear
brake levers pair $24 with brake inhibit $24

Airparts, inc. Good source for 4130 and aluminum tubing
Rod Ends for steering tie rod 1/2"-20 $11.50 * 4 = $46
36" 1/2" aluminum tie rod $6
1/4" 4130 @ 5/ ft = $50
1.25" 4130 tubung 5/ft = $50
1X2 aluminum square tubing (faring supports) @ 4 = $40
1/2 X 1/2" U channel @ 2 $40 (faring supports) $40
Clear faring $224 http://www.windwrap.com/


Cloroplast 4 X 8 $100 each

Walts Bike Shop (local guys who save my butt regularly)
SR Suntour suspension forks with disk brake mounts $90
Avid BB7 brakes @ 75 $150

Tires 3@$40 $120
Misc $50
Junk Bikes $0

Total $3552

Double the cost of my last E-bike, with double the speed! Not bad. Compare to a used car, but with two orders of magnitude better mileage.

I wonder if it will really live up to the specs I have set for it using the calculator?

[llile]

Here is how I will do the cycle analyst and use only one throttle. Thanks to Justin of e-bikes.ca

Hi Lawrence and thanks for the inquiry and nice to hear from a fellow addict!

The other option which is a little more accurate is to get the Stand Alone model of CA which will properly measure the combined current going into each controller, and then for the limiting features you hook up the CA as a current throttle as described in section 9.3 of the user manual. So throttle goes directly to the CA, and then the CA outputs its own throttle signal that goes to the two controllers.

In either case you don't really need any additional circuitry or hacks, it's just a question of doing the appropriate parallel wiring connections on the existing lines.

Justin

[Pure]

I apologize if I missed where you covered this, but what are you going to do about the problem of your motors not being sinked up? Running dual controllers, one motor will always put out more or less effort than the other.

[llile]

I apologize if I missed where you covered this, but what are you going to do about the problem of your motors not being sinked up? Running dual controllers, one motor will always put out more or less effort than the other.

Well, here is what I will try:

1. identical components
2. Short, and robust power conductors with oversized and short grounds linking everything up
3. Plenty of battery to reduce sags (needed for range anyway)
4. Meh? Maybe they don't balance, and the thing wants to pull to one side. I'll steer to the other side. Meh? Maybe it wears out a tire. That isn't so bad.
5. If I detect an objectionable imbalance, I may hack the circuits from the Cycle Analyst that drive the controller, maybe add a little series resistor or some other trick to balance them out. There are LOTS of ways the signal driving the two controllers could be processed or filtered or divided if it comes to that, and I'm able to buidl the electronics if need be. I don't expect anything too complex would be needed though.

[Pure]

On top of what you are already going to do, I wonder if using dual batteries would help. Something tells me, 2 motors on one battery, one is always going to want to pull more power from the battery than the other. Do something/mod to reduce the power one pulls and the other one will want to pull more. Running 2 complete setups, from battery to motor, should prevent that. So you have 2 things to plug in when you charge, big deal. Then again, I've never done this so I'm just guessing.

[michaelplogue]

Another option would be a front wheel drive trike with a single, mid-motor...



http://www.youtube.com/user/scbvideoboy

[llile]

Another option would be a front wheel drive trike with a single, mid-motor...



http://www.youtube.com/user/scbvideoboy

Wow, that is impressive. This guy must have worked for years on such a pretty, and complex, frame. I am going to say such a build is beyond my skill level, even though I am setting about building a custom frame.

[llile]

On top of what you are already going to do, I wonder if using dual batteries would help. Something tells me, 2 motors on one battery, one is always going to want to pull more power from the battery than the other. Do something/mod to reduce the power one pulls and the other one will want to pull more. Running 2 complete setups, from battery to motor, should prevent that. So you have 2 things to plug in when you charge, big deal. Then again, I've never done this so I'm just guessing.

THat is not a bad idea. Right now, I am fixated on the idea on one big battery, mostly because Justin sez that simplifies the Cycle Analyst setup, and only requires one CA hooked up the way he instructed. I am building a battery set that could be rewired to dual batteries if need be.

I am hoping the mismatch problem will just not be as much of a problem as we may think, as long as I am careful to make everything symmetrical physically and electrically. We'll see when I hit the road in a few months!

[llile]

Front end geometry gets interesting on a trike.

First, there is caster, or the relationship fore and aft between your fork axis, and the contact patch where the tire hits the pavement. In a regular bike, believe it or not, the tire contact patch is *behind* the projected axis of rotation. This makes the front wheel stable, that "ride no hands" ability. When I have built two wheel recumbents, I have noticed that this head tube angle is pretty sensitive, small changes can make a big difference in stability and rideability. I have ridden commercial recumbents that got this wrong, and were hard to ride at low speeds. But I am not sure how much difference caster makes on a trike vs a two-wheeler. I'll build it with some caster, unless I find out for sure that it isn't needed.

There is camber, or the angle between the tire's vertical axis and true vertical. If the bottom measurement between the tirse is a little "wider" than the top, it apparently improves cornering. However, if camber is zero, it improves traction. I am considering using zero camber, because when we;ve built pedal powered trikes int he past, zero camber worked OK.

There is toe-in. Too much, and the tires wear out fast, too little and the bike won't track. I will build an adjustable tie rod, so I can adjust toe-in.

I am hoping that building a leaning body (me, the back wheel, and the heavy batteries) will improive cornering enough to compensate for zero camber. Lotta variables.

Then there is the tie-rod bearing position. THis needs to be on a line between the front tire and the rear tire, or in other words, draw a triangle betwen the tires, and the tie rod bearing needs to fall on the triangle. THis makes the tires turn at the right angle on corners, the inside tire turning tighter than the outside tire.

THere are probably other front end geometry considerations as well. Any knowledge out there?

[Rassy]

This will be an interesting build. Several times different forum members have planned to put front two wheel drive on a tadpole, and as far as I know no one has done it yet, or at least it has not been reported here.

I have set up a number of two and three hub motor drives, always involving a pusher trailer, both one wheel and two wheel, and sometimes also having a hub in the front wheel of the bike or delta trike. Except for one recumbent bike where the front wheel motor/battery/throttle/controller were completely separate from the Bob trailer motor etc., I have always used a single battery pack and throttle with separate but identical (as much as possible) controllers and hub motors. I really think it is important to use a common battery (may be several batteries paralleled together), because that's the only way you can be sure both controllers are getting the same voltage all the time.

There have always been "experts" that fretted about the lack of a "differential" etc., but even a fairly powerful two wheel pusher trailer never had any adverse affect on a two wheel bike during cornering.

I was always looking for a better hill climbing solution and am finally happy with my through the gears mid-drive.

Enough rambling. Good luck with your build.

[Kingfish]

+1 on common battery. In the words of Samwise Gamgee: “Share the Load”

With a trike having two forward tires, keeping them in sync on forward motion will be a challenge: My tractor has AWD, though it can be disengaged. When I press the foot throttle, both tires turn at the same rate. The foot brake though is split; it shares a link which can be disengaged and used as separate Left-Right controls, or linked as one. Using that analogy applied to an etrike – perhaps it could be that there is one unified throttle, though when turning x-amount, the inside motor freewheels, and if turning tightly, brakes? (Sounds like an invitation for tipping) Otherwise it’s one throttle per wheel: One on each side of the steering, and I would think that could get very tiresome in short order on long straight runs/cross-country.

Another thought: Mid-drive with something akin to posi-slip. Not a true independant 2WD solution, but then it would prevent excessive freewheeling.

OK – I thought of a third option: One master throttle providing power to two slaved throttles. When heading straight-forward, both secondary throttles see 100%. If the steering wheel is turned left or right, the associated inside throttle decreases with the angle of rotation. This can be mechanically achieved through a lever, and there should be offsets at both ends of the spectrum so that a hard turn cuts out 100%, and a slight course correction has no effect.

This could actually be achieved using a single throttle in software with feedback on forward speed (accelerometer) and angular measurement of the steering (or perhaps another accelerometer perpendicular to the first) so that the effect is greatest at low speed. That I think would be the best solution, but then we’re back to the discussion of a design-built 2WD Controller, and in this case – for trikes!

~KF

[dnmun]

why is it required to have two motors on the front? why not just use one motor on the rear? i like rassy's build for power since it is a mid drive stoke monkey type rig.

all the extra weight to carry two motors could be used to carry more battery if it was a single motor on the rear. jmho.

[llile]

Well, here is why two motors:

1. I like a challenge.

2. It hasn't been tried often (ever?) before. A few people have tried multimotor 2 wheel bikes, or multimotor pusher trailers, but a 2WD trike has been done rarely if ever.

3. See ebikes.ca simulator numbers above - I think there are some advantages

4. High efficiency on uphill grades compared to single motor

5. Experts say it can't be done. You should always listen to experts, find out what can't be done, then go do it.

I *could* use a mid-drive chain drive system - however this would be a lot of power (1675W) to go through a standard chain and gear system. Would the internal hub or freewheel take the strain? I am sceptical it would. Most mid-drives are ~500W.

Here are some quick calcs:
9C 2807 single hub motor on the rear, 150 KG (rider, bike, groceries and 72V 30 AH batteries) 48 mi range at 2% slope, stalls on a 20% grade (I have some of these) 1146 watts total on the flats. 13MPH, 55% efficiency, overheat in 5 minutes on a 10% grade. This is only a mile long hill before it overheats. I have some hills that are longer than that. The single rear 9C 2807 hub motor won't handle the hills around here. I'd have to use a motor with more oomph and less efficiency.

Same setup, 2 - 9C 2807 motors. To account for the shared load on 2 motors, cut CDa to 50%, bike weight and battery AH to 50%. 61% efficiency on 20% grade. 44 mi range on an average 2% grade. 700 * 2 = 1400 watts on the flats. 32 mph/77% efficiency, 28.5 mph on 10% grade. Won't overheat before I can climb any 10% I will encounter around here. The dual motor has little penalty on the flats, since the motors run high up on thier power curve, but dramatically better efficiency on hills. Two motors don't use 2X the power, if you query the simulator, actually it isn't much more than one motor. The advantage in rolling hilly country is tremendous - you are scooting around at over 30MPH up almost any hill maintaining high efficiency.

Anyway I think it is an interesting challenge.

[Rassy]

Hey, I'm with you llile. I not only think it will work, but I think it will work great. I even think that as long as you don't enter a corner to fast, i.e. tipping speed, you will be able to accelerate through the corner without any real steering issues.

I'll be watching for your build and test runs. Good luck.

[llile]

Dislexia is hell.

Just spent about two hours machining a difficult cut - I have two parts that will hold the two head tubes at the proper angles. I'm cutting angled holes that will mate with the head tube pipe and make a tight gap where the weld needs to go. These parts need to be symmetrical, in a left and right version. Each part is made of 3" X 1.5" aero cromo tubing about 8" long, with an angled 1.5' hole cut hear one end. The tubing is hard and thick, the hole saws maybe a little dull, the holes 0ff-center making a hard cut, centering bit skipping off, but after a lot of careful measureing , clamping and squirting oil on the hot cutter, I got them cut. Identical, not symmetrical. Oops. Maybe this vehicle will work with two left wheels?

My brain just doesn't process left and right very well. One time I had a $1700 drill fixture machined that was a perfect mirror image of the fixture my factory needed....

I want a T-shirt with a big red fist on it and the slogan "DISLEXICS UNTIE!"

[llile]

I wonder if we can work out a general equation or spreadsheet calc that will tell when a tadpole three wheeler reaches the tipover point going around a corner?


Obviously the variables include the mass and CG of the vehicle, the width of the wheels, the straight-line speed (or tangent speed) and the radius of the turn. But me college physics has evaporated from my brain other than that.

I'd pretend that the friction of the wheels with the ground can be ignored, because for a practical vehicle the thing will tip over before the wheels skid along sideways. That should simplify things a bit.

We can also simplify it if we assume the CG is at the same point fore-and-aft as the widest point of the wheels. This isn't really super accurate, because the the CG really appears somewhere in the triangle formed by the front wheels and the back, and the width of that triangle right where the CG occurs is the true "width" counteracting the tipover forces. Let's save that little complexity for later.

[llile]

E-Bike Porn! Some views of current 2WD recumbent using mid-drive stokemonkey. Note the fine Mad Max style paint job, ready for any post-apocalyptic nightmare.

Current 2WD bike http://members.socket.net/~llile/bikes/P1020120.JPG

2WD tranny detail with Stokemonkey http://members.socket.net/~llile/bikes/P1020121.JPG

Cockpit View http://members.socket.net/~llile/bikes/P1020123.JPG

Left side of Tranny http://members.socket.net/~llile/bikes/P1020122.JPG

[llile]

Here are some early photos of the trike build:

The Heart of the beast - I am using the guts of a downhill bike head tube for the leaning frame bearing http://members.socket.net/~llile/bikes/P1020107.JPG

Duct tape mockup of the trike front endhttp://members.socket.net/~llile/bikes/P1020108.JPG

Rear end donor bike http://members.socket.net/~llile/bikes/P1020111.JPG

How many donor bikes does it take to make a trike? http://members.socket.net/~llile/bikes/P1020112.JPG

Conduit mock-up of trike front end http://members.socket.net/~llile/bikes/P1020113.JPG I make a lot of mock-ups.

Joint Jigger: http://members.socket.net/~llile/bikes/P1020114.JPG these are indispensible for making round holes in round tubing at precise angles. Buy from here http://www.jointjigger.com/

Tie rod parts - tie rod bearings and concentric tube clamps for fine adjustments http://members.socket.net/~llile/bikes/P1020115.JPG

Welding station in the shop http://members.socket.net/~llile/bikes/P1020118.JPG

Drilling and Clamping station in the shop http://members.socket.net/~llile/bikes/P1020119.JPG

[llile]

Here are the two tubes that were full of dislexia fail: they are supposed to be symmetrical, but instead were identical. Kinda burnt up too, I think I need a new, sharp hole saw.

[Kirk]

Cool build. that one is really going to turn some heads. You might want to read up on Ackerman Compensated steering. I ride a Catrike Expedition with a 5303 on the back and that thing hauls ass and corners like its on rails.Haven't met a DF bike, either pedaled or as an E-Bike that can corner near as well.

[corkscrew]

I've thought about going a similar route.

You might consider these motors? Would look more elegant than the double sided motors in two MTB forks I think.



http://www.ebikes.ca/store/store_motors.php

Also, have you looked into the plans on Atomic Zombie? They have forums over there for building your own trikes.

Cool project btw.