Commuter Booster - <1kg Friction Drive

Hey, folding bikes don't really work either, but that doesn't stop some people from riding them. De gustibus non est disputandum.

My shopping list for electronics:
http://www.aliexpress.com/item/For-arduino-Nano-3-0-Atmel-ATmega328-Mini-USB-Board-with-USB-Cable-Free-Shipping/591964167.html
http://www.aliexpress.com/item/Free-shipping-RT0802A-Character-8x2-LCD-Display-Module-Green-5V-Black-Character-Green-Backlight-1PCS/1758279136.html
http://www.aliexpress.com/item/IIC-I2C-Interface-LCD1602-2004-LCD-Adapter-Plate-for-Arduino-Free-Shipping-Dropshipping/1362338996.html
http://www.aliexpress.com/item/Allegro-precision-hall-current-sensor-acs756sca-050b-pff-t-acs756sca/1696636520.html
http://www.aliexpress.com/item/12-x-12mm-x-5mm-Momentary-Push-Button-Tactile-Tact-Switch-Non-lock-4-Pin/2035654735.html
http://www.aliexpress.com/item/3-pcs-lot-free-shipping-pcb-instrument-box-mounting-box-for-wall-switch-project-box61x36x20-mm/2016075634.html

Cheap as! Pity the shipping is so slow during this time of year.

PS. Adrian's software is over-complicated IMO. I'll be coding something simpler than that (if I have to). Maybe I'll even skip the wheel sensor.

Hi guys, I've made an approximate copy of Adrian's bracket and swing arm, and I have a problem: Sometimes, as the drive engages into the tyre, it starts rattling against the stops very hard. It then ends up with bent pin of the swing arm (best case), or the clamp block cracks because of excessive impact force applied to the screw stops. Does anyone know a solution to prevent this rattling?

Hello adrian_sm,
I am looking to build the Commuter Booster system for my mountain bike which I use to commute to work. It's 12 miles round trip.
Searching google led me to this forum. I like what you have done, it looks great and is much cheaper than the alternatives and hopefully for the distance I travel it will work well.

I have a question regarding the "Magic Box".
What would be the less comprehensive alternative/s please?

best regards
Blan

I’m coming to the conclusion that Adrian’s motor mount is obviously very well suited for road bikes, but is not very well suited for mountain bikes – if you don’t want to wrap the motor in sandpaper. I’ll explain:

With high pressure road tyres, the motor has no chance to go over the “point of no return”, after which it would go all the way up to hit the seat stays, because the tyre is stiff and will never permit that. If we use low pressure fat tyre, however, the motor can compress it and go past the no-return point. I’m already hearing Adrian saying “your motor sits too high on the seat tube!”, BUT I have to mount it this high, otherwise it will slip and disengage at speed. To mount the motor lower, one has to use sandpaper on the motor can (which I don’t want to). But even with sandpaper, some tyres are fat enough to have low enough pressure to permit the motor compressing them far enough to go past the point of no return.

All this means that the pin limiting the swing arm travel has to withstand significant forces. My motor at 1kW will pull an estimated 6kg tangential on its can. If the distance from the pivot axle to the tyre is 8 times longer than the distance from the pivot axle to the pin, the pin has to be able to survive 48kg(!) static force, and the dynamic force can be even higher. (Here I assume the swing arm is already close to the point of no return). I’ve already bent three M4 screws I use as the pin in the swing arm, and I’m very worried about the strength of the seat tube clamp plastic. So, the swing arm travel limiting mechanism has to be completely revised for my application. I remember the photo of an early CB prototype with a different travel limiting arm, but I’ll try to find another solution.

Folken said:

Hi guys, I've made an approximate copy of Adrian's bracket and swing arm, and I have a problem: Sometimes, as the drive engages into the tyre, it starts rattling against the stops very hard. It then ends up with bent pin of the swing arm (best case), or the clamp block cracks because of excessive impact force applied to the screw stops. Does anyone know a solution to prevent this rattling?

Click to expand...

It sounds the the geometry of you "approximate copy" might not be quite right. I have a video that describes the set-up process for my drive, this might help you decide which element is not right.
I think it is this one:
[youtube]TeDcXKZfDpM[/youtube]

But the key elements (from memory) are:
- correct initial contact angle between pivot point of motor mount, contact patch on tire, and axle of rear wheel.
- adjust counter balance spring to take weight of the motor
- adjust of deadstop to give minimal clearance to tire when not engaged
- limit maximum engagement before the "over centre" or "point of no return"

BTW The commuter booster has been successfully used on both road and mountain bike, with bare motor shells, ie. no sandpaper. But doesn't like knobbly tires.

I would advise you not to use screws as the limiting pin, they are not design for lateral loads.

Good luck!

Blanthegenius said:

Hello adrian_sm,
I am looking to build the Commuter Booster system for my mountain bike which I use to commute to work. It's 12 miles round trip.
Searching google led me to this forum. I like what you have done, it looks great and is much cheaper than the alternatives and hopefully for the distance I travel it will work well.

I have a question regarding the "Magic Box".
What would be the less comprehensive alternative/s please?

best regards
Blan

Click to expand...

Good question. The cheapest and easiest alternative to my fancy "Brain Box" is actually and very cheap off-the-shelf controller.

The S06P magically works very well, at both 24V & 36V.
The key is the ramp rate of the throttle, and ability to spin up these small motors well which is not the case with many controllers. There may be alternatives now available, but this is the one I know to work.

Here is a link:
https://bmsbattery.com/ebike-kit/545-s06p-250w-torque-simulation-square-wave-controller-ebike-kit.html

If you want a push button throttle, you will need to work out a quick voltage divider circuit.

Thanks for the reply adrian_sm

Thanks for your advice Adrian.
I'm pretty sure I followed the installation procedure, but the situation is further complicated by the fact that none of my tyres are perfectly round. About 2mm radial play on the best one. And the current limiting circuit is still misbehaving sometimes. But I'm working on that.

A longer swing arm mounting the motor could help. This will give more tire engagement for the same contact angle.

I know Kepler used a longer swing arm. But I have never had the need with the tire pressuresi have tested.

I can confirm that the rattling I reported was caused by my current limiting circuit misbehaving (shame on me!)
After the servo speed/direction regulator came, it became real easy to deal with. Still, with no torque limit at low speeds, and with my low tyre pressure, I prefer the good solid stop that I constructed using two rubberized stops from an old hard disk and an aluminium plate bolted on top of the seat tube bracket. I’m riding the bike for several days now with no problems.

With one 6S 3AH battery, the total weight of the system is about 1.5kg. I’m mainly turning it on up hills.

Although my particular motor model was sucking air through itself pretty well, just for peace of mind, I used 50mm fan blades from an ancient CPU cooler for some additional cooling.

I couldn’t find a comfortable position for the button of the same type as Adrian uses. On flat bars, my hand feels uncomfortable wherever I place the button. Any ideas, folks? Right now I had to use the horn button from an electric bike (a few $ from Aliexpress) which works pretty well. I’m not sure yet if I will use the hall thumb throttle.

The next step is the Arduino! (wish me luck )

"Hello, world!" :lol:


Some stats:
14km commute time shortened from 45min to 30min. Motor run time is 15min, which is, surprisingly, enough. Peak power is 500W, this is uphill power. On the flats, the power is reduced because I used a non-rail-to-rail output opamp. I'm very pleased so far, but my analogue circuit will have to go, eventually.

Folken,

I'm trilled to see you get so far so quickly. Thanks to Adrian for doing all the pioneering and sharing your knowledge. As has been mentioned here several time, it is a curious condition that we are in that nobody seems able to offer any new tech friction drive kit.

I'd be all for exploring this with some folks such as Adrian and Folken. Between the two there sure seems sufficient expertise to prepare the core kit and some options.

Core:
Machined motor mount
Pivot and stop mechanism mount

Options:
Controller (Arduino)
Motor
ESC
Downtube mount
throttle
battery pack
battery case
wiring

It seems that such a kit would get one up and running in a couple hours. If not fully operational, this would provide a huge boost to getting it all going.

Making a universal kit is a challenge that is not required to be resolved as part of the initial phase. Maybe a fitment template could be drafted to have the prospective buyer print and check their bike geometry to see if the kit fits as is.

Initial rollout would be smaller and maximize success with the fewest obstacles and installation issues. Only sell the initial kits as full kits so everything needed is supplied and can be configured and matched for the maximum plug and play.

For the targeted phase one buyer that validates the kit fits, then the initial kit features can focus on making the components as plug and play as possible and mountable and adjustable with the least effort and fewest tools.

If the motor mount and pivot mount can be preassembled do that. Provide all Interface connectors to connect motor, controller, throttle, and battery pack.

That seems to cover 90% of it. It seems so doable which makes many of us scratch our heads why it has not been done as yet. Adrian and EV_Todd and others have stood at this vantage point but nobody has taken the leap. Is there no way to openly discuss the blockers and address the risks and concerns to make this feasible?

Folken,
You listed the controller parts list. What motor and ESC did you use?

I fully intended to make a business of the friction drive, took time off work to develop the design, concept,business model etc. But ultimately I felt it was not sustainable as a business.

Here are a few reasons why:
1) Robustness / Ease of Setup
- Takes some skill/care to install and setup. Requiring good support, or result in frustration from user
- No good in the wet, or modification when in the wet. Fine for fair weather riding, frustrating otherwise, compared to alternatives.
- Doesn't fit all bikes. Again more effort for people to select install, with possibility of trying to make it fit, and being disappointed
- Cost hurdle of making throttle/controller/batteries etc. waterproof, easy to mount, and robust.

2) Regulations
- EU laws restrict speed to 25kph max assist.
- Friction drives are best suited to road bikes, which can already acheive those speeds without assistance
- This restricts the market will to pay significantly. Fine if you don't care about the law, but not god to build a business
- The high speed ebike laws the EU was look at introducing, didn't pan out to be ease for a friction drive kit to adhere to. In fact I dn't think they got approval at all.

3) Opportunity Cost
- personally I can make more money doing my "day job"

Technically the drive all works, suits some bikes, suits some people, but it is borderline viable as a business, and I was too chicken to back it any further.
A simple similarly powered hub motor is a pretty compelling alternative, cheap, widely available, robust, mass produced, easy to install, suits most bikes,

But maybe I am just too risk averse.

Thanks for your perspective Adrian. Your goals and stadards are much higher than what I had in mind. To be fully compliant with markets large enough to manage the liability and make a sustainable profit and wage, I'd be on the same path as you but a lot further ahead.

Realistically, I guess my vision is more like a serial group buy of standard products, a couple magic components and sufficient instructions and user guide to get some fully rolling and others pretty far. The remaining bit being the adaptation to the specific bike frame. This is all constrained to buyers who "pre-qualify" via actual test fit using the provided template.

So I'm thing batches of 10 at a time. Or some similarly small enough yet large enough to get a decent break on the components. One should never plan to do anything more than expecting to maybe throw off sufficient net profit to pay for one's own ebike hobby needs and wants.

The difference is clearly bona-fide vs hobby business. Personally, it would also be a tax avoidance vehicle so the business net profit could even be negative yet still advantageous as a beneficial operation.

windtrader said:

Folken,
You listed the controller parts list. What motor and ESC did you use?

Click to expand...

Motor: http://hobbyking.com/hobbyking/store/__31154__NTM_Prop_Drive_50_60_270KV_2400W_AUS_Warehouse_.html
ESC: http://hobbyking.com/hobbyking/store/__20555__Turnigy_dlux_100A_SBEC_Brushless_Speed_Controller_w_Data_Logging_AUS_Warehouse_.html

The motor is 50mm, has no skirt bearing, but weighs below half a kilo. Adrian blew his 50mm motor once, however it was a different model, and most importantly, he had no power limiting at that stage. Let's see how mine goes. I'm not risking a fortune here anyway.

The ESC has a very useful datalogging feature. So with a matching programming box, I can see max amps, max temp, motor run time, etc. And it's still cheap.

By the way, guys, be careful when you're testing motors. When I just started experimenting with mine, it tried to chop my fingertip off with its sharp edges. There was lots of blood, but luckily not that much of an injury. And now, with fan blades mounted on the motor, I have to be even more careful.

windtrader said:

Realistically, I guess my vision is more like a serial group buy of standard products, a couple magic components and sufficient instructions and user guide to get some fully rolling and others pretty far. The remaining bit being the adaptation to the specific bike frame. This is all constrained to buyers who "pre-qualify" via actual test fit using the provided template.

Click to expand...

Been there, done it. Required a lot of support time for 50% of the people, the rest were fine on their own.
I couldn't justify that level of support unless it was paying its way, which it wasn't. So either I up'ed the cost to cover the support, or provided a more complete kit. I prefered the later, and so did most people I asked.

Now I am thinking of more of a open source approach. I share everything t kick things off. Full design of all parts, bill of materials, guides etc.
Someone coordinates group manufacture/buy of custom bits, community provides the support.
It will only self sustain if enough are interested, which is kind of the point. If people aren't interested it ain't worth it.

Hey guys, I was wondering about the best ESC settings for our application. In particular, frequency and timing. My ESC has two freq settings (8kHz/16kHz) and three timing settings (Auto, Low, High). It’s more or less clear about the frequency – 8kHz will give me better ESC efficiency, so I use this setting. But what about timing? I tried googling it, but the opinions are contradictory. And the answers come from RC crowd, so the application is quite different. I tried experimenting, but I couldn’t really feel the difference. I didn’t monitor motor temperature though. Your thoughts?

No more analogue stuff, only Arduino! I’ve made an MCU box with Arduino Nano and a Hall sensor, and a similarly looking display box with an 8x2 LCD and parallel-to-i2c adapter board:


Fits nicely between my aero bars using cable ties:

The button throttle is connected to the display box to minimize the number of cables.
I get Watts, Volts, Amp-hours and Amps. (“970” in the photo is the ESC pulse width I displayed for debug purposes before).
The current limit is hard-coded at the moment. And I’m limiting current, not power so far.
Even though I have no wheel sensor, I can still program the maximum assist speed by adjusting the top pulse width limit for the ESC. The speed limit will vary with battery voltage a bit, but I can’t feel it. This can later be compensated in software if desired, anyway.

I’ve published my code here:
http://endless-sphere.com/forums/viewtopic.php?f=2&t=61004&p=999375#p999375
Enjoy!

Anyone have a Commuter Booster or equivalent mechanical setup for sale? I can handle the electronics but would prefer not to re-create the mechanicals. Planning on using on a road bike with good tire pressure and plenty of space in the frame (XL size Specialized). Just want uphill assist but also want to be able to move it to my wife's city bike for rides together.
Thanks, Steve

It has been almost one month since I posted on the forum about having a crack at this build.

This is a short video of my test earlier today.
https://instagram.com/p/yju6jig0RL/

This seems to be the right place to post about ongoing maintenance for the Commuter Booster and Brain Box.

Last week my wife's commuter had a failure while riding in the rain. (What's wrong with it? " It doesn't work anymore." )

I was expecting to find blown FETs on the ESC, but the ESC powered up OK when plugged directly in to a battery.

So, I had to open the Brain Box. The Arduino also seemed to have lights when a USB was plugged in to it. Next was the little circuit board inline on the power leads - the current shunt/voltage monitor circuit. I had to remove the heatshrink from around it.

Turns out the solder had melted and open circuited both surface-mount shunt resistors (one on each side of the PCB).



It's a bit of a mystery to me why this occurred. The shunts are rated for 90A, and I've got the brainbox set to run at around 50A. No sign of water ingress. Anyway, after resoldering, the system sprang to life again. Happy Ending!

That's why I like my Hall sensor. It never even gets warm! I guess 7 bucks is not a terrible price to pay for peace of mind.

A little bit off-topic, but...
I have a heart rate monitor / cycle computer (VDO HC12.6). When I turn the motor on, the heart rate reading often freezes, or sometimes it even reads numbers like 240 BPM. The chest sensor is analogue (the radio signal is not coded). Did you have such issues with RC brushless motors and ESCs interfering with heart rate monitors? Will it help to switch to Bluetooth 4.0 or ANT+ based heart rate sensor?