The Grinhill Mk2 RC-motor drive

[Grinhill]

Here's my most recent refinement, a wiring tidy-up. It was prompted by the need to increase battery capacity, so I will be going to two water bottles in parallel.

So if I need to use both bottle cage mounts, the power switch and other bits will need to move. The top tube would be ideal, except for the gear cables being in the way. Maybe a long channel for the wiring to sit in, hanging from the tube. What can I find in the garage...

rc_proj 079s.jpg (110.24 KiB) Viewed 1261 times

It's a diffuser off an old fluoro lamp. The size and shape seems ideal for the job. Stuff can sit in it and be kept clear of the moving cables. So I relocated the switch box, and added a bit of paint to match the bike.

rc_proj 075s.jpg (71.8 KiB) Viewed 1258 times

rc_proj 076s.jpg (47.86 KiB) Viewed 1262 times

rc_proj 077s.jpg (60.14 KiB) Viewed 1260 times

Now with everything reinstalled, a few dabs of hot-melt glue to keep it in place, a few cable ties, it looks pretty tidy.

rc_proj 078s.jpg (58.54 KiB) Viewed 1273 times

[Grinhill]

Here's some data since I've been using the 8S2P battery pack: (I did post something about this a few weeks back but have since found that I wasn't running at full speed!)

New top speed on the flat (unassisted):
55 km/h (35 mph) at 1300W power cunsumption

New top speed up 10% grade (with assist FWIW):
45 km/h (28 mph) at 2000W power consumption

For my return commute (12 km) I use around 6Ah of the 10Ah.

I have had some slight movement of the sliding plate mounting arrangement after hitting potholes. Although only a few mm, this has reduced the belt tension sufficiently to allow some belt jumping at high power. I will have to add a couple more bolts to lock it in place.

[grwsaltspring]

I'm following your build with great interest as I'm curious as to how the "simple" build will turn out compared to some of the more exotic ones here.

Cheers Greg

[Grinhill]

I locked the plates together with two additional screws, just a couple of the self-drilling sort.

Still had belt slippage at the 2000W mark, it's just some very quick skipping, not continuous.

Maybe I'm just reaching the power limit in this configuration. The long belt and small motor pulley are less than ideal for this type of belt. You can see the belt slapping around a bit at high speed.

[grwsaltspring]

Any idea on the direction you will take this next in your ongoing refinement?

Greg

[swbluto]

I locked the plates together with two additional screws, just a couple of the self-drilling sort.

Still had belt slippage at the 2000W mark, it's just some very quick skipping, not continuous.

Maybe I'm just reaching the power limit in this configuration. The long belt and small motor pulley are less than ideal for this type of belt. You can see the belt slapping around a bit at high speed.

I had continuous belt slipping at 2 kW from a start and ended up busting the belt. I'm thinking that somewhere around 1.5-2kw is the limit for 5mm 15mm HTD belts in conventional direct drive configurations, but I think that my axle oscillations probably contributed to the slipping. It seems that also matt had belt slipping at 3kW, so it's definitely somewhere in the 1.5kW - 3 kW region. I've heard it's actually a torque limit and a belt can transmit more power than this, but it has to be spinning at a high RPM to ensure that it stays below the torque limit (As power = torque * (RPM, or technically, angular velocity)).

[Miles]

Ratcheting point (for a given belt) depends on static tension and torque/teeth engagement on the driver pulley.....

[swbluto]

Ratcheting point (for a given belt) depends on static tension and torque/teeth engagement on driver pulley.....

So true, but generalities are sufficient as guidelines as not everyone has the resources readily available to calculate the ratcheting point for their particular belt. And "conventional direct drive" seems to be narrow enough a category to generalize.

(Although, I'm driving an 8 inch wheel and he's driving a 24+" inch wheel. Maybe not.)

[Miles]

The point is though, that it's essential to take into account the number of teeth engaged on the driver pulley.....

[swbluto]

The point is though, that it's essential to take into account the number of teeth engaged on the driver pulley.....

True, but for conventional direct drives on bikes, the amount of engaged teeth doesn't vary by that much, does it?

I assume that people need to maintain a gear ratio of at least 10 for this configuration and that the "chordal angle"(Just taking a guess here) is usually between 45 and 60 degrees. Since the driving pulley isn't likely to be less than 14, and the wheel pulley isn't likely to be more than 200( Or would it?), it seems that there's probably a variation of upto 2 or 3 out of a mean of 6 engaged teeth.

So, what's the difference between 5 teeth engaged and 8 teeth engaged?

Ok, so it seems it might matter by a little bit. I guess it's a linear relationship and 8 teeth would have 60% more static loading capability than 5 teeth?

[swbluto]

Well, I'll go out and say that I was putting in 1.8 kW into the system while probably engaging 5 or 6 teeth. According to my simulator, it appears the motor was probably putting out 10.5 N.M. when it continuously slipped and broke. It seemed like it didn't continuously slip when it was 4 N.M. and less.

(By the way, before people go "Oh, that's a sucky torque", consider this: It has a gear ratio of 4, that's 42 N.M. at the wheel. Also, the wheel is 8" which about three times less that of a 24" wheel, so the torque that would correspond to a similar force on a 24" inch wheel is about 126 N.M. According to ebikes.ca's simulator, that kicks butt, but it seems the 530x are similarly powerful in the 40-50 volts region at 100 amps. My scooter is consuming 60 amps at 33 volts.)

[grwsaltspring]

Are the number of teeth on the drive pulley and ones ability to successfully mate this pulley to the RC motor of choice the limiting factors with a build like this one (IE Grinhills build)? Particularly if we are trying to use off the shelf components for the most part.

[Grinhill]

A little more info to add to the debate.

My belt is the 9mm wide one, limited by the large pulley, which has about 12mm between flanges. My estimate of the tooth count on the large pulley is 266 (422mm diameter, 1330mm circumference divided by 5).

An unusual feature of the large pulley is that the teeth aren't full profile, they are fairly shallow ridges at 5mm intervals. Then, the pulley is divided into quadrants around the circumference, and only two of the four quadrants have the ridges, the other two are smooth. Apparently the friction from having such a large contact area reduces the need for having all teeth.

I'm not able to determine whether the slipping is occurring on the small pulley or the large one.

[Miles]

Grinhill,

It will almost always slip on the small pulley. What size is yours?

The sections with the missing teeth are probably there because of the need to separate the four mould sections radially in order to be able to mould the double flange.

[Grinhill]

20 teeth. (ratio = 13.3).

[Miles]

A 20t pulley should be able to handle at least 5Nm torque (around 4kW at 8000 rpm) with a 9mm wide high-torque belt. How are you setting the tension?

The problem is usually the high torque levels you can get when accelerating at low speeds ......

[Grinhill]

A 20t pulley should be able to handle at least 5Nm torque (around 4kW at 8000 rpm) with a 9mm wide high-torque belt. How are you setting the tension?

I have the M6 screws slackened a little, push the motor assembly upward with one hand then drill some smaller holes for the locking pins using the other hand. The belt had about 10mm movement with light finger pressure (in the centre of the chord). I repeated this process yesterday and managed to get some more tension, now only about 5mm deflection.

The problem is usually the high torque levels you can get when accelerating at low speeds ......

I don't use WOT from low speeds, I always ease the throttle up gently.

I still had a few slips this morning on the steepest hill section, but this time I noticed it coincided with bumps in the road. I think the whole drive unit is flexing ever so slightly, which is releasing the tension enough to let a few teeth slip. Another factor is this was towards the end of the ride, so the belt would be warmer. Does this make a difference?

This morning was a good test - a headwind of around 40 km/h.

I was burning 2400W to maintain 50 km/h on a level road! (note no slipping during this section).

[heathyoung]

Yep it was crazy windy this morning! Seen your bike a couple of times around Cardiff, I'm usually in the car otherwise I would say hello.

[E-Ride]

...
This morning was a good test - a headwind of around 40 km/h.

I was burning 2400W to maintain 50 km/h on a level road! (note no slipping during this section).

So how much amps? (2400/22.2 ~ 108?)

[Grinhill]

So how much amps? (2400/22.2 ~ 108?)

I have the Cycle Analyst set to Watts, and I wouldn't take a hand off the bars to press the button at that speed.

This is while running on 8S (29.6V), so say 80 A.

I checked the peak amps recorded on the CA, it was 116A. Probably during acceleration.


Heath, you are welcome to PM me if you would like to take a closer look some time.

[Grinhill]

I came up with an improvement for the belt-slip issue a few weeks back, and it's working well (yes, I know it's a bit "ghetto" ).

rc_proj 092s.jpg (66.76 KiB) Viewed 1332 times

I have seen power up to 3200W on one occasion! It's an old skate wheel (actually came from a suitcase) and the bearings are rubbish, but still works. It's fixed, no spring.

Some further info on the belt - I have observed a clicking noise when running the motor at low revs, which turns out to be twice per revolution. This appears to be the source of my slipping problem as well - the large pulley, not the small one! Because of the pulley design, not using a proper HTD profile for the teeth, and having two quarters of the pulley smooth, the following happens:


There must be a slight mismatch in the alignment of the belt teeth to the pulley teeth, for whatever reason. After the smooth section, the belt rides on the ridges for a while instead of seating down to engage the pulley. This will cause a momentary increase in tension due to the increased diameter. The belt then engages with a snapping noise, slipping back to where it should be for a while until the next section with no teeth.

This means that it is my large pulley slipping, not the small one. And it's worse at high speed because the belt is travelling at higher speed. It also means that there are some fundamental design issues with the pulley. One possible solution would be to use a longer belt to increase wrap around the large pulley by using an idler. This will keep the belt engaged with the pulley for almost the whole way around, not giving the opportunity for a mismatch to occur.

[Grinhill]

One possible solution would be to use a longer belt to increase wrap around the large pulley by using an idler. This will keep the belt engaged with the pulley for almost the whole way around, not giving the opportunity for a mismatch to occur.

Well, today I tried this solution, and it worked brilliantly - not a single slip!!!

I already had a longer belt (327 teeth) which I bought by mistake a while back. I also had the other skate wheel from the old suitcase. Cut a length of aluminium from a shower screen moulding, a bit of grinding, filing and drilling, and here it is:

rc_proj 099s.jpg (71.95 KiB) Viewed 1224 times

I checked the belt tension and alignment first, adjusted the angle of one a the rollers a bit to keep the belt on track, then gave it a spin with the wheel off the ground. There was slightly more drag, my full-throttle wheel-in-air power increased from 170W to 190W, but it's well worth it. (interestingly at the end of the ride with everything warmed up this figure dropped to around 130W).

It was a little quieter than before, and has alleviated the belt rubbing on the frame in one spot as an added bonus.

Figures from my test ride down the hill and back up:
6.6 km
Max speed 60 km/h (37 mph)
Avg speed 42 km/h (26 mph)
Batt used 4.6 Ah
Efficiency 21.5 Wh/km (34.6 Wh/mi)
Peak power 3200W

[amberwolf]

I like this build...I may borrow some ideas from it for a future version of the DayGlo Avenger.

[heathyoung]

Thats interesting - the cheap version of the golden eagle bike engine system as available here (sold for $450 as a blackstar on feepay) - has exactly the same style of snap ring - these used to slip as well... Golden eagle use an idler wheel as well, sprung onto the belt to try and combat this.

[spinningmagnets]

Grin, I've been following every post of yours. I'd like a bike to be 30-MPH (48-kph) capable, even though I'd probably ride at 25 most of the time (40-kph). Would it be better for range and torque if I just...

A. Copy your build and use less throttle
B. Get a smaller motor pulley (might not be possible)
C. Use a lower system voltage. (29V now, yes?)
D. Spec a motor with a different kV (if yes, which one?)

Any input is appreciated...