My simple sim motion chair

Ever since I saw the Frex Simconmotion I've wanted my own motion driving sim cockpit. So I decided to make one and am finally making some progress.

http://www.youtube.com/watch?v=R5cuLQnAu9w

Here's just a test of the RC servo driver I'm going to use.

It still needs some fine tuning, but even with the prelimary software configuration and sloppy paper clip and cardboard mockup, it works pretty well. Now to work on the full size model.


I'm going to try to drive the full size model with power window motors from a 1:1 car. They're essentially going to be piggy backed off the servos in the video via an H-bridge motor driver. Here, I drew a picture:




Here's the main base frame of the moving part of the chair. I'm trying to keep the whole thing as low to the ground as possible, so I'm mounting the sliders low on the inside of the frame rather than just sitting them on top. Please ignore the ugly welds, at this point I was still using my stick welder.. the new mig is a million times better for thin metal.


Here's another shot of the base frame. You can see I kept the sliders down low.



Here are a couple of shots of the u-joint where it's gonna get bolted. In keeping with the effort to keep things low, I've cut down the u-joint to make it as short as possible. I welded two pieces of flat bar to the joint for mounting.


This is what the seat will actually be bolted to. Just a simple frame. I later added arms to the frame for where the motors will actually move the seat frame.


Here's the subframe placed on top of the u-joint.


Here's the chair mocked up on the frame.


These are power window motors from a 80-83 Ford Mustang. Some other motion sim builders are using windshield wiper motors.

Wiper motors were made to run in one direction only, which means that if you try to reverse direction by reversing the leads, you'll often find that one direction runs faster than the other. They probably optimize the motor timing for the single direction and thus timing is all wacky when reversed. Power windows are obviously made to run in both directions. So both directions should run about the same speed. And indeed these motors tested at 100rpms almost exactly both ways.

And since wiper motors were designed for one direction, one of the motor leads is often grounded to the case, since it's always going to be the negative lead. If you try to reverse the motor without isolating the leads from the motor case, you'll dead short your power supply. So one has to modify the motor to account for directional speed changes as well as isolate the motor leads from the motor case. Window motors have no such problem.

Another advantage is that window motors come in pairs, while there is typically only one wiper motor in a vehicle. That means I can buy two window motors that are mirror images of each other. This helps aesthetically and simplifies mounting.

One disadvantage is that window motors typically have gears attached, while wiper motors typically have a convenient arm attached ready to go for a motion sim. But I have a welder, and I can weld on an arm.



The panel will contain all the motors and electronics. I'm designing it so that I'm able to unbolt the entire panel from the rest of the chair to make it easier to maintain. and upgrade if neccessary. You can see the Futaba S3151 digital servos in place. These are going on the chopping block!


This is all the electronics involved. Two servos, an RC BEC used as a 5v switching voltage regulator (9 bucks from dealextreme), a Pololu VNH2SP30 dual motor driver, and the Pololu serial servo controller. That's it! Pretty simple eh?


I have removed the gears and the motor from the servos. The idea here is that the servo drives the motor using PWM on the positive motor lead, and keeps the other one at ground. THen it just reverses the leads when it wants the motor to reverse. Looking at the VNH2SP30 datasheet you can see that for each motor channel, it has two direction pins (INA, INB) and a PWM pin.  Drive INA high, INB low, and pulse the PWM pin and the motor will turn one direction. INA low, INB High, and it'll turn the other direction. So in theory, I can run wires from the servo motor directly to the INA/INB pins and keep the PWM pin high, and it should work!


But of course, it's not that easy.. when i hooked it all up this way, it sorta worked... but the motor was running full speed all the time. It could reverse directions, but there didn't seem to be any speed control. Hmm... There must be something wrong with either the signal going to INA/INB pins, or the motor driver wasn't fast enough to keep up with the PWm signal. I suspected the former since someone else reported that the motor driver works fine when pulsing the inputs instead of the PWM.

So how do I determine if the signal is ok? I need a oscilloscope or logic analyzer and I don't have one!

Or do i?


Well I didn't.. but I bought parts to build a simple logic analyzer for the parallel port! I found several plans on the web and they were all pretty much the same.

http://www.xs4all.nl/~jwasys/old/diy2.html
http://akikorhonen.org/projects.php?action=view&id=59
http://tfla-01.berlios.de/

The plans called for a 74HC245 octal bus transciever. I was in berkeley last weekend an found Al Lasher's Electronics. Imagine an electronics store in the 70's and crammed packed with components and parts and all kinds of misc stuff and this is it. This is what a radio shack should be, and exactly what we need in the city. *sigh* I kinda miss living in San Jose just because there are a TON of places to buy components. Here in San Francisco there's nothing!

Anyway. They didn't have a 74HC245, but did have the 74LS245, which according to wikipedia is.. different... I don't really understand all the ins and outs of logic ICs, but apparently one is TTL and the other is CMOS, and they have different input/output levels. The pic above shows the 10k resistors to pull all the inputs right, but that didn't work right. The inputs sort of latched high when it saw a high. I don't really understand why, but things were ok after I removed the pull-downs. The inputs just float high, and that's ok I guess. I put a 7805 voltage regulator on to run the 74ls245. I have it running off my old K6-300mhz laptop in XP. It has 8 channels, and I was able to get about 660ksamples/s. Which is very slow as far as real logic analyzers go.. but it's good enough for me. Especially since it cost about 5 bucks to make..


Look at that! This is part of the sample where the motor switches direction. So going left to right, Line 7 is high, 8 is low, the motor is spinning full speed in one direction. THen it starts to slow.. but instead of starting to pulse 7 and keeping 8 low, it pulses 8 high! It finally transistions to spin the other direction by keeping 8 high, and slowly pulsing down 7. Well this is weird, but according to the truth table, if both INA and INB are high, it should brake to VCC, meaning both motor leads will see 12v and the motor will stop. So theorectically it should work.. but it doesn't.. who knows..

So I looked for more answers. I tried looking for the datasheet for the PT2422-X chip in the servo..


I couldn't find it, but did find a reference to a FR6461 in a google search for pt2422 servo

Googling FR6461 servo led me to the FR6461 datasheet and a Russian page with a reference to a Renesas part number: M64611FP.

It looked like the FR6461 is a clone of the Renesas part, and the pins on the PT2422 looked like they were going to the right places, so I think it's pin compatible with the Renesas M64611. So it looks like pins 5 and 6 are the outputs of the chip. So I call my brother to see how real engineers hook probes up to tiny surface mount chips and he says that they usually design in accomodations for logic analyzer probes. And that if there's nothing convenient, you have to just solder a wire somewhere and hook a probe on that. So that's what I did. I found two somewhat convenient places to solder to pins 5 and 6









SO I soldered some wires on and hooked it up to the Logic Analyzer.


That's more like it. One side's held to ground while the other is pulsed. Let's hook it all up!


Looks just like a normal servo except for the two extra wires. The feedback pot is still connected to the output shaft. I'm going to be using a linkage to connect the window motor to the pot.



This is how everything is gonna sit. I was thinking about using those terminal strips on top just to distribute 5v and 12v power. But I decided it would take up too much space, and there are only like 2 boards in there.. don't need a power distribution board.


I poked holes in the radio shack project box and put grommets in for the wires. I drilled a hole for the neat radio shack lighted switch.


It's all done, except for the heatsinks for the motor driver. I cut down an old Slot1 Pentium 3 heatsink and glued them permanently to the driver chips.


I cut a hole in the cover of the box and attached this spare 12volt fan.


Here's another look. I used an imitation Deans connector for the main power. I got them at hobbycity.com. They're better than real deans because they have a grippy surface, unlike the completely smooth exterior of real Dean connectors. And you can see I ran the PWM wires from the servo chips through the coiled up servo wires just to keep things neat. I thought about inducing a current in the signal wire, but if the coil is made of both the power and the ground, doesn't the magnetic field cancel? I dunno, but it doesn't present itself as a problem for now..

And now here are some vids of the thing in action!


http://www.youtube.com/watch?v=OTf_Er6ZCog

http://www.youtube.com/watch?v=A0BPqQrHibA
I taped a piece of carboard to the servo to raise the arm, and a bent a paper clip to clip onto the window motor gear. The linkage is a 6-32 threaded rod with two ball link joints from a t-maxx I think. Bought that at the local hobby store and just bought whatever looked like it'd work.

http://www.youtube.com/watch?v=Y67QtHo-ux4
The motors that I got run about 100rpms, versus 58rpm of the servo. I originally wanted to get motors that matched the servo, but I thought I'd take into account loading on the motors and I wanted to see how much faster I could move the motors over stock before the gain on the feedback loop would be too much. In this video, I'm just moving the servo from posistion to position, and you can see how it overshoots everytime and ocillates a bit and then settles. In the other videos, the problem isn't as apparent, but I'm sure it's there. I think it'll settle ok once the motors are fully loaded. If not, I'm gonna try to decipher the Renesas M64611 datasheet and try to lower the gain.




Here's the chair mounted to the frame in a pseudo mockup. It's almost done! Of course I couldn't help sitting in the chair and making racecar noises as I began to think about the wheel mount table.




The table is done.

Now I needed to think about powering the chair. In all the test videos you've seen before, the system was temporarily powered by some drill batteries. But that wouldn't do for the long term. So what I needed was 12-15volts, at maybe 10 amps? I don't really know the current draw on the motors, but 10 sounds about right. So I tried using my modded PC power supply but it shuts down as soon as you bump the motor. The inrush current of a big DC motor probably triggers some kind of protection on the power supply... bummer. So I had to come up with a new solution. How bout just a big transformer, rectify it with a diode bridge, and smooth it out with a capacitor? Sure why not...


What we have here is a cheap car battery charger/engine starter. Harbor Freight to the rescue! Well not really, because the charger pictured is the second one I got from them. The first one almost burned down my house due to a short in the secondary winding of the transformer. After turning it on and verifying the output, after a while, it got super hot and started smoking! So I brought it all the way back to Newark, and yelled at the kid at the counter, and got a new one. Anyway, buyer beware.

So the charger has two of the components I needed: a transformer, and a diode bridge. A rectified AC waveform looks lumpy like this (on my new oscilloscope!):

It's all peaks and valleys because the rectifying diodes merely invert the negative part of the AC wave. To get a decent DC output I needed to smooth it out with a capacitor. With large currents I thought I better go with a largish capacitor.


This is a cheap car audio capacitor that I got from Ebay for about 25bucks. It says it's 1 Farad ±5% but who knows. You can see it has some circuitry and a LCD display on the top. The circuit turns on a LED when you first hook it up, then 10 seconds later, a relay trips and THEN it connects the capacitor to the voltage source. I think they do this for safety. If Joe Blow hooks up a big capacitor to a big battery backwards, he'll *really* blow. This way you have ten seconds to realize that you did something wrong.


And this is the big cap taken apart. I needed to override the relay because I think I'm smarter than Joe Blow (not really, i exploded a cap on the motor driver ). And because the circuitry wasn't playing nice with my battery charger.

What happens is this: Plug charger in, cap initiates countdown, after ten seconds the relay connects capacitor, empty cap acts as a dead short initially which causes large current draw, large current dump causes massive voltage sag in battery charger, voltage sag freaks out the capacitor circuitry and causes relay to disconnect cap, now voltage returns to normal, and relay reengages, and starts the vicious cycle  again. So I just jumpered the input and output of the relay.


This is the finished power supply with cap attached.


And here's the output of the charger/cap combo. Smooth as butter. It's about 14.6 volts DC (using a 10X probe, so it shows 1/10th the value on screen).


I thought I'd show you my homemade ceiling projector mount. All the commercial ones hang down too low. so I had to make my own out of a cover of a big electrical junction box.

Back to the chair now.





It's pretty much done now!

And now some videos!

http://www.youtube.com/watch?v=uQfnXuT1oSo
I was able to keep the sliders working with the motion system.

http://www.youtube.com/watch?v=vyY-KcsntVs
See how much it flexes? I should have used a thicker mounting plate for the motors, And probably better motors too! These ones have a squishy nylon output gear that just spins about a stationary shaft.  A support bearing would have been a good idea too.

http://www.youtube.com/watch?v=BOufJuZHUOU
Here it is in action!

http://www.youtube.com/watch?v=EcL_HYl6LC4
This time I'm driving a powerful car with sticky tires, and it's a handful!

I have to work on the profile some more, and also try some other games like Race 07/GTR Evo. But for now it works well enough in Live For Speed and the entire project didn't cost very much money at all. Thanks for reading!

WOW !!!! One of the most HUGE posts I ever seen!!!

Most people here start posting their progress one step in time. You did make a jump to showing it all the steps and the cockpit in motion at once !!!! Awesome!!!! :clap: :clap:


Back to your stuff... You should found out yourself already, that any RC servos cannot handle smooth acceleration because they don't have programmed PID algorithms inside their chips. The turn the motor at FULL SPEED until they reach their proper position where they just microstep the motor Forward-Backward to keep it in torque. This is the buzzing from the motors that exist in all your video!!!

A true PID motor controller like the AMC1.5 in conjunction with DSMhb1.2b motor driver would be probably more cheap than all that stuff you used and the hard work making it work. I admit that I have tried in the past almost the same you did, but I had no quality in motion and that project was aborted before even start it.

Keep up the good work!!!!

Best Regards, Thanos

Hi Thanos,

I've admired your work with the BS2 and the AMC controller and DSHB for a long time now. For some reason getting a printed circuit board made here in the USA is expensive... so it worked out cheaper to do it my way. I thought it would be a lot less trouble than it was. I didn't anticipate the problem of hooking the motor driver to the servo..

Though of course having a true PID controller like the AMC is better than my solution

I found that the I used RC servo is a Proportional controller (the P in PID) and not a very good one. The motor is speed controlled with a PWM signal so it is not going exactly full speed to the destination; it does slow as it nears the destination. But it only had maybe 4 distinct velocities.

The particular RC servo i used was a Futaba S3151 Digital servo. It differs from standard servos in that its motor PWM signal is about 300hz, rather than just 50hz. I think the buzzing from the motors in the video is just the 300hz PWM signal, since it is very much in the audible range. It's not bothersome when racing.

Thanks for the compliments!

BTW, My wife recently visited greece and said it was a beautiful country. I hope to visit some day!

-Kelvin

Hi Kelvin that was a serious amount of quality work and info you just released here to the members of X sim , have seen some projects and have to say it never ceases to amaze how many ways this project can get built and looks bloddy awesome Kudos my man truly excellent work Big Thanks for sharing : :clap:

I agree its excellent work and the best of all is that is complete!

I think that it should be put on the wiki next to the other projects as well !!! :clap:


Regards, Thanos

Hi Kelvin,
very cool :clap: . It is so small, compact and finally adjustable. Great, i like it :thbup: .
How long did you work on that simulator?.

I've been planning this project in my mind for a LONG time now, since I first saw the Frex Simconmotion years ago... The discovery of X-Sim just made it all possible.

Actual work started when I posted the first video of the servo test on youtube in March, so not too long. Another reason I did my project with modified servos, was I knew I would make a mistake if I tried to build the AMC controller, and that would slow me down, and I just wanted to race. I'll be making improvements to my cockpit slowly, and I'll update this thread.

Thanks guys!

-Kelvin

yeh ya simple chairs making me tear hair out.
1. bought - pololu serial 8 card , 2 hitec hs-55 servos
plugged in and setup card only lights yellow led.
no movement from servos in animaltronics software

Basiclly trying to the same model shown in start of post hooked up to rfactor.

Please help, have personal ventrillo channel to talk through process cheers mate
plus eventually need a good running profile for chair. cheers again

Hello all.

I stumbled upon this whilst looking up 'Window motors' for another planned project and wanted to try to understand how it works.

Now I'm just confused because I think you attach pins 5 and 6 on the chip in the servo to somewhere on the motor driver. Have I missed an important step out?

I would also like to know how would this connect to a Mosfet H-Bridge like this - http://www.learn-c.com/basicmosfethbridge.gif - as I can't figure out which wire(s) from the servo should go to which gates on the H-Bridge.

Any help would be greatly appreciated.

Excelent work!!!
I like the position of the motors... is very cool you are not using the T in the back of the chair!!!

Now, if you move the seat from below, the motors has to do more strength??? or is just the same?

You could put the pedal and steering wheel to the structure of the chair and move all together! It's just an idea ...

Congratulations! :yippiee:

Awesome work!!!!

Congratulations ApacheXMD on such a detailed post and for coming up with such a brilliant idea.

For one thing you can still use the seat sliders to accomodate different drivers and it also looks to be a less complicated mechanical assembly so would perhaps appeal to more simmer's.

It would be good if a few of the more advanced members could pick this up and possibly improve on things like feedback control and the connection to the PC etc. This could result in another Simforce GT type project but without the cost of actuators.

This is a fantastic effort and I will be copying your setup all the way. Thank you for an incredibly eye opening post. It is very detailed and i'm greatful that you have included so many good photos of your hard work.

Regards,

Matt

The best Job!!!

What kind of potenciometers do you used?
What Value?

Thanks

Agreed! A fantastic effort, beautifully executed and described. It was nice to see the window motors in action.

Regards, Erich.

You definitely get the light and tight award. Simple and genius. And you're my neighbor too. How cool!
Congrats from Oakland...D

First off, Awesome work!
But since I am new here, can you help me out? I'm not understanding how the motor linkage setup is actually moving the chair.
I'm probably overlooking the obvious, but can you explain it in a little more detail?
Thanks, this project is a true inspiration!

Thanks for the compliments guys. Sorry I haven't been around to answer some questions, been busy with work and life.

Pins 5 and 6 are the outputs of the servo chip. The signals from these two pins tell the servo motor which direction to spin. Example, if pin 5 is high (5volts) and pin 6 is low (0v), it will spin one direction. If 5 is low and 6 is high, it will spin the opposite direction. I needed these two signals to go to the Inputs of my motor driver (h-bridge), which will run the power window motor based on whether 5 and 6 are high or low.

If you're going to be using a simple H-bridge like the one in your picture, you're going to need some more circuitry to tie it to the servo. You need to be able to control the 4 transistors with only two inputs and make darn sure you don't create a condition where it dead shorts your voltage source. Circuit design is definately not my strong suit, so I'd look elsewhere for more guidance with using a simple H-bridge like that.

How hard the motors have to work depends on the load, and the moment arm it has (leverage). I tried to give the motors as much leverage as possible by placing the linkages at the far corners of the seat base. It's still not enough, but it's the price I paid for having a more compact design. There's definately room for improvment.

Moving the pedals and steering wheel would defeat the purpose of the Frex-like design, and would require much greater degree of movement.

The pots are whatever the ones that came with the servo's. I think they were 100kohm.

The seat is mounted to a pivoting subframe (pivots on a universal joint). The motors are attached to the subframe with the large linkages you see in the picture.

The motors basically lift and drop the corners of the subframe, thus moving the seat in two axes.

-Kelvin

Hello ApacheXMD,
Very strong work, congratulations ! Very easy concept and cheap possibilities. Awesome.

Just a question : How is you card Pololu (servo controller) recognized by the software X-sim ?

I have a K8055 (vm-110) which is recognized easliy by X-sim. But I can't see in the list of supported hardware the pololu card.
A plug-in is needed ? Or the card pololu is recongized easily because it is RS232 connection ?
Thansk for your reply.
See you
Adbgg

The Pololu servo controller interaces with X-Sim's Universal Serial Output (USO), so yes, it's just standard RS-232. That's what makes X-Sim so great, it let's you use a very large number of hardware, and makes interfacing relatively easy.