Adding 2DOF (Seat Shaker) to Aluminum Rig

Look to be 25:1 to me based on how easy they move

I hope not, but I get that feeling too. I tested the no load speed of the gear-motors and they were spot on at 75 rpm which based on the listing is correct for a 50:1. pgsaw doesn't have them listed anymore only the 25:1 so I pulled up my order history. I had one of them apart to flip the shaft and I wish I had taken a picture at that time. I might pull the motor off and just count rotations of the input vs the output to verify. I think I can do that without disassembling my whole setup.



here is his listing for the 25:1 version- http://www.ebay.com/itm/GEAR-MOTOR-...-25-1-ratio-/351066992384?hash=item51bd392f00

I was able to test one of the gear motors and they are 50:1. I removed the motor and spun the input shaft, exactly 50 rotations of the input resulted in 1 rotation of the output shaft. It's good news, but I will have to add a lockout system for non motion driving.

I was able to rebuild the seat base, now I'm waiting for some new 3d printed parts to finish the assembly.





I was having some rotation issues with the u-joint so I added some star washers and changed the bolts from button head to socket head, the socket head bolts allowed for more torque and I think I fixed the problem.





By removing the rear crossbar I no longer have an attachment point for a spring tensioned seat belt system. I will have to get creative when I get to that point.

You can move a 50:1 gearbox by hand?!? I wouldn't have bet 1cent it could ever be possible. My 60:1 is as stuck as a rock even without motors connected.

Looking forward to seeing your finished rig

I finished printing a whole new set of parts, these are beefed up from the original designs. Also I changed the mounting bracket for the rear lever.



Another thing I did was make the nut cavities shallower, I picked up some jam nuts which are lower profile then standard nuts. This allows a little more material the mount to.



Here is the new mounting bracket, much simpler to mount, and much more rigid because of the mounting location.



With the rear arm installed, you can see the difference between the new and old design.



very little clearance between the mounting bolts and the rear arm, when I build this in aluminum i will probably countersink the bolts to give me more room for flex. I almost guarantee that with the printed parts I will get some rubbing here.



Here is the small linkage installed, this was a short linkage so I had to buy a male and female version of the swivels and thread the into each other.



The final linkage installed, I chose to go with the underneath mount for simplicity sake. I eventually want to mount on top of the extrusion, but I have to design a system so the swivel rod end wont crash into the seat frame.



hear is a shot from the back, when the seat base is in the level position the linkage is vertical and there is no toe in or out.



at the limits of the motion I have some binding, I should be able to fix that with a trip to the belt sander to knock the corners off my bracket. I know 8020 sells a rounded bracket so I will look into picking a few of them up.



Here are a few videos of me standing on the platform and moving it around.





I have a few items left for assembly and then I will move over to the software side.
Install seat
Install shifter and e-brake
Install switch panel
Change center tablet to new amazon 7"
Install transducers
re-install side monitors
Install Oculus sensor
sort out racing harness

I started working the electronics/software side of the build, I wanted to run the XPID software by thinsel with the tft lcd system. so I had to picked up a few parts (some of these I teased earlier)

Arduino Mega 2560 R3 - amazon.com/gp/product/B00JTBMD7E
2.8" tft Touch Screen - amazon.com/gp/product/B01EUVJYME
Mega Screw Shield - amazon.com/gp/product/B00UT13YXA
I also wanted to do some bench testing prior to hooking up the big motors so I picked up two mini gear motors
Mini Gear Motor - amazon.com/gp/product/B01N1JQHPD

here is the assembly of the screw shield, some solder required





and adding the arduino and tft screen, all plug and play

Next I booted up the arduino and took a crack at installing the XPID code(https://www.xsimulator.net/communit...rduino-uno-mega-and-sabertooth-with-lcd.6747/), it seemed to load fine, but I ran into all sorts of issues. This isn't meant to be a full guide, however if needed I could probably write one from all the testing I've done.



so I reached out to vthinsel and he was great, responded almost imediately and provided me with the latest version of the code (3.17 vs 3.15 in the forum post). I explained all of my issues and where he could he helped me debug the issues. not sure if I can share a conversation post but here it is

https://www.xsimulator.net/communit...duino-uno-mega-and-sabertooth-with-lcd.35484/

So I loaded up software version 3.17 and here were the issues I had

Issue #1 - No Touch
I scanned the XPID code versus the sample coded for the LCD and found the pinout to be wrong for my lcd so edit 1
//HODDEM UPDATE 2-23-2017//
#define YP A3 // must be an analog pin, use "An" notation!
#define XM A2 // must be an analog pin, use "An" notation!
#define YM 9 // can be a digital pin
#define XP 8 // can be a digital pin
//HODDEM UPDATE 2-23-2017//
//#define YP A1 // must be an analog pin, use "An" notation!
//#define XM A2 // must be an analog pin, use "An" notation!
//#define YM 7 // can be a digital pin
//#define XP 6 // can be a digital pin

Issue #2 - Touch is not aligned with the screen, always mirrored across the x or y plane
Edit 2 - had to align the touch with the actual screen
if (touch.z > MINPRESSURE && touch.z < MAXPRESSURE) {
touch.x = map(touch.x, ts_xmin, ts_xmax, 0, 240);
//touch.y = map(touch.y, ts_ymin, ts_ymax, 0, 320);
//HODDEM UPDATE 2-23-2017//
touch.y = map(touch.y, ts_ymin, ts_ymax, 320, 0);
//HODDEM UPDATE 2-23-2017//

Issue #3 - at boot the program asks if you would like to calibrate, if yes then it goes into a calibration cycle, if no I get a blank screen
Edit 3 - found a snip-it of code in the calibration sequence and applied that same code when not calibrating

void TFTcalibrate() {
unsigned long start;
UpdateButtonPressed();// to restore PIN states
tft.setTextSize(2);
tft.setCursor(0, 130);tft.print(F("Press TFT corners"));
tft.setCursor(0, 150);tft.print(F("within 10 seconds"));
start = millis();
ts_xmax = ts_xmin = ts_ymax = ts_ymin = 512;
while (millis() - start < 10000) {
touch = ts.getPoint();
if (touch.z > MINPRESSURE && touch.z < MAXPRESSURE) {
if (touch.x > ts_xmax) { ts_xmax = touch.x; }
if (touch.y > ts_ymax) { ts_ymax = touch.y; }
if (ts_xmin > touch.x) { ts_xmin = touch.x; }
if (ts_ymin > touch.y) { ts_ymin = touch.y; }
}
}
//HODDEM UPDATE 2-23-2017//
UpdateButtonPressed();// to restore PIN states
//HODDEM UPDATE 2-23-2017//

Issue #4 - getting strange values, nan and cannot adjust the values at all
Edit #4 - I did some research and arduino now used nan (not a number) values for empty eeprom slots vs what used to be treated as 0. the code was comparing to see if the value was 0 and then would write the default value, with a nan value the code did nothing. I added a hard write section for first boot, there is no logic though so it has to be performed manually by un-commenting all of the forced values and then re-commenting them after the first run. you can force an eeprom update by changing the sketch version in the beginning of the code
int XPIDVersion = 317;


void InitEEValues() //Used to initialize variables and EEPROM content in case of new arduino, or updated EEPROM schema
{
if (M1.getMax() <= 0 || M1.getMax() >1000) {
M1.setMax (1021); // Maximum position of pot 1 to scale, do not use 1023 because it cannot control outside the pot range
}
if (M1.getMin() <= 0 || M1.getMin() >1000) {
M1.setMin(1); // Minimum position of pot 1 to scale, do not use 0 because it cannot control outside the pot range
}
if (M2.getMax() <= 0 || M2.getMax() >1000) {
M2.setMax(1021); // Maximum position of pot 1 to scale, do not use 1023 because it cannot control outside the pot range
}
if (M2.getMin() <= 0 || M2.getMin() >1000 ) {
M2.setMin(1); // Minimum position of pot 1 to scale, do not use 0 because it cannot control outside the pot range
}
if (M1.getDeadzone() < 0 || M1.getDeadzone() > 10) {
M1.setDeadZone(5); // +/- of this value will not move the motor
}
if (M2.getDeadzone() < 0 || M2.getDeadzone() > 10 ) {
M2.setDeadZone(5); // +/- of this value will not move the motor
}
if (M1.getK() <= 0 || M1.getK() >= 5) {
M1.setK(1);
}
if (M1.getP() <= 0 || M1.getP() >= 5) {
M1.setP(1.000);
}
if (M1.getI() <= 0 || M1.getI() >= 5) {
M1.setI(1.500);
}
if (M1.getD() <= 0 || M1.getD() >= 5) {
M1.setD(.300);
}
if (M2.getK() <= 0 || M2.getK() >= 5) {
M2.setK(1);
}
if (M2.getP() <= 0 || M2.getP() >= 5) {
M2.setP(1.000);
}
if (M2.getI() <= 0 || M2.getI() >= 10) {
M2.setI(0.500);
}
if (M2.getD() <= 0 || M2.getD() >= 10) {
M2.setD(0.300);
}
if (DeadZone <= 0 || DeadZone >= 50) {
DeadZone = 4; //increase this value to reduce vibrations of motors
}
if (ReadAnalog <= 0 || ReadAnalog >= 8) {
ReadAnalog = 8;
}
if (M1.getStandby() <= 0 || M1.getStandby() >= 1024 ) {
M1.setStandby(512);
}
if (M2.getStandby() <= 0 || M2.getStandby() >= 1024) {
M2.setStandby(512);
}
if (SabertoothType <= 0 || SabertoothType > 2) {
//HODDEM UPDATE 2-23-2017//
SabertoothType = 2;// sabertooth 2*25 =>1 sabertooth 2*32 =>2
//HODDEM UPDATE 2-23-2017//
//addrSabertoothType = 2;// sabertooth 2*25 =>1 sabertooth 2*32 =>2
}
ts_xmax = ts_ymax = 500;
ts_xmin = ts_ymin = 500;
//HODDEM UPDATE 2-23-2017 Uncomment these values for first eeprom update to fix NAN (not a number issues) it will force values into the eeprom. once the values are burned the first time it will work correctly from then on. to force an eeprom update change the xpid version.
//M1.setK(1);
//M1.setMax (1021);
//M1.setMin(1);
//M2.setMax(1021);
//M2.setMin(1);
//M1.setDeadZone(5);
//M2.setDeadZone(5);
//M1.setP(1.000);
//M1.setI(1.500);
//M1.setD(.300);
//M2.setK(1);
//M2.setP(1.000);
//M2.setI(0.500);
//M2.setD(0.300);
//HODDEM UPDATE 2-23-2017//
WriteEEValues();

}

After I made all of these changes the arduino and lcd appeared to operate correctly.

now I was ready to do some testing so I grabbed my new mini motors and 3d printed some pot brackets and lever ams for visual feedback.





now I was ready to start the wiring process, this is where having the terminal breakout for the mega is a dream.
so I soldered up the pots and the motors and connected everything per the diagram in the XPID thread.



Then I hooked up my bench power supply fired up simtools and tested the outputs. I got absolutely nothing to work, so back to the drawing board.

after a lot of reading up on simtools (which I had done very little of before) I realized I had to configure both the axis assignments and the interface settings before anything would work.





so Test 2 after setting up simtools and Nothing again. Well actually not nothing, I was only getting motion when turning output off for the 3 seconds during shutdown. that was a clue because the shutdown output appeared to work, but the interface output did not. so back to the forums, and of course most posts are for the old simtools, I'm using simtools 2.0. I did find a thread with the key though the new simtools has axis A and axis b assignments the output needs to know a or b. simple enough just add an a to the axis.



So I tested again and the motors went haywire and crashed into the side of the pot bracket. I decided to remove the pots and brackets and test the motors by turning the pots by hand. the expected behavior would be turn the pot to the home position and no movement by the motor, move from the position and the motor would start turning in the opposite direction. The further from home the faster the motor should spin. here is what I got.



something wasn't right, it seemed like the home position was jumping randomly or the pot was jumping randomly. seemed like interference or a bad pot, but after checking out pot voltages everything was rock solid so I started looking at the program again sure it was a bug.

vthinsel suggested I use the builtin graph feature to see if the target position, current position and power were stable or not.
green line is pot position
black line is target position
red line is motor power


when the position and target are overlapping the motor should have zero power, all 3 lines should overlap. They did, but if I moved the pot slightly the power spiked and the spike was always an offset so if I moved the pot slightly left or right of setpoint the power was still on in a single direction. When moving the pot through the range, green line and red line should be mirrored across the black line and should cross each other on the black line, again that was not happening. so I started playing with the PID settings real time and was able to determine it was just bad parameters. after more testing I reassembled everything and performed another test.



Success!

I played around with the small motors for a while, but I don't have any way to test game output (shame on me while doing the build my iracing subscription expired).
so I moved everything over to the large motors on the rig and did a test with pots and levers disconnected, it was an instant fail again.

I did a little debug and found during testing I had adjusted some sabertooth settings that may or may not have had an effect. Then I remembered that I needed voltage clamp resistors when using a power supply. from the manual

Power Supply

Sabertooth 2x32 connected to a power supply and voltage clamp resistors.

Sabertooth 2x32 can be used with a suitably sized power supply. Power supplies used with a Sabertooth 2x32 should have output voltages between 7V and 30V. It is important to make sure the power supply being used can produce enough current for the load. Power Supply mode is selected by setting DIP switch 3 to the ON position. One major improvement in the third generation Sabertooth motor drivers is the addition of the Power outputs P1 and P2. These can be used with an appropriate resistor to


Also I'm using a 5a max bench supply. Even with no load reversing direction on two motors will max that out. so I set the current limit in the sabertooth software and did another test. things were working much better and again after some live tweaking with the lcd screen I was happy enough to test with everything connected.

so here it is



I'm really happy so far, hopefully the XPID software will workout. vthinsel is planning on updating the code to fix some of the issues I found and I need to read up on my PID (ie read the PID without PHD doc) so I can tweak in the motors.

If anyone is interested in my final code or any of my settings I will share them, but I would like to get everything working and tuned on the rig before sharing too much in-case I have a setting way out right now.

last thing, I think I will eventually have to remove the rear linkage assembly and move the motors back 3" just for simplicity sake. I will also end up extending the frame at least 3" to compensate and to give myself a place to mount the seat-belt. I wont do that until I get the build further along to make sure I don't need more space, but I'm so OCD about things like that, that I want to completely replace the rails instead of extending them. that would require a complete tear down of the rig, re-spacing all of the supports, legs and vibration dampeners, probably a full days job. no bueno!

Now that I have tested the motors I need to work on my power supply setup, install the seat to test the 3d printed parts, and get a test run in with iracing. This was such a milestone that I'm actually having a hard time with where to go next.

Nothing secret, you can share Thanks to your tests and great feedback Ill release a new version soon ideally with code handling your TFT requirements

I couldn't leave well enough alone, I ended up mounting my seat to the the seat base and doing some additional testing.



Everything was going well, but I could see the system struggling with the 5 amp limitation of the bench power supply. I decided to take it to the next level and try some human testing. Of course I had to start small so I threw my 3 year old in the chair for a little rodeo ride (don't judge me) and it was working great with gradual motion. I started to increase the speed and snap, all hell broke loose. It actually wasn't that dramatic, the right rear motor over rotated slightly and broke the pot mounting arm.





So I need to do two things, print a handful of these brackets for future accidents (maybe redesign for full 360 rotation) and work my server power supplies so I'm not limiting the system.

What is the value of the resistors you use ? Your text got messed up and I'm interested in those

oops, I missed a whole section. But it doesn't say what resistor to use.

here is the manual
https://www.dimensionengineering.com/datasheets/Sabertooth2x32.pdf

and the full section on resistors
Sabertooth 2x32 can be used with a suitably sized power supply. Power supplies used with a Sabertooth 2x32 should have output voltages between 7V and 30V. It is important to make sure the power supply being used can produce enough current for the load. Power Supply mode is selected by setting DIP switch 3 to the ON position. One major improvement in the third generation Sabertooth motor drivers is the addition of the Power outputs P1 and P2. These can be used with an appropriate resistor to dissipate the regenerative current generated when the device slows down or stops. This allows the use of power supplies without batteries or capacitors to absorb the regenerated energy. A calculator on the Dimension Engineering website will assist with choosing appropriate resistor packs. At least one of the power outputs should be configured as a voltage clamp when using a power supply.

The resistors I am using came in the box with my 2x32, I believe they are 5.1 ohm and 10W

I updated the 3d model with the extended frame, removed the rear linkage assembly and added the electronics.
Next up for the model is getting a seat that is the correct size and figuring out the shifter and ebrake attachment

check out the model here
https://grabcad.com/library/aluminum-racing-simulator-r2-03-1

I cut some 3" long extension pieces, I just cut these with my chop saw using a carbide wood blade.



I had to drill out the center holes to 5/16 for the mounting bolts to fit through, I am using 4" long bolts and they will just attach to the existing frame.



Here are the extensions installed
The alignment isn't perfect, but its not even noticeable from a distance.





I removed the rear linkage assembly, I will no longer need it now that I can move the motors further back.



Next I moved the two rear isolation bumpers as far back as possible. they are now entirely supported by the extensions.


While I was at it, I changed the angle bracket on the seat base to a more robust and round bracket.



Finally, I moved the motors further back and reconnected the swivel rods. because I essentially replaced the rear linkage assembly with the motor, the length of the rods didn't need to change and I was able to reuse the old rods.



I did have to pull my pots off and recenter them, the new configuration has the motor lever horizontal vs vertical before. I put everything back together and did a little testing and I was having issues with my second motor oscillating and overshooting. After some trial and error I found that I had set the current limit to 10A in the sabertooth software. For some reason motor 1 runs just fine with the 10A limit, but Motor 2 does not. So I setup my sabertooth software for the max 64A and things started working much better.

I also did a test using a single server power supply and the added power over my lab supply had everything working very smooth. Just for giggles I added the second supply and it seemed like the motors were in fact more responsive.

Here are my two supplies Jerry Rigged in parallel, they are DPS-600PB server supplies.


My motors are listed at 180 watts, I swear I had read somewhere that the motors can actually pull closer to 360 watts during hard use so way back I bought two 600 watt power supplies. I did open the sabertooth software and watch the input voltage drop when using a single power supply, so I think the second supply while overkill is helping.

And here is the latest test video with the chair installed, the jittering is from me losing control of the sliders while trying to hold the camera and run the mouse.

Quick question to anyone that has a seat shaker, did you mount your shifter and e-brake to the seat so it moves with the seat or static mount them to the frame?
just looking for how others have done it

Thanks

I am of the opinion that I prefer it mounted to my seat, Is it Ideal for weight balance, Absolutely Not. Tho anything technically can be accounted for if you plan it right.

That's the way I was leaning, but I wasn't sure how people are doing it. It could get challenging mounting my stuff in a way that I don't hit anything, especially with a long E-brake arm sticking up.

Big night tonight, I got iracing setup and did some seat testing. Everything works and my 3D printed parts survived a couple of aggressive crashes.

One thing I noticed is I get very little movement from driving around the track, I'm really only getting movements when the track is not flat. Almost as if the seat is following the track regardless of what the actual forces should be. I expected accelerating and braking to have more behavior. I played with a bunch of settings and I'm going to need some help getting that part working correctly.

If anyone can share their axis settings in simtools, specifically all of the tuning stuff like gain smoothing etc.

Also I have too much movement from limit to limit. when I'm pitched forward I get crammed into the pedals. When I'm pitched backward I can barely reach the pedals.