Front hub motor on SWB recumbent, advice sought

[mrbill]

A friend of mine is planning to convert one of his short wheelbase recumbents into an ebike for commuting.

He wants to be able to keep up with traffic on a city boulevard, so he figures that a top speed of 30-35 mph would be good. When he's feeling energetic he may occasionally take a hillier commute, so he wants a system that won't get too bogged down on hills up to about 12-15% grade. If he rides in the hills he will be pedaling significantly. He weighs about 140 lbs (64kg), so I'm aiming for a total system weight of around 220 lbs (100kg). The commute distance is 22 flat miles round-trip or 37 hilly miles round-trip. If this project goes well, he may wish to ride longer tours on this bike.

The proposed bike is an HP Velotechnik SpeedMachine: http://www.hpvelotechnik.com/produkte/spm/index_e.html
with the added rack (58lb capacity) so that batteries can be carried.

He wants to keep the conversion simple, to retain the original 27speed XT drivetrain. To that end I have recommended the following:

Infineon controller that supports regenerative braking, possibly adding cruise control and ebrakes
Nine Continents 2805, front hub, with disk rotor mount, possible addition of torque arm(s)
Ping 48v/20Ah (or 2x 10Ah) battery

Questions:

1) Is this a good design? If not, what are the problems and possible solutions?
2) Are there other direct-drive front hub motors I should consider?

Thanks.

 

[amberwolf]

i don't see any problems with the bike itself, assuming that front fork is steel. might want to make torque arms for it, though. if it's aluminum, you definitely would.


for motor wind choice, id' go with one that is meant for slow speed, and use a high voltage on it. seems to work better for hills and such based on reports from dogman/etc. just throttle down for the hills if necessary to keep total power levels down to safety zone on the motor, and more throttle for faster speed on flat areas.


a fast winding motor, if i understand correctly, will have higher currents at slower speeds more than a slow wind motor will, because it's back-emf will not raise enough to cancel them out if it's bogged down on hills and such.


for battery, you'll want to calculate the wh necessary to complete the trips he needs, and add a significant percentage to that, to make sure there will be enough.

you'll also want to calculate the amps the system will draw on hills and at speed, etc., to ensure the battery can sustain that kind of load without sag or damage.


the simulator at ebikes.ca may help with some of this, and there are calculators around the web for figuring out watts needed for various weights, aero, terrain, and speeds.

 

[mrbill]

the simulator at ebikes.ca may help with some of this, and there are calculators around the web for figuring out watts needed for various weights, aero, terrain, and speeds.

Thanks for your reply.

I found the 9C 2805 after playing around with the simulator at ebikes.ca. I could get a similar profile with the 2806 at 65 volts (60-volt Ping), but the battery cost is higher per wh.
The front wheel on this bike is a 20" (ISO 406, I think).

I'd like to be sure he won't be suffering front-wheel spinout too often. Also, I have no idea how the handling will be affected by putting a hub motor in the front wheel of a bike.

 

[dogman dan]

The inevitable problem here is wanting to go fast, yet climb a wall. The general solution is to ovevolt or overwatt enough to reach the speed you like. Then you can make so much heat climbing the hill you melt the motor down. So after a few motors you end up with something really beefy. Something that was intended for a 5000w scooter, not a 500w bike. Well, there went lightweight and pedaling, and most places, street legal.

It's a dilemma. I'm not sure which winding would match this in 20" rim. But in 26" a 2810 motor can be run on 72v 30 amps, climb some very steep hills without overheating, and still reach around 30 mph. 30 mph is damn fast for a long distance btw, very hard on everything, tires, brakes, spokes, rims, etc. You don't have Le Tour service with new wheels daily.

I'd say run some simulations of the avaliable 2807 winding in 20" rim. It should climb for you, and you should get some speed out of it at 72v. Try to run it at less than 2000 w most of the time. Idealy, you'd cruise at 1500w at 30 mph or so. Spinout will be tolerable up to about 2000w, at 3000w you'll start running through front tires fairly quick. You'll only be spinning on startup.

Somethings got to give, you just can't take fast winding motors and go climb walls with them, unless you put so much watts into them they get super hot. That means real short rides, not long tours. So you have to adjust something, either leave the hills alone and climb only 0-8% grades, or ajdust the motor and live with a bit less speed. 30 should be fast enough.

You mentioned wanting to keep the battery cost down. That's easy, toss the idea of grades above 10%, toss the idea of speeds above 25 mph. Then you just need a 48v 15 ah ping, and the 2806 winding. Basicly, to do a lot is simply going to cost more than typical ebikes do.
Lipo is the cheapest way to get more power, similarly priced or cheaper than lifepo4 once you add in good chargers and other equipment, but less long lasting eats away at the savings. More complicated, slightly hazardous to charge, but the tradeoff is smaller lighter battery packs that really put out the power to climb those walls.

Seriously, it doesn't sound like that much a performance increase from 25 mph and 10% grades, but your wallet will disagree. It just seems to make costs go crazy to build a performance ebike.

 

[mrbill]

The inevitable problem here is wanting to go fast, yet climb a wall. The general solution is to ovevolt or overwatt enough to reach the speed you like. Then you can make so much heat climbing the hill you melt the motor down. So after a few motors you end up with something really beefy. Something that was intended for a 5000w scooter, not a 500w bike. Well, there went lightweight and pedaling, and most places, street legal.

It's a dilemma. I'm not sure which winding would match this in 20" rim. But in 26" a 2810 motor can be run on 72v 30 amps, climb some very steep hills without overheating, and still reach around 30 mph. 30 mph is damn fast for a long distance btw, very hard on everything, tires, brakes, spokes, rims, etc. You don't have Le Tour service with new wheels daily.

I'd say run some simulations of the avaliable 2807 winding in 20" rim. It should climb for you, and you should get some speed out of it at 72v. Try to run it at less than 2000 w most of the time. Idealy, you'd cruise at 1500w at 30 mph or so. Spinout will be tolerable up to about 2000w, at 3000w you'll start running through front tires fairly quick. You'll only be spinning on startup.

Somethings got to give, you just can't take fast winding motors and go climb walls with them, unless you put so much watts into them they get super hot. That means real short rides, not long tours. So you have to adjust something, either leave the hills alone and climb only 0-8% grades, or ajdust the motor and live with a bit less speed. 30 should be fast enough.

You mentioned wanting to keep the battery cost down. That's easy, toss the idea of grades above 10%, toss the idea of speeds above 25 mph. Then you just need a 48v 15 ah ping. Basicly, to do a lot is simply going to cost more than typical ebikes do.
Lipo is the cheapest way to get more power, similarly priced or cheaper than lifepo4 once you add in good chargers and other equipment, but less long lasting eats away at the savings. More complicated, slightly hazardous to charge, but the tradeoff is smaller lighter battery packs that really put out the power to climb those walls.

Seriously, it doesn't sound like that much a performance increase from 25 mph and 10% grades, but your wallet will disagree. It just seems to make costs go crazy to build a performance ebike.

I hear you.

The rider of this bike is not heavy and is a strong cyclist, so he's used to pedaling hard. Still, I'd like to design the system to function adequately with no pedaling.

According to the simulator at ebikes.ca, the new Crystalyte HS3548 laced into a 20" wheel has more than enough speed at 48 volts when mounted on a recumbent (cdA around 0.35). It also can climb a 10% grade full-throttle and never overheat (according to the simulator). It actually overheats in 44 minutes if the grade is 12%. This is good for about 17mph up a 10% grade at 30Amps, using something like the Ping 48v 15Ah battery. According to the simulations, the system should do O.K., although it will be at its limit when the rider is not pedaling. Most of the time, the load will be much easier. Flat ground speed tops out at 30mph, which is enough.

For now the emphasis is on the flat commute and keeping costs reasonable and the installation simple. I just want to make sure that the system will not bog down horribly on hills. His alternative is ponying up $2800 for the Bionx kit on this bike, and I think we can do better by at least a factor of 2 and enjoy better performance with similar system complexity.

 

[dogman dan]

Well, It could well be that in 20" wheel the motor winding is a better choice than I think. Supposedly the H model clytes heat soak a bit slower too since there is a bit more mass to heat. 20 inch wheel sure helps everything on the climbs, but all my experience is on the 26" wheel I prefer for both road and dirt rides.

I just couldn't help thinking Woah, when you put up 15%. Beleive me, every 1% past 10% makes serious difference in how quick your motor heats.

The rider makes a huge difference too. The key thing when climbing is to keep the motor within it's comfortable RPM. Let rpm drop too much on a climb and the motor starts taking 2000w, and turning 1500 of it into heat. You can smoke a motor in a short time if you are riding up a wall too slow.

The slower winding motors have a slower minimum speed, so they naturaly have an ability to climb with less % of the wattage making heat. It ends up more efficient on long rides, and also gives a snappier take off from the stop sign. The trade off of course is a slower top speed, so you may end up with a more expensive 72v battery to get top speed back up. But in the end, you still get some tops speed, and more important, even when at 72v, the motor still has that same ability to climb the really steep hills slower, and cooler, and more efficiently.

If you can chart a 2810 9 continent 26" wheel at 48v 20 amps, and then find a 20" wheel motor that is fairly similar at 48v 20 amps, then you will be on the right track for a real climbing bike. It just depends on what you really want, a climber or a zoomer.

There is a way for a good pedaler to slog a fast winding motor up a steep hill. It's rarely talked about here, because we learned to build bikes better for hills. Basicly, you use a wattmeter to limit watts to about 800w max. So when you see it climb too high, you roll back on the throttle. Then you pedal your guts out to keep speed at at least 10 mph. With a good pedaler, this method may work fine for you. It's a pedal assist approach instead of a I still pedal anyway approach.

 

[pdf]

What grade you can climb is a strong function, of course, of the length of the grade (Assuming a fixed weight, speed, etc.). What dogman says is good advice. The one thing I think I might say differently is that LiPO is a great option, in my opinion. If you bulk charge it regularly and take a look at the balance occasionally, it will perform well. I am not sure running a Ping-like battery at 2C is going to work, but that is my experience. Others might say differently.

I rode a mid-drive (and still do) with a 36v 15 amp-hr duct-tape LiFePO4 daily for over two years, skeptical that a hub motor would climb hills like this one. Eventually I got a 2810 with a Lyen controller at 18s LiPO for kicks. I have been continually surprised at the capability of that setup. It will only do about 26 mph or so wide open in a 26 in. wheel, which is less than dogman experiences. Not sure why. Have checked wheel rollout calibration and it is right on. Need to GPS verify, I guess. This thing will eat a 10% grade for a mile or so at 20mph or so. The only place the mid-drive I have exceeds the hubbie is if I am carrying an extra 100 lbs up a big hill. Then the mid-drive can be geared like a tractor and chug right up. For serious hauling in hills, a mid-drive will be hard to beat.

Unless you have miles of grade over 10%, I think a motor of 9c-like quality is going to do fine. My experience is that going to higher volts on a motor of moderate to high turn winds makes a great system. If your friend is only 140 lbs, a setup like this is going to climb like a scalded cat.

If you have long hills of 15% grade or more, you need to plan more carefully, but with 10% grades of a mile or so and occasional short 15-20%, a setup of the type dogman suggests is simple and efficient. I would go with LiPO to be sure I can supply the current I need though.

Advice is worth what you paid, etc. etc.

 

[dnmun]

bill, you need to see if the dropouts on the hp velotek will be wide enuff to take a hub motor. i suspect your mid drive system or brent's would be more effective and more powerful on hills without the resistance of a hub if he is pedaling without electric assist.

 

[StudEbiker]

There's no reason he can't put a motor in the rear and still keep his original gruppo. He'd lose a few gears in the rear, but he'll never miss them. Put a MAC in the back and let it rip.

 

[mrbill]

bill, you need to see if the dropouts on the hp velotek will be wide enuff to take a hub motor. i suspect your mid drive system or brent's would be more effective and more powerful on hills without the resistance of a hub if he is pedaling without electric assist.

How wide should the front fork dropouts be if they're aluminum? Should we use a torque arm on both sides of the fork?

Also, has anyone found a hub motor that is machined to take a freehub? Failing that, is there anyone who supplies replacement motor covers with a freehub mount instead of a threaded spindle for a freewheel? This as been one of the things keeping me from exploring hub motors in the past.

Brent's system is a nice turn-key solution, but it's very expensive. Also, not sure if the MtnDrive Type 2 would fit over the front wheel on a HP Velotechnik SpeedMachine or Optima Oryx 20.

 

[amberwolf]

the newest motor from i think crystalyte is made with a freehub spline and mechanics. there is a thread about it from last month, i think, maybe late november. some controversy since it's design is such that you must unlace the motor from rim to open it up, and it requires custom-shaped spokes more like really thin motorcycle spokes, so if any break you can't fix your wheel.


there was one more motor but i forget what it was.

 

[dogman dan]

FWIW, I ran the 2810 on 20s, 8 v more than the 18s battery. So I was seeing close to 30mph.

Wanting to climb even steeper stuff than 15%, I then went to a 2812 motor, and get 25 mph max now, from 20s lipo. (72v)