Need Help!- strange move!!???

[zavateandu]

hi all
i have a YOUJI VTC 800 with FANUC 18Tand i have a strange move from the Z axes.
if u put a clock for measure on the chuck and then run a small program for the Z axes to go on origin and come back to the clock that is set to 0 i see after 10 times going up and down that the z axes comes with 0.01mm more that on 25% rapid feed.\
if i put on 50% rapid feed i can see after 2 times gets the 0.01mm so after 20 times i will get a 0.1 more.
this i discover cause on the tool change position was too low and was touching the magazine support and after some hour of running i always have to change the second zero return of the Z axes for the tool change.

first we changed the hole driver for Z cause we thought is something electronic but the result is the same.
than ,because we had some noise from bearings of the Z axes,we changed the couple between the z motor and the screw and also the bearings from both ends ,the results is the same.
we checked the backlash for rapid move and normal move is very good.

untill now i can not get it form were can be this movement come from.
as i can see is almost a constant value for every up and down move so i think is coming from somewhere but i do not know form where.
maybe some of u experienced this cuase for me is first time seeing this and is hard to find the cause.

any suggestion will be apreciated i can try many ideas.
HAVE A GOOD DAY ALL!

[tc429]

try the moves in slow feed.

Ive never seen it on a serial encoder yet, but twice in the past 30 years have seen Fanuc incremental encoders lose counts in one direction only(and never throw a disconnection alarm), both times the cumulative error increased with speed. both cases also would return to zero if power was cycled to clear the grid and manually homed. not a common issue though for sure.

the ONLY fanuc ac servo ive ever seen run away was a serial encoder- still dont understand exactly how they work, but know they can run away and maintain commutation with zero feedback errors...had one that was only a few days old run into the stops so hard it brinelled the screw(servo had to be at 2000 rpm or more), we had to literally put a pipewrench and big pipe on it to break it free...sounded like glass crunching. as it moved, was a bear to get apart...

I wonder if perhaps your 'feedgear' parameter might be off- dont know if it could be possible to get like a count of error per rev that doesnt divide out evenly or something...is this a new machine, or a old machine with a new problem? if it was OK till recently, not a parameter issue...has anyone adjusted backlash comp recently? the way the serial servos have different types of backlash acceleration comp, different backlash for feed/rapid, maybe theres a bug in the overcomplicated firmware

if Id have to guess(as youve already verified mechanical stuff) I would expect a flaky encoder to be most likely, but would try both making the same moves at a low rapid rate, and fast/slow G1 rates to verify if it might be a comp issue, and /or speed related...speed related would point to encoder- but only as long as the same speeds in g1/g0 modes showed the same condition...if go/g1 show different(with g1 programmed to same rate as low/high rapids) then I would dig into the backlash acc/dec, rapid/feed settings...
good luck, I'm very curious to see what you find here

[cricricyuf]

how bizarre ...

[underthetire]

So, we are talking Z right ? The quill not the ram? Our Youji has some serious cooling problems sometimes, so we can see growth. Does it come back after sitting with no movement? We also have the 18T control on this thing. Youji did do something funny, when you zero the machine in the program, it looks for the zero return switch every time, and does a grid shift. If everything is absolutely correct, this works ok, but I would think you would see larger errors if that was the case.

[zavateandu]

after chabgng the bearings from the both ends of the screw and the cuopling between the motor and screw we made a full day test and we had this results.
after 1 hour of up/down moves the z increase with 0.05mm and than after 6 hours it did not increase anymore so i think the screw was warming up and than was at the normal temperature of work.
i think i should made a little more the pretension of the screw.
so the problem with the strange move disappears.
thanks a lot for your help and sugestions
see you next time.