Industrial joystick repair challenge, tips needed.

[tru3533]

Got an industrial joystick that is defective.
The joystick was replaced with a new one, so this joystick has a fault.

Testing the joystick by applying 12VDC to the terminals J1, J5

Measuring the output, terminal J3 with a Fluke, no load, with the following result:

Joystick in neutral = 6,0V       OK
Joystick full forward = 9,0V    OK
Joystick full backward = 3,0V OK

Adding a load to the output with an resistor 1,2K  between J3 and power supply ground.
resistor draws 4,9 mA on 6V
new measure in circuit and output J3,  new values indicated in (). See included drawing.
Joystick in neutral = 4,25 V
Joystick full forward = 4,25 V
Joystick full backward = 3,0 V

Is it correct/OK to test this way?

I have replaced LM258, LM317LM and relay, any idea what can be wrong?


Drawing is reverse engineered by me, with hopefully no faults.
Resistors are measured on PCB.
Capacitors not tested.
All diodes tested.

Any tips highly appreciated!

 

[Armadillo]

Try measuring the 200ohm[located on plate?] value or bypass it with another resistor.

[tru3533]

The resistors on the Ceramic slider plate seems OK.
Armadillo: Shorted out the 200 ohm resistor with 100 ohm resistor, sorry to say no difference.
 

[stj]

you arent supposed to put a load on such things.
it's junk btw, if you need industrial reliability then non-contact sensing is the way to do it.
usually using linear hall effect sensors and x&y magnets.

[Armadillo]

The resistors on the Ceramic slider plate seems OK.
Armadillo: Shorted out the 200 ohm resistor with 100 ohm resistor, sorry to say no difference.

Now how did you bypass the 200 ohm resistor?

Please plug in the module, measure the ohm from Pin 4 of LM258 to anode of D5. What is the ohm?

[tru3533]

Resistor was bypassed by soldering a 100 ohm resistor over pin 2 and 3 on the ceramic slider plate. (counting from left)
Notice that the ground level is not changed when applying load.

measure the ohm from Pin 4 of LM258 to anode of D5. What is the ohm? 800 ohm
K1 coil measure is 500 ohm.

[Armadillo]

Resistor was bypassed by soldering a 100 ohm resistor over pin 2 and 3 on the ceramic slider plate. (counting from left)
Notice that the ground level is not changed when applying load.

measure the ohm from Pin 4 of LM258 to anode of D5. What is the ohm? 800 ohm
K1 coil measure is 500 ohm.

Hi, your pin 4 is the Actual Ground for LM258, elevated by relay ohm and the 200 ohm. So when you say you put 1.2K resistor at J3 to ground, what ground are you using,  cannot use power supply ground? The elevated ground itself drop somewhere 3v.

And if coil is 500 ohm, why you measure 800 ohm [ 500 + 200 = 700] or [500 + 200||100] = 566.67 ohm]. Not that I am petty, but I suspect some poor connection, dry joints of the power to the IC.

Please also check your 12v power supply by putting some load to it. Test it to write off the suspect for the power supply first to reduce complexity.

If you are able, change the LM258, start over measurement using pin 4 as ground.

Oh yeah, while you at it with the load, measure pin8 to pin4 of LM258. We see why voltage drop so much.

[tru3533]

Actual ground for LM258 pin 4 is anode D1,D3 elevated by the RSlider 200 ohm.
Actual ground for LM317LM pin 4 is anode D1,D3 elevated by the K1 500 ohm + R15 100 ohm

LM317LM is not powering the circuit, only function is to supply stable power for feedback of potmeter

Measuring pin 4 LM317LM to pin 4 LM258
K1=500 ohm + R15 100 ohm + RSlider 200 ohm =800 ohm
As you see ground levels is not the same.

I'm using my trusted old HP lab supply for this measurement, so supply is OK
Bridge D1,D2,D3,D4 makes the plug to the joystick reversible when you power up the joystick.
Ground level will always be D1,D3 anode.
Power + level is always D2,D4 cathode.

On drawing you see Power+ level = 11,77 V. (12V supply - D2 drop) = pin 8 LM258
No drop on power level

LM258 is changed.
Resistors R10, R11 is removed from circuit, and measured = 2,5K each

[Armadillo]

I relook the circuit, the ground should be marked in Red to Pin 4.

Measure all from Pin 4.

From Pin 4 to Anode of diode D1, you should actually only read 200 ohms not 800 ohms.

[Armadillo]

On drawing you see Power+ level = 11,77 V. (12V supply - D2 drop) = pin 8 LM258
No drop on power level

no change because you measured wrong ground away from the suspected resistor.

Measure Pin 4 to Pin 8, very specific.

Measure ohm from Pin 4 to Anode of D1, cannot be larger than 200 ohms, very specific.

[tru3533]

This is correct 

Your question was "measure the ohm from Pin 4 of LM258 to anode of D5".
So I guess this was an typing error.

[tru3533]

I made new drawing to simplify the problem.

Looking only at DC level

[Armadillo]

The challenge is not the circuit.

[tru3533]

Interesting...
But this level on pin 4 is not available on the plug for the joystick.
So how is it possible that the joystick can send out the correct value? and be read by the controller?

I know for a fact that the joystick is defective.
It has been replaced by a new joystick and then everything works.

[tru3533]

Well back to the drawing board :-)
I think it is possible to connect the load resistor as I did.

I just watched EEVblog #600 - OpAmps Tutorial - What is an Operational Amplifier?
Here Dave demonstrate LM358 who is in the same group as LM258

The reason why pin 4 is lifted can be seen at @43:22

 

[Armadillo]

You are confused with the virtual "INPUT" ground and the power ground.

"NO VOLTAGE CAN BE LESSER THAN THE VOLTAGE APPLIED AT THE GND PIN"  Specifically for LM258, no voltage can be lower than Pin 4.

[tru3533]

OK, understand. Thank you for holding out 

So the big question is now how can I test the joystick measuring only on the terminal connector?

By measuring only with Fluke it seems OK. But it is broken.

[Armadillo]

Try focusing on somewhere else, for example the jack connector, mechanical side mating, could those be loose, wire ends etc....

You have 2 of them, can make the comparison. Try plugging it in to see any loose contacts or something like that.

I am sure you can find it out. Good Luck with the repair.

[tru3533]

Found the obvious solution, the circuit works with mA.
with reference to ground.

The joystick circuit is constructed so it gives out 0 mA in center position and 20 mA range in each direction.
Failure was misalignment, with overshooting on the positive side.

Took some time, but all in all learned something.