How to modify a BPM hub motor to drive in the reverse directionUsing 2 BPM front motors to drive a cargo trike
A friend recently asked me to modify a BPM motor to run backwards. The motor was one of a pair to be installed on the front of a Zeitbikes cargo trike (a Chinese box trike - I don't know if they are sold under that brand any more).
These box trikes are very heavy, not stable at speed or on uneven surfaces, and this one was to live at Highgate Hill (Brisbane) and would need to do a lot of climbing with loads.
An attempt had been made to drive the trike with a single rear motor, but this had proven to have inadequate torque at the low speeds the heavy trike was able to achieve safely.
Here's the trike with a single rear hub motor: too fast, not enough torque |
The frame was not suitable for installation of a mid-drive motor, having extra tubes around the bottom bracket.
The new plan was to use 2 less powerful, higher torque, lower speed motors in the front wheels, one on each side of the cargo box.
The Zeit trike uses hydraulic disc brakes on the front wheels (which work very well), with calipers on the box side of each wheel, so making the left side wheel turn opposite to normal was necessary so that the brake disc could be on the right hand side.
Reversing the mechanics
It is relatively easy to make the internal motor of a geared hub to run backwards, just by the way it’s connected to the controller - as many of us find by accident when first wiring them up. However when the internal motor runs backward, it doesn't turn the hub backward, you just hear a whizzing inside a stationary hub. This is because geared motors use a freewheel, the main purpose of which is to allow the bike to coast along without resistance if the motor is not powered. In order to change the rotation of a geared motor, the freewheel needs to be reversed mechanically, as well as reversing the motor electrically.
If you've opened a geared hub motor, you've seen the freewheel. It's a steel disc that mounts onto the main motor shaft thru its centre (with a slot and key to lock it in place).
Here's the freewheel straight out of the motor |
Here's a BPM freewheel with one gear removed |
The freewheel disc carries the 3 nylon planet gears on 3 stub shafts attached near its perimeter. The "freewheel" part of the freewheel is 3 (not visible) sprung rollers that allow the circular plate of the unit to turn one way around the centre boss but not the other. To reverse the motor, we need to rearrange the freewheel so that the rollers allow movement in the opposite direction from normal.
Reversing the freewheel involves:
- removing the 3 stub shafts
- removing 3 small rivets that hold the freewheel plate together
- dismantling the freewheel body
- reassembling the body with the centre boss and the 3 stub shafts reversed to the other side of the plate.
Step by step
First the 3 nylon gears were removed from the freewheel disc. This is easily done by removing the circlips from the stub shafts, I used circlip pliers (whose tips open when you squeeze the handles) and some rag around the job to catch any flying circlips. The nylon gears, along with their sealed bearings, slide off pretty easily.
Freewheel without gears, stub axle side |
Freewheel other side, with centre boss sticking up in the centre |
Next job was to punch out the stub shafts, the serious part of the job. It's important to be very careful: you can't afford to have any impacts on the outside end of the stub shafts, or you risk damaging the circlip groove and making reassembly difficult.
I have ground off the short riveted end of the stub shafts in the past, to open a freewheel and repair the sprung rollers (link). But this loses a lot of metal from the rivet head, making reassembly difficult. Having recently been riveting a few chainsaw chains, it occurred to me that the stub shafts could be un-riveted the same way as a chainsaw chain is unriveted: by punching. When you punch apart a chainsaw chain rivet, you only lose a small ring of metal from the perimeter, leaving enough metal in the rivet usually to be able to peen and re-use the rivet to make a chain loop.
Here is the freewheel on the anvil, having a stub shaft punched out.
Suitable punch on top of riveted axle end, the stub axle is inside the old nut below |
The freewheel plate is supported by large nuts, bigger than the flange in the middle of the stub shaft. I used a punch a little smaller than the rivet head, driven by a 1.2kg blacksmithing hammer.
Here's the first stub axle punched out |
Once the stub shafts have been driven out, 3 small rivets hold the laminations of the freewheel plate together. These are punched out, like the stub shafts but with a smaller punch.
Once all the rivets are punched out, the 3 plates of the freewheel come apart - I carefully kept things together as much as possible. The centre boss (which fits onto the main motor shaft) is then lifted out and turned over to be re-inserted.
Installation requires the freewheel rollers to be pushed back into their spaces to let the centre boss in. You'll work it out.
Axles punched out, small rivets punched out, now the first plate can come off and reveal the freewheel rollers in their tapered slots |
Here the centre boss is lifted out. The small steel rings at rear are the old rivet head rims, left on the punch after punching the axles out |
Once the centre boss is re-installed, the freewheel plates need to be riveted back together, with both the small rivets and the stub shafts. The stub shafts are inserted into the opposite side of the freewheel plate from where they came. With the centre boss reversed and the stub shafts reversed, the whole freewheel unit looks just the same as it did before being dismantled - the only difference being the plate freewheels in the opposite direction around the centre boss.
Be careful! Don't reassemble the whole thing the same as it started. I liked having a spare freewheel unit on the bench to check I was doing things right.
Here's the freewheel reassembled, with heads peened onto the axle ends and small rivets with a ball peen hammer |
Reversing the electrics
The internal motor is very easily reversed electrically if a sensorless controller is being used: simply swap any 2 of the phase wires and the internal motor will reverse. With a sensored controller (which I don't usually use), it takes more trouble: swap 2 phase wires, then try different arrangements of the 3 coloured sensor wires (leave the red and black alone). This has taken me some time in the past.
It works!
The motor I reversed this way now works perfectly, in reverse. It and a normal non-reversed motor have been installed as the 2 front wheels of the box-trike. The motors are Bafang/8FUN BPM 500w code 13 front motors, laced into 20” rims.Here's the 20" wheel where a 24" was |
The box trike on a test ride, climbing a hill with the new motors |
If you want to buy a spare freewheel unit before you start knocking the original apart (like I did), check that you get the right sort: there are 2 different freewheel types in BPM motors. Have a look at my BPM page for more info.
The customer has been very happy with the trike, using it daily in hilly Highgate Hill, Brisbane. It has loads of torque, and powers up to about 20km/h. This is about as fast as these box bikes should go: their non-independent front steering - turning the whole box to steer - is very vulnerable to any bump on one side pushing that wheel back and forcing a turn. To me, the relatively high rider position also feels unstable when the trike tilts to one side due to the road camber or crossing a slope.
The promise of 2 drive wheels
I think the 2-wheel drive system we installed on this trike has a lot of promise for cargo cycles. I know that a cantilevered axle (where the wheel is attached only on one side, like a car wheel, or a wheelchair wheel) appears a more elegant way to put wheels on each side of a vehicle. But using normal bike wheels bolted into dropouts on each side has many practical advantages. Most importantly: there are heaps of wheels like this available, in all sorts of formats: electric hubs of various types and speeds, disc braked, etc..
Imagine how much help some electric assistance would be to these hard-working charcoal makers:
Phnom-Penh charcoal delivery trikes (sorry I have no credit for this photo) |
I quite fancy some sort of delta, recumbent cargo trike format, like this:
(sorry, no photo credit again) |
Hello Bruce,
ReplyDeleteFirst of all I would like to thank you for sharing your experience and the effort you do by publishing on your blog.
I have a question about this creation with 2 motors; In theory, if the 2 motors have a slight difference in rotation speed or power, this would make the bike go around in circles or at least make it pull to one side... Did you do anything in particular to prevent this from happening or is this unnoticeable? Also, when going around corners, the inside wheel turns slower than the outside wheel, cars have differentials because otherwise they are hard to handle because of this. I'm really curious to know if the handling of the bike is still manageable, it must be harder to steer? Hub motors gain in torque at lower rotation speed so the wheel on the inside of the corner pulls harder, doesn't the bike force itself to go straight on?
Thank you in advance for your reply and kind regards,
Knutsel
Hello Knutsel,
ReplyDeleteYou have the physics correct, but in practice the forces you describe have no noticeable effect. The trike steers very easily, turns and goes straight easily.
However these trikes are very strange beasts to ride for various reasons, with our without motors. Steering by turning the front box feels very strange to me, and the trike is easily knocked out of direction by a bump or resistance on one front wheel. I also found it hard to accept the tilt of the trike, due to the camber of the road - same feeling on upright trikes with the 2 wheels at the back. Going faster than about 20km/h, especially down hill, felt dangerous to me as I was worried one wheel might hit a hole or bump, slam the box to one side, and roll the trike over. Independent front wheels, like on a Greenspeed, wouldn't have this problem.
The good news is that the owner is very happy with it and uses it every day on steep hills, and many owners of Christiania trikes (with the same format) seem to love them.
What the independent motors would offer, unlike a vehicle with a differential, is that if one wheel was to lose traction and spin, the other wheel would still have power. Unlike a car where one wheel losing traction will stop the whole car (which we see all the time on our steep dirt road).
Best wishes from Bruce.
Hello Bruce,
ReplyDeleteThank you for your quick and clear reply.
I sell and rent bikes of this type and even transform non electric ones to electric. I have always used back wheel motors because I was afraid that it would be impossible to handle with 2 motorized wheels in the front, I never really tried it though.
Your experience makes me want to try it out, I put it on my "to do during the low season" list.
About the weird handling, it helps to put the saddle lower, that makes it a lot easier. Also they get much more stable with heavy charge.
It is important to use your body weight, hang over to the inside of the corner.
Contrarily to a normal bike, it is impossible to touch the ground with the pedal if you got the foot at the inside of the corner at the lowest point (dunno if that is still the case on yours as you put smaller wheels on the front) so you can push down with your weight (stand on one foot), this lowers the centre of gravity to the inside of the corner and helps a lot.
It also helps if the pivot under the box is a bit "tight", that makes it a lot less sensible for bumps, one can even use 2 springs or gas loaded (can't find the English word) "rods that hold up the back hatch of a car" to make it a lot more stable, those are quite easy to install.
Kind regards,
Knutsel
Thanks Knutsel, this is very useful advice for me and others!
DeleteMy questions:
If you put the saddle lower, doesn't this make your knees bent, and cycling is more stressful on your legs and knees, or is there another trick?
We call those car back window rods "gas struts" here. That sounds like a good idea. I would be very interested in seeing a photo of how you do this.
Where do you sell and rent these bikes? Is it a flat area?
How fast do you cruise in a box trike? How fast can you still feel safe?
What sort of motors do you use on the back? We did the front wheel thing because we couldn't find a suitable back motor.
Yes indeed, it makes the knees bent especially if the rider is tall like me . Sliding the saddle backwards compensates a bit but there is not much that can be done. Bending your body forward in corners will also lower the centre of gravity. One has to find a compromise between confort and easier handling.
ReplyDeleteThere are no high hills here, so there is not much power needed generally. I have used XD 2 speed motors, 800w Q128 (set to max ~500w to prevent overheating) and even a Q100 (low speed winding set to max ~250w) and the clients are satisfied.
The gas strut idea is not mine, I sell Babboe (Dutch manufacturer) and they have it installed on their bikes. I will send you a picture of their solution.
I uploaded some pics here: https://downloads.zelectrification.com/Accès/Libre/Images/Bruce/
ReplyDeleteThanks Knutsel, these struts could make the trike much more steady. Thanks for your knowledge!
DeleteYes, the advantage is very noticeable. The box moves easily when it's turned slowly but in case of a shock, a fast movement, one can feel and hear them doing their job. I do ovrer 50km/h downhill with my most precious charge in it (the kids) without any fear, as long as there are no tight corners of course. I think it might be quite hard to implement though because the position of the fixations (the length of the levers) should correspond to the "braking power" of the struts.
ReplyDeleteKeep up the good job on your blog, and thank you too.
Sharing knowledge makes the world a better place for everyone, especially for those willing to learn ;)
After applying what you describe above...will i be able to rotate the motor in both directions, considering i am able to electronically control the wiring of the motor terminals
ReplyDeleteHello Bijen, Knutsel is right: the process I describe produces a motor that is totally the same as the original, except that the motor drives the wheel in the opposite direction from normal. If you wanted to drive this motor in both directions, you would need to lock the freewheel. Then you could only go down hills with the motor engaged (this happened to me once when a freewheel broke, I wrote a post: https://bruceteakle.blogspot.com/p/bafang-freewheel-failure-and-perhaps.html )
DeleteNo, the motor only runs in one way, that is what the clutch is there for.
ReplyDeleteOnly direct drive motors run both ways.
Okay, so help me out on this, what if my application is okay with no freewheeling and a geared slow down while free running is expected.
DeleteCan i modify this assembly somehow to make it run in both directions or else is it possible to replace the sun and the planets with another larger single gear which connects directly to the outer ring of the planetary system?
What torque and speed changes should i expect in the later changes?
Locking the centre boss of the freewheel to the outer assembly of the freewheel would enable the hub motor to run either way (if you get the electrics right - pretty easy with a sensorless controller: just reverse 2 wires). The rough, quick and irreversible way to do this would be to weld some tacks between the boss and the side plates - I'd make sure I had a spare freewheel first. More sophisticated methods would require disassembling the freewheel (the first stages described above) and using epoxy or some other method to lock the centre boss to the outer assembly.
DeleteIf you were to lock the motor to the hub case (perhaps by stopping the planet gears from rotating) the hub would turn at motor speed: about 5 times normal geared speed.
What are you thinking of making?
Watch out though, disabling the clutch will transform your motor in a generator in the cases where it would be freewheeling, thus sending back current into the controller. This risks the controller to malfunction.
ReplyDeleteI didn't think of that! During the brief period when I had a motor with a seized clutch, the controller seemed to be ok, but that was only for a one hour ride, only partly downhill. I believe some controllers can do regenerative braking - so they would be happy with reverse current - but I don't know where to get one. I expect ebikes.ca would help.
DeleteA lot of good quality controllers can do something with it. But I have had several cheap supermarket ebikes to repair in my shop with stuck clutches that turned themselves off every time the rider stopped pedaling downhill (pedal assist systems). The first time I really had a hard time understanding what was going on :p
ReplyDeleteOkay we are discussing on whether to weld the studs or not... But meanwhile understanding the freewheeling assembly I have come up with a thought ,
ReplyDeletewhat if we remove the studs...which will allow the centre boss to rotate in either direction and let the carrier assembly of the planets be stationary as their link is the studs which don't exist now and we might apply some grease to let the centre boss rotate like a bearing and allowing the carrier to be stationary and getting rotation in both directions as sun is being moved in both directions by the motor's shaft and the carrier is stationary allowing the planets to rotate at their axis and making the ring rotate
I'm not clear about your idea, but if the centre boss spins on the shaft - if the key is removed - then no torque is able to be applied to the ring gear in the hub case. Then the motor and gears will whizz around inside a stationary hub case without driving the wheel. The 3 planets push the ring gear around the shaft - drive the wheel - because they can push against the shaft with the boss and its key.
DeleteHello guys, we opened up the motor today and we found that our gears are a little bit different
DeleteWe don't have a centre Boss and studs, we have a centre boss (with two rachet stopping pins or rachets i don't exactly know the names) and the carrier's connection to the boss is of rachet teeth. The temporary solution that we have got is that we have removed the keys of the centre boss and reassembled the motor, we are gonna test it tomorrow just wanted to know your thoughts about this
Hi Bijen, I think the motor will have no torque either direction with no key. You'll find out!
Delete