Newbie- BAcklash

[servicetech]

hi all,
my first post here.
have a fanuc control on a machine. and getting alarm 421 on the Y axis. The machine has linear scales and when the machine is moved in Y its ok but when the y axis stops i gives alarm 421,
Fanuc's Alarm Description: The difference between the errors in the semi-closed loop and closed loop has become excessive during dual position feedback. Check the values of the dual position conversion coefficients in parameters No. 2078 and 2079.
2078 and 2079 are the same as they always have been.

i suspect backlash is the problem ? any thoughts.
And im not at the machine to test backlash.
i think that increasing value 2079 will get the machine running tempoarily .Would increasing the backlash value param in Y have the same effect. or am i way off track.

thanks

[vlado]

Hi,
You must measure the backlash with indicator and then set parameter 1851.
Write again if you do not fix the problem.

[servicetech]

the customer has changed the ballscrew and bearing, so its not backlash.but still have alarm 421 when the axis stops. the only way they could get it to run was to reduce the rapid traverse from 40000 to 10000.
Any thoughts?

[MysticMonkey]

bearings and ballscrew have been changed OK

Has the backlash been measured ?

Is the backlash compensation value correct ?

[servicetech]

well im told its been measured and set correctly. they disabled the linear scale and measured it conventionally. and seems to only happen after decelleration in rapid. thats why rapid they reduced rapid from 40m to 10m/min.

[servicetech]

backlash is ok and parameters are original.
rapid traverse was reduced from 40m min to 10 and has now been changed to 20 m min. new linear scale. motor coder cable been swapped ampluifier for scale been swapped and is ok. when run at full rapid still get alarm 421.
have found #203 DTE Y=1 when it stops with alarm 421. this is making me think its the pulse coder from the motor. although possible the amplifier.
X axis had alarm 421 3 wks ago. X and Y share a duel type amplifier so, Any ideas. Anyone?

[underthetire]

I have an idea,

Ballscrews and bearings need to be set up correctly. My guess is there is too much bind on the screw at this point and you are getting a diff from command position to actual on at least one of the feed backs. I would check to see if the axis moves freely by hand by turning the screw. Then I would turn off the scales and try the rapid with just the encoder, then if you get the alarm you know its in a drive/motor.

[ben_heinman]

I would try scaling back the velocity loop gain magnification. Make parameter 2107 smaller. Try a value around 50-150.

[CNCZILLA]

I would agree with the binding theory, especially if this has started since the screw and bearing replacement. This is not just simply a parts replacement. The screw could be misaligned or the bearings could have improper load. Besides that, if the screw and bearings were replaced then whoever replaced them would have or should have physicaly checked the backlash with an indicator and then parameter 1851 changed to give you proper repeatability. If this wasn't done properly then it is very possible that this could be all you need to address. It could also be the gibs causing the problem if you are getting what we call stick-slip at lower traverse rate such as stopping. Check the load in the load meter when the axis is moving and when it stops, this will give you an indication of some bindage somewhere.

[servicetech]

the alarm was there before anything was changed. thats why they started to change things, thats why im starting to think its not the mechanics. but still need to confirm.
they also new there was a problem with backlash.
the axis did originaly have backlash but now it is ok. the screw was not changed i have since found out but the bearing was.
The machine doesnt have gibs it has linear rails/bearings.
i will check the load but the screw turns easily even with the motor attatched.
we tried to change the coder but it was faulty, aparently it was a reco item.
thanks again and i will keep trying.

[CNCZILLA]

OK good you know its not gibs. Since you know the bearings were changed but not the screw then it still may be backlash because it could be in the screw. There is a seperate test for this to determine if the backlash is in the screw or the bearings. You have to place the indicator base on the moving part of the saddle and then put the indicator stylus down in the thread of the screw in the rounded part. This way the indicator moves with the saddle and you will show lost motion if any in the screw as you use the pulse wheel to move the axis back and forth. If there is play in the screw then this will be shown in the indicator. I would also be interested in how the backlash is being checked and as you may or may not know if you have scales they must be disconnected before you can check backlash. also you cant accurately check backlash using the pulse wheel by moving it back and forth, you have to write a program and place the indicator in the right place and let the machine run in rapid and see that result and then in a slow feed rate. There will probably be a difference in the two. You set you backlash parameter somewhere in the middle as a guide line then check your repeatability.