High end motion simulators are really freaking expensive. Like not a little bit expensive but insanely expensive.
See this engine bay?
*That* was expensive. It had lots of custom work done to fit things where they shouldn’t fit. That *should* be expensive. It was a bit of an adventure into the unknown. The invoices still put me into a cold sweat.
Now let’s turn around to the other toy in the garage – I wanted to add motion to it. Let’s look at the top end of the market.
|Brand||Type (Actuator)||Vertical Acceleration||Vertical Movement||Cost (not including shipping, customs, GST ie. add a good 25%)|
|A||Seat Mover (SCN-6)||200mm/sec||?||$70’000 NZD Complete rig – just in here for the craziness aspect.|
|B||Platform Mover (DBox)||100mm/sec||100mm||$50’000 NZD for the actuator upgrade|
|C||Platform Mover (Inhouse design)||280mm/sec||100mm||$15’000 NZD for the actuators|
Okay I take my toys seriously. But that takes the cake. That’s really in the territory of “you’ve got a Veyron in the garage and have some space and can’t think of anything else to spend your money on”. Okay maybe Brand C might be viable, but even then…
If I were running one of those companies right now I’d almost be worried. Almost, in so far as at that price point I’m probably selling to people who have no DIY interest. Because DIY up to now has been fairly much a “one off” and cobbled together, requiring fabrication and a fair amount of pioneering along with trawling through forums to find answers.
But now there’s something different in the market. It straddles the line between the two. Let me show you something that has 100mm movement, with 250mm/second max velocity, that costs a fraction of the top end. It’s this the “SFX-100”. Designed by a guy with the online name “Saxxon66” he’s released the plans & software on Github (https://github.com/SimFeedback/SimFeedback-AC-Servo)
It’s absolutely not a “ready to run” solution. You still need some DIY/maker skill, but not nearly as much as some of the other DIY type seat movers. You do need a healthy level of awareness and respect for mains voltage (some wiring is required), aside from that thanks to 3d printing (more on that later) it’s more a case of assembling parts, as opposed to fabrication. What do you need to be able to do :
- Follow along instructions
- Basic soldering/wiring for the arduino controller
- Understand mains wiring for the AC drives, you need to get from a mains plug to the terminals on the drives.
- Assembly, loctiting things etc
- Quick learning on 3d Printing (it’s really fun) or get someone else to do it for you.
Basically if you have the patience to learn anything that’s foreign to you via Google/YouTube you should be okay.
For the price of 1.4 RTX 2080 TI cards in NZD, you can achieve similar levels of top end performance of the big boys without needing to fabricate parts, nor sell a car to finance them.
The system comprises the SFX-100 (Actuator) + SimFeedback (PC Software + Arduino Firmware) system. You get the parts list, the spec of the critical pieces, 3d printing files, wiring diagram, vendor list is all provided and up on github. The design is open source, free for non-commercial use.
Almost all the bits to make up one actuator. Missing from photo is o-ring, six screws and some elbow grease. In summary :
- You source the parts (An hour so depending where you live and what’s available to ship to you, weeks of waiting)
- Print some components (An hour or so of work. 200 hours of printing)
- Wire up, test, and configure the motors (An hour or so)
- Assemble the actuators (6-12 hours, if you’re really fast an hour an actuator)
- Wire up the controller (2-3 hours), flash the firmware (5mins), test it (5 mins) and fit it to your rig (????) and have fun.
We’re looking at a total of around 20-30 or so hours of work. A big cautionary note – you have to wire your servo drive controllers (x4) to mains power there is no standard 3 pin IEC connector. Obviously this is your own responsibility. The servo drives come with Chinese manuals so you do have to figure this bit out for yourself and is probably the single biggest barrier to entry.
You have the benefit that dozens of others have come before you and are very keen to help out if you get stuck. Although please do RTFM, as most often the answer is there. It seems like a person a day is joining the Discord – for everyone who’s donated. Which tells you this things getting some traction. Oh and hey I wrote a build guide as I went along.
If someone had said “you’re going to buy a 3d Printer that can go down to 50 micron resolution, and you’re going to print 1.5 km worth of filament by Christmas time” in September, I would have told them they were joking.
One key aspect of this project, that makes it truly accessible, is that the major component pieces aren’t CNC machined, but 3d printed. You wouldn’t think it’s strong enough – but it is with regular old PLA (don’t worry you’ll come to know what this term is and what it smells like). You’ll then also discover this wonderful world of 3d printing. I can’t begin to think how expensive/time consuming/problematic it would be to get this stuff made via CNC for Joe Public. Whereas Joe Public, with a bit of methodical work can get a good functional 3d print.
My advice – if you can afford it, get a 3d Printer. If you’re anything like me you’ll find a bunch of other things to print…
Now of course to get some good output, you need a good physics engine and track mesh data. To this end the videos and review I’ll refer to using my favourite sims – Assetto Corsa, and Assetto Corsa Competizione. I’ll save comparisons with other sims for another post as they can be quite different.
First of all tracks come alive. You thought you got feedback with that Direct Drive wheel didn’t you? This is taking it up a notch or 10.
Firing down the front straight at Hungaroring in the Ferrari GT3 car, the reason why you want heave becomes very apparent – the rig starts cycling through up/down motion following the contours of the laser scanned road. Not the pitch of the road – no, but the actual race track surface – hollows, crests, that are in the laser scan mesh – the better the sim is at representing this and outputting to telemetry, the more you’ll feel.
Below: Exaggerated levels for demo purposes, definitely wouldn’t use this setting all the time :
Get out of the GT3 into an F1 car and those seemingly gentle oscillations turn into brutal vibrations. Don’t believe me? Let’s compare the two….
Hit a kerb and you expect to move around a bit. This you can tell the difference from a flat painted surface, through to a solid car launching sausage kerb.
Traction control hunkering the car down for it scrabbling for grip. You’ll feel that too.
Pitch as the track surface goes up a hill or down a hill. Yep
Hit the brakes and feel that ‘surge’ forward. Yep. Hit the accelerator you go backwards Yep. How quickly will you find out your braking to the apex is untidy? Oh about Turn 1 I reckon. If you’ve picked up bad habits the motion platform will tell you about them.
What happens when you go off track – well, it’s a basic math equation – if on track you’ve got the rig tuned to give you a good amount of sensation, go off track and you’re hitting sand traps, walls, etc. They’re easily 2-3 x what a kerb would be like. If your sim racing game of choice models that effectively – you’re *really* going to feel it in the platform. Whilst it’s not “whiplash” consequence it is uncomfortable being jolted around violently. So you kinda get that extra emphasis to be a better driver.
Essentially anything that the simulators output in telemetry SimFeedback picks up and translates into movement. The better the physics engine of the title, modelling of the track surface/conditions, the better your experience. The best way to put it is there’s not an output from the game to say “I’m changing up a gear” there’s a series of values for all of the car velocity/pitch/direction that represents what’s happening so many times a second. Eg velocity longitudinally along the car 0, 0, 0 [clutch in], -2, [engage gear], [clutch out], 3, 2, 1, 0 [back to steady speed]. For instance, lets go crash sideways into a barrier…
How is it for “power”? Well it has more than enough headroom. Would you believe me that I only run it at 50% intensity level with the motors set to max speed. No? Well if I turn it up to full intensity level (a) it’s really hard to see/concentrate going through kerbs in corners and (b) I’m just waiting for something on my rig to break. It’s honestly *too* much. Great to show off to people. Too much to lap with.
How about noise? Check out the videos.. it is definitely quieter than SCN-6 style actuators. In my case there’s actually more noise as a result of what the actuators are doing to the rig, the MDF is twisting/squeaking in various places that I’ve got to address. You don’t hear the motors, and you hear a slight noise from the ball screws/sliding. That’s it. There is a high pitch electrical ‘whine’ from the motors when they’re holding the rig at a static starting position. Most of the time I have open back headphones on and nothing noise wise breaks the immersion.
One thing that I didn’t really understand when seeing someone else’s motion sim platform was setting up different “profiles” for car and track combinations. Now I get it ! Some cars can be absolutely violent to the point of unpleasant driving. So you do need to play with settings to tune the platform to your liking.
What I did find a little odd with the default AC profiles that I changed. Again this is my personal preference. YMMV :
- Needed to reduce the -ve surge axes. It was too much (particularly in ACC) and I was actually getting jolts on downshifts under hard braking which doesn’t happen in real life.
- Reverse the sway effect as it was backwards (at least for non VR user)
- I turned off the traction loss effect
- Increased the pitch and roll axes slightly
That got a good baseline I then tweak up/down depending on mood and car/track combination.
After 2 weeks of running I suddenly thought to myself, if I was running the 3kw heat pump this much our power bill might be crazy. I found a cheap plugin power meter that allowed me to breath a sigh of relief.
- At idle 30 watts
- Holding the rig at ‘static’ ride height 50 watts
- Normal track running 60-110 watts
- Violent crash peaking around 260-300 watts tops
Your mileage may vary of course, it draws powers the harder the servos need to work – so a heavier rig, heavier human, higher intensity levels all will contribute to bigger power draw.
As an aside I measured my PC + OSW + Screen and that’s more like 700-800 watts.
You probably should for safety sake look at high quality power boards and I chose to put my Servo drives behind an isolating transformer on a separate power circuit/outlet than my main PC.
Servos: 3000rpm Intensity : 1 Max Speed: 170 Pitch modified to +/- 64
Note this is the maximum I’d to run it at; the F1 chassis seriously shakes the rig – don’t know how long the Ultrawide would last….. Servos: 3000rpm Intensity : 1 Max Speed: 200 Acceleration: 1 Smoothness: 0 Pitch modified to +/- 64
Look the inputs to the sim games aren’t changing. Just as if someone with a used 2nd hand G25 wheel for $100 lap faster than someone with a $2000 OSW direct drive wheel, same rules apply here. A motion platform doesn’t change your inputs.
What it will do is let you know you’re untidy. It will let you know where the track surface is rougher. It *might* give you more queues through motion you’re going to loose traction. So you should in theory learn faster through the platform than if you had no platform.
At a certain intensity level you will get slower as it will be harder to concentrate/see/keep precise control inputs as you’re being bounced around.
What it will do?
It's all about immersion
The more queues your body/mind gets, the more immersion you have the better the experience is. Just like an OSW wheel that suddenly gives you real feedback of the car loading up in corners. The same way the SFX gives you queues what the car and track are doing.
Rocketing through the green hell is something else on this rig. Here’s my first video after getting the rig running, in a seriously janky state, but you get the idea…
- Unbelievable performance
- Great value
- Awesome learning process
- Quiet operation
- Low power consumption
- Many, many people before you
- Joining the Discord community of nutters
- *Now* you get why laser scanning is a thing
- No IEC connectors on servo drives means mains wiring
- Sourcing some parts a little bit time consuming
- Lots of cabling/boxes with the 4 separate servo drives & controller
- Once your mates get in your rig, it will be even harder to get them out.
- You won’t want to drive anything that’s not laser scanned anymore.
”A laser scanned track, with a good physics engine is seriously eye opening experience - one that you'll want to repeat over, and over, and over.
This system has brought a high performance platform mover design, into something that doesn't cost that much more than a single RTX 2080 TI*. Do you want 30% better graphics performance, or do you want to feel the road surface ? I know which one I'd pick..
It's brought high end professional equipment down to prosumer level pricing in the same way the OSW movement has. Judging by the activity in only a few months since the designs were released I think we all should shout Saxxon66 a few beers!.
* RTX 2080 Ti is approx $2300NZD right now. Lets say 1.4 cards and you’ve got all your components purchased