Servo control of a stub turnout

[alhoop]

I want to use a 9g micro servo driven by a Picaxe-14M to control a stub N-Scale 3way turnout.
While the two endpoints are position repeatable the center is not due to servo backlash(?). If I approach the center position from the same direction the position is repeatable. Is there any way around this?
Here is my code which reads a 1 pole 3 position switch and uses case statements. The highs out control relays that provide correct track polarity to the three frogs.
Thanks for any help or new code.

 

[bluejets]

How about using an optical sensor as an input, (slotted type using a vane)such as is used in printers, copiers, all sorts of equipment. This way the servo would drive to the exact spot regardless of backlash. They are very accurate and I have used them many times. There are a few different types and I usually set them up with an additional output transistor for inverting.
If you want details, call back.

 

[inglewoodpete]

It would be wise to look at all possibilties before going down the sensor-and-software path. This sort of solution can give you a lot of work.

One alternative would be a mechanical solution. A detent track/spring may be simpler to implement. What method is used on a manually operated pointset (turnout)?

Another solution could be a small stepper motor.

 

[westaust55]

If as you indicate the servo has the desired repeatability from each side then can you not just use a different value from each side/end as the mid point?

 

[allison1947]

Can you send the servo to one side eg "left side" then send to centre? That way it will always be approching from a known side.

case 2 'center
low 2
low 4
high 3
servopos 5,75 'go left first
pause 100 'pause long enough to get to "left" from any position
servopos 5,150 'now go to centre position

It may work.

 

[alhoop]

My bad - my problem was caused by not enough current from a 5 volt supply to reliably drive the servo.

More Info:

I am using the Picaxe-18M2 with the attached code.
Without the pause 1000 command I was getting servo jitter about every 2-3 seconds. The pause command seems to have fixed it - I don't know why.
Servo can onlly be used on port b and I am using b.4,b.5 and b.6 for switch inputs and b.1,b.2 and b.3 for outputs with servo out on b.0.
When I read pins b does that effect servopos out on b.0?

Al

 

[alhoop]

Servo contro of a 3way stub turnout

Guess I need to bump the above post up.

 

[hippy]

Guess I need to bump the above post up.

I was wondering why, re-read your post and noticed the question tucked at the end. Sorry for missing it !

Servo can onlly be used on port b and I am using b.4,b.5 and b.6 for switch inputs and b.1,b.2 and b.3 for outputs with servo out on b.0.
When I read pins b does that effect servopos out on b.0?

Reading the pins shouldn't have any affect on the servo output.