Originally posted on the 1C forum here but Sokol advised it to be posted here too. There are some awesome builds on this site - very humbling!

The X52 Pro has been relegated to the cupboard since I’ve now finished my custom u-joint stick (except for future modifications, of course!) It is a standard universal joint arrangement with Neodymium cube magnets and hall sensors although the centering mechanism is, perhaps, a little unusual in that it is based on magnetic repulsion. If you have seen this done elsewhere please let me know since I would like to link to it.

From Sokol:
- http://www.sukhoi.ru/forum/showthread.php?t=59583
- http://translate.googleusercontent.com/t...yWA#post1625474
- http://www.sukhoi.ru/forum/attachment.php?attachmentid=135284&d=1306861313

I decided on magnetic centering since I do not like the friction of a spring. I did not want to tighten the bearings to stop the stick falling over since I like movement to be as unrestricted as possible for fine control in the centre. Magnetic repulsion centering allows for frictionless movement (apart from the u-joint itself) with resistance that is almost zero in the centre but increases markedly toward the edge. The magnets were bought from [url="http://www.aussiemagnets.com.au"]http://www.aussiemagnets.com.au[/url].

Note that this is a relatively low force centering mechanism. It centers the stick fine but does not provide kilograms of resistance.

You can click on the photos for larger versions on Flickr.






Stick construction:

1. Bought a used u-joint from a quad bike (a little smaller than a car u-joint) and knocked out the spider.
2. Stripped and regreased the needle bearings and drilled out the bearing cups to accommodate a 1/4" bolt.
3. Reassembled but only loosely tightened the bearing cups so the movement is extremely loose.
4. Glued (araldite) some cut-down 1/4“ bolts into the spider as pictured below. These provide the platforms for the two axis cube magnets (10x10x10mm N42 Neodymium) which I also glued in place.
5. Reduced the play of the stick by judicious use of some Selleys 'Knead-It Steel' metal epoxy ([url="http://www.selleys.com.au/putty/epoxy/knead-it-steel"]click here[/url]) and thin gal plate strips. I chose this method since it is easier to make slight adjustments with the epoxy putty than when welding.
6. Aligned the axis magnets as described (thanks very much for the guide) and added the hall sensors. These were two old Allegro UGN3503UA although the newer Allegro A1321EUA-T could be better.
7. Hooked the axes up to one of Leo Bodnar’s BU0836X cards which I will also use for the control box I am building.
8. Reversed the inevitably backwards hall sensor on the x-axis!
9. Replaced the circuit board in an old Logitech Extreme 3D Pro stick with a custom circuit (a few extra buttons) and used this for the handle.








Centering mechanism construction:

Materials were a galvanised steel plate as a base for the stick, 100 mm square tubing and twenty 20x20x5mm N42 Neodymium magnets with 9.2 kg pull. These could be stronger if you want more resistance but N42s are already really challenging to place correctly.

10. Welded the angle brackets (well, bits of cut tubing) on the 90 mm high square tubing. This section is a guard for the hall sensors and a platform for the centering ring to sit on.
11. Welded the 20 mm high square tubing at a 45 degree offset to the guard section.
12. Glued (araldite, clamps and a lot of bad language) the magnets in place, all with north facing to the centre (doesn’t matter which orientation of course as long as they are all the same).
13. Glued the magnets in place on the stick itself in an opposing polarity to the ring magnets (i.e. stick magnets repel ring magnets). Alignment is important: you want the magnets on the stick to be as close to the magnets on the ring as possible at full stick extension so they will actually be slightly higher on a centered stick than on the surround ring. Note that I had to make a platform for the magnets to sit correctly on the stick shaft. More epoxy plus 20 mm square tubing although I could have welded this.








Assembly:

14. Drilled out the holes in the plate to accept both the stick and cage bolts.
15. Bolted the u-joint stick to the plate.
16. Slid the cage down over the stick and bolted it to the plate.
17. That’s it. I have the stick attached to my chair using more square steel tubing of two diameters which slide into each other.

Tips:

If anyone wants to try this design here are a few tips I learnt the hard way:
- Attached the magnets last! These strong magnets attract steel shavings or filings like you wouldn’t believe and are really difficult to clean up.
- You will need to glue the magnets in place over a couple of days since placing adjacent magnets is pretty tricky. Use clamps (plastic) to hold them in place if need be.
- After trial and error I made the bolt holes used to attach the cage to the plate significantly larger than the bolts since I needed some fine position adjustment to get the stick to sit dead centre.
- Sliding the cage down over the stick is difficult since the magnets on the cage ring attract the stick shaft. Best to wrap the shaft in something thick to reduce the attractive force (I used a piece of drilled out wood).
- You need an awful lot of force to push the bearings from the spider. The good news is that the cups are strenthened steel so you are unlikely to damage them with a little ‘encouragement’.
- You’ll notice that I added some extra 10x10x2.5mm magnets on the stick (see the photo). These just increase the resistance a little.

I am really pleased with the outcome. The loose centre is just great for fine adjustments when lining someone up but you don’t lose out by not having centering. I had decreased the strength of the X52 Pro spring by cable ties but the effect is nowhere near as good as this since the spring tension increases rapidly at a particular point. I will, no doubt, replace the stick handle at some point although I have to admit that the internal build quality of the Logitech is actually better than the X52 Pro.

Finally, apologies to anyone offended by my rubbish welding skills but this was my first attempt. My rudder pedals look much better since I was getting the hang of how to build up the weld joint but that's another story!

Very interesting centering idea. Keep it distant from magnetic media and put a sock on it, otherwise seems like the optimal centering setup for stuff like helicopter cyclics. If you want to remove the last of those iron filings try using some adesive tape, or some putty to inglobate it and pull it away.

Good ideas to get rid of the filings, thanks. I actually thought the magnets would disrupt the hall effect around the cube magnets but they must be distant enough in this design since there's no discernible difference in response with or without the cage.

Nicely done! Thanks for posting it here.

Very inventive idea, jimbop! I'm with you that I would have thought the magnetic centering construct would have affected the hall sensors...I'm guessing the proximity of the hall sensor to the neodymium magnet (and it's strong field) prevents this??

Thanks for posting!
Derek

Yes, I think the Hall effect is weaker (or the sensors less sensitive) than people generally expect. The first thing I did was test this by just getting one of the 20x20x5mm magnets and waving it around a few inches away from the sensor. When I saw there was no response I figured the idea was worth a shot.

I'd almost be tempted to build a new base that used a toroid electromagnet to give the centering mech some real fight to it....

A shield above the hall effect sensors should damp out any extra oerstads that get in there.

g.

How do you shield from a magnetic field?

with a piece of magnetically permeable material. It will "shortcut" the field lines

Some experiments with magnetic as spring:

http://translate.googleusercontent.com/t...yWA#post1625474

Movie:
http://www.sukhoi.ru/forum/attachment.php?attachmentid=135284&d=1306861313
In spite the horizontal position of handle, this is a joystik: "...is my personal opinion but for me it is more convenient, the hand doesn't get tired because the palm rest over the handle..."

Sokol1

This magnet-spring solution is very innovative and inspiring. Actualy I was close to finish the design of my stick mechanism with extension springs, but I decided to rework it with magnets.

There is a lot of advantage here:
- Very compact size
- Easier force adjustments
- No material restrictions like with the springs (max/min extension of them)
- No spring wear out, break

There are two main issues here to solve:
- The magnets should have a perpendicular travel to use their max. spring effect
- Magnetic shielding of the sensor units, and the whole base unit.


IMHO Combined with rotary fluid dampers can result a fantastic stick mechanism.

Thanks for the feedback, everyone. Definitely some work to do in order to optimise as VO101MMaister points out. Things I would try include stronger magnets (electromagnetic ideally which would also make placement much easier) and correct alignment. You are obviously correct in saying that the alignment should be perpendicular relative to the closest position, just a little more difficult to do in practice.

Not convinced that shielding (or, more correctly field diversion since you can't block a magnetic field) is necessary unless you are using stronger magnets since the Hall effect of the centering magnets does not seem to be detected by the sensors lower down. I was certainly surprised at this. Wouldn't hurt to add though, especially if you were using strong magnets.

And it would be awesome in combination with dampers!

Now if you wanted to get really creative, you could use 8 electromagnets and 8 neodymium "target" magnets to create a dynamic force feedback system.

The PWM system and software is of course left as an exercise to the reader.

g.

I had another idea (besides the use of electromagnets) occur to me when I first saw this.

If one wanted to have a Force Trim capable stick (ala DCS BlackShark), it would seem that making the fixed magnet ring (or electromagnet ring, if that can be worked out satisfactorily) able to move (in the X and Y axis) when desired, and locked down when desired, could make that feasible.

For those not familiar with Force Trim systems such as the KA-50's, a trim button is used to shift where the stick is centered. It doesn't change the maximum possible deflection of the stick in either axis, but it does allow you to place the stick where you want it to stay, Trim it out there, and then be able to take your hand off of it and have it stay put. If you trimmed it at half way up Y, and halfway over on X, the stick would be outputting those values, but the centering forces would now be holding it in that physical position. You could still move the stick from there as it is not physically locked in that position... it is just being centered there.

Having thought about it a bit, you'd probably have to make the magnetic ring physically larger (spaced further from the stick) with stronger magnets to compensate for the increased distance. The reason being, that you don't want the magnetic ring physically preventing the stick from being fully deflected, no matter what position it is in when trimmed.

As a possible alternative to having the outside magnetic ring change position to effect trim, if one fashioned an electromagnet-based outside ring, and had the elctronics know-how... it might be feasible to keep the outside magnetic ring fixed in position, and just electrically vary the magnetic field to achieve the new trimmed center position. This would basically entail the same, or similar electronics that Gene refers to for accomplishing force feedback, although I suspect it would be less complicated.

This is entirely feasible and I almost implemented it in a crude manner in the form of a horizontal bolt system that can shift the position of the ring. A few mm is adequate to change the centre point of the stick. You could do this with a quick release clamp system or a rail mechanism that supplied adequate friction.

I would be very happy with a simple elevator trim. The magnets could be moved relatively to each other to create the effect.
I feel the electromagnets are not easy to work with. They have high consuption, they often have a limited runtime (they must be switched off after a while), and they must be cooled!

Great stick and thread! Some good ideas indeed.

Yeah, I found the same out about electromagnets today. Possible but more difficult. Stronger neodymium magnets are available anyway but I am happy with weak resistance as long as it centres.

The easiest way to do the trim would be to adjust the lower cage relative to the stick. Pretty straightforward I think.

Well my first enthusiasm is faded a bit. I did some more research.

Tha main problem, that the distance between two magnets and the repulsion force is not linear at all. Actually it is seriously parabolic. Have a look at this:


On the other hand the repulsion force between two identical magnets is much lower than the pulling force. Look at this calculation page:

Click here

So to have a considerable stick force on a full size control stick, we need super strong magnets (what cost a fortune) and very short travel between them to stay on the more linear part of their repulsion effect.

Well, let's examine that a bit...

How many Kilograms of force do you need / want? (Multiple magnets along the same vector can add more force.)

How much movement do you need / want? Keep in mind that the closer you place the magnets to the pivot point, the lesser the amount of physical travel is. Of course, if you get really close to the pivot, you may need to look at using something other than direct mounted hall sensors for the measurement of the stick movement. That could be either potentiomenters, hall sensors connected via linkage or gears, or ???

This is why I suggested using electromagnets. You can get some insane force levels out of a good, hand wound electromagnet.

g.

Well, I would never say never...:)

Could you recommend some good manufacturers?



CyBerkut,

I would like to have min.3-4kg (at max deflection) on a 500mm long stick. Required deflection is +/-20deg.

I have an idea how to maintain the distance between the magnets, I will make some sketches and come back with it later.

I've never purchased a commercial electromagnet VO101MMaister, I've always wound them by hand. I think Google would serve you well here.
g.

Aren't we discussing about reinventing the wheel, though? You can get pretty good force out of an electromagnet attached to a shaft and perhaps geared down... hey, what about just using an electric motor? There's a few interesting links here: http://www.simprojects.nl/diy_force_feedback.htm
I think I'd just use a couple of servo arrangements to move the centering device, limiting the maximum force to that provided by it.

Yes, you could be right. That link is interesting, thanks. As I mentioned in OP the magnet-method is low force and I am happy with this since I always prefer low force on my sticks anyway. There are probably easier ways to achieve kilograms of force if desired.