Notes on my first build, rear belt drive

M

mike_o

10 mW
Joined
Mar 9, 2017
Messages
20
Just wanted to share a few details on my first build:

- Scott Aspect MTB
- 3D printed 440mm diameter belt wheel, fixed to 16 spokes. No teeths, just friction tape
- NTM 4248mm 650kv outrunner w 20T GT2 timing pulley
- 10mm wide 1524mm GT2 belt
- 30:1 gear ratio
- 4S 16Ah Multistar LiPo pack
- 50A VESC
- Tensioner by rubber band loaded dual ball bearings
- Type of riding: Power supported daily commute.

bikelowres.jpg

So far I've done 6-7 short trips, and today I took the bike for work - 10km each way. It worked well, my average speed was are 25km hour +- 2km/h depending on wind. Recharged during work with 6004mAh, so consumption around 600mAh @ 4S per KM. Perhaps less for the return trip, as this had tailwind. This would translate into 200 watts of assist power. The motor temp is around 20*C above ambient, so electric efficiency would be at least 80%, I guess.



The bike does more or less what I hoped, perhaps it's more power hungry than I expected.
Is 200W to be expected for light/medium assist? I was perhaps expecting 100watts, but there are losses, ofc.

thx / vbr
 
That's a smart solution with the grip tape on the very large diameter belt wheel. :thumb:
Curious to know how the belt will hold up and if the wear is acceptable. Please update after a few hundred miles/kms.
 
Nice and simple execution.

For power, 200W seems high. I figured your ride time was .39 hours, and if you used .6AH, that's 1.6A for average battery current. Given a 14.4V battery, that's more like 25W?

Anyway, on my hubmotor bikes, at a little slower speed (13 mph) my display says I'm using around 100W.
 
Wow, very interesting implementation. I like it! How did you 3D print a 440mm wheel? This reminds me of a clothes dryer.
 
doc, falcon, slowco,

thanks :D

Efficiency
The recharge was 6Ah, or 6004mAh, not 600. The 600mAh is per km, resulting in the 200W. I just finished recharging after the home run, this was 4400mAh for the 26min drive, or 152W average power, so the wind has indeed an effect.

3D belt wheel
I persuaded one of my employees to draw and print the drive wheel. Initially we decided to go for a tooth-less approach as there's more than 120cm of belt contact. The surface is convex to help the belt line up, but the material turned out to be very slippery, so we decided to go for the friction tape. And, yes, I'm keeping an eye on wear. I can share the file, if anybody wants to use it. I am getting another, larger wheel shortly, but that had to be printed in several pieces (3D printer size limitation), that will need to be glued together.

Free wheeling
Originally, I had envisaged the belt wheel riding on 8 or 16 ball bearings (fixed to the spokes) with a ratchet mechanism for free wheeling when pedalling hard or coasting. I'm not sure if I want to implement that because it will prevent me from picking up brake energy. But the belt tensioner is a problem, because it only works in the drive direction.

Two way tensioner
I consider replacing the single action tensioner with a double pulley tensioner that will allow force to travel in both directions. The two pulleys will be pushed/drawn togehter by the rubber band, but they will swing freely together to allow for shifting power directions

vbr

Mike
 
docw009 said:
...Anyway, on my hubmotor bikes, at a little slower speed (13 mph) my display says I'm using around 100W.
Click to expand...

Thanks - good information. I used the 6Ah consumption, 27min drive and 15V to calculate.

The return drive was a more moderate 4400mAh, and if I use 14.4V volts instead of 15V, I'm getting below 150W. More like 145W. Now subtract the belt loss (5-10%), a little higher speed, and my knobby tyres, and I think were almost comparable.
 
Tommm said:
How is that big white rim mounted to the wheel?
Click to expand...

1mm wire through dual holes around the spokes. In 8 spoke positions I cut small recesses as half of the spokes are outside/higher than the other 8. This way, the rim/ring rests nicely on 16 spokes. I don't expect to transfer monster power this way, but it seems to cope effortlessly with the 2-300watts, I have on tap now.
 
My current tensioner does not allow braking. It will only make the belt slip on the motor pulley.

oneway_motor_tensioner.jpg

To overcome this, I'm designing a dual way tensioner:View attachment 1

and brake mode:
two-way_motor_tensioner_brake.jpg
 
SlowCo said:
...Curious to know how the belt will hold up and if the wear is acceptable. Please update after a few hundred miles/kms.
Click to expand...

I've done 60km so far. The belt looks nice and there is no slipping whatsoever, but the 20T alu motor pulley shows sign of wear. It has rotated 900.000 times now, so thats perhaps to expect. I have ordered a 25t brass pulley to better cope, but going 5mm module belt is probably a better solution.
 
I wonder if a child's bicycle rim could be used in place of the 3D printed pulley? Possibly with some cogs bolted through the spoke holes to engage the belt.
 
Grantmac said:
I wonder if a child's bicycle rim could be used in place of the 3D printed pulley? Possibly with some cogs bolted through the spoke holes to engage the belt.
Click to expand...

I believe that is normally how its done. I never paid much attention to it, I would guess its attached through the sidewall of the rim so there is nothing rubbing on the belt but I am sure you could make a component that can use both sidewalls and attaches to the spokes.

If its a wide enough smaller rim, flat plate
https://i.pinimg.com/736x/f5/7b/b6/f57bb62b23de729dfdf9bf4abea96beb--indian-motorcycles-vintage-motorcycles.jpg

https://3.bp.blogspot.com/-XcbT49boV7s/U0gpb8AqJAI/AAAAAAAAB1o/VeNcWff4pnc/s320/Picture155.jpg

I see older bicycle using spokes, that would be an odd lace job.

Non related but check this out
https://i.pinimg.com/736x/1f/f0/72/1ff0729f09c2123908871509b609d95c.jpg

Interesting concept. Belt stays flat and easy to implement
https://motoredbikes.com/media/idler-arm-setup-drive-pulley-is-1-5-inches-and-rear-sheave-whizzer-is-14-inches.39569/full?d=1450156917
 
Pulley wear at 70km:

pulley_70km.jpg

Still no slippage, but it doesn't look like it will make it past 100km. Waiting for the brass pulley to arrive. I wonder what causes the wear: Just the fibre inforced belt, or dirt particles picked up from the non-slip surface or from the street?

3M or 5M belt and pulleys most likely a better choice, but the GT2 belt was the most readily available when I started out.
 
I passed the 100km mark today, and celebrated by replacing the 20T pulley with a fresh 16T Alu pulley (the only one available with the brass pulley sitting in a slow boat somewhere). I also swapped the 4240mm 650KV motor with a 4258 mm 500kv and went from 4 to 6S. The 10mm GT2 belt looks rather sad, but I have yet to experience slipping.

To eliminate the risk of grains of sand from the friction tape transferring from the large wheel to the Alu pulley and to reduce belt wear, I replaced the friction tape with an ordinary 10x4mm adhesive rubber seal/gasket (for windows and doors)

gummibaand-paa-metermaalj_280_280.jpg


That works remarkably well. I had feared derailing the belt, but the 10km drive home from work on a mix of high speed paved roads and bumpy dirt roads didn't affect it.

The ride home was in a light head wind and ended up "costing" around 4500mAh @ 6S, so the efficiency is fairly unchanged from the smaller motor on 4S. The current is lower, though, and that is welcomed by my 50A SK8 VESC.

So basically - stretching it, I know - I can ride to work on the 5000mAh 6S, I recently got for $35 on HK's Black Friday sale :) With a little pedalling and recharging at work, ofc

175545_9067000235-0.jpg


Not half bad for a 760gr battery that easily fits in the bottle holder
 
The 25T brass pulley arrived after quite some shipping time. Meanwhile, some othe stuff also arrived, including 5 pcs of new 1524mm GT2 belts and a 280kv 5055mm SK3 motor:
18175.jpg


It was on sale ($30) at HK, so I grabbed one to get the current down (low KW = equals more cells = lower current), as my SK8 ESC is rated at 50A only. I mounted it, replaced the rubber belt tensioner with a normal spring and tested on the 24v/6S pack. Smooth and powerful, but down to 26km/h topspeed.

Now I searched my basement, and found my old 3 x 3S 5200mAh pack used for the FPV channel flight (https://youtu.be/BwYqwspPHhM), re-did the connections from parallel to serial and, voila, a 9S, 5200mAh pack that still fits the bottle holder :)
 
On FOC efficiency: After having run different motors in different modes (FOC/"BLDC") and configurations, it seems that the motors always run at a higher temperature and with lower efficiency in FOC mode. I may be loosing up to 20% range in FOC mode. I wonder why? Maybe because I'm running the motors at near full RPM at 50-60% of the time?

Anyway, I'm reverting to traditional PWM ("BLDC") control mode for a while to get results for comparison.


EDIT: I have just concluded a simple 3km test with the exact same set-up and near identical condition. In FOC, the recharge was 1020mAh, and in normal PWM ("BLDC") mode, the recharge was 860mAh. The motor also seemed noticably cooler in "BLDC" mode.

So, in my case, "BLDC" mode outperforms FOC mode in efficiency :shock:
 
I'd like to suggest that you replace the rubber band in the tensioner with a silicone band. Much tougher and little if any degradation over time.
 
LeftieBiker said:
I'd like to suggest that you replace the rubber band in the tensioner with a silicone band. Much tougher and little if any degradation over time.
Click to expand...
Thanks - Actually I have just replaced it with a metal pull spring :)
 
mike_o said:
On FOC efficiency: After having run different motors in different modes (FOC/"BLDC") and configurations, it seems that the motors always run at a higher temperature and with lower efficiency in FOC mode. I may be loosing up to 20% range in FOC mode. I wonder why? Maybe because I'm running the motors at near full RPM at 50-60% of the time?

Anyway, I'm reverting to traditional PWM ("BLDC") control mode for a while to get results for comparison.


EDIT: I have just concluded a simple 3km test with the exact same set-up and near identical condition. In FOC, the recharge was 1020mAh, and in normal PWM ("BLDC") mode, the recharge was 860mAh. The motor also seemed noticably cooler in "BLDC" mode.

So, in my case, "BLDC" mode outperforms FOC mode in efficiency :shock:
Click to expand...

I’ve been trying to work this out also. There are several ways to implement FOC mode in hardware and some of them are more efficient than others. It is possible that the VESCs we are using implement FOC mode such that the motor never sees full battery voltage like it does in BLDC mode. Since the drivers are switching on and off many more times in FOC mode there is additional heat loss in the VESC. You might find it also runs hotter in FOC mode.

I would like to use an oscilloscope to monitor the voltages in the different modes and see what’s actually happening at the motor input. This would require setting the bike up on a trainer and wiring in some probes. Not something I’m able to do right now. I asked about this in another thread and nobody seemed to know.
 
falcongsr said:
...Since the drivers are switching on and off many more times in FOC mode there is additional heat loss in the VESC. You might find it also runs hotter in FOC mode....
Click to expand...

My VESC is equally cool in both modes. I can't really feel any temperature increase. I'm afraid that I don't have access to an oscilloscope that can sit on the bike while driving...
 
A little update: Currently running the bike with an 4258mm NTM 500KV motor on a single 6S pack (the 3 x 3S packs were just too troublesome to charge - one at the time).With a 16T sprocket the gearing is something like 16:780 or 1:48. I have enlarged the tensioner bearings for lower friction and longer life and I've updated the VESC firmware to the latest available, and gone back to FOC configuration. Clearly less noisy.

My latest 13km ride to work cost only 3030mAh @ 6S, equalling 235mAh/km, corresponding 353mAh/km @ 4S which is the best I've got so far. I think the ride took some 36min, so it should be around 130W of assist power. Maybe there are so many other factors to the efficiency that it's impossible to distinguish the "BLDC" and "FOC" modes...
 
How much do you weigh, and whats the added battery + motor + esc weight ?

As for the more Watts then expected, I think any little thing can swing the #'s one way or another, even at 1% grade and wind. Being you thought it would be 100W, I guess you somehow calculated that, yet it was really 200W real world, its still not too bad.

Can you post a full bicycle side shot, on the belt side of the bike. I would like to get a feel for how everything looks together, on the bike.
 
markz said:
How much do you weigh, and whats the added battery + motor + esc weight ?

As for the more Watts then expected, I think any little thing can swing the #'s one way or another, even at 1% grade and wind. Being you thought it would be 100W, I guess you somehow calculated that, yet it was really 200W real world, its still not too bad.

Can you post a full bicycle side shot, on the belt side of the bike. I would like to get a feel for how everything looks together, on the bike.
Click to expand...

markz,

sorry for the delay. I'm about 210lbs and the weight added by motor/drive/ESC/battery is very low, like 4lbs.

a12559225-97-IMG_20190916_225642439.jpg
 
You must log in or register to reply here.

Similar threads

Reid Welch
"Extreme MY1018" archival, closed out
Replies
11
Views
15,704
GIGATT
G
Back
Top