Seat mover vs DBOX? No...seat mover + DBOX!
1 1

23 posts in this topic

I had a random idea last night and figured to ask if anyone had experience or an opinion one way or another.

I know people tend to fall into camps: seat mover fans and DBOX fans.  I have a seat mover (stage 4+ SimXperience) but I always keep an open mind to what is out there.  I'll admit to being intrigued by how quiet and compact a DBOX setup can be, but I'm very happy with the way my rig forces me to tighten my core (i.e. basic seat mover principles).

My question is this, could we not combine the two?  Why not have a DBOX rig, but have the wheel deck supported stationary above the moving rig, essentially creating the same type of setup as a seat mover, but using the DBOX actuators in place of the SCN5 arrangement?

Good idea?  Terrible idea?  Doesn't work because ____?  Any thoughts?

Share this post


Link to post
Share on other sites

I have often thought that combination would make for the ultimate rig. I asked Berney the following question on iRacing and this was his reply. I am glad they are playing with the concept so you never know what might eventuate.

 

 

Through your experimenting did you ever try combining a seat mover with a platform mover? I have not seen anyone try this but I have often considered what would happen if I sat my SimXperience Rig onto a D-Box system. I don't even know if SimCommander and the D-Box software could work in unison and although expensive, would that not give you the best of both worlds?



Berney wrote:


This is indeed something that we have and continue to put a fair bit of energy into. I don't think it would be fair of me to draw a final conclusion at this point, however, I can share a few of the concerns and a tidbit of background on each: 

Is a D-Box Representative of 4-Corner Independent Suspension or Pitch/Roll/Heave 


If you think of your rig as a solid axle and your D-Box actuator as a shock/tire that interfaces between the solid axle and ground, it quickly becomes apparent that you can't control the actuators independently so as to represent a 4-corner independent suspension or you would either try to tear the rig apart of lift actuators off the ground. 

To solve this, the control software calculations basically wind up giving you pitch/roll/heave or something very close to it with maybe a VERY small tactile/vibratory representation of the independent suspension movements blended in. Any remaining tactile effects wouldn't be able to be independent either in theory. 

In short, I question what a D-Box can do that any other Roll/Pitch/Heave solutions couldn't do with fewer actuators and lower cost. 


If for the purpose of this conversation, we say that the D-Box really only provides Roll/Pitch/Vertical Heave, this presents an interesting challenge: the combination of Seat Mover Style G-Force simulation with Roll/Pitch simulation. 


ride_roll_anim1.gif


G-Force vs Pitch and Roll 

If you're not familiar with the concept/principles of our Stage Series motion, I'll spare you the pitch here and suggest that you wrap your head around it by visiting our website and reviewing the 'Motion Principles' tab on any of the Stage Series product pages. 

In short, we choose to focus on the most extreme component of what a driver feels and have concluded that to be G-Forces and related muscle tension/pressures followed by traditional motion simulation which aims to manipulate your inner ear/balance, etc... To simulate the impact of G-Forces on the human body in a convincing manner, you need to be able to make darn quick movements relative to simulating the impact of pitch/roll on the mass of an entire vehicle convincingly. 

That said, we've always prioritized flexibility/tunability so as to facilitate widely varying customer preferences so a Stage Series sim can do either G-Force simulation or Roll/Pitch simulation and even some interesting combinations of the two depending on how you setup your profile. 

The problem becomes that since the D-Box motions are very much like pitch/roll style motions, depending on how you setup your Stage Series (G-Force simulation or Roll/Pitch simulation) the D-Box movements simply become additive or canceling but with a bit of added weirdness since due to the amount of mass involved with moving an entire Stage Series, the D-Box movements are relatively delayed and probably would be perceived as such even with a lighter rig due to actuator performance differences. 


image001.jpg 


Timing 

Not all motion simulation software is created equal, in fact, the majority on the market induce an amount of lag that is just plain sloppy workmanship/unacceptable in my humble opinion. 

I find it just as important, if not more important than the motion style, that all motion devices on a simulator be in sync to an error of within a few milliseconds or less. Software lag aside, systems that move a large mass freqently have mechanical lag to go with it, or they have steep gearing that limits their top speed which makes them feel awkward when combined with a device that doesn't have these speed limitations. As an example, most (if not all) D-Box actuators as said above to have a 100mm/sec max speed, but the Stage Series actuators have a max speed of 400mm/sec. 


Bottom line, I beleive we can provide tuning options to to address most of these concerns and in fact we already do for several of them. The jury is still out on value proposition and whether or not there is method of achieveing the D-Box Pitch/Roll/Heave at a lower cost and with less compromnise when combined with other devices. The jury is also still out on whether or not a D-Box configured to do only heave so as to add what it can without adding anyting negative to the Stage Seres equation is cost justified. Those aren't the sort of things we make a final ruling on. I see this question as a customer preference type of thing. Our focus is on giving you options regardless of our opinion on what is or is not cost justified. 

Share this post


Link to post
Share on other sites

A subject near and dear to me.   I have Dbox and considered SimX at the time.  I currently have Dbox and GS-4 working together, with Accuforce wheel, and the systems do not "compete" with each other (yes, it's two programs to turn on/off, but not interfering).  Works very well imo. However, the Dbox tilt does likely add to the weight/load on the GS-4, which resists the momentum created by Dbox movement.  

Oh, and I think "lag" with Dbox is a myth.  Very powerful and (to my perception) immediate feedback. 

Share this post


Link to post
Share on other sites

Happy to hear everyone's thoughts and interesting that Berney is thinking of new ideas.

Perhaps I worded my thought poorly, because I see a few of the comments are wondering why I'd want to combine the two systems or that the heft would make the DBOX work harder.  That wasn't my idea.  

My idea was to utilize the seat-mover principle with a DBOX system.  To explain further, in a seat-mover, the seat moves (hey) independent of the wheel, thus creating the core-tightening forces Berney described.  The seat is being moved by two (not including my rear traction loss) SCN5 actuators.  What I am saying is that the entire rig, except the steering wheel, moved via DBOX.  There would be no SimX setup.  Imagine any DBOX style rig, say the Sim Labs P1, but instead of the wheel deck being mounted to and moving with the rig, it was suspended on standalone feet with a DBOX rig under it.  This way, the entire DBOX rig would move but the wheel would stay put, creating the seat-mover effects, but by using DBOX actuators instead of SCN5. 

Does that make more sense?

Share this post


Link to post
Share on other sites

Beef36 - when I tried DBOX it was many years ago, so I'd love to try a more modern setup...but I actually preferred the seat mover feeling to the DBOX feeling.  With DBOX I felt like I was on a fun suspension ride with no consequences.  With the seat mover, my whole body tensed as I felt like my actions had consequences.  Thats the best way I can contrast the two.  The reason for my idea was that I think the DBOX actuators are superior to the SCN5 actuators, mainly in size and with how quiet they are.  This is assuming they both have lag that is imperceptible.  I was thinking we could take the feeling I like (seat-mover) with the superior equipment (DBOX) to have the best of both worlds, but I see your point that they'd probably have to recalibrate how they deliver motion queues.

 

Lasanga Smoothie- Great name.  I can see what you mean in a way (even though he is not moving independent of the wheel which is paramount in a seat-mover), but that motion looks awful.  Its way too dramatic, poor guy in the vid looks like he wants to vomit a lasagna smoothie everywhere!  Lol  

 

estranged_coma - this seems to be a very popular setup.  I've yet to try a G seat, but the new GS5 Berney has been showing off looks slick.  We're really getting up there in cost with a GS5 + DBOX though, I sure hope its good.  I don't like iRacing, so I don't have an account which means I can't read your post on that forum.  Would you mind pasting it?

Share this post


Link to post
Share on other sites

Very interesting thread. I currently actually play with the idea of changing from a 3dof seat mover to either 6dof hexapod or a 3dof setup using 4 of the Prosimu PRS200 actuators.

Especially Berney's explanations make me think...again...whether it's worth it at all. There's Barry Rowland's review on his d-box setup, but, unfortunately, as much as I like and appreciate his videos, he's leaving some very important bits out. It's been a while since I watched it, but I remember his statements about rear traction loss are let's call it lame for the lack of a better word.

On the other hand, nowadays with dd wheels most of the cues in regard to what the rear end of the car is doing comes through my wheel anyway.

In regard to the question in the original post I think the d-box actuators would be too limited in travel in order to make this work...unless you position them very close to the centre...then (depending on your rig design) you get maybe too much leverage/load and that may slow down your system. I don't think that would work. However, you could just move the seat instead using the tri-motion concept. Then you deal with lateral force on your actuators...

 

Share this post


Link to post
Share on other sites

(pasted from my iRacing thread)

I’m the lucky owner of two rigs. The first is a GT Rig which is equipped with the ECCI 7000, CST F1 pedals, Butt Kicker 2, Fanatec CSL base with custom mounting for UltraForce GS-4 driven by Sim Commander 3. It is completed with 3x27” screens. 

My newest one is a Formula Rig which is a custom designed MDF cockpit that simulates an open wheeler seating position, Fanatec CSR Elite wheel & pedals, 2x- (2250i) D-Box actuators and completed with a SFOV screen from a projector. 

I’d like to focus on comparing my GS-4 and the 2x D-Box and in using them in the iRacing world. Over the past 2 years I’ve been lucky enough to also try other motion setups including DIYs and 4x actuator d-box systems. I will (try my best) to refrain from saying ‘this is better than that’ but rather comment on what I think which motion system is more suitable for what type of experience and user preference. In the end I will say what I prefer and why. 

I’ll structure my writing by going over questions that many of my friends and fellow sim-racers have asked me as I think it’s the best way to describe the comparisons. I won’t go into technical details of the GS-4 and the D-Box as those info can be found plentiful on the internet. Lastly, please keep in mind that this topic is a very subjective matter, what one person calls beautiful the other could describe as ugly 3b63d1616c5dfcf29f8a7a031aaa7cad.gif 

Before I begin it might be helpful to describe myself. I’ve been sim-racing since the late ’90s and have tried most mainstream race simulations. In real life, I have competed in amateur kart races in various tracks across Asia. I have also experienced sitting in race-ready street cars driven by professional drivers on circuit tracks and off road.  

Then let’s get onto the questions: 

How does my UltraForce GS-4 compare to my 2x D-Box system? 
I describe the GS-4 as a system that simulates how my body feels when being in a car and the D-Box as something that simulates how the race car reacts in its interaction with the road environment. In the GS-4 when going around corners the back and bottom panels pushes my body in a sustained manner that I feel the G-force present throughout the entire corner. Same thing during acceleration & braking and other forces. The D-Box on the other hand gives a very clear feeling of how your car’s suspension moves, how the tires rub against the track surface, how the car chassis rolls and bounces during over and understeers. When the gear changes I feel a thump, the RPM rumble of the car vibrates my rig impressively, and yes when I go off track and hit a wall I hold on for dear life. I’ve coupled my GS-4 with BK2 so that I feel additional effects of the car but the D-Box has this excellently built-in. 

Where does the GS-4 outshine the D-Box and vice versa? 
The G-forces that I feel on my body is where the GS-4 shines. I do not get this with the D-Box. On the GS-4 I also still feel the bumpiness of the track surface and the effect of the gear changes, etc. but where it lacks the most is in feeling connected to the 4 tires of the car. This is where the D-Box shines. The D-Box lacks providing me with the sustained g-forces so when going through corners, for example, I simply feel my body tilting in one direction. Recalling back when I was still using my fanatec wheel on my GS-4 GT rig and comparing it with my setup now where the GS-4 is coupled with the ECCI 7000 I have to say that I’ve had more of a workout in the GS-4 than in the D-Box. In other words the GS-4 it takes more energy on my body than my D-Box. (I will make a different statement later on when I describe 4x D-Box systems). 

 

If I only had to buy one today, which one will I choose? 
This is a hard one! The GS-4 was my first motion system (my first baby) so I have sentimental feelings about it. If I were rational, and the leading factors in my consideration are price and compactness then I’d still choose the GS-4. The D-Box is expensive, takes a much bigger space and more tricky to mount. But the motion systems is (oh so much) more complete despite the lack of sustained g-force because I can feel exactly what the car is doing. I can even say that the physics of iRacing really comes alive and really makes a lot of sense with the D-Box. It is almost like iRacing was meant to be experienced with a D-Box. Maybe this says a lot about how realistic iRacings physics is, and perhaps it may even be ‘too realistic’ that if you only experience it with a wheel and pedal you are only ‘seeing the tip of the ice-berg’. In hindsight this analysis makes sense when looking at the endless debates about ‘how realistic is iRacing’s physics’. 

How do I compare 2x D-Box and 4x D-Box? 
This is an interesting and highly subjective topic. Let me first state my personal opinion and preference. Taking away the consideration of price, and only focusing on the motion immersion, I prefer the 2x instead of the 4x. Why? Firstly in my opinion the 2x gives enough accurate cues to what the car is doing like what I described above. The 4x gives me ‘too much’ sensation that my brain and body can’t process it fast enough to handle it. Bear in mind not only would I have to process the motion effects, at the same time I have to also manage what the wheel FFB is doing, visual cues from the screens and the audio effects from the speakers. But this is not the failing of the 4x D-Box system but rather the failing of my lack of focus to be a real life racer. The reason I say this is because the 4x is as realistic as it can be, it really is. If you want a complete immersion experience and the ultimate setup and to feel what it really feels like to drive a real race car, then the 4x is the (only) way to go. A 4x D-Box system is the cheapest way to get a real (insert your favourite race car name here) inside your living room! But you also have to have the focus and skills of a real race drive in order to tame the beast that is 4x D-Box. 

Where do both my GS-4 and my D-Box lack? 
To this day I’ve yet to experience a motion system that gives me the sensation of the elevation of a track. Consider tracks like Brands Hatch, Spa, and Suzuka where in certain sections of the track goes steeply uphill like Eau Rouge or downhill like the last corner of Suzuka. Although these systems are advertised as being able to give you these sensations I still maintain that they do not. Even the 4x D-Box system does not do this, or maybe it did but as described above there were so much coming out of the 4x that I probably could not notice it. 

Both also still lacks a clear sensation of 2 other things: traction loss and slip angle loss. I did an experiment where I turned off the FFB of both my wheels and if I were to close my eyes I wouldn’t be able to tell. Maybe it has to do with fine-tuning the software side of the systems but so far I’m led to the conclusion that traction loss and slip angle loss come better simulated through the wheel and the visual cues. 

How loud are they? 
Easy question. D-Box is much quieter. The GS-4 squeaks audibly when the panels move. 

Others points? 
Achieving an equilibrium balance of how strong your wheel is in combination with the motion system and others such as how strong your audio is is paramount. I’ve experienced setups, including my own when it was not yet properly configured, where the motion system overpowers the strength of the wheel (imagine combining a D-Box with a G27 wheel). The key to getting the maximum benefit is to ‘orchestrate’ all the components so that they all play a beautiful symphony. 

The D-Box is heavy duty. It will break your rig if yours is not a rigid heavy duty rig. My formula rig is constructed out of MDF but they are properly secured at the joints with industrial bonding adhesives and reinforced with bolts and nuts. I then connect the rig to the D-Box using 80/20 frames to achieve the desired rigidity. I’ve set the D-Box software to be just slightly above default (around 35-40% strength) but I suspect if I go beyond 65% my rig will simply break. 

I’ve ordered the DK2 Oculus Rift and I’m curious and excited to try it on both of my systems. I wouldn’t be surprised if I arrive at slightly different conclusions when I’m completely inside a virtual world and where I’m not ‘planted’ to fixed references around my rig, i.e. the wall, the door, the windows, etc. 

Which system makes me faster? 
Neither. I'm still equally as slow. But I think the D-Box makes me more consistent. 

What would my ultimate motion setup be? 
If money, space & time weren’t an issue I’d combine a stripped down GS-4 and install only the panels onto a seat then onto a compact no-flex rig (preferably formula-type) and then mounted onto a 2x system. And then all then mounted onto another axis which could simulate the elevation of tracks. 


Well, thanks for reading. Feel free to comment, agree, disagree, or ask other questions. 

 

Share this post


Link to post
Share on other sites

Thirsty - what are you referring to by "tri-motion concept"?

estranged_coma - thanks for sharing the detailed impressions.  I think (Thirsty cover your eyes!) Fast Track Sims is centering their new rig motion theory around a rear 2 actuator DBOX setup with a front pivot point.  I'm interested to learn more about this as I'd imagine this setup with the GS5 would be a killer setup.

Errr...that setup, bolted to a platform with rear-traction loss...that would be a killer setup. 

Share this post


Link to post
Share on other sites
2 hours ago, switchface said:

Thirsty - what are you referring to by "tri-motion concept"?

estranged_coma - thanks for sharing the detailed impressions.  I think (Thirsty cover your eyes!) Fast Track Sims is centering their new rig motion theory around a rear 2 actuator DBOX setup with a front pivot point.  I'm interested to learn more about this as I'd imagine this setup with the GS5 would be a killer setup.

Errr...that setup, bolted to a platform with rear-traction loss...that would be a killer setup. 

Ah, well, I finally got over this Fast Track Sucks rig purchase nightmare ;-)

Have a look at this video showing a tri-motion rig:

It's basically a set mover with heave - just the seat moves, everything else is static.

I like this idea, but there are two things I'd be concerned about. Noise seems to be more of an issue than with conventional seat movers. The other thing is lateral forces being applied to the actuators. That can come from operating your brake pedal if you choose to work with high pedal forces.

Anyway you could combine that with a GS4/5 seat as well. I, for myself, decided against using it, but I think it may provide great immersion.

Share this post


Link to post
Share on other sites

Glad you were able to not let that bother you further (hey, I did say cover your eyes!).  To bring them up again, they have a new P1-R2 rig using 2 DBOX actuators with a front pivot point.  This is interesting and I'd like to learn more about it.  I messaged Mark on one of their Facebook posts about it but he never responded.

This concept above is exactly what I was describing in my original stoney idea!  Hmmmm.  I'm curious why noise would be an issue - the SCN5 are squeaky whereas the DBOX type actuators are usually quieter.  To offset your lateral force actuator, what if we simply put the pedals on the chair platform as well (I guess this is my exact idea from the OP). 

Well, all these ideas are coming at a weird time.  I may be selling my house, which means my beloved rig may have to go :(

Share this post


Link to post
Share on other sites

The problem with this front pivot design is, it, depending on the configuration, reduces not just travel but speed as well. That can be an issue in case you do not use very fast actuators.

You can get full speed out of them, but in this case you sacrifice pitch.

In regard to noise, our 2 SCN5's are sort of audible enough I would think. The 3rd one is not always active. With 3 of them permanently working and carrying load it may be different. It's hard to judge from these videos...only saying it's a concern.

Adding the pedals to the moving part increases weight with quite some leverage. The thing would have to be re-balanced.

Doing this with a 2-actuator d-box system may not give you enough speed. On the other hand, if you position the pivot very close to the seat and actuators, sort of in a triangle shape configuration like it the video above, it may actually work. Having the pivot somewhere at the pedals may not be good.

Sorry to hear you may have to sell your rig...would be a pity. Is it because of its size?

Again, I'm considering trying out something else as well...the bloody rig builder in me... ;-) . I tried a pretty expensive 6dof hexapod platform, which I had in mind changing to. That was a disappointment though. So I'm thinking too much downtime & money...for what? In the end these 3dof seat movers controlled by Simx's Simcommander are damn good...I love it :-)

Edited by Thirsty

Share this post


Link to post
Share on other sites

Yeah, I finally was able to watch that video with sound and the actuators are definitely loud there.  A lot of these newer motion platforms seem to use SCN5s in a setup that looks like a DBOX setup (similar to the one you just posted).  I love my current setup, but wouldn't want it any louder.

I'd imagine Mark/FTS has the pivot in the ideal spot, otherwise he wouldn't do it.  Its no like he hasn't build/designed/sold 4 actuator setups before, so their recommendation to save money and go with 2 has me at least intrigued to learn more.

Would be selling the rig only because of the regular added complications of a home purchase/move.  Its a high dollar kit that weighs a lot and takes up a good deal of space.  It would just be easier to not have to deal with all of that when I'd have a lot more on my plate.  I think I have about a dozen emails of people that have asked to purchase it over the years, but I have a local friend thats been wanting to get into sim racing and hasn't pulled the trigger yet - I'll give him a local/friend discount to pass along the fun.

6Dof Hexapod?!?  Wow, it would be something special, thats for sure.

Share this post


Link to post
Share on other sites

There are some new options that could potentially compete with a D-box setup to a large extent, and at a significantly lower cost while offering more travel and adjustability. I think the combination of a D-box type setup combined with the GS-5 seat would be really excellent for my use. I love the heave motion (although, 1.5" is not enough for me) and require a quiet setup for racing at night when my family is sleeping. I also need the rig to be compact and the D-box works great for that. Sustained G-forces are what is missing for me so I really think a G-seat is the last piece of the puzzle I'll need.

Share this post


Link to post
Share on other sites

I read in part of a thread that someone played on a dbox and gs4 and felt no sensation - motion of going up and down hills and camber (Brands Hatch).

interested as granted I have only lowly 2 DOF DC motor (seat mover )and DIy actuator rigs(full balanced frame 2 DOF) , however,  I can feel the hills and camber very well. 

Maybe its what the motion software puts out in forces available.

I run SImtools and pitch and roll are almost in all race game plugins. 

Even the games that only have sway and surge (flatout2) as example, theres still the tilting forces of pitch and roll.

Maybe adjust the motion settings on the dbox if it has it

 

 

Share this post


Link to post
Share on other sites

 

On 10/16/2018 at 4:22 PM, motiondave said:

 

Chassis Mover

I agree. I tried Dodge simulator (4x d-box) at 2017 Chicago Auto Show. Not that impressive at all. The iRacing guy used 2x d-box posted above seems make a sense, as described he is pretty happy with. This is consistent with others conclusion. In the comment section of SRG d-box review video https://www.youtube.com/watch?v=bU7TJG8PUSk,

anasmsk11 who owns a racing center having both 2x and 4x setups said he and all others tried both found the 2x is far better and more realistic feel than 4x.  Thus I dont think that 4x is more realistic than 3x for 3dof motion. Most of people are trapped by a myth of believing a 4x d-box style is better for 3DoF. I updated this post to describe in detail to show you that using 4 actuators on 4 corners drive for 3DoF motion design actually is wrong. Also, I share my thoughts on this subject per my experiences in the past years on different types of racing simulators.

People are initially overwhelmed by 4x d-box style probably not because of the 4x configuration instead likely due the fact that who might have never tried linear actuators rig before.  Like those who had tried both of 2x and 4x actually prefer the 2x config. This is due to the fact that both 2x and 3x can do combination of roll and pitch movements. A simple common knowledge tells a truth about this. A turntable record player designed with 3 legs to move in 3dof. It would be difficult to level the player if designed with 4 legs. Often you end up with one leg lift up. This issue can be clearly demonstrated by comparing 4x vs 3x of 3DoF setups, see below two videos: 

vs

In the first video, 4x case (SFX-100), at relatively large movements (say high speed cornering) you can see a leg lift up. That not just means one leg was being too short at given a moment of time, but also means other leg(s) could be moving too far (or short), which cannot produce 3dof motion. It looks OK to roll one side to other, or pitch back and forth solely. But if you want to do roll, pitch, and heave together at same time (combination), then you will see the problem. Other example: https://www.youtube.com/watch?v=EElowvFEpZQ. Also: you see all 4 drivers are restrict/constrained each other cannot freely/independently move. Even with very shorter spacing between rear and front actuators to enhance pitch movement, like this D-box setup: https://www.youtube.com/watch?v=dlRjOJt7Nq0  There is no evidence to show it can do a combination movements (roll/pitch at same time).

In the 2nd video, 3x case, no such problem at all. You can see smooth combination movement (roll/pitch/heave). Thus, such 4x configuration just cannot produce a realistic chassis movement at high speed cornering. Note: the SimTool SW does NOT support 4x for 3DoF either (it supports 3x for 3DoF). In fact there is no such motion SW in the market can support 4x 3DoF. For using Prosimu or PT actuators to configure such 4x D-Box style, you will see the same problem as shown above video. Below I describe exactly why.

SimTool allows you to assign any dof to any axis with arbitrary weight (just like other motion SWs), as long as you configure rig as n drivers/axises for nDOF (2x for 2dof, 3x for 3dof, 4x for 4dof, 5x for 5dof and 6x for 6dof —topological Stewart platform), in which the axises are totally free and independent each other so they can freely move. As an example to show the 4x config will not work, let’s fix 3 actuators’ positions at zero. In other words, send zero to them. Then try to move the 4th actuator in any of amount x while keeping other three at zero. You will fail on this trial. You just cannot do it with SimTool. Sending signal x1=0, x2= 0, x3=0, and x4= non-zero, then 4 actuators will struggle to balance hence introduce false movement, of which the vertical component motion so-called "Heave" is particularly sensitive to human. Mathematically speaking, given arbitrary 4 points, you cannot form a plane. The 4 points have to be constrained by an equation in order to form a plane (https://www.youtube.com/watch?v=t4Dk39N8_mM). Otherwise one of four points will be out of the plane. This is why we see one of the legs lift up in the first video above. Can one make a motion SW to implement 4x for 3DoF? Mathematically possible, but just not effective and also against principle of motion platform design. Basically, given in any moment of time, upon receiving a data point in 3DoF (roll angle, pitch angle and Z distance), the SW needs to find (calculate) two constrained planes. One is on the ground in which all 4 points of actuators’ legs need to line in the plane. Other plane is on the chassis so that requires all 4 points of actuators to line in. Such calculations carried out in real time and the 4 actuators need to move collectively in according with the inputs and two planes found in real time. 

Now, for 3x case, you would not have this problem. This is well-known fact. Given arbitrary 3 points, you can always form a plane. This is why you can freely move any one of thee actuators and independently regardless others. You can send arbitrary values x1,x2, and x3 to the actuators, they will accurately reach there (pitch/roll/heave positions).  For 3dof motion, only need 3 actuators (motors): 3x. Adding extra one will cause actuators struggle for 3dof, and mess up the 3dof motion. Similarly, using 7 actuators will mess up 6dof motion. Later, I show you the video to demonstrate the how wrong of 4x for 3dof motion can be.

Even with small movement of 4X d-box style, say with zero heave input, it will produce false heave movement. See below link of the video, since 4x cannot produce combination of roll and pitch at same time, it will struggle to balance and hence introduce vertical (heave movement) motion that is not supposed to be there. Such false heave movement is very sensitive to human that may lead one feels like "floating":

https://www.youtube.com/watch?v=XXz8Do6zT0g&t=14s (check all the 4x D-box style vides, all the same like: https://www.youtube.com/watch?v=Zkfj5v_r89E&t=12s)

Usually, chassis height is very stable, you don't feel such "floating" effect (because of suspensions, shock and absorbs), see real life:

Https://youtu.be/Lf67POsSSi4  and  https://www.youtube.com/watch?v=DnGdA6qjr1c

Unlike a boat which the water wave can heave the whole boat up down, when a car hits bump or pothole it hits front tires first, then rear tires. So chassis does not heave, but pitch. Chassis is relatively stable due to suspension and absorber and shocks. Watch the video below carefully you will fully understand why you shall never use suspension telemetry data to feed your actuators (always use chassis motion telemetry data apply to your actuators), and why chassis does not heave mostly:

The suspension can drop/bounce considerably but the chassis remains relatively stable, even for classical race cars: https://www.youtube.com/watch?v=ThsG0Byn9ko Using suspension reflect telemetry data for your actuators is wrong because the seat, pedals, and steering wheel all are mounted on chassis. Using chassis data makes sense. The chassis pitches and rolls mainly (front tires hit bump first) within very fine limit range: https://www.youtube.com/watch?v=VvkTpYH9AUI.  All 4x d-box style rigs found in youtube show such floating behavior. In particular cases, the whole chassis may heave if hits very large bump or valley, of which the size is comparable to the distance between front tires and rear tires, like rally car. Thus, one shall use 3x for 3dof instead of 4x for 3dof motion rig. It could make day and night difference in terms of feel. Watch carefully 3x vs 4x with same actuators (PT actuators with SimTool): 

vs 

Watch the 1st (3x case) one very carefully to see how the dashboard (on the screen) is exactly sync with the rig movement! This shows how good and precisely the 3x can reproduce chassis 3DoF motion in according with the sim motion data!!

Now, watch the 4x case video carefully. The floating feel and incoherent feel of roll/pitch (due to lack of roll/pitch combination movement) of 4x config just kill the sim racing realism. Seeing is believing! Same actuators with same motion SW, but day and night diff between 3x and 4x configuration as you can see!

"Heave" is a vertical distance movement. In the chassis mover, it is Ride Height (distance between track surface and chassis bottom). Racing car, the heave or ride height varies very little, as shown above videos. Below, I show the front/rear ride heights from rF2 Corvette GT car on Silverston I ran. As you can see the front and rear ride heights move opposite thus the chassis is pitch back and forth as supposed to do. The max pitch angle is < +/- 2 degree:

896910280_ScreenShot2019-08-18at12_04_21PM.thumb.png.7e25b7884b922a422885f2e7e7da5d6f.png

The bottom chart is the vertical-g value, which is vertical impact force (not vertical distance movement). However, motion SW, like the SimTool, actually uses the Vertical-g value as input for Heave. This is the problem! User would not feel road surface texture and micro bumps without rideHeight data being applied. On race track, we will have huge vertical-g (vertical impact force), but not necessary large heave (vertical distance movement) though. Apply vertical-g telemetry data to actuators (heave) vertical movement will give false feel of floating. Correct implementation shall apply vertical-g to the actuators' stroke acceleration instead of stroke distance, and apply ride height data to actuators' stroke distance: heave. This is why people who tried both 4x and 2x actually are prefer 2x (no heave dof at all). This is main reason exactly as anasmsk11  said on 4x vs 2x in the comment section of SRG review video: "feels like its floating in the air, and theres too much going on..."; "The 2250i system feels much more composed, predictable and realistic. Might be hard to believe, but it truly is much better"; "With the 2250i, there are only 2 actuators, the cockpit feels much more solid, predictable, and more precise in the movement, you can tell exactly whats going on so it relates much better to the visual. Overall, the amount of motion you get is no less than the 4250i at all, you get plenty of motion and vibration from 2 actuators, just much more precise, which as a result, is more enjoyable and feels more realistic." To me for Frame Mover, as a home motion simulator, to simulator chassis movement is a goal. But racing car chassis movement limit in very narrow ranges. The chassis roll and pitch angles limit < 5 and <2 degree, respectively.

793476921_ScreenShot2019-08-17at5_42_18AM.png.37d04e45942c4712f15dca380fd00ebe.png

To move the chassis (whole fame) beyond such limits, you would not feel realistic. If you want to simulate motion cue for g-effect, as for a home motion simulator, you are better to add something else instead of increasing chassis motion intensity, such as seat mover or g-seat, etc.

4x drivers for 3DoF is very very difficult to implement although mathematically maybe possible.  Practically speaking, it is not possible. In other words, it is basically not feasible to use 4x drivers to do 3DoF having a combination movement of roll and pitch. Searching through all published papers and  robotics & motion platform industries offered, I just could not find a single example of 4x for 3DoF. Today, all motion platforms designed as n x drivers for nDoF config. I never have seen such n+1 drivers for n DOF, except from early suggestion by d-box users through internet. How such scheme of using 4 actuators on 4 corners for 3DoF motion rig could widely spread through internet today is beyond me. Such wrongful design just cannot be effective. I found only the most expensive sim NADS/Lexus/Toyota uses 4 actuators on the corners, but their purpose is for vibrations only, not for motion. The max vibration amplitude limits under < 0.2 inch though:

NADS and Lexus motion simulator

The frequency cut off at 20Hz, that says any possible chassis vibration above 20Hz will not be reproduced. I looked at the rF2 chassis vertical movement data again: front left and front right ride height data. The cut-off makes a sense to me. Below are the results for front left and right chassis rightHeights spectrum analysis I did with rF2 GT car data on Silverston track:

Front-Height-fourier-analysis-rF2-GT.thumb.JPG.2e02877f5f06498a7914304655a0f6ab.JPG

As you can see there is not much above 20Hz. You don't need an actuator to be capable of vibrating at very high Freq. If you want to place 4 actuators on 4 corners for vibration purpose, there is no need to go beyond 20Hz. 

Note: In real life, cockpit can have its own resonance frequencies due to additional cabinet structure resonance and engine vibrations, etc. Racing sim SW (rFf2, AC, iRacing,etc) do not model cockpit/cabinet resonances. However, Motion SW such as SimXperience and SimTool also offer tactile SW so-called “SimVibe” and “GameVibe”, respectively, that use suspension reflection and engine rpm, etc telemetry data to mix output as sound wave to feed your computer sound cards. Such audio signals can be amplified to apply to your transducers, with which you can place wherever you like on your rig. This is purely subjective though. Since Motion SW like SimTool does not offer vertical chassis vibrations, it is good idea actually add transducers to gain some vibration and micro bumps feel. But you really dont' want to over-play with them (https://www.youtube.com/watch?v=54d39emV0AY).

As 2x for 2dof although not 3dof, but would be correct for combination of roll/pitch motion, probably better though (no heave dof, so I would not feel false floating). I took anasmsk11's words and  build mine with 2x with SCN6 4-inch actuators and add a seat mover on the top.  One most important aspect of sim racing motion rig is to offer a "correct" motion cue on high speed cornering. This is all we go after. No body wants to spend $$$ to build a sim for daily driving experience at 50 mph. I drive bumpy road at 35 mph to work every day. I dont need a sim to experience that. We all want exciting racing experience, thus coherent feel on high speed cornering is most important aspect on racing sim. While cornering at high speed, due to weight transferring, the car will roll and pitch at same time. 2x conf for 2dof will give correct movement and hence offer more realistic sensation feel of car and tire dynamics. With 2x, You would not get confusion and struggle feel. But 4x, as the iRacing guy pointed above "The 4x gives me ‘too much’ sensation that my brain and body can’t process it fast enough to handle it"  . He also said:  “I’veset the D-Box software to be just slightlyabove default (around 35-40% strength)but I suspect if I go beyond 65% my rig will simply break.” Why? If it moves rig in true 3DoF (degree of freedom), why will break the rig? It seems to me that implies the 4x struggle.  Again as anasmsk11 said in the comment section of the SRG d-box review video: "...myself and everyone who tried both, all said the 2250i felt much better. You really need to try it. The 2250i system feels much more composed, predictable and realistic. Might be hard to believe, but it truly is much better. ....only 2 actuators, the cockpit feels much more solid, predictable, and more precise in the movement, you can tell exactly whats going on so it relates much better to the visual......"

People often has misconception thought because car has 4 wheels then the 4x must be better than 3x in terms of 3DOF. No, it is not. Your rig is a rigid body and you still get 3dof (roll/pitch/heave), nothing more, but in a struggle way (all 4 drivers are restrict/constrained each other cannot freely/independent move),  Such fashion of 4x actuators struggle for 3DoF could yield a fuzzy feel leading one to believe “better” or more “realistic” psychologically. If you have a real car dashboard build on top of such 4x style rig and with large curved projector screen in front with true 1:1 life FOV, then you will easily notice the problem. I built mine on 2x configuration: 

I feel chassis movement is coherent and predictable, and feel dynamics of rFactor 2 cars pretty realistic.
The correct  implementation of 3DoF motion sim, of course, has to be using 3 drivers: 

3x for 3DoF &nbsp;D-box setup

To me it is worst to think of putting 4 actuators on 4 corners. If I turn up the motion intensity I  will get floating feel and cannot feel roll/pitch coherently while high speed cornering. If I turn down the motion all the way, then I got tactile effect only. To whom already had 4x configured on the rig with powerful enough actuators, my recommendation is to change to 3x configuration and add 2 transducers on 2 corners with tactile SW if needed. This is far better, you will feel correct dynamic behavior of car and tires while cornering at high speed due to weight transfer. 

Recently, Ford made a GT simulator using 3x for 3DoF, with two motored rods in front and one motored rod on rear:

also: https://youtu.be/ftes99dVEdY

43036934_ScreenShot2019-09-10at7_09_30PM.png.0502661cc53f20d40dba82af09d0dae2.png

Because use rod joint end, the roll and pitch are free in movement that yields precisely movement with high resolution. (Unlike the D-box and it’s style such as Prosimu Prs200, SFX-100, and PT-Actuator solutions they all fix the actuators on the frame).

Under, there is a front rotation track for traction loss and two sliding rails for surge and sway, respectively, see video: Sway, yaw, and surge motionIt is full 6DoF chassis motion simulator.

Traction Loss

The Ford sim system uses 25 computers. It is believed that it probably implemented true or quasi real-time calculations. So the chassis sliding simulated with yaw and sway combination with proper motion  cue algorithm could be good enough

Per my experience, this is not the case for sims we use such as rF2, AC, and iRacing, which are not real-time calculation for tire model. They all use pre-calculated cached table-lookup methods. Adding yaw and sway motion or a dedicated TL actuator for chassis trying to simulate sliding or drift state of the car is not worth IMO.  One of the main reasons is that the physics of a sliding/non-grip state in tire model is not known and not understood well, see this video: (Non-grip state (sliding state).

A sliding state is a state that tires lost grip and slide. Often occurs when the slip angle is larger than a max slip angle while over steering. Where the max slip angle is the tire has max grip or friction coefficient. At max slip angle one shall feel max lateral force or FFB from the steering wheel (Slip angle curve , also see demoSlip angle vs tire grip level). A controlled sliding state is so-called “a drift state”. At racing, to archive fast lap time, one has to avoid sliding. The tires can slip but not necessary slide though. Commonly, as slip angle is larger than max slip angle the car will lost grip to slide. There is no way from physics point of view to proof that if a sim SW can produce a realistic sliding state (how and when to start to slide), as said in the above video: “It is a black art”.  With limit stroke distance of actuator, one needs a good Acceleration OnSet cueing scheme. You can swing the chassis left and right, but by no means of realistic. I tried both SimTool and SimXperiene Traction lost for rF2 and AC, and I gave up. I see some YouTube videos of DIY TL and the NLR traction plus system showing chassis swinging even in neutral turning at corners. This is totally unrealistic! Unlike plane, car chassis hardly yaw, swing or sway unless at a point when car starts sliding. Below illustrates 6DoF motion of car:

2109513088_6dofmotion.jpg.81e5e9c4e16a6c14d41850adfe443084.jpg

The bottom three motions (sway, heave and surge) are linear motion of chassis (moving in space)! But motion SWs like the SimTool and others actually applied three accelerations (lateral g, vertical g and longitude g) to spatial coordinates, respectively, to create motions. While cornering at high speed, you will have a lateral g force upon you. But the car will hardly sway (lateral motion) though, because the tiers’ grip force is balance out the centripetal force. Below is the Chart of Sway velocity vs Lateral-g I ran on rF2  Corvette GT1 car on Silverston:

671212142_Sway-velocityvsLateral-g.thumb.JPG.2c82fcb7cd2859c16473c8f5f548adf6.JPG

See car moves little in lateral direction! Be careful when add a sway or/and TL motion on your rig, since Motion SWs Like SimTool etc actually use Lateral-g telemetry data (centripetal g instead of net lateral g) applied to your actuator for sway motion. So your rig sway motion just cannot be realistic! It needs very cautious to use chassis sway and surge trying to simulate g-effect on driver's body. It probably will never work if try to diy at home (See: this video).  I see Lexus simulator use very long sway and surge rails, which are longer than cross of whole football field, in order to have enough range for speed, acceleration and de-acceleration, and smooth reverse with Washout algorithm. At home, other options like seat mover and g-seat approaches are far more effective than sway/surge rails, which are not to generate acceleration/de-acceleration as for g-force, rather "g-effect".  I will explain later in this post.

With current sims such as AC, rF2 and iRacing, it just not possible to yield a realistic feel with traction loss and sway motion. It is easy to find this out yourself by comparison between drifting  a real car and a sim car with TL. Learn how to drift a car in 3 easy steps, see: Chris Harris, how to drift .

In our brain norm, the sway or yaw or swing motion on chassis is anomaly, so such motion would be very sensitive to inner ear, in which the semicircular ducts are responsible for detecting rotational movement. If chassis is purely sliding without rotational component (all 4 tires sliding exactly same), then such motion cannot be sensed by our inner ear. Just like our inner ear cannot detect constant linear motion. As all our rigs being lack of offering speed sensation (constant speed sensation is from visual and sound not from motion), wrong swing motion could easily override “true motion cue”. So even with very tiny wrong swing or yaw or sway motion on rig will kill a feel of realism. When rear tiers start to slide, the TL is supposed to move. Then right after, the TL moves back in order to center. This is a huge problem as your inner ear senses reversal motion. An advanced adaptive Washout algorithm for OnSet Cueing is needed to slow down the reverse. It needs to be slow just enough so you don’t feel motion but fast enough to return the center in time. I rather not have TL if cannot get it right. Thus I turn my TL motion way down trying to "ensure" the TL moves only while over-steering. Otherwise it would be more harmful than benefiting in most of time, as demonstrated and described in the video: Demo of Traction Loss.

Seat Mover

On motion cue for g-effect, many think of increasing the chassis 2dof motion (roll/pitch) to simulate g-effect. To me this is not good idea since real car chassis motion is very limited (< 5 and < 2 degree, respectively, for roll and pitch).  Excessive motion of whole chassis causes totally unrealistic feel and little motion cue for g-effect on body.

A term of "g-force" is often misunderstood in this regard. A driver, in a car which is accelerating or de-accelerating or changing direction of velocity (cornering), is in a non-inertial frame of reference who will observe forces accordingly. The reason people often call these forces is "g-force" because of Equivalence Principle. How can we simulate such forces? Naturally, we can use gravitation force due to Equivalence Principle. Such method requires large degree of roll and pitch. The max sustainable force is only 1g if roll or pitch in 90 degree. The whole rig needs to place in a platform such as a large 6DoF platform. Plus, rig itself needs to be simulated with vibration and road bump and texture too. For instant/burst acceleration or de-acceleration, long surge and sway rails can be added, see: NADS and Lexus simulator.  Ofc, this is not for home use.

A simple Seat mover solution is not to simulate g-force. Rather, it is to simulate g-force effects!  I added Seat mover (Simxperice SCN5 kit) on top of my 2x chassis mover, and found this scheme is pretty effective. Most people often made a mistake, IMO,  to use dc gear motor solutions for seat mover. The dc gear motor solutions often suffer two shortcomes: nonlinear and backlash. Even with very little backlash, ~0.5mm you will loss fidelity. This is why all earlier sim makers such as SimXperience, CXC, Flex, etc, never used dc gear motors, they all use SCN actuators. SCN used stepper motors that have higher torque at zero speed, hence very fast reverse with little or almost no backlash. 

Seat mover is to simulate a motion cue and effective neuromuscular tensions on the driver's arms, body and neck while cornering at high speed or braking. The muscle memory works just as if racing a real car. Such g-effect is simulated by body roll and body, arms, neck muscles tension and subconscious mind due to dynamically change distance between your body and steering wheel, and to fixed screen, so that your body will fight to balance subconsciously (keep in center--- screen center) while cornering just as if drive a real race car. On straight, the seat mover also yields pretty good effect of jerk force on body while braking or changing gear (jerk force:  Jerk physics).

See the video below Seat mover vs real life:

Seat mover looks unrealistic but the effects are pretty "real" to me (https://www.sciencedirect.com/science/article/pii/S0268003315001680). A real pro racing car driver at high speed cornering, his body neuromuscular strength and heart beat rate can reach to very high level (subconsciously), just like any other pro sports men: published paper. (note: the motion set very high in the video for demo purpose).

Driving seat mover rig for 20+ laps you probably will feel such stresses on your body and heart. Seat mover is not for simulating daily driving experience. It is for high speed racing experience with similar body neuromuscular stress effects. One of reasons that people discount seat mover I can think of is the fact that today people use direct drive (DD) steering wheel with very strong FFB, of which resultant muscles tension on arm and wrist could overwhelm the effects caused by seat mover. Other argue seat mover looks unrealistic (real car seat never moves). To me this argument does not hold. We are talking motion simulator and it is not supposed to look like a real. Because of motion, we then have cueing. There is no such simulator that feels and looks real, except for a real car. In above video, the motion of seat was tuned too high, that is for demonstration purpose though. Keeping in mind that racing is just like other sports. Driver shall not consciously pay a particular attention on seat motion, which is part of whole motion cueing system on rig.  Just like a real racing car driver who never consciously pay attention on a particular feel, such as FFB, G-force, etc. , which all are subconsciously built in muscle memory.  In practical, as I had experimented on seat mover, you need to turn down the motion intensity (roll/pitch/sway/surge) and increase  washout filter gradually, until at a point you just seem not feel seat motion if you pay attention to. Then add small amount of heave (vertical-g). The results combined with chassis mover can be phenomenal!

G-seat

Other option is the g-seat, which can be pneumatic (Lateral-g effects, and demo) or mechanic. Technically speaking, a g-seat is not a motion sim, rather a “Tactile” device, which does not move driver and the rig.  Hence it will not affect inner ear sensory. In this regard, I agree that g-seat is better than seat mover. It creates push forces upon the body to simulate a cue as reaction forces from seat due to g-force (Newton’s 3rd law). Especially with pneumatic g-seat for F1 style, the driver confined within a cabinet cannot mover, some professional racers highly value a g-seat: F1 g-seat. For GT style, combined both  g-seat and seat mover would yield the best result of g-force cueing. On seat mover, driver's body unconsciously intends to lean on side while cornering, thus the feel of pressure from side air bladders is far more nature, similarly for braking. 

There are other solutions for “g-seat”, for an example, a hybrid method: by moving bottom of the seat only to create side and  back pressures as cornering and accelerating, respectively: Seat bottom mover g-seat .

A lessen I learned is there is always disappointment no matter what. In the end, it is not possible to simulate a race car truly realistically at home. 

Edited by Joe Extraknow

Share this post


Link to post
Share on other sites
On 8/13/2018 at 7:51 PM, switchface said:

I had a random idea last night and figured to ask if anyone had experience or an opinion one way or another.

I know people tend to fall into camps: seat mover fans and DBOX fans.  I have a seat mover (stage 4+ SimXperience) but I always keep an open mind to what is out there.  I'll admit to being intrigued by how quiet and compact a DBOX setup can be, but I'm very happy with the way my rig forces me to tighten my core (i.e. basic seat mover principles).

My question is this, could we not combine the two?  Why not have a DBOX rig, but have the wheel deck supported stationary above the moving rig, essentially creating the same type of setup as a seat mover, but using the DBOX actuators in place of the SCN5 arrangement?

Good idea?  Terrible idea?  Doesn't work because ____?  Any thoughts?

https://www.youtube.com/watch?v=yktALbmZZsQ&t=45s

Made an account just to show you someone did it. :o)

Share this post


Link to post
Share on other sites

 

On 7/4/2019 at 8:47 AM, Jaco said:

https://www.youtube.com/watch?v=yktALbmZZsQ&t=45s

Made an account just to show you someone did it. ?

 

Super cool of you Jaco to start an account just to post this, so thanks.

 

On 7/4/2019 at 3:48 PM, Bushchook said:

Here is a SimX one too. He says it works well.

 

 

Steve Bernstein Combo.jpg

 

Both of these examples you've each posted do accomplish the concept of combining the two.  My concern here is that you now have multiple/different motion software running at the same time, which I'm sure that adds to complexity or latency timing issues (or maybe not?).  The 2nd example with the entire SimX rig on actuators...I can't even imagine how its doing that as I know how heavy my rig is (granted I have more components, but still).

I attended Rennsport Reunion late last year and was able to try the Porsche RSR simulator (where the rig is the actual RSR!).  It was using a similar concept to my idea, in that the race seat was on DBOX but the wheel/pedals were stationary.  Funny, but not funny story, I mistakenly depressed the clutch (thinking it was the brake) heading into turn 1 at Laguna at 130+mph, and beached the car.  No top time for me!

Share this post


Link to post
Share on other sites

Very very interesting thread and has really got me thinking about 3 acutator, not acutator for 3dof.

In terms of the 4 acutator, leg up issue - has any tried 4 acutators but instead of the corners of the rig rectangle, positioning the 4 acutators on the respective axis?

ie just like a 3 - so one acutator each side - but unlike a 3 where you place one at the front or rear, you position one at the front and rear.

So unlike the common practice of positioning the acutators as "wheels" - they are more on the respective planes of movement you are seeking to reproduce?

Hmmmmmm??????  Thoughts?

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
1 1