Lifespan of contactless / Halls Effect sensors?

[Yardstick]

I'm having a controller nightmare at the moment with components failing one after another.  Having replaced a spiking throttle pot on my Cougar with Rel4y's excellent drop in digital sensor, I then had to replace the X-axis springs on the Cougar base (U2:NXT modded gimbal).  I can live with those failures, as both are mechanical components and are expected to fail over time.  However, within a day or two of replacing the springs, the Y-axis sensor is now playing up.   It calibrates OK but with use it quickly loses the centre and the range of travel drifts seemingly at random with the min and max moving all over the place.  At full mechanical deflection, one minute I might have a registered min of 0%, with max of 90% and a centre at 45%, then seconds later the min is 5% with a max of 98% and so on.  I thought it might be a loose sensor but the Halls Effect sensors used in the U2:NXT are both firmly bolted in place.   

 

I had been lead to believe that contactless sensors were bomb proof but now I'm not so sure.  So I'm left scratching my head trying to work out why the Y-axis is playing up and conclude that the sensor might be going bad.  Does this sound right?  How long do contactless sensor typically last?  If they fail I would have expected them to do so completely not behave like this one is, hence my initial thought that it might be mechanical.

 

Although the sensors are simple to replace and appear to be still in production (Midori QP-2H/QP-2HC Halls Sensor) they are crazy expensive - $100 each + shipping and only available in pairs.  So a replacement is not really a viable option.  

 

 

[RedKestrel]

8 minutes ago, Yardstick said:

I'm having a controller nightmare at the moment with components failing one after another.  Having replaced a spiking throttle pot on my Cougar with Rel4y's excellent drop in digital sensor, I then had to replace the X-axis springs on the Cougar base (U2:NXT modded gimbal).  I can live with those failures, as both are mechanical components and are expected to fail over time.  However, within a day or two of replacing the springs, the Y-axis sensor is now playing up.   It calibrates OK but with use it quickly loses the centre and the range of travel drifts seemingly at random with the min and max moving all over the place.  At full mechanical deflection, one minute I might have a registered min of 0%, with max of 90% and a centre at 45%, then seconds later the min is 5% with a max of 98% and so on.  I thought it might be a loose sensor but the Halls Effect sensors used in the U2:NXT are both firmly bolted in place.   

 

I had been lead to believe that contactless sensors were bomb proof but now I'm not so sure.  So I'm left scratching my head trying to work out why the Y-axis is playing up and conclude that the sensor might be going bad.  Does this sound right?  How long do contactless sensor typically last?  If they fail I would have expected them to do so completely not behave like this one is, hence my initial thought that it might be mechanical.

 

Although the sensors are simple to replace and appear to be still in production (Midori QP-2H/QP-2HC Halls Sensor) they are crazy expensive - $100 each + shipping and only available in pairs.  So a replacement is not really a viable option.  

 

 

Could something like that be caused by a pinched wire or something other than the sensors inside the controller? It seems suspicious that this would all be happening right around the same time. I thought the contactless sensors were supposed to last longer simply because there is very little friction or wear involved.

[Yardstick]

Agreed the behaviour suggests it mechanical and the wires exiting the sensor have to make a tight 90 degree turn - I identified it as a potential weak spot when I first got the modded gimbal and added some electrical insulation tape to re-enforce the wires.  However, I have inspected them and at least externally they look OK.  I have tried wiggling them to see if that produces any movement / spiking but nothing.  The problem with the Midori sensors are that they are a closed unit and you can't open them up to see if anything is awry.

 

The attached pic (not my Cougar), shows how tight things are but there is room even at full deflection for the wires.

[JimTM]

Does the function that you have the Y-axis bound to in IL-2 have a duplicate bind from some other control? Look for an orange icon in the control bind box in IL-2 and mouse over it to find the duplicate.

Edited July 22, 2020 by JimTM

[Mitthrawnuruodo]

1 hour ago, Yardstick said:

Although the sensors are simple to replace and appear to be still in production (Midori QP-2H/QP-2HC Halls Sensor) they are crazy expensive - $100 each + shipping and only available in pairs.  So a replacement is not really a viable option.  

 

With only one axis behaving strangely, perhaps you would be able to swap the X and Y sensors as a test? If the problem follows the sensor, you could be certain that the sensor is faulty.

 

Although I have no experience with these particular sensors or this stick, hall effect sensors are generally very reliable. I wouldn't expect them to fail unless they are abused electrically or mechanically.

Edited July 22, 2020 by Mitthrawnuruodo

[jollyjack]

I modded my warthog throttle (thus an other thrustmaster unit) with this, and the guy selling it seems areal expert: deltasimelectronics

 

[Yardstick]

1 hour ago, JimTM said:

Does the function that you have the Y-axis bound to in IL-2 have a duplicate bind from some other control? Look for an orange icon in the control bind box in IL-2 and mouse over it to find the duplicate.

 

Sadly, its not a duplicate bind - I'm testing it in the TM joystick analyser app.

 

This might help to illustrate the erratic behaviour.  It's a trace of the joystick travel following a restart.  I have first moved both axes to their limits, then moved the Y-axis up and down tracing a horizontal line at the limit of its travel (as you can see, it is all over the place).  Periodically I've moved it left and right as well, since that seems to have an effect on the Y-axis travel. 

 

The more I think about it, the damaged wires theory may be worth exploring further.

 

Edited July 22, 2020 by Yardstick
typo

[FlyingH]

Are yoour wires connected the same way via the orange connectors? If so I guess there is the problem. Otherwise I´m out of options.

[Yardstick]

7 minutes ago, FlyingH said:

Are yoour wires connected the same way via the orange connectors? If so I guess there is the problem. Otherwise I´m out of options.

 

No problem with how its wired, it was modded 14 years ago.  The fault has only just emerged.

[Yardstick]

Well after further experimenting, I am pretty certain that it's a mechanical issue.  The below trace shows me moving the centre axle of the Halls sensor in my fingers (blue line), so I am confident now that the sensor is good.  Moving it by hand the trace is perfectly smooth and moves through the full range of motion.  Once I took the sensor off its mount I noticed that the lock nut was just a little bit loose which might have allowed the body of the sensor to move a little.  So I've tighten that and got rid of some of the play in the joystick output but unfortunately not all. 

 

I now have about 5% play rather than the 12% - 13% that I had previously.  However, I am at a loss to work out what is still moving independently of the gimbal.  The axle of the sensor has a half moon profile which fits into the gimbal very snuggly.  There is absolutely minimal force needed to rotate the axle (obviously since it is contactless), so it can't have worn its mounting hole but I still have that small amount of play.

 

I also tested the wires thoroughly and they are all sound.

 

 

Edited July 22, 2020 by Yardstick
typo

[DD_FT-]

Did you test the wires while moving the stick?

[Yardstick]

Yes and moving them independently - no spiking or deviation.

[Sokol1]

The issue start after you replace springs - with different type than early used (since is not longer produced).

You have tested sensor without springs, and work OK.

Try test with some spring bands instead the new springs, for assure that are not a play in the gimbal.

 

Seems the issue is with the new springs pair, may one are more strong than the other... or something with their "hook" point.

 

 

 

 

[Yardstick]

Update: 

 

Having spent hours with the joystick base opened up, I think I have worked out the problem(s).  Hopefully setting it out here with my possible fix, may help someone in the future. 

 

The U2: NXT is a clever design and it probably as good as you can get in a non-cam based system. Each axis has resistance applied by a pair of springs, which also act against each other to centre the stick.  However, since some sticks may have a centre of gravity that is off-set from the vertical, the U2: NXT also has additional arms that can be independently adjusted to ensure that the stick returns to an upright central position when released.  These adjusting arms are made of the same high grade aluminium as the rest of the gimbal mechanism and each is secured by a single M4 socket head machine screw (see pic).  This is both the beauty of the design but also a potential weakness.  When secured the gimbal arm only moves about 15mm over its full range and in my case the Y-axis seems to have worked loose.  The movement was very slight, probably only 1mm or so and seems to have been present even after I tightened the screw although it was impossible to see. 

 

So finally after considerable testing, I eliminated every other possibility and returned to this arm.  I decided to use Loctite on the screw as well as the surface where the adjusting arm contacts the main gimbal arm.  I tightened the screw as much as I dared (steel screw into aluminium) and gave the Loctite a few hours to cure.  Now 3 days down the line I have a stick that stays reasonably centred and what drift from centre I do have, may be down to the having stock springs fitted on the Y-axis and aftermarket on the X.  Although the springs are a reasonably close match, they are not identical.  

 

The other thing I discovered was some wear in the gimbal bearings (press fit cartridge type).  The horizontal play that this allows is small but twisting the stick does induce a few degrees of movement in the Y-axis.  I could try and replace these but finding an exact match may be more trouble than it is worth.  When using the automatic calibration this can allow the joystick to register travel that takes it past the normal extremes of movement and can cause centre point drift.  The fix I have found is to use manual calibration, which saves the calibration to file and does not register extremes of movement past the pre-defined calibration.

 

So I now have working stick and thankyou to everyone on this and an other forum who came up with suggestions.  However, I've decided that relying entirely on a 15 year old piece of kit is probably not the most sensible approach and have ordered a VPC WarBRB.  Hopefully with the extra heavy spring option I can get a feel that is close to the U2:NXT.

     

Edited July 26, 2020 by Yardstick
Formatting