Help design a DIY middrive for this heavily customized trike

[amberwolf]

Amberwolf's Schwinn Trike Rebuild / Conversion to Heavy Cargo Hauler

From here: https://endless-sphere.com/sphere/threads/the-sb-cruiser-amberwolfs-2wd-heavy-cargo-trike-dog-carrier.67833/post-1827806 I got an old Schwinn Tri-Wheeler Town and Country (probably the 1971 or 75 or 78 model, based on info from bikehistory.org) from someone I knew from the...

endless-sphere.com

There are mulitple possibilities; I don't presently have a set design in mind, just a goal of being able to drive both rear wheels from the system, without preventing the pedals from still doing their job the way they presently do. (the front wheel is taken care of separately).

The requirements are that it reaches the max of 20mph pretty quickly (say, 5-6 seconds worst-case) from a stop even with say, a 400lb load (me plus the trike plus some cargo; this one doesn't have to be as hefty as SBC which could do double that loading), and that it leaves the pedal drivetrain still working as-is, as much as possible, as a backup in case of all motor system failure. It has to be cheap, preferably built entirely with stuff I already have (just buying tires for this build is more $$ than I really have available, and I'd like to get a "new" battery to use on it).


The simpler the design and build is, the better, as it's more likely I'll actually finish it in a useful time.



The trike currently has cranks that drive the input to a Shimano 333-variant 3speed IGH that then drives a (peerless?) differential on the live rear axle to the 24" rear wheels.

The trike will change so that the rear wheels aren't on the axle ends anymore, but use 26" wheels on stub axles mounted to an extension frame to lengthen the trike and expand the cargo area (and move the seat back). Sprockets will go on hte live axle in place of the wheel hubs, and a chain from each of those to sprockets on the new wheel hubs on the stub axles will then drive the wheels from the pedal system.

The motor can input to any part of this system, but I expect that doing it to the IGH shell (effectively the diff input) will be mechanically the easiest; I can add whatever sprocket to the shell easier than I can to the diff.

I wouldn't mind inputting to the IGH input along with the pedals, but that would force the pedals to turn if the motor is turning (without adding more freewheels to the system, complicating the design and creation, which I am trying to avoid--the more work it is, the less likely I'll ever get it done).

If the motor would be better off with the ability to shift gears between it and the wheels, I can do one of several things:
--add a cassette/derailer (have IGH's but don't think I can use them in stub axle mode; see **** later in post) on each rear wheel, with the shifters tied together so they shift to the same gear. That would also add gearing for the pedal drivetrain if I needed it (unlikely).
---add a shiftable drivetrain between the motor and the diff input.
--probably other options i'm too tired to think of ATM.


I have a number of common ebike brushless hubmotors (some DD, some geared) that can be used to drive the system in one way or another..

I also have powerchair brushed motors with gearboxes that already have stub axles that can probably take the necessary loads, but tthis wouldn't be a middrive, and would preclude using the pedals to drive the wheels as backup to motor failure, unless I built a clutch system (like many geared motors have, etc) between the stub axle and the wheel hub, and put the pedal chain input on the hub side of that. There is one option that wouldn't preclude it, and that's if I can use the older gearbox type (I might have two of) that has a mechanical clutch inside to disconnect most of the gearing from the output shaft (to allow a human to push the powerchair they came from, if the power failed).

It would probably be easier (to preserve the pedal drive train usefullness) to use the gearbox output to drive the diff input.

Both the gearboxes and the hubmotors are easy to gear to the same rrpm range that the pedals would have; I did this for CrazyBike2's "best" middrive system that used this type of motor.


**** If it were possible to use them as stub-axle wheels with the heavy load they'd have in this use, I've got a couple of SturmeyArcher (or SRAM?) IGH + drum brake wheels from the ex-JumpBike sale a few years back. That would give me brakes *and* gearing on those wheels. But I can't imagine those axles would handle this kind of abuse.

The stub axle design is not an absolute requirement...but I would like to explore it and test it in this design, as it makes wheel/tire work roadside MUCH easier--I can literally just pull the wheel off and swap it out with a spare to get going, or I can easily access the tire/tube/etc without having to unload the trike and roll it on it's side to get the wheel out of the frame like I do on the SB Cruiser trike, etc.

If I can't use a stub axle, I can build the frame so it is a tiny bit like the old cadillacs with the body covers over the wheels, though in my case it would be a structural frame that can be unbolted with a QR on one end and swung out off the outboard axle end to get it out of the way of the wheel. (another feature I'd like to build and test anyway, as I always wanted to do this for the SBC but haven't had a spare transport while doing the experiment in case it didn't work out and had to be undone or revised repeatedly).



If I had two of them, I'd use the brushless powerchair motor that mounts to the frame and has it's own stub axle inside, and the motor is the hub. Then I'd just build bolt-on spoke-flange adapters for that to lace the wheels up on. But I've only got one, and I haven't found affordable (super-cheap) ones I could get another identical one of (or a pair of that are different from mine). But all this is getting much more complicated than I'd like...I'd much rather just drive the pedal drivetrain and have a lot lest custom stuff to do.



I'm sure this is not enough info yet to come up wiht something useful, but my brain is turning off (still have a head/chest cold that's no fun), so will just post it and see what responses there are,a nd work from there.

 

[offGridDownUnder]

The motor can input to any part of this system

Your other mentioned preferences may contradict this, but perhaps consider the concept I use for my Greenspeed trike with stoke-monkey motor (see sig).

I have differential sprocket sizes on my motor and wheel, in my case lowering the ratio as I want hill climbing ability despite possible large loads (so I always get home). As a consequence, I find my trike accelerates very well, while topping out at a locally legal speed. It's quick, not fast. As top speed where I live is limited, I have the room to play with the lower ratio.

I started with a chain connecting the motor output (from the 6 bolt ISO mounting pattern on the left of the motor) to a corresponding sprocket on the left side of the rear hub. Given your idea of stub axles you have some options with a similar approach. One I can think of is that these stubs are on a frame, hinged at the bottom, that bolts (or quick-releases?) in the tensioned position, but releases to permit easy wheel removal. Alignment accuracy is required at the bolted position, but not in the released position, as that is just for changing tires/wheels.

Really contradicting your other constraints is the idea of doing this with two Gates Belts. More $ up front, but they are excellent for long term tensioning, with full ability to provide regenerative braking, and great environmental resistance and low maintenance. In either case, you would have two ISO mounting points on your drive axle, leading to the same on each of your actual rear wheels on their stub axles - I don't see any reason you couldn't use two front wheels with one turning "backwards" to make the arrangement symmetric, or both conventionally with just an offset chain line on one relative to the other.

 

[amberwolf]

Thanks for trhe reply and thoughts and ideas! See below for specifics:

Your other mentioned preferences may contradict this, but perhaps consider the concept I use for my Greenspeed trike with stoke-monkey motor (see sig).

I have differential sprocket sizes on my motor and wheel, in my case lowering the ratio as I want hill climbing ability despite possible large loads (so I always get home). As a consequence, I find my trike accelerates very well, while topping out at a locally legal speed. It's quick, not fast. As top speed where I live is limited, I have the room to play with the lower ratio.

I don't need fast, just 20MPH top speed (legal limit) to cruise at on roads (back streets and no-traffic areas I often ride slower, 10-15mph). But I do need quick acceleration (I think I already posted it but about 5-6 seconds 0-20mph). It's nearly all flat around here, so I don't really require hillc-limibing ability. Going up driveways into parking lots, and up the shallow slope of the bridge over the canal at the south entrance of Metrocenter are the only common slopes I deal with. Once in a VERY long while I have to deal with the North Mountain slope starting over at dunlap/7th(st?) up to the top of NM before it goes back down the other side; the SB Cruiser does that under motor power just fine, though it is running hard and using a lot of battery to get up that long kinda steep road. Hopefully this trike wouldn't have to be used for that, but it'd be lighter and take less power to do so maybe it could anyway. (I could never pedal up that so wouldn't even try--I couldn't do it even way back in my regular bicycling days)

Something sort of like the SM system is likely what I'll end up with, except it won't be driving the crank at all--if I drive the pedal drivetrain it would be post-pedal-chain. At the IGH shell is the "easiest", as I can add sprockets there easier than anywhere else prior to the transaxle.

If dual motors are used, then they could either drive the transaxle itself at the outboard ends with the sprocket-hubs I would be building to replace the wheel hubs that wont' be mounted on it anymore, or be inserted "under" part of the chain from those hubs to the actual wheel hubs so they simply pull the chain along there. In that event they would ahve a freewheel on them so they wouldn't add drag to the pedalling in the hopefully-rare-as-never event of completely powered system failure. (has only happened once in the decade I've been using SBC, so I can hope for the same or better with this one :lol: ) .

I started with a chain connecting the motor output (from the 6 bolt ISO mounting pattern on the left of the motor) to a corresponding sprocket on the left side of the rear hub. Given your idea of stub axles you have some options with a similar approach. One I can think of is that these stubs are on a frame, hinged at the bottom, that bolts (or quick-releases?) in the tensioned position, but releases to permit easy wheel removal. Alignment accuracy is required at the bolted position, but not in the released position, as that is just for changing tires/wheels.

The stubs will be on plates bolted to the extension frame; they have to be able to support the trike weight single-ended, so it'd be too complicated to add pivoting to those. (it's complicated enough to add a pivot and QR to the brake caliper mount on the plate to allow wheel removal without removing the rotors or calipers).

I don't need to move the stub axle to do wheel on/off. That's done by just undoing whatever hub-retention method I end up with (probably a big nut and washer setup, onto the threaded outside of the axle end, or a bolt going into a nut welded into the end of the axle.

Chain tensioning I'd probably do by slightly oval boltholes on the extension frame where it mounts to the trike itself--set it up for sufficient tension when isntalling that subframe, then use a master link on the chain to remove it if necessary.

Alternately slightly oval holes on the mounting plate for the stubaxles, and bolt them to the extension frame instead of welding them, and adjsut taht way at install, then use the masterlink as above.

Really contradicting your other constraints is the idea of doing this with two Gates Belts. More $ up front, but they are excellent for long term tensioning, with full ability to provide regenerative braking, and great environmental resistance and low maintenance.

Can't imagine going that way; aside from no money for the belts, sprockets, etc., installing or serviceing/changing them would require at least some disassembly of the trike to get them on and off. Chains I can just "break" or use master links (that I already have), and I also already have chains (off the scrapped bikes, etc; probably have to splice multiple different chain sections together, but I've done that before when necessary, not optimal but it works).

I don't plan on regen braking back there, so not worried about the drivetrain handling that.

I use it on SBC because I have it, and it's handy, but even with the Phaserunners on both DD motors in the rear it's not as controllable or useful as simple friction brakes would be. (it also doesn't behave the way it needs to--it only has hard braking available after it drops down below like half of my cruising speed, above that it's way too gentle, and the transition is quite sudden when I drop below the cutoff point if I'm holding the lever already maxed for a quicker stop; there are no settings I've found that correct this; it's jsut the way the controller appears to work).

The only controller I have that has propoprtional regen other htan the PRs is a signle older grinfineon, and it is very weak, not useful for anything past a typical bicycle-weight bike with medium-sized rider, if even that. It'll probably be used to run the front hubmotor (possibly the gmac which might be able to amplify the braking force neough to be hepful). Everythign else I have either has no regen at all, or it has just on/off that isn't useful.

If I use a brushed motor then it's easy to make a variable regen controller even if the contgrollr for the motor itself doesnt' have regen, just by using a brushed controller design in reverse powered by the motor and feeding a resistive load, throttle controls braking force. (a four=quadrant controller could do it too, but I don't have one and am not interested in building one though its fairly simple compared to bldc controllers and i have good (if old) designs from 4qd).

In either case, you would have two ISO mounting points on your drive axle, leading to the same on each of your actual rear wheels on their stub axles - I don't see any reason you couldn't use two front wheels with one turning "backwards" to make the arrangement symmetric, or both conventionally with just an offset chain line on one relative to the other.

The wheels will be built for this out of rims/spokes I already have, and i'll probalby hve to build the hubs out of tubing and plate and bearings from other things, so it doesn't matter if the parts started out from front or rear wheels (the rims / spokes i have are for one of each off a scrapped bike).

 

[offGridDownUnder]

But I do need quick acceleration

My hill climbing ability is due to torque from motor choice and then significant down-gearing. I find that translates to acceleration. That's what I thought would be useful for you. My trike is not fast (high speed), but it is quick (to get up to speed).

 

[amberwolf]

That's what i intend, and use this now to some degree on sbc by using the 20" wheels. But for this one i need the larger wheels (29s if I had them) to get smoother ride becuase of the pavement we have (small rear wheels is my one seriously-regretted choice in sbc's design for this reason), and I'm limited by whatever I hapen to already have in batteries, controllers, motors, etc. for cost reasons.

Motors I have stuff that would certainly spin up faster under higher voltage than I use now (52v 14s), and some of the big heavy ones (mxus 450x) definitely spin up faster than I use now even under this voltage. Smaller ones probably do too; I have to check each possible one I have that works with whatever mechanical design of middrive comes out of this and fits in the space available for it.

Except for the PRs I dont' have any controllers that go up above a 52v battery reliably (except the Lebowski/hondas that I haven't finished, and those are rather gigantic for the purpose).

Batteries I hve are modular (EIG cells) so I could rebuild one into whatever's needed....but then all controllers/electronics on the trike powered by the traction battery have to support that voltage. So that requires at least two controllers (one front, one rear) or two separate traction batteries at different voltages (and different chargers, etc).


There's no perfect way to build this without buying new things, so like most things I do it will be a bunch of compromises between what I already have and can make out of that, and what would really work better.

 

[amberwolf]

From the main thread:

 

[amberwolf]

So some of teh design is now "set", in that parts are on their way or decided on for some of it.

As in the compposite impage abve, the wheels (on stub axles mounted to the rear extension frame) will be driven by chains from the ends of the diff axles.

The diff will be a peerless type, I think it's the 100 model, that will replace the one built into the schwinn.
assembly video
https://www.peerlessgear.com/wp-content/uploads/2020/09/100-series.pdf

The new diff is much larger and tougher than the little thing it came with. bercuase this diff is bigger, it progably wont' fit in the enclosure the trike has for it, so I might hafve to build and entirely new frame piece for this. I'd ratehr not modify the schwinn's frame anwyhere i don't have to, but we'll see once it arrives this week how much mod i owul dhave to do to make it fit.

the axles on this diff are much bigger than needed for the power transfer to the chains, since it doesn't have to support any wheight, so i will probably cut them off near the diff*** and use the original axles or some pipe to do the transfer to the chain sprocke touptuts. then use the cut off larger axles tomake the stub axles out of.


***they[re part of the gearing so can't just take them off like i can with the schwinn's diff that uses cotter pins for that. i thin kthe gears slip on with either doulbe d or splines depending on the version. if it's a double D i can make my own mating end for new ouptut shafts, but i can' t make a spline to match it. then i can use the entire thicker axles on the rear if i need them, and don't have the complexity of cutting and mating output shafts to this diff's axle remains.



The original might survive eferything i need it to do, but the new one should be virtually indestructible for my usage, which is a requirement since both pedal and motor power go thru it to get to the wheels. the motor won't go thru the IGH though, it'll go direct to one side of the diff (the igh output goes to the other, for just the pedals).

I think I will try the GMAC as the motor for this, it should be usable at a lower quieter rpm. I might give it it's own shiftable gearbox (probably via another IGH). to give it a low gear and a high gear, so I can let it spin faster for better efficiency at startup from stops then shift it to quieter rpms for normal cruising speed with the hgiher gear.


still pondering and looking for ideas / parts to mount the motor itself, etc.

 

[neptronix]

I was just about to recommend a GMAC
If this wasn't a super heavy bike, i'd recommend a Shengy SX2 because it is ~4% more efficient.. and somehow, the more you spin it up, the more efficient it gets, so if you can get a big reduction ratio from 400rpm to crank (~85rpm), you're in business.

On the MAC, because the low end torque is so powerful, i'd run a phase to battery amp ratio of 2:1 to 2.25:1 or so to clip off most of the initial torque. This is how a lot of mid drive motors are programmed to prevent overloading the drivetrain.. from this ratio, you tend to get a fairly flat torque curve.

 

[amberwolf]

I was just about to recommend a GMAC

Well, I already have the GMAC from this:
https://endless-sphere.com/sphere/t...vy-cargo-trike-dog-carrier.67833/post-1776196
but even damaged as it is it should run the trike.

It worked in the 20" wheel to drive SB Cruiser, with two caveats:

--Because it would spin up really fast if it ever left the ground even for an instant, the Phaserunner that was driving it would error out and I'd have ot stop, in the middle of traffic, before the top of a sloped bridge, to power the whole systme off, wait a few secondsand then power back up, wait for everything to get ready, then ride up the rest of the slope and get out of the way of traffic. As long as I also had the second motor on the left side (an mxus 450x dd) then I wouldn't have to stop right then, but i'd have less climbing and acceleration power, etc. Since I don't have a third PR to use on the new trike, it'll be run by (most likely) a 2018-ish grinfineon.

I didn't like the noise it makes to get 20MPH in a 20" wheel; if I can run it at half that rpm (whatever that is) as a middrive it'll be ok.

If this wasn't a super heavy bike, i'd recommend a Shengy SX2 because it is ~4% more efficient.. and somehow, the more you spin it up, the more efficient it gets, so if you can get a big reduction ratio from 400rpm to crank (~85rpm), you're in business.

Yeah, but teh GMAC is too noisy at high rpms because of the gearing, so it has to run at lower rpms for me.

On the MAC, because the low end torque is so powerful, i'd run a phase to battery amp ratio of 2:1 to 2.25:1 or so to clip off most of the initial torque. This is how a lot of mid drive motors are programmed to prevent overloading the drivetrain.. from this ratio, you tend to get a fairly flat torque curve.

I won't have a programmable controller, so I will have whatever it comes factory-set as. Most likley it'll be the old grinfineon I used to use on the SB Cruiser before I put the PR's on there.


some stuff in the main thread
https://endless-sphere.com/sphere/t...onversion-to-heavy-cargo-hauler.125947/page-2
about the diff that arrived today, and some other bits for the trike, some relevant to middrive design. lots of pics.

 

[amberwolf]

Some more potential drive parts in the main thread

Amberwolf's Schwinn Trike Rebuild / Conversion to Heavy Cargo Hauler

I think you'll break the axles off if you try that. Plus, keep in mind you have to shift them both from the right side, and drive them both on the right side. That wouldn't work at all for a wheel mounted on the right.

endless-sphere.com

 

[amberwolf]

and more potential drive parts. Amberwolf's Schwinn Trike Rebuild / Conversion to Heavy Cargo Hauler

also a mockup of a potential layout using the powerchari lift bed as the cargo bed. it's still a few inches to a foot too long, once the trike basket holder frame is replaced with a suport frame fo th seat that moves it forward enough to match the sb cruiser's seat/crank relationship, there'll be romo for the cargo bed to move that much forward too.


this is where the old diff went, and the new one would ned to go there or just behind it, with the axle ends with sporckets on htem to feed the rear wheels via chain. whichever motor gest used that drives the diff can go in a number of places, but the most likely is in the rear triangle among the seatstays and chainstays.

 

[amberwolf]

Still looking for input on the design of the drive system.

Right now the "givens" are":

--"regular" bike wheels on single-sided axles at the rear, one on each side, with hubs that have a driven sprocket on them as well as a disc brake rotor.

--front and rear wheel designs the same, so the same wheels can be used on any part of the trike, to allow keeping a spare wheel onboard for rapid roadside repairs in the summer heat or other bad situations.

--peerless diff with the drive sprockets for the chains that go to those hubs on the ends of the diff axles.

--needs high torque to get going very quickly from a stop even when the trike is loaded down with several hundred pounds of whatever.

--doesn't need high speed, 20mph max (though the system should be capable of more than that to ensure good acceleration all the way to max speed, it will be limited to that 20mph in the Cycle Analyst).

Lowest possible ground clearance to still allow going over


Open options:

Several possible motor types and designs; see the various ones in previous posts. (controllers available also listed).

Motor mount design completely open.

Motor placement partly open, it can go "anywhere" from behind the seattube to the front edge of the cargo area. Alternately it could go under the cargo deck right between the wheels, since the wheels would provide it clearance over anything the trike would ride over.

Peerless diff mount and axle support.

Drive method from motor to diff

Drive method from pedal drivetrain's IGH output to diff.

Rear wheel stub axle mount, and design; needs to allow for rotating disc caliper off the disc by undoing one quickrelease fastener.

Front stub axle mount to match rear ones; can be for a single-sided fork or a potential suspension triangle design.

Front "fork" itself is open for design; just needs to be a single-sided axle for the above reasons.

Battery choice (including system voltage), design, and placement.


Framework for the cargo area and seatbox area is stil up for design; these would be what the motor mounts and rear wheel mounts will need to be mounted on. If possible this framework will bolt or clamp to the Schwinn framework, but welding is an option if necessary.
Also keeping the Schwinn original as much as possible is also a goal but not a requirement.



I'm sure I left stuff out; keep dozing off, so just ask questions and i'll add whatever info's needed.

 

[amberwolf]

Grr...Xenforo ate my post with a security error and somehow my clipboard is also cleared. :/

So...retyping it, dunno if I'll remember most of what I meant to say.

*****************
I keep forgetting about this conversion:

Converting a hubmotor to a middrivemotor

Hi folks, The idea is to use the axle of a geared hubmotor as output shaft. No one ever did post anything about this in the past ( I searched a lot), that's why I wanna show how it might work. I use a 24V 250W front driving MBG Goldenmotor for this mod (reviewed the 36V rear version here...

endless-sphere.com

that I've always wanted to try, but never had a good enough reason. But this trike might be a good experimental testbed. I have the GMAC, which could generate more than enough power for the whole trike this way, but would be kind of "expensive" to experiment with. I also have two or three litle 350w Fusin geared hubmotors, one of whihc is clutchless like the GMAC.

They're physically small so will fit easire eiwthin the framework, especailly once converted so the shell doesn't spin, only the axle. They're not all taht much bigger around than the Peerless diff itself, and narrower. Perhaps the same volume.

pics of the motor below, in a 26" wheel, and variuos inside parts, including one of the clutch on the first one after it fractured (which is why I would use the clutchless version).