VMC won't hold position in Y after shutdown

Interesting problem: When I fire up our old (1996, Fanuc O-M control) Leadwell 610AP VMC in the morning, it is always off in Y by a little bit. Sometimes it can be off by thirty thou and sometimes only a few, but y never ends up in the same position as it was the night before. Once I reset the Y work offsets it makes good parts all day long with nice surface finishes, so I don't think there is anything loose mechanically.

Now, for a little background on the machine: A year ago, the Y ballscrew seized up and was replaced with a factory new one. The first problem was that the new ballscrew was a different pitch, so we changed around a bunch of parameters and at least got the machine to work properly. The second problem was that the new ballscrew was 20 years old and was "sticky," so the surface finish on the parts was unacceptable.

So, out came the "new" ballscrew and in went the old (albeit reground and rebuilt) ballscrew. Parameters were changed back, the machine makes good parts, and all appeared right in the world. Until I fired up the machine the next morning.

Anyone have any ideas as to why this could be happening? I have a theory that the machine is "relaxing" overnight and when I turn it on in the morning the servo motor doesn't engage in the same spot every time. Maybe a parameter I overlooked? FWIW, backlash is a couple of tenths max, and like I said, it makes good parts once the Y offsets are reset.

Thanks in advance

Don't you have to reference return the machine after power is turned off?

Could it be that your slides are a little sticky after 20 years? The servos would still pull the axis into position but maybe they're having to pull a little extra to overcome the drag and when you drop the power the machine "relaxes", as you said. Try going past your shutdown position and coming back to it before you shut down. Load the screw in the other direction. See if the direction you're off changes.

So let me get this straight.. A 96 Fanuc doesn't index to a pulse on the encoder? and you have to rely on the machine
not moving overnight while powered down to maintain position?????

Its a good thing Fanuc doesn't make cars... They would still be using a flat leather drive belt and have a pedal for reverse.

I'm not sure why, but every time I hear the word "Fanuc", I think of a Monkey Fucking a Football.

If the machine is running all day and holding size that would indicate the ball screw is good and also the encoder. I would suspect there is something wrong with the homing of the machine. Possibly a bad microswitch or if the drives have a position memory that is screwing up. Maybe a weak battery.

John

Have a good look at your limit switches that tell the machine to start searching for it's home position. The easiest way is to home the machine loads of times AFTER powerering up and see where it's position lands up. The reason I say this is because your machine only uses those switches once when it is homed and after that it knows where it is.

If it is faulty change all of them and do not go cheap. Buy the best limit switches you can... It is well worth the money.

bucktruck said:

Interesting problem: When I fire up our old (1996, Fanuc O-M control) Leadwell 610AP VMC in the morning, it is always off in Y by a little bit. Sometimes it can be off by thirty thou and sometimes only a few

Click to expand...

Are you only going off the brain dead positions the control is showing first thing in the morning? Wayne asked you at Post # 2, You need to reference return, first. That 30 thou can be a kick back when you hit the E.Stop / Control off, to shut it down. That's a tuning error, balance is wrong.

Bobw said:

So let me get this straight.. A 96 Fanuc doesn't index to a pulse on the encoder? and you have to rely on the machine
not moving overnight while powered down to maintain position?????

Click to expand...

Fanuc encoders have an index pulse and have had one since at least 1978. Could have absolute encoders too if the builder opted for them. Whether one needs to reference the axes before auto operation is possible is determined by the machine builder, not Fanuc.

The OP needs to tell us a bit more about how he starts his machine up in the morning. I'm guessing here, but I'm thinking he is commanding G92 at power up.

As machtool says, it could be a tuning issue with balance. I'll put up another guess. The OP has chips packed under his way covers that push the axis back when powered down.

Thanks for the responses guys.

Start up procedure:
1: Turn on main power
2: Turn on control power
3: Home machine. Home > XYZ > Home Start. All of this at 25% rapid.
4: Zero relative position menu. XYZ > CAN

I was thinking servo issue as well but I just checked gain values and all 3 axis seem to be about the same. (Not that I really understand this; my tech showed me how to check this last time he was here.)

As far as chips being packed up under the way covers: I wish it was that simple. I've had the way covers off multiple times and it's clean as a whistle under there.

I understand that it's a 20 year old machine and it's going to have issues. Heck, I only use it for prototyping and tooling so resetting the Y offsets each morning isn't going to kill me. And sometimes it sits for weeks at a time while I play "Manager." Just thought I'd run it by the brain trust on PM as it has me a bit stumped.

I have a Fanuc OMD and when I home out and watch position is always counts down to a couple thou, stops counting for a instant, and then resets to zero. I used to worry about it but it never showed up in parts. It this what you are experiencing or, if you edge find the same surface the next day that position has shifted by .03?

Sticky limit switches? I have had to lube them before because they were sticking. 1984 Robodrill with a 3m control. I think the positions started to drift off each time we started in the morning, after homing it out on power up. Worth putting some oil or some light grease on the plungers that activate the limit switches. This is just part of my lubing ritual for my machines so I don't have these problems.

lowCountryCamo said:

I have a Fanuc OMD and when I home out and watch position is always counts down to a couple thou, stops counting for a instant, and then resets to zero. I used to worry about it but it never showed up in parts. It this what you are experiencing or, if you edge find the same surface the next day that position has shifted by .03?

Click to expand...

When I edgefind the same surface it is off by a bit. Sometimes it's zero, other times its more than that. It's never been more than .040.

DavidScott said:

Sticky limit switches? I have had to lube them before because they were sticking. 1984 Robodrill with a 3m control. I think the positions started to drift off each time we started in the morning, after homing it out on power up. Worth putting some oil or some light grease on the plungers that activate the limit switches. This is just part of my lubing ritual for my machines so I don't have these problems.

Click to expand...

I think I'll lube the limit switch and see if that does anything. If not, I think I have another good one laying around that I can throw in.

With the machine off push them all down one at a time to see if they are sticky. While lubing them push them down a few times to get the lube in, and free up the sticky ones. The plungers that you only use when you overtravel should be the worst.

bucktruck said:

I think I'll lube the limit switch and see if that does anything. If not, I think I have another good one laying around that I can throw in.

Click to expand...

Hello bucktruck,
As you're saying that you have to reset the Y axis Offset, the position of the Reference Return (Machine Coordinate Zero) is changing. Its unlikely to be a sticking switch when the error is relatively small. Your description of starting the machine indicates that its not using an Absolute encoder. Accordingly, the Reference Return Position is determined by the occurrence of the next index pulse of the encoder after the falling edge signal of the Reference Return switch is detected. A sticking switch normally results in over-travel, or the Reference Return position being off by the Lead of the Lead Screw. Also, a change in the Reference Return Position won't be caused by a Servo Motor issue. Its more likely to be with the encoder, or the coupling of the encoder. However, such problems normally persist during the continued running of the machine and not a one off when executing a Reference Return on start up.

Regards,

Bill

bucktruck said:

Thanks for the responses guys.

Start up procedure:
1: Turn on main power
2: Turn on control power
3: Home machine. Home > XYZ > Home Start. All of this at 25% rapid.
4: Zero relative position menu. XYZ > CAN

I was thinking servo issue as well but I just checked gain values and all 3 axis seem to be about the same. (Not that I really understand this; my tech showed me how to check this last time he was here.)

As far as chips being packed up under the way covers: I wish it was that simple. I've had the way covers off multiple times and it's clean as a whistle under there.

I understand that it's a 20 year old machine and it's going to have issues. Heck, I only use it for prototyping and tooling so resetting the Y offsets each morning isn't going to kill me. And sometimes it sits for weeks at a time while I play "Manager." Just thought I'd run it by the brain trust on PM as it has me a bit stumped.

Click to expand...

My guess is bad encoder or cable on the Y axis drive. If there is stray voltage on the marker pulse channel it would cause the condition you are describing. I would put an oscilloscope on the marker pulse channel to make sure there is only one signal per revolution of the Y drive motor.