First attempt - 6DOF and learn to TIG weld at the same time.

I really think you should post it to public as I had kind of the same problem and anybody may have the same problem in the future. But its your call

@Squonk
I hope you don't take it the wrong way but I'm gonna be a bit brutal.

First off I would suggest that you reroute your wiring, looking at what I can see in some of your pics.
You don't use shielded cables for the feedback and worse yet they are bundled with the motor power cables.
I'd say it would be a candidate for "how not to wire up your rig" picture.
Separate all your cables and use 3 core shielded wire for the feedback or atleast plait the 3 wires together.
You can bundle all the motor power cables together and bundle feedback cables together just try not to run the 2 bundles parallel to each other.

I can see the reason of the continuous rotation of the motors.
There is 2 main things wrong.
You are NOT using a hall effect pots if the mouser link you posted is what you have, they are resistive pots.
They are 360deg mechanical meaning they can continuously rotate without damage however they are only 345deg electrical, which is the reason of 15deg dead rotation.
You said that the feedback pot reports a range of 03 to 95.
I haven't looked up what settings mean on the thanos controller but by the looks of it the "Sensor Direction" parameters are wrong and are the reason of cont rotation.
The values of positive=135 & negative=246 to me means that the values are max and min values and 190 would be neutral position.
I take it that the controller moves all motors to neutral at startup but your feedback only is able to attain a max value of 95 before passing through the deadspot only to get a value of 03 which is why it keeps rotating never to get to the commanded 190.
If you don't change your pots you would need to the "sensor direction" parameters to something like positive=90 and neg =5 which would leave you neutral as 42. This is correct if you use full 360deg rotation of the motor arm but in reality you only use a max 180deg. That will give you a something like positive=45 and negative=5 and a neutral of 20.
This is a pretty damm coarse for feedback, is it really only 0-255 for feedback control or is theses values for reference only?
I would suggest that you feed the pots with atleast 10v instead of the 5v, that way you can have 0-5v for the first 180 deg and the 180-360deg being 5-10v. This will double your feedback resolution, better still would be to use an adjustable regulator so that you can calibrate you 180deg movement to span 5v or even 10v.

@SilentChill
@OZHEAT

The issue was wrong connected ribbon, as it has two rows from which one row only is used for data. See this video after minute 10:00

Just to answer a few of the above... the numbers 03 to 95 are Percents not actual positions (03% and 95%). No need to have to display the whole numeric value that is 4 characters, 1024 for 10bit or 4096 for 12bit.

The Sensor direction setting shows which of the six sensors will be negative reading or positive reading direction. The direction is changed in software as many sensors output analog voltage that cannot be reversed mechanically. So after expanding the Sensor Direction parameter it means:
POS135, NEG246 = Positive direction Sensor1 Sensor3 Sensor5, Negative direction Sensor2 Sensor4 Sensor6
There is not much space on the LCD for otherwise fitting this setting





Again, in many locations I use percent as representation as for example on the 6DOF Start Position (Home position of the motors when offline). Again, if you change the value with the menu encoder, it will give you 255 steps to change it. I could use 65535 steps to change it but then you would have to make countless turns on the menu encoder to set it to the exact value you want, so I use a percent approximation.

@OZHEAT, thanks for your observations. I think @Thanos addressed most of them. The problem I was having had to do with the fact that the controller arrived with a ribbon cable that connects the main board and expansion board installed in a way that there was no continuity between the two boards - hence no control output to the Sabertooths and continuous rotation of the motors. I didn't include that in my "found the problem" post because it was a simple oversight in how the unit was assembled and I didn't want to imply a problem with the controller itself.

As for the wiring, the mess you see in the photos I posted was never intended to be the final arrangement. As I got further into troubleshooting, it got messier, and none of the photos I put up were meant to illustrate how or how not to wire a rig. The final arrangement will separate signal from power as you suggested. And, I am using Hall effect sensors, though I might inadvertently have linked to a different component in one of my posts. I'll remove that link.

LOL you have no idea what your talking about

Ok, I was wrong. only trying to help.

I think the bigger issue would be that currently it is so easy for the motors to runaway.
Not too bad for a worm gearbox drive with 360deg pots but a linear drive could be easily damaged.

What happens if the sabretooth driver is ready before thanos controller is ready?
What happens if you hit reset on the controller?
What happens if the thanos controller is not powered on?
What about a computer lock up?

I know the main problem is the sabretooth driver not having an enable/disable signal and relies on seeing 2.5v on the input for the motor not to move, a non ready/operating controller will put 0v on the sabretooth's input telling it to go full bore one direction.
Surely it is the motor controller's job to make sure that such runaways don't happen.

Maybe a solution would be to put in a relay to switch the 5v supply to the sabretooth driver after you know the controller is ready and would also take care of the no emergency stop issue.

This is why I stopped using it until there is a full user manual available because all the things you mention made my actuators go crazy. My controllers are on a different power circuit so I turn them on first and then I turn the power to the sabertooths. However this didn't help as with no software limits and the amount of rewiring I would have to do to make the hardware limit switches work it would be much trouble. Its hard also with no software utility like SMC3 has so you can see which way everything is going and set the PID easily, because of this I have gone back to SMC3 for now until all these problems are resolved.

The one thing that is missing from the sabertooth is an enable input. Only way to disable it fast would be cutting power or disconnect the motor.

As far as the AMC1280USB if its powered off there is nothing you can do.
If the AMC1280USB resets, the 12bit DAC chip will still hold 2.5v outputs until receives new values.
If your PC crashes, nothing will happen to AMC1280USB and sabertooth, it will hold last position for a few seconds and roll back to home position as soon it detects its offline.

As far as I know (I don't know the AMC1280), you have only one interface, the AMC1280.
Sabertooths are connected to it. And you will use only Axis (a).

Will have to check in AMC documentation thou.

Thanks. That's what I thought originally, but then started to convince myself otherwise.

I'll give you some example links for setting up the AMC1280USB with simtools as soon i get near a computer...

Quick update. I reverted to Arduinos and SMC3 this weekend and managed to get all motors set up and responding to SMC3 Utils. Hope to move on to Simtools and figure out the three interfaces this week, then it's just clean up, attaching the seat/wheel/pedals and trying it out. Thanks to all for the help to this point. I may try again with the 1280 at some point, but it was just kicking my butt.

Gingerly (very gingerly) testing out the DOF for the first time. It ain't pretty compared to most of the rigs I've seen here, but it seems to be working. The final control solution was three Arduinos and three Sabertooth 2X60s. I couldn't get the AMC1280 board to work, but will revisit that someday when I have a better intuitive understanding of all of this.

The motors were quite jerky when I started, and I found I had to set the Fpid divisor to 6, and add 75% smoothing filter to all axes to calm them down. That may be excessive, and I don't know exactly how that will affect, say, a racing sim vs a flight sim (though I suspect it will be OK for flight and not great for racing). I'm sure I'll be playing with those settings a lot in the coming days.

Next step is to get a platform and seat/wheel/pedals on it. I've read the threads about center of rotation, and I might make a v2 upper frame that lets me recess the seat in it. But for now, I just want to see if I can get it working with a game.

I can't thank everyone enough who chimed in with suggestions and support. I'm especially grateful to:
@SilentChill
@noorbeast
@Thanos
@Xyclone
@Tim McGuire
@RandomCoder
@Zed
@llluis
@Nitros54
@SeatTime

... and of course to @yobuddy and @RufusDufus and anyone else who had a hand in creating the software that makes projects like this accessible to mortals. I don't have much to add to the knowledge base, because I basically used the work and ideas of others, but I will make some measurements and post those details along with my BOM, because that was the information I most wanted when I was thinking about starting this project in August.

Thanks again, everyone!

Thanks, @noorbeast. I'm thinking of giving No Limits 2 a try tomorrow, and I'll do that.

@Squonk, I’ve overlooked reading your thread for a while. Just thought I’d say that now you have arduinos setup to run your sabertooths there’s really no reason to go back to anything else. The problem with the amc 1280 is that in it trying to be all things to all people, instead, its not working properly for almost everyone as it seems to be just to complicated for its maker to provide timely updates to its firmware to fix all its issues. The sabertooths analog control method it uses was really designed to be used with a joystick that has self-centering springs for driving robots etc. anyway. I assume your using my modification of SMC3 for sabertooths now? If so then you don’t have to worry about all the issues people have stated above since it will automatically stop your motors if it doesn’t have a control signal coming in or if it looses it while in operation. The analog method does provide more motor speed resolution over SMC3-SPS but from my experience having a higher motor speed resolution doesn’t provide much if any benefit.

I am, and thank you for that great work! I agree with you, and will stay with my current set up, but will watch to see if the 1280 matures with time.

First test rides on No Limits 2 with human passengers - three kids of varying sizes/weights (youngest son shown) and then me. Still have some learning to do about profiles, but all the welds held, motors stayed cool, no breakers were tripped. My ride terrified me slightly and I'm now sipping ginger ale to try to calm my stomach. So, mission accomplished, I guess!

Next is building a frame for pedals, wheel, shifter and mounting that along with the camera for the Oculus.

Really impressive build! I know you referenced other builds, but any chance you could share your parts list? I got a little confused how parts changed when you went from the AMC1280 board to Arduinos.

Thanks!