Warren's 350 MAC mid-drive (finally) recumbent

As I understand it, once it is moving the correct direction, a sensorless controller can sync from the motor's back-emf.





Edit: removed landmine. ES remains independent!
 
Check out the Castle Creations line of ESC (controllers in RC world) they're tiny!! You could fit two of the littler ones in a pack of cigarettes.

Good luck, how about some pictures?

Katou
 
Hi Warren,
Warren said:
Folks,

I have been in correspondence with cell_man about a 350 watt, geared, rear, MAC hub motor. It sounds perfect, except for an aluminum threaded freewheel mount. He is checking into availability of the newer version with a steel freewheel mount.
I'd prefer the aluminum. Matt had FW adaptors machined in both aluminum and steel. The reason he had some made in steel was not that the aluminum wasn't strong enough. He said it was because with steel he had the option of removing the set-screw-shoulder (narrower overall width) and welding the adaptor to the shaft. The reason I would prefer aluminum is it is quite a bit lighter. Here's Matt's description:
http://endless-sphere.com/forums/viewtopic.php?f=28&t=7180&p=122330
...he steel adaptor weighs 4.6 ounces. The aluminum adaptor weighs only 1.6 ounces!

The thread shoulder can be machined off allowing the set screw hub to be inside the FW thread section. That narrows the setup by 1/2 inch. But, it hides the set screws. However, in some situations, that may be desireable...

One of the three pictures:
file.php
 
Mitch,

My concern was not that the threads wouldn't hold up, but that the aluminum threaded portion might break off the cover, as apparently has happened on several different brands of hub motors. The problem has been not a large enough radius between the threaded portion, and the side cover causing stress riser cracking.

My latest idea is to use two sprockets bolted to the disc brake mounting holes, on a 350 watt, front motor. It is almost a half pound lighter, mostly from the shorter axle.

I figure I will make a couple of spacers to separate the cogs from the motor and each other...enough to allow chain clearance.

http://www.velosolo.co.uk/shopdisc.html

http://www.tomicog.blogspot.com/

I will flip the motor so the cogs are on the right side. The inner cog will be a 16 tooth for the chain running to the rear wheel. The outer cog will be a 15 tooth for the chain to the front single chainring. This is to get my cadence down to a reasonable level. A 13 tooth would be my preference, but 15 is as small as the disc brake bolt circle will allow.

To compensate for the motor running backward, I will cross the chains (figure eight) running to the crank, and to the rear wheel. Since these are long runs on a recumbent, the side load where the chain crosses is small. A short section of teflon chaintube should allow the slack side chain to deflect around the tensioned chain. It will look essentially like the normal chain run on typical recumbent.

http://www.ransbikes.com/Enduro09/Enlarge09.html

Since I want to be pedaling, any time the bike is propelled, this should be about the simplest setup.

Cell_man, Paul, must be on holiday. He hasn't gotten back to me. :)

Warren
 
Hi Warren,
Warren said:
My concern was not that the threads wouldn't hold up, but that the aluminum threaded portion might break off the cover, as apparently has happened on several different brands of hub motors. The problem has been not a large enough radius between the threaded portion, and the side cover causing stress riser cracking.

My latest idea is to use two sprockets bolted to the disc brake mounting holes, on a 350 watt, front motor. It is almost a half pound lighter, mostly from the shorter axle...
It sounds like it breaks due to the pounding the hub motor gets mounted in a wheel every time the bike hits a bump. Assuming its a decent quality hub, a hub in a mid-drive configuration, which isn't subjected to the same stresses should be fine (unless I'm missing something :) ).
 
From http://www.ebikes.ca/troubleshooting.shtml item number eight in reference to the threaded collar shearing off on the C'lyte 400 series motors
(emphasis mine)
#8
Threaded Collar Shearing Off

The aluminum side covers of the Crystalyte hub motors aren't engineered by NASA. Sometimes there are voids or weak spots in the castings and sometimes this can have consequences. We have seen a few instances where riders have managed to shear off the threaded shoulder that the freewheel screws onto. This usually happens when they are standing pedalling in the easiest gear, putting a maximum amount of torque on the freewheel. Because the threaded portion has a sharp inner corner rather than a proper fillet, it experiences a high stress concentration and is prone to fatigue here as a result. The 500 series rear motors use steel threads and we've never seen one of those fail.
 
Nice idea Warren, but won't you have to reverse the internal freewheel in the motor? Oh. Wait. Crystalite. Well if Crystalite hubs break perhaps you should be looking at a different motor supplier? :wink:

My ideal is still a geared hub with two sprockets. One freewheel to the chainset, one fixed to the rear gearset. Do you think there's enough meat on a typical rear hub motor to mount these on the thread? The fixie sprockets on those sites look like they use all the thread width for the mount.
 
jbond,

> won't you have to reverse the internal freewheel in the motor?

No. Because it will work exactly as it does mounted in a wheel.

> Oh. Wait. Crystalite.

No.....350 watt MAC.

> Do you think there's enough meat on a typical rear hub motor to mount these on the thread?

One fixie cog, or one freewheel is all that will fit. There is not room for two.

Warren
 
Warren said:
> won't you have to reverse the internal freewheel in the motor?
No. Because it will work exactly as it does mounted in a wheel.

Ah, I didn't allow for the chain cross. So you mount a twin sprocket on the disk brake side, reverse the motor and then twist the chains. The end result is the hub is driving in the same direction.

Warren said:
> Oh. Wait. Crystalite.
No.....350 watt MAC.
I got confused by the quote above.

Warren said:
> Do you think there's enough meat on a typical rear hub motor to mount these on the thread?
One fixie cog, or one freewheel is all that will fit. There is not room for two.
Shame. Cyclone have a set up like this with a 2 sprocket assembly where one free wheels. but I think the shaft diameter is much smaller than a typical hub gear cluster thread. These rear motors typically take a 6 or 7 speed shimano freewheel so there's plenty of width available for two sprockets but the actual screw thread doesn't sound wide enough. Looking at those pictures of the fixie gears on Veloshop, they all use a lockring. Couldn't you use a freewheel as the lockring?
 
jbond,

> These rear motors typically take a 6 or 7 speed shimano freewheel so there's plenty of width available for two sprockets but the actual screw thread doesn't sound wide enough.

Yes. The RH, 1.375" diameter thread is typically no more than 9 mm long.

> Looking at those pictures of the fixie gears on Veloshop, they all use a lockring. Couldn't you use a freewheel as the lockring?

The inner RH thread is 1.375" diameter, and only 6 mm long. The outer lockring thread is LH, 1.29" diameter, and only 3-4 mm long.

If you placed a big enough order, you could get hub motors with low enough rpm, no spoke flanges, a LH threaded 1.375" inner for a 16 tooth southpaw freewheel to drive the rear wheel, and the smaller 30 mm RH threaded size for a 13 tooth freewheel to be driven by the pedals. This is presumable what Hanebrink and Pi Cycle do/did.

Warren
 
Hi Warren,
Warren said:
If you placed a big enough order, you could get hub motors with low enough rpm, no spoke flanges, a LH threaded 1.375" inner for a 16 tooth southpaw freewheel to drive the rear wheel, and the smaller 30 mm RH threaded size for a 13 tooth freewheel to be driven by the pedals. This is presumable what Hanebrink and Pi Cycle do/did.
For his mid drive bikes Marcus (Pi Cycle) used standard off the shelf rear (geared, except possibly when he was experimenting with regen) hub motors. Its very important to him that buyers will always be able, as much as possible, to maintain their bikes with off the shelf parts. He is really opposed to our throw-away mentality and wants his bikes to be heirloom quality, which requires maintainability.

I'm sure on his mid drive bikes he had a fixed sprocket between the hub motor and rear hub and a FW sprocket on the hub motor driven by the pedals. With an Alfine 8 speed, with the built in FW that meant none of the chains were back-driven. At one point, using Nuvinci hubs, on which it was easy to disable the FW, he tried that and experimented with regen.
 
Mitch,

> Marcus (Pi Cycle) used standard off the shelf rear (geared, except possibly when he was experimenting with regen) hub motors. I'm sure on his mid drive bikes he had a fixed sprocket between the hub motor and rear hub and a FW sprocket on the hub motor driven by the pedals.

He must have used Loctite to keep the fixed sprocket on. Every time you hit the throttle, the motor would try to unscrew it. There would be an eighth inch of thread left, at most, for the freewheel. You would definitely want to Loctite that as well. In this case, to prevent stripping those three threads engaged.

I'm not saying it won't work. But it a had better be a steel thread. Even then I wonder how long it would last.

Warren
 
Mitch said:
Marcus (Pi Cycle) used standard off the shelf rear (geared, except possibly when he was experimenting with regen) hub motors. I'm sure on his mid drive bikes he had a fixed sprocket between the hub motor and rear hub and a FW sprocket on the hub motor driven by the pedals.
Warren said:
Mitch,
He must have used Loctite to keep the fixed sprocket on. Every time you hit the throttle, the motor would try to unscrew it. There would be an eighth inch of thread left, at most, for the freewheel. You would definitely want to Loctite that as well. In this case, to prevent stripping those three threads engaged.

I'm not saying it won't work. But it a had better be a steel thread. Even then I wonder how long it would last.

Warren
I'm not sure exactly how he implemented it but I'm absolutely sure he used a standard hub motor. If you used a track cop for a fixed sprocket:
31qjjQwnXEL._SL500_AA300_.jpg

inside the FW the FW and the sprocket would tighten against each other. He could have used some kind custom shaft extender that would fit a standard hub but I think he would have mentioned that if he did. Here are some pictures so the rest of our viewers will have some idea of what we are talking about. Early version:
http://www.bikerumor.com/2009/04/29/spy-shot-raw-pi-mobility-commuter-bike-frame/
pi_0228.jpg

When I saw it the entire drive section was an independent piece that bolted to the frame. It contains the bottom bracket and a place for the hub.

What you’re looking at here is A) the first production Pi Mobility frame to roll off their line and B) not technically a spy shot…the owner of the company sent these images to us. But, they’re pretty cool. Check out the suspension mech on the fork, totally recessed into the legs.

If you have no idea what a Pi Mobility bicycle is, read our other post for details, but basically, it’s an electric motor bike with Shimano Alfine 8-speed hub and a 20mph unassisted top speed.
pi_0225.jpg


A later iteration of a completed bike. Its a little hard to spot but there is a "fender" that does double duty as a structural piece of the frame supporting (triangulation) the drive section:
pi1.jpg
 
I'm just imagining this, as I haven't got a hub, fixed sprocket and freewheel to hand. I think it should work if you put the fixed sprocket on first and use the freewheel as a locknut. Rather than or as well as loctite, one possibility would be to do it all up tight and then put a grub screw in the shoulder of the fixed sprocket. Apart from the strength of the hub itself, the big question is whether there's enough thread to get a standard track fixed sprocket and a standard freewheel on there.

One question is about gearing of this setup. If you have a 14-28 cluster on the rear wheel, you could use 17t sprockets on the motor. There'd be no effect on pedal gearing, but the motor would be geared up (17-14) for a bit more top speed, while still having 17-28 gearing available for hill climbing. By using different sized sprockets you could perhaps reduce the chainset big ring (from 52 to 48 say) and still keep the same max speed cadence or the same cadence for a higher top speed.

The disk bolt attached sprockets are a neat solution, but I can't see how to bolt on a second freewheel. And crossed chains are not elegant even if they can be made to work.
 
jbond,

" I think it should work if you put the fixed sprocket on first and use the freewheel as a locknut.... put a grub screw in the shoulder of the fixed sprocket. Apart from the strength of the hub itself, the big question is whether there's enough thread to get a standard track fixed sprocket and a standard freewheel on there."

Yes. It should work, in theory. The big problem is the threaded portion is too short. You only have an eighth inch if thread left for the freewheel, with a 1.375 x 24 tpi that means three threads. I have no doubt that, on a steel thread, a half inch long, it would work. However, with only an eighth of an inch to engage the freewheel, I believe it will fail pretty quickly. It happens to guys on fixies , if they run cheaper, thin cogs. If you hammer, you put huge loads on the threads, twice a revolution.

"By using different sized sprockets you could perhaps reduce the chainset big ring (from 52 to 48 say) and still keep the same max speed cadence or the same cadence for a higher top speed."

The inner cog size will have no effect on your cadence. It will only affect the speed of the bike. Your cadence is set by the the speed of the motor. For maximum efficiency, this is about 80% of no-load speed. The smallest freewheel you can fit on a 1.375 x 24 tpi thread is a 16 tooth. The biggest non-custom chainring is 60 teeth. 60/16=3.75 Typical cadence for most humans is 80 rpm. 80x3.75=300 rpm 300/.8=375

You want a hub motor with a no-load speed about 375 rpm. This is not a problem...unless you want to run 48 volts. Most hub motors are wound for 24 or 36 volts. The Crystalyte 408 is good. The C 4011 is even better. You could run much more common chainrings...even get 60 rpm cadence for loafing.

"The disk bolt attached sprockets are a neat solution, but I can't see how to bolt on a second freewheel."

You can't. I don't want the option of not pedaling. I would quickly turn into a passenger. If I want that, I'll get a motorcycle.

"And crossed chains are not elegant even if they can be made to work."

Absolutely true...which is pushing me back to driving the left crank. These came yesterday!

http://tinyurl.com/25jhqvb

Warren
 
Hi Warren,

"The disk bolt attached sprockets are a neat solution, but I can't see how to bolt on a second freewheel."

You can't. I don't want the option of not pedaling. I would quickly turn into a passenger. If I want that, I'll get a motorcycle.
Like a Stokemonkey? Always seemed unsafe to me. If your foot slips off a pedal you have about 3/4 of a second to get it out of the way. Clip-ons would be even worse. And what happens if you have to, in an emergency, cut off the power, if you are not quick enough?
 
Mitch,

'Like a Stokemonkey? Always seemed unsafe to me. If your foot slips off a pedal you have about 3/4 of a second to get it out of the way. Clip-ons would be even worse."

I have been riding a tandem bicycle for 14 years. It, like most tandems (but not all) doesn't have a freewheeling crank. A foot comes off a pedal occasionally. It is no big deal. Just pedal with one foot, until you can put it back on. Actually, clipless is much better. Now getting a foot on and off on rollers is a bit trickier.

http://www.youtube.com/watch?v=wME5NTriTco&feature=player_embedded#

Warren
 
Here's a dude that set up his bike with a stokemonkey drive, and he's pretty happy with it:

http://endless-sphere.com/forums/viewtopic.php?f=4&t=15294

Katou
 
Like a Stokemonkey? Always seemed unsafe to me. If your foot slips off a pedal you have about 3/4 of a second to get it out of the way. Clip-ons would be even worse. And what happens if you have to, in an emergency, cut off the power, if you are not quick enough?

I've got about 2600 miles on a Stoked Xtracycle. It seems like a very natural way to drive the system to me. I actually prefer to have the pedals driven. It is good feedback for what the motor is experiencing. Even with the Stokemonkey driving with no load, the pedals don't spin faster than you can pedal. And usually, the motor is not spinning at its max RPM when you are riding, depending on what gears you are using. I guess how often your feet come off the pedals depends on how you ride (going Gonzo down a mountain or cruising a beach). I'm a bike commuter (no clip-ins) and it just nearly never happens. In the two or three times I can recall, I just naturally took my thumb off the throttle. Like someone said earlier, its like riding a tandem and most tandems don't have freewheeling cranks either.

I've never been it by a pedal with the Stokemonkey. It just doesn't happen so I wouldn't worry about that. However, if you use your ebike more as an electric moped, it might be annoying to have your feed going 'round for no reason. I can see both sides of it. For me, I like it the way it is and would not put a freewheel on the cranks even if someone offered to do it for free. I think some people would prefer it the other way, but which ever way you have it, you will probably get used to it very quickly.
 
I'm sure you can get used to anything. And maybe driven cranks works for some people, but I don't think it would work for me, so I'd want to avoid designing it in as a requirement. My current bike has a throttle-pedelec switch (either/or). I find myself using the pedelec most of the time as a kind of cruise control. Even though pedelecs are a bit crude and take 1/2 a turn to start and 1sec to stop, it makes a very natural cruise control. I only use the throttle for fine control in heavy traffic or in the woods. In both cases, there are times when I want to use just the motor with no pedalling while standing on the pedals. Perhaps with a recumbent, this is less of an issue and if you're used to a 'bent standing on the pedals is obviously out and if you're moving you're pedalling.

I've been thinking a lot about the user interface to E-Bikes. There's a lot of options from twist-thumb, Pedelec, Panasonic style bottom bracket torque sensors, Cytronex buttons, 3 speed controls on the controller, current limiting via a CA, RC style Servo pots and probably many more. If we're thinking about light weight and assist, then eeking out battery capacity and efficiency is really important. And if that's the case, then we need to try and create a UI that is very intuitive and doesn't require any thought but which also encourages minimal use of the battery-motor system. The RC guys are heading down a route where intelligence in the controller is added for current limiting and soft start just to make the motors and esc survive. This opens up possibilities for fine tuning battery capacity use as well.

Note that the Panasonic system is mid-drive, pedelec-torque controlled and gets very high mileage from the battery capacity. This seems to come from two places. First, the rear gears mean that the user naturally keeps the cadence fairly constant, which helps the low powered motor climb hills by going slower with greater torque. Second, the UI is very simple. Pedal harder, get more assist; don't pedal at all, get no assist; and that's it. I think we can get close to that with a DIY mid-drive but add some features as well. eg. Motor only with no pedalling and a variable or multiple choice assist level. As I said at the beginning, there are some situations where I want complete fine control. But most of the time, I really want something more like a cruise control where I can dial in 100w assistance (say) and then not think or do anything about it at all, just slowing down for the uphills and speeding up for the downhills.
 
I think I chimed in on this topic once before, but here's my 2 cents.

When setting up my mid-drive trike I read the FAQ's on the Stoke Monkey site. Always pedaling and being in touch with the motor sounded good, so I started with no freewheel at the crank. Didn't like it at all. Even though I like to always pedal, I would do something on every ride to make the motor "jerk" on the pedals. Sometimes I would simply quit pedaling before releasing the throttle, or stop pedaling for a moment to adjust myself on the seat, etc. Even caused a couple of chain derailments. I am 70, and it's hard to teach an old dog new tricks.

Anyway, once I put the freewheel crank on the trike I loved it. No more jerking no matter what I do. No more derailments. And I still feel just as in tune with the motor, since I am always pulling along with it and as long as my cadence is in the right range the motor is also in the right range. And if I really need to take a break from pedaling I have that option.

Anyway, I'm a big believer in each his own, so try different things and do what works best for you. :D
 
jbond said:
But most of the time, I really want something more like a cruise control where I can dial in 100w assistance (say) and then not think or do anything about it at all, just slowing down for the uphills and speeding up for the downhills.
FWIW, the Cycle Analyst has an optional setup you can wire up that uses the current limits you can set in it to do exactly this, using any motor system.

Regarding cranks freewheeling vs non, I'd say that like Rassy I prefer them freewheeling for the same reasons. I had non-freewheeling cranks on CrazyBike2 with the powerchair motor/jackshaft middrive setup, and while it worked well there were a bunch of times bad things happened because the motor forces the crank to turn. Destroyed chains and chainrings, derailments, torn pants, almost a crash once because of pants caught in chain; didn't tear and got motor stopped quickly, but unable to free my leg to put my foot down on the left side that I was starting to tip towards (as I was trying to make a left turn), and ended up having to sharply turn and lean right instead so I could stop myself with my free right foot instead.

I now have most of the parts needed to convert that to a freewheeling jackshaft so the pedals and motor are totally independent of each other, but both input to the same regular bike rearwheel drivetrain. Have to make a few other parts for it, and experiment with them, to be able to have my middrive back. Then run it from the CA so it is current-limited for a constant-torque capability, with both presets and throttle control.

So, it's down to preferences as to freewheeling or not, but I think that there *can* be safety issues with a non-freewheeling setup, if it isn't designed right (mine wasn't).
 
Regarding cranks freewheeling vs non, I'd say that like Rassy I prefer them freewheeling for the same reasons. I had non-freewheeling cranks on CrazyBike2 with the powerchair motor/jackshaft middrive setup, and while it worked well there were a bunch of times bad things happened because the motor forces the crank to turn. Destroyed chains and chainrings, derailments, torn pants, almost a crash once because of pants caught in chain; didn't tear and got motor stopped quickly, but unable to free my leg to put my foot down on the left side that I was starting to tip towards (as I was trying to make a left turn), and ended up having to sharply turn and lean right instead so I could stop myself with my free right foot instead.

Well, color me dumbfounded. I am not sure why people have problems with the Stokemonkey setup. As I wrote earlier, I've put 2600 miles on mine (plus 13 today) and haven't had any problems.

I think the reason I've never had a problem may be due to the way I use it. I noticed today that, unless I am stopped on a steep hill, I always pedal first and last, so I never use the throttle without pressure on the pedals. I don't really think about it, that is just how I do it. So there is no jerking or unexpected motion. Even with the bike geared down, there isn't enough torque to jerk the bike or pedals since they are already loaded, just a constant, strong tug. Maybe that is the difference? I don't know. I've never torn a pant leg (of course, you have to tie down your left pant leg also, and that is something I forgot to mention), messed up a chain or other hardware, or nearly crashed, but the experience of amberwolf is a cautionary tale.

I did have to replace my bottom bracket last week, but I could tell something wasn't right with it 3000 miles ago. I heard an occasional binding-like noise right after I bought the bike and took it back to the bike shop. They first said it was a pedal bearing and replaced a pedal. Took it back a week later and they said it was coming from the suspension fork and was not a problem. It got more pronounced as the miles rolled over. When I took the bearings out a week or so ago, some of the balls on the right side were nearly flat on one side and the "cage" was mangled. Hard to believe the crank would even turn. The bearings on the left were still fine. However, I would not be surprised if the bottom bracket on a Stokemonkey setup is more vulnerable. It is able to apply more torque than I can and it goes into the bottom bracket. However, I have my CA current limited to 15 amps so it is not like I'm about to wrench it off with the motor. And the cranks are turning pretty slowly relative to most electric motors so the bearings really have it pretty easy, I would think. The impressive thing to me about a bottom bracket system is that I can still do steep grades with the current limit set to 15 amps. Of course, at some point, I will be limited by the torque I can generate with 15 amps, but I've had it that way the whole time I've owned it and have hauled several bags of groceries up some pretty steep hills and haven't even started breathing hard. It really is amazing what a relatively small electric motor can do for your ability to haul things. I have thought perhaps I'll turn up the current limit to 20 amps and live on the edge a little :wink:.

Anyway, I'd always advise trying a system before you buy if you can. It is how I ended up with a Stokemonkey but may be how someone else ends up without one. As my Dad used to say, that is why they make chocolate and vanilla.
 
pdf,

"I have my CA current limited to 15 amps so it is not like I'm about to wrench it off with the motor"

You are running 36 volts, correct? With a 15 amp max, that is maybe 400 watts coming out of the motor, and another 100 watts from you. What kinds of speeds do you average with that power?

A question for the electronic wizards. Is it always more efficient to run higher voltage?

I am thinking about going with a 36 volt, 30 ah pack, instead of 48 volts, and 20 ah. The lower motor rpm means I have more motor gearing options, with off-the-shelf bike sprockets and chainrings. Being able to run a bigger cog on the motor would be quieter, and mechanically more efficient. The lower motor rpm means the planetary gears should run quieter.

For a given watts/speed I would be pulling more amps from the pack, controller, and motor...electrically inefficient. But at the same speed on 36 volts, I won't be chopping the voltage back as much with the controller. Anybody know if that wouldn't that be electrically as efficient?

My head is about to explode, Warren
 
Back
Top