Why are DD wheels so powerful?

[Boby Kim]

Why are DD wheels so powerful? They can breaks your hands and come with an emergency stop button. Hence they are so expensive. I run a TS-PC fairly light (75% in TM software) en around 50% in game settings, as i learned that it is all about the fidelity of the forces, not the strength of the force.
So i would like to have a DD wheel with the max strength of a consumer wheel like to TS-PC for the same money. Why arent those DD wheels made?

 

[stigs2cousin]

Funny you mentioned the Cobra, I had the pleasure to drive one for while. Great car.
Acceleration were brutal but they steering, except for parking, was not particularly heavy.

Nice, that one is till missing on my bucket list.

As stated above the wheel forces on the DD are not only to replicate to real forces ( which they don´t do too well) but also to add to the "visceral experience" .
I don´t run ( and don´t plan to run) motion so the only way to " rock my world" are the probably excessive wheel forces. My rig is on rubber mounts to spare the neighbours from my tactile feedback system and as a side effect the whole rig can rock a little in its feet. The high wheel forces bring quite some movement which I like. It doesn´t replicate driving my car on the Nordschleife exactly but it´s a lot of fun and a little workout non the less.

MFG Carsten

 

[Cote Dazur]

but also to add to the "viceral experience"

This is such an important factor, often I think we all go a little too far in the search of the most “real” simulation at the expense of the fun factor. To keep enjoying our outstanding hobby, we need to balance the quest for ultimate simulation and the quest for fun, as one without the other is a little bit empty.
Your lucky you also have tactile as I find it compliment FFB so well, preventing to depend too much on FFB to provide aspect of the driving that have no business to be felt in the wheel.
My next move is motion now.

 

[RCHeliguy]

This is such an important factor, often I think we all go a little too far in the search of the most “real” simulation at the expense of the fun factor. To keep enjoying our outstanding hobby, we need to balance the quest for ultimate simulation and the quest for fun, as one without the other is a little bit empty.
Your lucky you also have tactile as I find it compliment FFB so well, preventing to depend too much on FFB to provide aspect of the driving that have no business to be felt in the wheel.
My next move is motion now.

Then you'll want better tactile and better motion

 

[metalnwood]

I like having the headroom in my SC2, before that it wsa the large mige and the small mige before that.

I dont have unusually high forces in gt cars but I do have more in the open wheelers. I dont think I want less.

The DD wheels most are using now were not developed by multimillion dollar companies with the ability to have bespoke motors made for them in their own houseings with suitable bearings. We are starting to see that now as companies like fanatec get involved and can do the CSL DD.

Where it all started, regardless of the torque rating of the servo, the size of the servos in the 20Nm+ range were ideal as a base to attach a wheel, with bearings that would be fine for the application. So all things considered, I think the overall package of the larger servo was better suited and many companies still use off the shelf servo's.

If designing from scratch, you could have a larger housing than necessary for the windings, one that would have a larger bearing and good size shaft for mounting a QR/wheel but what is the point when you get that package off the shelf just with a bit more power - which you don't have to use.

If our DD's were not put together by small companies and single guys but designed from scratch by fanatecs/logitecs then we might have lower FFB values in bespoke housings to support the other needs. So I think its a bit of a legacy thing and those with the budgets will have to match the power ratings of the other guys now too.

 

[Tom_Hampton]

I appears that you can now hang on to the wheel if you hit a wall and not pay for it. I've been tentatively hanging on more and more. I may be getting into bad habits.

Just don't take that habit to an actual track. It hurts, IRL too.

Then again...don't take the habit of running into walls to an actual track.

 

[dave kirk]

I used to have the AccuforceV2, absolutely adored it and the software was amazing, but with a larger rim there was not enough holding torque for my taste, Now I have a SCpro2, there is plenty of headroom to run whatever torque level i want

 

[Tom_Hampton]

the worst thing is with current ffb technology that:
in real car there is car that weight 1500+kg and steering wheel weight 1 kg, the forces on front wheels is gyroscopic force that makes car go straight and centers steering wheel (same force that make motorcycles or bicycles go forward and not fall on side when they are moving as they are two wheeled vehicles)

You have a very strange understanding of how things work. Gyroscopic forces are not what "centers" the steering in a car. Gyroscopic forces resist change, but once the change has occurred, they don't try to revert that change. Besides, gyroscopic forces are pretty negligible in cars because of the very small changes (and low rate-of-change) in wheel angle. Centering forces are caused by caster and trail (both static and dynamic). These forces are simple torques that are proportional to the steering angle---the greater the angle the greater the force. As long as that angle is held (and the wheel hold traction), the force will remain.

Cars and bikes are fundamentally different in their steering physics.....best not to mix the two.

The mass of the car and the wheel are completely irrelevant in this discussion. The only thing that matters is does the DD system have the torque and bandwidth to replicate the forces in sync with the real physics. Generally, these 20-30 NM Mige motors have both. Bandwidth probably suffers a bit, but not in a significant way.

I find the FFB simulation to be reasonably accurate up to and a few degrees beyond the ideal slip angle of the tire. I can find the slip angle by feel with a DD...something that is much harder to feel with a non-DD wheel.

However, I find the dynamics of breakaway and spin recovery to be completely off. I find it nearly impossible to overcome the DD forces in a simulated spin, whereas in a real car once the tires are in a complete skid there is very little force coming back through the wheel. Its as if the tires never fully break away in the sim, they always have grip.

But since the point isn't really to spin, I don't fret that nuance a lot. It is annoying to not be able to catch a few spins that I know I would be able to catch IRL---having done so. But, its just that a nuisance.

 

[Deleted member 963434]

You have a very strange understanding of how things work. Gyroscopic forces are not what "centers" the steering in a car. Gyroscopic forces resist change, but once the change has occurred, they don't try to revert that change. Besides, gyroscopic forces are pretty negligible in cars because of the very small changes (and low rate-of-change) in wheel angle. Centering forces are caused by caster and trail (both static and dynamic). These forces are simple torques that are proportional to the steering angle---the greater the angle the greater the force. As long as that angle is held (and the wheel hold traction), the force will remain.

Cars and bikes are fundamentally different in their steering physics.....best not to mix the two.

The mass of the car and the wheel are completely irrelevant in this discussion. The only thing that matters is does the DD system have the torque and bandwidth to replicate the forces in sync with the real physics. Generally, these 20-30 NM Mige motors have both. Bandwidth probably suffers a bit, but not in a significant way.

I find the FFB simulation to be reasonably accurate up to and a few degrees beyond the ideal slip angle of the tire. I can find the slip angle by feel with a DD...something that is much harder to feel with a non-DD wheel.

However, I find the dynamics of breakaway and spin recovery to be completely off. I find it nearly impossible to overcome the DD forces in a simulated spin, whereas in a real car once the tires are in a complete skid there is very little force coming back through the wheel. Its as if the tires never fully break away in the sim, they always have grip.

But since the point isn't really to spin, I don't fret that nuance a lot. It is annoying to not be able to catch a few spins that I know I would be able to catch IRL---having done so. But, its just that a nuisance.

You little not understood me , i dont mean to compare cars to bikes and i may not be Einstein on physicks so what i said may be not correct 100% those gyroscopic forces i mention. your right i may not have good understanding what i say. but most important thing to me is that what i sayd what happen in real life VS in sim when you let off wheel:
worst thing in current ffb is
in real life if you not only let off wheel but you can also turn it slightly and wheel goes back to center and car is still stable
in sims at speeds lets say 200+ or 300+ if yuou let off wheel at bumpy road it want to transfer those bumps into wheel so it make it shake, but thing is in current ffb technology (outdated windows direct driver or something) that forces shake wheel so much that it starts oscillating more and more and thus make car go out of its path and crash.
i tell again, its not a big deal and no problem if you just hold wheel always, but it just a small inconveniece how ffb works in sims, that its not always 100% connected, like wheel is slightly unconnected and actual oscillations on wheel side can shake whole car in sims and make it crash when driving straight.

 

[alfye20]

I find it nearly impossible to overcome the DD forces in a simulated spin

Servo Friction is your friend here, not too much not too low

 

[Cote Dazur]

Servo Friction is your friend here, not too much not too low

What is/does servi friction? How will it help in that situation?

 

[alfye20]

What is/does servi friction? How will it help in that situation?

On Simucube wheels, and some others, you can use constant operated filters like Friction.

Description From Granite Devices WIKI

The friction parameter adds artificial dead weight to the feel of the SimuCUBE. The friction filter adds resistance across the entire signal bandwidth. Going too far on this parameter can actually override the forces applied by the software and cause the wheel to not move. In essence this parameter slows the reaction of the wheel. This filter is best used to mimic what would be steering hardware in a real car and all the friction in the steering caused by the steering rack, ball joints, bushings, ect.. Including tire contact patch drag. This filter will also limit a feedback by-product that feels a bit like a rubber band. This effect comes from the fact that force feedback for the most part is designed to return the steering wheel to the next position point no matter what you do and to do that it adds power in the opposite direction that you are turning.. The rubber band feeling comes from when you quickly turn in the direction that the force is wanting you to go and then back into the force.. When you do this with a high speed system, such as the OSW, that can move the wheel faster than you can turn it yourself, you effectively lose all feeling of force when turning with the power making the turn motion overly light. if you weave like this around a constant radius corner you will find that you get this heavy feeling and then no feeling, heavy and then none, like stretching and releasing a rubber band. By adding a friction % you will slow the wheels ability to return to center in a more progressive manner allowing you to turn the wheel back toward center and retain some feeling of force in the return motion. Note, that the higher you set the friction the more friction you will have in both directions of movement so by adding friction you can increase the force required to turn the car to a possibly unacceptable level and as well diminish the speed of wheel return which can affect your ability to catch a slide. This should be adjusted to a level that you feel is realistic. This setting would be best used if you feel the forces of the wheel are correct but the steering feels overly light in nature during certain movements, by increasing it you are not seriously affecting the way the feedback translates to the wheel but you are adding weight to the steering mimicking the drag that would be in a real life steering rack.

 

[tahir51]

I also believe the market in will eventually come down to 12-15nm Wheel bases being the "standard" DD as developers like Fanatec will optimize and develop new servos specifically for simracing. So the new 12-15nm wheelbase will probably have the same acceleration as the now 20nm base.

They already did a step in this direction with the CSL DD, having a very light (carbon I believe) inner shaft which reduces the inertia thus allows faster acceleration. Combine this with a light 500g wheel rim and you would have an already very responsive wheelbase even if it is only 8nm.

 

[RCHeliguy]

I also believe the market in will eventually come down to 12-15nm Wheel bases being the "standard" DD as developers like Fanatec will optimize and develop new servos specifically for simracing. So the new 12-15nm wheelbase will probably have the same acceleration as the now 20nm base.

They already did a step in this direction with the CSL DD, having a very light (carbon I believe) inner shaft which reduces the inertia thus allows faster acceleration. Combine this with a light 500g wheel rim and you would have an already very responsive wheelbase even if it is only 8nm.

Nm = torque. Torque is what makes the motor accelerate quickly.

Hp = torque(Nm) * rpm/7127 and we are dealing with nearly 0 rpm at all times driving.

The inner shaft has almost no moment arm of inertia and has very little impact on acceleration unless you are dealing with an extremely weak motor.

Lower Nm bases have always been available, but they can not accelerate as quickly from 0rpm with a load on them. That is purely a matter of physics and defined by the Nm the motor is capable of.

If the weight of your wheel has much of an impact on the direct drive's performance, that wheelbase is underpowered.

The best you can hope for is that a motor is more efficient using less wall outlet power, but you can't take away torque and expect the same performance. That isn't a matter of what they come up with.

The software driving the motor is where the secret sauce is. There are many ways to create a feeling at the wheel and if they are creative about how they drive the motor, it is possible that they create a better feeling wheel that isn't as powerful.

 

[Cote Dazur]

Nm = torque. Torque is what makes the motor accelerate quickly.

Very true. I read you post, very well written. It made me wonder.
A 20nm motor when operating at 50% is actually 10nm, does it feel slower when operating at 50%?, 50% slower?
Is the feeling of speed when using a DD base really so closely related to strength?
Like with many other aspect of anything, at one point of the diminishing returns is the theoretical advantage not so important anymore?

 

[RCHeliguy]

Very true. I read you post, very well written. It made me wonder.
A 20nm motor when operating at 50% is actually 10nm, does it feel slower when operating at 50%?, 50% slower?
Is the feeling of speed when using a DD base really so closely related to strength?
Like with many other aspect of anything, at one point of the diminishing returns is the theoretical advantage not so important anymore?

That ALL depends on how the software driving it works and how your profile is configured.

For example if you have the wheel loaded up on a corner near 10Nm with it set at 50% will it allow the detail to come through past 10Nm? I'm going to guess yes.

The software guys would let tiny impulses and spikes through using the full capabilities of the motor just because that would add feeling that people want. This is all signal processing.

There is a lot going on in the secret sauce that they are using to control the motor. The signal processing and control loop of the motor is where most of their effort goes. People get hung up on motor specs because that is the simple part that they understand. If the secret sauce is bad the motor specs mean nothing.

 

[RCHeliguy]

Let me go off on a quick tangent.

RC Helicopters are VERY strongly impacted by rotor speed. Ideally you want an absolutely steady rotor speed while you are flying 3D acrobatics so that you have the same tail control and your stick inputs give you the results you expect. If the motor bogs down you can lose the tail and crash the heli.

Motors are about power and efficiency, but they are EXTREMELY simple devices that are very well understood. Assuming your motor is powerful enough, the ESC ( Electronics Speed Control) is the most critical part of this equation.

In the old days, the throttle was associated with a V curve based off the collective inputs. That sort of worked, but not very well.

Then ESC's got governors which would decide how much current to dump on the motor to maintain a steady speed. Better ESC's could keep the rotor speed +/- 25-50rpm.

Then they added FBL Controller intervention based on a rotor speed sensor. The FBL controller has an advantage over the ESC in that it knows what you are trying to do from the sticks and it can preemptively dump more power on the motor as it is applying a hard collective or cyclic input to the swash rather than just reacting to the motor slowing. Not only that but you could configure it to goose the motor more or less depending on your flying style to get performance that may work better for you than simply having a steady head speed.

What we found out in practice was that FBL controller input was more efficient and gave you longer flight times than having the ESC riding the motor hard to keep it at the same rpm.

This is similar to DD wheelbases in that the motor is basically the dumb part of this. These DD wheelbases have precise sensors to let the software know how much motion is happening in the wheel. That way the software knows if the wheel is doing what it wants it to do and as it is reading the telemetry values and then filtering that through your profile, it can try to create the experience you are after. That is the part that really matters.

 

[blekenbleu]

Some wheel base controllers attempt to measure installed steering wheel inertia during power-on calibration
to more precisely control amounts of power needed for accelerations.
One trick is measuring so-called back EMF during pulse width modulation "off" times
to better estimate speeds and loads, even for nominally open loop motors.
At least in theory, this could help alter perceived steering inertias
to better approximate that for some simulated vehicle.

 

[Fennario]

Another way of thinking about things is in terms of equal energy/force delivered to the end user in specific instances such as hitting a sausage wrong, bumps in the road while cornering, etc. but where the duration of the force is modified per the platform.

1. Belt driven. 1nm of consistent force over 1/5 of a second.
2. Direct Drive. 10nm of peak force over 1/500 of a second followed by a quick tapering.

In both instances the total force applied to the end user is the same but the increased amount over a shorter duration for the DD is what provides the greater fidelity (adjust figures if my physics is off).

 

[Deleted member 197115]

Some wheel base controllers attempt to measure installed steering wheel inertia during power-on calibration
to more precisely control amounts of power needed for accelerations.

Which wheel bases are these? With SC2 there is no self calibration on start up as with older encoders, plus you can always change steering wheel without shutting down wheel base.
And most of the times it's not really the weight of the wheel but diameter (lever) that affects how hard it is to overcome particular torque resistance.

 

[Deleted member 197115]

Another way of thinking about things is in terms of equal energy/force delivered to the end user in specific instances such as hitting a sausage wrong, bumps in the road while cornering, etc. but where the duration of the force is modified per the platform.

1. Belt driven. 1nm of consistent force over 1/5 of a second.
2. Direct Drive. 10nm of peak force over 1/500 of a second followed by a quick tapering.

In both instances the total force applied to the end user is the same but the increased amount over a shorter duration for the DD is what provides the greater fidelity (adjust figures if my physics is off).

I am pretty sure they refer to this as Slew Rate, this is how simucube 2 models compare based on their corresponding max torque.