What are the best rear shock settings with hub motors?

[peters]

The hub motors are heavy, the unsprung weight of the rear wheel is much more than without a hub motor, so the proper suspension adjustment is not easy. But there are opinions that despite the extra weight it can be set properly and the weight is not a drawback. Today many offroad ebikes are manufactured with heavy hub motors. I suppose it is not the unsprung weight alone that matters, but the ratio of the sprung and unsprung weigths.
I did not find a particular thread for this topic, so I intend to discuss it here and find the best solution for offroad use.

Some questions I think about:
How the shock should be adjusted compared to a bike without hub motor?
Motorcycle shocks are better than bicycle shocks for this purpose?
Which extra features of the shock is useful for better suspension? E.g. high-speed/low-speed compression? Two separate damping settings? ("beginning" and "ending stroke rebound" at Rockshox Boxxer)
Progressive spring?
Did anyone do comparative tests with different rear shocks and different settings?
What are the experiences?
Please give as much technical explanation as possible (can be equations or anything).

I can tell about my experience with RS Kage RC, but I have nothing to compare to:
I found the best is to set the spring preload to zero and increase the damping to more than the usual, because the weight of the wheel pulls out the shock when it moves outwards (not only the spring pushes), so with a lighter damping there was a knock at the end of the rebound. With more damping the rebound seems quite slow, but still better than the end-of-rebound knock. There is also a low-speed compression knob on this shock, I set this to minimal, because my spring rate is 600lbs/in., probably I should change it to a softer one.

Here are some informative videos explaining the suspension features on bicycles:
http://www.pinkbike.com/news/technical-tuesday-fox-dhx-setup-2010.html
http://www.pinkbike.com/news/technical-tuesday-rockshox-world-cup-adjustments-2010.html
http://www.pinkbike.com/news/Tech-Tuesday-negative-spring-air-shocks-2012.html

 

[docnjoj]

I have had 3 different means of determining what I like in terms of rear motors and suspension reaction.
First the ride is subtle and smooth with just the original 8 speed cassette wheel. This suspension on the Steintrike is not at all soft, since it is designed for large Europeans and 50 kilos of extra stuff for touring. It does have about 5" of travel at the wheel and 600 in/lb springs.

Second the reaction with a 9C hub required tightening the spring (no separate shock adjustment but it is a high quality nitrogen/oil damper. Very harsh ride and lower suspension travel, of course.

Third using a lightweight (2kilo) geared motor. Almost like using the standard freewheel in point one. Smooth and able to react to large bumps and speedbumps very rapidly and fully. No adjustment to coil spring tension needed.

So from my perspective, unsprung weight sucks and any response to it really compromises the ride.
otherDoc

 

[motomech]

Well, every combo of motor/shock/rider weight/riding style would determin the settings, but I can tell you what worked for me.
On my 2WD, my rear motor is a mini, light at less than 2 Kg.
The rear suspension is rather old skool, A Fox Float R with a simple single pivot link.
I am heavy, more than 250 Lb.s.

Rather than setting the "sag", like a dirt bike, I use an external O-ring on the shock body, which tells me how much travel I'm using and set the air pressure accordingly. Usually around 140 psi.
Rebound-Since I am riding an the street, I am more interested in comfort than control and I have found that adding much rebound adversly effects the ride, so only one or two clicks on rebound.

There is no compression setting on these units.

The biggest difference came when I installed an "air spring volume spacer" from this kit;

http://service.foxracingshox.com/consumers/Content/Service/QuickTech/FloatAirSpringSpacerProc.htm

The difference was dramatic and I recommend the spacer for heavy riders. It stiffened the mid-stroke, adding resistance to "bottoming out", while keeping the initial movement soft.

If I was riding off-road, I suspect I would want more rebound damping.

As a side note, when I used to ride dirt bikes, the way I set the shock was;
Compression-I would add compression till I got the rear tire "hooking up" the best. Lack of compression damping would let the tire bounce under power(probably not many Edirt bikes where this would apply).
Rebound-if the seat would come up and smack me in the butt in the "whoops", I would add rebound.
On the old Yamaha Monoshocks, the only thing that would midigate this was to be under power when hitting large bumps.

 

[cal3thousand]

For me, I had to increase my spring rate due to the weight of the batteries up to 600lbs from 550lbs. I set my rebound tighter than without a hub motor and sag is about the same for me.

 

[d8veh]

To control the wheel properly with the additional mass of a hub-motor, you'd need to increase both the spring-rate and the damping, but, as you've already figured out, the ratio of sprung to unsprung weight will go down, so the suspension will never be as good.

 

[Drunkskunk]

I run one of the heaviest production hub motors on the end of a long arm rear suspension, so I'm very familiar with the problems of unsprung weight.

Of course, one of the big factors in a suspension is the ratio of load VS unsprung weight. The is a big difference in handling between 50 LBS of unsprung weight and 100 LBS of load, VS 50 LBS of unsprung weight and 500 LBS of load.

But that's not the only factor. Mass in motion has kinetic energy. Even if the ratio is high enough to make control easier, you still need a shock that can handle dissipating the higher energy. You'll need larger reservoirs, more surface area for cooling, larger valve surfaces, and components that can run under higher stress than many MTB shocks. Motorcycle shocks may be overkill, but luckily there are DH shocks, that work well.

I've found I can tame the Monster by keeping the rebound set high, the air pressure lower, and the valving on the compression set to soft. That works in most cases, and keeps the back end planted even in rough conditions.

 

[cal3thousand]

I run one of the heaviest production hub motors on the end of a long arm rear suspension, so I'm very familiar with the problems of unsprung weight.

Of course, one of the big factors in a suspension is the ratio of load VS unsprung weight. The is a big difference in handling between 50 LBS of unsprung weight and 100 LBS of load, VS 50 LBS of unsprung weight and 500 LBS of load.

But that's not the only factor. Mass in motion has kinetic energy. Even if the ratio is high enough to make control easier, you still need a shock that can handle dissipating the higher energy. You'll need larger reservoirs, more surface area for cooling, larger valve surfaces, and components that can run under higher stress than many MTB shocks. Motorcycle shocks may be overkill, but luckily there are DH shocks, that work well.

I've found I can tame the Monster by keeping the rebound set high, the air pressure lower, and the valving on the compression set to soft. That works in most cases, and keeps the back end planted even in rough conditions.

Do you have a process that you follow when tuning a shock to the bike? (I'm going to run lower air pressure and test that out, but wanted to know more about your method)

 

[docnjoj]

I've found I can tame the Monster by keeping the rebound set high, the air pressure lower, and the valving on the compression set to soft. That works in most cases, and keeps the back end planted even in rough conditions.

This is the standard method in European sport sedans to get a good ride and good bump control. It seems to work on trikes too, but as stated above, the unsprung/sprung weight situation always has to be controlled. Less unsprung weight always leads to better ride/handling. It's the law!
otherDoc

 

[peters]

Thanks for the comments so far.
I also found the distribution of weight between the front and rear wheel is not negligible. When I sit forward or lean then the rear suspension is worse than when sitting back. More sprung weight over the rear wheel is better, so a shorter swingarm is better from the suspension point of view. When this weight is too low, the rear part of the bike pops up on a bump instead of compressing the shock.

 

[docnjoj]

Thanks for the comments so far.
I also found the distribution of weight between the front and rear wheel is not negligible. When I sit forward or lean then the rear suspension is worse than when sitting back. More sprung weight over the rear wheel is better, so a shorter swingarm is better from the suspension point of view. When this weight is too low, the rear part of the bike pops up on a bump instead of compressing the shock.

That may be true but you sure don't want that rear weight high up like on a rear rack. This really screws up the handling even on trikes. Center the weight with a battery in the middle of the bike or trike. Much better handling.
otherDoc

 

[peters]

I found this guide about the features of suspensions, although does not answer the question of compensating the unsprung weight, but useful to understand a few things:
https://www.sram.com/service/techdocs/rockshox-suspension-theory-guide

 

[peters]

Instead of the rear shock (that is a passive device) I could imagine an active suspension with force/torque sensors, motorized pivot and electronic control, this could compensate any wheel weight and could be tuned to any characteristic.
Just brainstorming...

 

[Drunkskunk]

Do you have a process that you follow when tuning a shock to the bike? (I'm going to run lower air pressure and test that out, but wanted to know more about your method)

Not much of a process really. I always start off with everything set at it's softest setting, and then work to up to firmer. I make only 1 adjustment per test ride, and for a coil spring shock, I work in the order of:
Spring rate,
Compression,
Rebound,
then Air pressure.
If it's an air spring shock, the air pressure is first.

 

[docnjoj]

Instead of the rear shock (that is a passive device) I could imagine an active suspension with force/torque sensors, motorized pivot and electronic control, this could compensate any wheel weight and could be tuned to any characteristic.
Just brainstorming...

Corvette and some Cadillacs have that magnetic fluid that thickens and thins with application of electric current. Nice and simple (not)!
otherDoc

 

[cal3thousand]

Instead of the rear shock (that is a passive device) I could imagine an active suspension with force/torque sensors, motorized pivot and electronic control, this could compensate any wheel weight and could be tuned to any characteristic.
Just brainstorming...

While we're brainstorming...

I've been waiting for someone to take advantage of the properties cornstarch (or something similar) that is solid at high speed and liquid at low speed, and put that into bicycle use:

http://en.wikipedia.org/wiki/Non-Newtonian_fluid

 

[peters]

I replaced my shock and made a little calculation and graphs to understand the situation better.
The new shock is the same RS Kage RC, but it is 240x76mm (that I already purchased for my next frame) instead of 216x63, but the spring rate is the same (600lbs/in.). To mount the new shock properly I cut a few cm from the shock console on the swingarm.
The suspension is really much more comfortable now and the bike is much easier to handle, although the change on the graph does not seem very significant.
The graph shows the shock travel vs. the travel of the rear wheel axle (x). The travel of the rear wheel axle is the travel on the arc (rotation around the pivot), not the displacement. I don't know which one is usually taken into account, but the difference is less than 1%.
Red curve is with the old shock, blue curve is with the new one. The curves are close to linear, although they are a bit progressive in the most of the travel and degressive at the end.

This is my equation, based on a little trigonometry:
y = S-sqr(PS1^2+PS2^2-2*PS1*PS2*cos(acos((PS1^2+PS2^2-S^2)/(2*PS1*PS2))-x/PA))

y: shock travel
x: rear axle travel
S: shock length (old: 216mm, new: 240mm)
PA: pivot - rear axle spacing (535mm)
PS1: pivot - shock upper screw (272mm)
PS2: pivot - shock lower screw (old: 198mm, new: 185mm)

with the actual parameters:
old suspension (red):
y = 216-sqr(272^2+198^2-2*272*198*cos(acos((272^2+198^2-216^2)/(2*272*198))-x/535))

new suspension (blue):
y = 240-sqr(272^2+185^2-2*272*185*cos(acos((272^2+185^2-240^2)/(2*272*185))-x/535))

I plotted the graph here: http://rechneronline.de/function-graphs/

On the red curve the maximum of y to be taken into account is 63, because it is the shock travel, and then the rear wheel travel was ~170mm.
On the blue one the max is 76mm and rear travel is 221mm.
The leverage is the inverted slope of the curves, 2.7 for the old one and 2.91 for the new one. Interesting that this small change makes a huge difference in the suspension. The curves are very similar, so a softer spring would have had almost the same effect, although the travel would have remained 170mm.
Now the sag is only 20%, so I would still need to change the spring to a 500 or 450 lbs/in for 25% sag, then I expect it even better.

 

[cal3thousand]

Impressive bit of math you did there!

This deserves a write-up to explain all the science/geometry behind it.

 

[peters]

Not a big science, just applied the cosine rule: c^2 = a^2+b^2-2*a*b*cos(g)
Sorry for the Paint drawings
Ok, at first I calculated the angle included by PS1 and PS2 at the pivot, with the cosine rule (figure 1):
g = acos((PS1^2+PS2^2-S^2)/(2*PS1*PS2))

When the suspension is compressed, the rear axle moves on an arc, marked 'x' on figure 2. The angle of the arc is 'f', so the length of the arc is x = PA*f, (f is in radians), so:
f = x/PA

The PS2 line also rotates around the pivot by angle f, creating another triangle PS1, PS2 and S2, where S2 is the length of the compressed shock and the new angle between PS1 and PS2 is g-f. Applying the cosine rule for S2:
S2 = sqrt(PS1^2+PS2^2-2*PS1*PS2*cos(g-f))

The shock is compressed by y, that is the diff between S and S2:
y = S-S2

And combining all the formulas above:
y = S-sqrt(PS1^2+PS2^2-2*PS1*PS2*cos(acos((PS1^2+PS2^2-S^2)/(2*PS1*PS2))-x/PA))

(A few things could be calculated differently, for example if not the rear axle, but the point where the tire touches the ground is taken into account for the rear wheel movement. But it would be more complex, especially if the point of view is that this tire-ground contact point is fixed and the pivot moves down - then the distance of the pivot and this point also changes at compression.)

 

[litespeed]

Nice work Peters!

Tom

 

[docnjoj]

I can't say I understand all the math but as a rule of thumb, if you increase suspension travel you can decrease spring rate. This, with a proper shock valving almost always gives a better ride and better control over large bumps and holes. You do want to decrease jounce spring rate and increase rebound shock rates. Think of driving an older Mercedes or Citroen over bumpy ground. They both just "soak up" the bumps. My trike has springs that are too stiff, but it is sprung for massive amounts of heavy riders and lots of luggage. Spring/shock only works well over large holes and bumps on the Steintrike. They use stiff ones cause they race them downhill on luge tracks in the summer.
otherDoc