ROKU problems

[CDL2042]

Our ROKU 658 has developed a problem. It is cutting off center about .0006 in the x-axis only.

You can pick up a ground block on center, mill a pocket, and pick up the block again. The block will still be on center, and the pocket will not.

 

We had it checked with a laser, and a ball bar, and everything looks great.

 

I was wondering if anyone has experienced the same thing?

[Kevangel]

Could this be from spindle taper being slightly cocked so that different length tools move your centerline ever so slightly, maybe try picking up part with cutter at same length as indicator ball? I have the same .0005 problem in my Moore jig-bore except its the Y axis. My spindle is dead square to the table etc, and the only thing I can think of, is the above:headscratch:

[mold100]

CDL, is your machine completely warmed up when you pick up your part, AND cut the pocket. Also if you adjust your fixture offset and recut then what happens and how long does it stay right on?

[CDL2042]

Yes the machine is warmed up, the guys are pretty good at doing that.

 

Here is what they told me. They cut the pocket, it was shifted x-.0006, so they tried it again with a x +.0006 offset hoping that would put the pocket on center. It didn't. The pocket moved to x+.0006.

 

That is messed up.

[mold100]

are they using a holder system such as 3r or erowa?

[CDL2042]

We use a 3-R system most of the time, but in this case the block was in a vise. We have the same problem with the 3-R though, because we had to rework a mold with .0006 mismatch from the electrode. That is how we discovered this situation.

[mold100]

there are alot of variables, clamp pressure in both situations if not with in a few lbs of each other could be a few tenths, is the location offset being properly stored and saved in location (I have seen operators type a number and hit insert instead of input and move to the next screen) roku roku are pretty acurate machines and with out a major crash being off .0006 is a big problem. I would first cut a trode on center out side on like a 1 by 1, check it, adjust take more stock off and recut, if they are not adjusting the comp to remove more stock then it will never be on each time they recut. When dealing with a higher end machine you have to look at everything, if it were a old fadal or haas you could say .0006 that is good. I would start with process and work holding being 100% confident in them and then look at the machine tool. IMO.

[PeterM]

We have a similar problem on our 2002 Kitagawa HX400. I ran some tests and even compensated for tool length, our error is a whopping .0014 (inch)in the X axis. Nobody has been able to tell us why either. There is a global shift available in the workshift page, we just input the error into there and everything works fine. I do wonder what the heck is up though.

[mold100]

that is a good point, is there a fanuc on the roku? the first g54 is global pos, you could maybe put it in there. .0014, thats huge.

[CDL2042]

I guess we will keep testing. Our operators are very good, and we have never noticed this problem until recently.

 

You are right though, the ROKU is extremly accurate, .0006 is HUGE for us, and we wonder how we will be able to mill critical things with confidence anymore.

 

As far as work offset, I already explained that when we try to offset to compensate, the result is not predictable. I wonder if Kevin was on to something with the spindle taper?

 

Thank you guys for the advice, I will keep you posted. Maybe its just something stupid that we are missing.

[?Mark]

quote:

---

spindle taper being slightly cocked so that different length tools move your centerline ever so slightly

---

+1

Looks like that's the case here also, your spindle head is not square to the table and ball bar test didn't show it for us either.

Couple relevant discussions

1

2

 

hth

[CDL2042]

Thanks, after reading the links, maybe a test bar would be the way to go.

 

I wonder why the service tech didn't suggest that? If that is the problem (spindle taper), is that something that can be corrected?

[?Mark]

The spindle head was adjusted on our mori by some screws ( I wasn't at work then...), but the effect is not perfect.

It's a little better then before, but when running some really close positional tolerance I still have to lie to a probe

It's supposedly very hard to dial it in.

And btw, I never crashed this machine (any machine for that matter...)

 

hth

[Brad St.]

Just curious if you've tried different tool holders to do the different steps? Such as a different holder for tramming and a different holder for the cutter.

 

Do you by chance have scales on the machine also?

[CDL2042]

Not sure if they have tried different tool holders.

 

We have glass scales.