Build thread for my 3 dof simulator platform

Looks great, I look forward to you finishing your project

Today I secured the MMs and arduino to a MDF-sheet and rewired all the connections to check that everything with electricity works. I will still put some northbridge coolers to the MMs (I cut those too today, but didn't install the coolers yet). Of course I need to disconnect the motors and PSUs to be able to work with spring bases and anti twist etc.

Also I installed one potentiometer to a motor and half of the required gears. I still need some small diameter hose to connect the pots with my gears (if my idea will work).


Next step will be the bases for the springs and anti-twist system, but for those I need some 4"x4".

EDIT: the hose connection of the small gear to pot didn't work.
Actually it was quite obvious when I installed it that it wouldn't work as the gear itself was not fixed, so it kept skipping some teeth.
Or the connection worked as I intended, it spun the pot and in case of a motor runaway it would just slip protecting the pot, I just need to figure out how to fix the gear to something.

Note to self:
Don't use cheap chinese pots.

Soldered 7 and 3 of those turned out to be readable by arduino code after soldering and only 1 worked with other pots in SMC3, as the two others confused the arduino and caused some kind of shorting. Luckily this time I didn't burn anything. As soon as I changed the 10turn pots back everything worked flawlessly.
I think I'll go with 180 degree hall effect sensors.

Almost a month since I wrote anything the last time.
The reason for the silence is simply that I haven't have that much time nor motivation to continue with the build.

Stuff that has happened with the build:
- Stiffened the motor bases with wood blocks and screws.
- I received the Hall effect sensors ( I should still solder them, well I soldered one, but didn't install it yet)
- Got some on/off switches for the PSUs (soldered the wires to them so that I can install them next to the seat, to be able to switch the powers from there if something goes wrong)
- Installed the bases for the springs on the bottom plywood, still some misalignment with one of the springs: it seems straight when there is no load, but with load it tilts ( have to adjust it a couple of centimeters)
- Installed the 3/4 bases for anti-twist



Stuff to do:
- Correct the spring
- Finnish the anti twist and modify the top plywood so that it wont touch any of the anti-twist "bases"
- Install better cooling system to the MMs (almost burned them with 1/3 load while trying to load the springs. Luckily thermal throttle hit the spot. I had small heat sinks and a case fan. I threw the heat sinks away and modded the northbridge coolers for MMs. I have to make the wiring for them and attach them.
- Solder and install hall effet sensors
- Modify the top plywood so that while tilting to the max positions nothing will touch it (now front motor will slightly touch if front is down and back is up)
- Try the whole thing with motion and hope that it will work. (And then fix something so that it will work even at some extent.) I do fear that my mechanical build is not good (just a hunch) or the motors are not strong enough.
- Build and install some HOTAS holds.
- Build and install something to hold a Vive basestation.

I understand; I go like the devil when I have the urge and then lose interest for months at a time. Great project, I look forward to more updates (when you feel like it, lol).

Some stuff I managed to do today:
- MMs northbridge coolings are now installed.
I accidentally dropped some of the thermal compound on the MMs chips. Took a while to scrape it out between the pins. I cleaned the rest with alcohol and a cotton swab.
Also as I fixed the northbridge coolers in place I managed to almost break one of the coolers while making the fixing irons / installation thingys. Luckily it was rather easy to bend the steel plates inside the sink back to original positions.

- All hall effect sensors are in place and working. I used clear hose and hose clamps to attach them to the shaft. This way the coupling will be flexible.
I was just too tired to realize that 2/3 of my motors are installed so that I have to reverse the leads of the motors.
I started to make some changes to the SMC3 code to reverse the motor directions, but fortunately that is not necessary.

As soon as I have reversed the leads of the 2 motors everything on the electrical side should be complete.
I still have to make sure that the northbridge coolers are contacting the MMs effectively.
At least the one motor that was already moving to correct direction could not overheat the cooled MM. We'll see how they will work under real load.

I'll post some photos probably on sunday as I won't have time to work on the project before sunday evening.
The project is progressing slowly as I don't live even in the same city where I build the project

I installed LFS and SimTools to verify that I get them working with SMC3, which was pretty easily done.

I switched the leads and everything worked just fine.
Until I put the top platform to the system...
A few points observed:
- Springs: will cause oscillation and as the worm gear boxes do have a little loose that will cause the hall effect sensor to grab that oscillation.
- As I use ball joints on both ends of the levers the top platform is highly unstable. Basically the springs are the only thing keeping the platform in place (and off course the anti-twist system, but I didn't install that yet).
- The motors are strong enough to move me and the chair on the platform with the springs at least when enough power is used.
- One of the hall effect sensors started clipping, probably a dead lead on the GND.

As I started stress testing the system (I sat on the chair, only thing missing would have been the wheel and pedals) everything seemed some what acceptable. Except for the instability when the springs were extended.
I started to make more aggressive movements with one of the motors: I noticed immediately the oscillation. Bad thing about that was that the MM was pushing full power through itself and as the oscillation only occurred after fast movement, I managed to burn one of the MMs No smoke but I could smell it easily.
The MM was properly cooled, the thermal compound well spread and the contact to the cooler seemed to be good. No external damage could have be seen after I unmounted the cooler.

Right now I'm pretty unmotivated to continue, as I struggle to resolve the problems that I'm having.
Mainly the oscillation and instability, as the springs are the main culprits and the motors won't be strong enough without the springs.

As far as oscillation is concerned, try increasing the deadband in SMC3 utilities. Also, you may have a bad solder joint(s) on you hall pot(s). You can get the leads pretty hot without damaging the pot so make your you don't have and cold joints. I have had some joints where the wire looked well attached, but I would have erratic behavior. Re-soldering has fixed this problem for me more than once.

Thanks!
I'll try the deadband increase as soon as I have time to tinker with the build. Probably next week.

Today I removed the burned MM and soldered a new one. Also I fixed the pot that was malfunctioning, there was a cold joint.
The pot on the motor that did oscillate is completely fine. It' s (my build and) the SMC3 parameters that cause the oscillation. I ordered some rosejoints to minimize the looseness in the levers.
I didn't have time to test run today more than that that the MM is working.
Accidentally almost burned the new MM as I rewired it wrong.... by mistake turned 180 ° one of the bundles of leads that I made for MM and that basically commanded the MM to push + and - to same output, kind of internally shorting the board. Luckily the thermal shutdown kicked in. After rewiring the MM it seemed work just fine (tested with motor and to both directions, not with load yet)

Hi again,

I moved the project to my place if I had more time to work with it during evenings.
I started to investigate the oscillation problem and found out that motors 1 and 2 are affected by it and motor 3 isn't, or at least the oscillation of motor 3 is negligible compared to motors 1 and 2.
The three motors were supposed to be similar, but the motor three has a bit different label (27A motors 1 and 2 are 25A) and the arm is on left hand side as motors 1 and 2 arms are on the right hand side.



You can see from the photos that the motors 1 and 2 tend to oscillate and draw current spikes and motor 3 doesn't do that.

Do you have any suggestions that I should do?
The looseness of the gearboxes feel pretty similar with all the three motors.

Increasing deadband to the motors does help but that doesn't solve the problem.
It feels like one of the motors is a little loose from the gearbox when it changes direction, at least the motor moves more than the others (the gearbox stays still as it is fixed to the plywood), but when tried with hands the motor seems to be well fixed to the gearbox.

Maybe to get a bracket for your motors?

Thanks for the input, but i don't believe that it can be that as the three motors are (almost) identically installed. The motors themselves are only connected to the gearboxes, which are secured to the plywood with brackects and blocks of wood.

I believe that it has something to do with the internals of the 2 motors or the connection to the gearbox.
I was thinking to remove the motor from the gearbox and see if there is anything loose.
I do know that worm gearboxes do have inherently some looseness, but it shouldn't result to this...

Of course there is the option that I have to further fiddle with the SMC3 utility's control values if the two other motors do need different tuning parameters (for example smaller proportional (Kp) value, which would make them slower but less likely to overshoot), but the oscillation is so severe that I wouldn't believe that would cure the problem.

EDIT: While reading about the same problem that other people have seen, I believe that I have to recheck my pot mounts to make sure that they are equally stabile for all the motors.

@E1m0 could be your pot mount but could be wrong settings in SMC3 take a look at this post http://www.xsimulator.net/community/threads/a-laymans-analogy-guide-to-pids.7649/#post-91827 it as made for PID tuning a JRK but you will understand how PID works on a ardino too.
as you are using SMC3 you can monitor your motion while playing a game as described here http://www.xsimulator.net/community...-and-windows-utilities.4957/page-2#post-49056
http://www.xsimulator.net/community...-and-windows-utilities.4957/page-3#post-49624 without building anything.
and one other way is doing this http://www.xsimulator.net/community/threads/simple-digital-scope-for-motion-sim-testing.5485/

Thanks!
Actually I tested my computer connections and axis/dof settings etc. exactly like that: using udp connection with LFS without plugging the sim in.
And I run a lap of LFS with motors plugged in. There was very severe oscillation wit hthe motors 1 and 2, motor 3 worked fine.

I will have to do more testing with the PID values. I hope that it will solve the problem eventhough I'm not that sure....

I the oscillation was (all but) completely removed from the motors with change of the PID values.
I had to put the proportional value down to 210 and integral to 2 to get rid of the oscillation, for motors 1 and 2.
For motor 3 I could have much higher values and it tracks a little bit better the set point.

I still have to try that with load.

Here is a video of my motors working with just the top plywood and springs.
You will see that the platform is wobbly, that's because of the lack of the stabilizing anti-twist that I have not installed yet.
This is just for you guys to see some of the performance at this moment.



I didn't run the motors all the way down, because there is no load at the moment and the back motors (2 and 3) would have had to work against the springs and I believe that the springs are a bit too high tension without any weight.

I need to change the length of the rods connecting the levers as now the middle point of the motion is not where the highest torque is (90° rule) and also some of the hardware will hit the plywood if I bottom the top platform.
Also the pot of motor 2 (they are in chronological order in the video) is shaking, I need to secure the base of the pot (or actually just make a new base).

Looks like a good step in your work!

Do you know the spring specifications? (or did you got them from recycling?)
it should be interesting to calculate the max and min force in your setup

The springs are these: http://www.ebay.com/itm/Clutch-Pull...asy-Upgrade-/191872124741?hash=item2cac784745
According to the seller:
---------------------------------------
Spring lenght is 155mm and a spring tension of 2000n. (although I dont really understand the tension unit begin newton and not newton/meter....)
Dimensions

• Overall length: 159 mm (6.25 inches)
• Greatest width of coil section: 58 mm (2.28 inches)
• Thickness of wire: 4.5 mm (0.17 inches)

-------------------------------------------
I also ordered some of the smaller springs if these are too high tension, but changing the springs would also require change the height of the spring supports.
The springs are placed so that there are 2 springs 20cm from the COG towards the either end and as the front motor is working by itself it it twice as far from the COG than the back motors.
Also the front springs are closer to the center as there are no motors on the sides in the front, the roll effect would have more "pitch" to it if the springs were on the sides.

When the whole top platform (20kg) is loaded (with me (80kg) and wheel (10kg) and seat (10kg); roughly 120kg) the springs are compressed down to around ~5cm from the ground.
The CTC is 6cm (or I can change it to 4cm). Right now I cant use a bigger angle because of the rod lenght (calculated it wrong when I didn't have the anti-twist bases) and I should increase the rod lenght with at least a couple of cm. Probably also the springbase height is a limiting factor atm.

EDIT: If I use 90 degree angle the motion will be 8,4cm total, which is higher than the max safe travel "Potential True Maximum Travel w/ Longer Free Length, True Travelmax : 72.630 mm"
I'm not sure about that because I don't know the exact material of the springs.
Using this: http://www.acxesspring.com/maximum-spring-compression-calculator.html

your springs are 20cm long and can be pressed between xx to 8,4cm

I think the longer, the better. If you use shorter springs, the effort will vary too much between high position (not enough force) and low position (to much force)

I've studied the gas springs a while ago... I'll post something about it soon

Ok I actually ordered 3 pcs of gas struts in case I don't get these working the way I like.

I don't really understand what you mean by 20cm long?
If I measure them they are 16cm long.

By the 8,4cm I meant that with my CTC (6cm) and total movement of 90° the traveled distance will be 8,4cm ( one direction: sin(45°) * 6cm = 4,2cm )