Slightly O/T - Integrex

[chris m]

Gentlemen,

 

I know that several of you have Integrexes out there and I would like to pick your brains, if you would be so kind. We purchased a 300SY with a 640MT Fusion control last year with the machine acceptance contingent on an in-plant runoff of the part that we [more or less] bought the machine to make. The machine is now here in my facility, and has been for several months now, but they can't make it work.

There seems to be a problem with the accuracy of spindle sychronization during the handoff that they can't solve. The job is a casting, not a bar job, so the machine hands off with both spindles stopped. There are a couple of hole patterns that are critical to each other that are done from opposite sides and must be held to .008" true position [not a particularly tough thing, you wouldn't think] but the machine won't hold it.

 

Any ideas why?

 

The Mazak guys say that our "full C-axis contouring" on the second spindle doesn't work the way it does on the first spindle so they don't have the same level of control on the second side; does this make sense? They also say that the fact that we specified ISO programming ONLY for the runoff part [no Mazatrol] is hamstringing their efforts.

 

I will say up front that we have no knowledge whatsoever of Mazak machine tools, their controls, or their software; but this doesn't make any sense to us at all. Are we getting a snow job? Is the machine itself the problem? It seems to us that it is; but many of you know much more about these machines.

 

My apologies for the novel-length post; but we are extremely concerned and figured that this might be a good place to seek input.

 

Thanks

 

C

[PE @ IHS]

The sub spindle on the Integrex is not a full rotary C axis, but rather an indexing one. (It is incapable of simulataneous interpolation). With that said, they are great machines. The ISO code excuse from Mazak is ridiculous. It is an excuse from people that generally don't understand CNC programs unless they are in Mazatrol.

 

Have you tried returning the main spindle to C home position (G00 G90 G28 C0) prior to part transfer? Both chucks will repeat position perfectly, regadless of which type of C axis they are equipped with. I would mill the soft jaws on the sub-spindle at C home position locked. Thus, if you transfer your part with both C axes in home position, you should be fine. Most people keep the chucks running during transfer, but this might not be a good idea with a cast part, since the lack of concentricity will affect the transfer repeatability.

 

1.) home BOTH rotary axes and mill your soft jaws on the subspindle

 

2.) tell Mazak to keep coming back until the parts come out right, or they can take the machine back.

 

Peter Eigler

[chris m]

Peter,

 

Mazak isn't going anywhere until they make a good part [15 good parts in a row with no defects, actually] but this doesn't help me much with an impending production deadline and no machine to make parts.

 

As far as the spindle home issue is concerned; while I haven't actually paged through their program I would hope that their apps guys know enough to do this. It sounds to me like their problem is that the C axis doesn't repeat, but no amount of programming is going to fix that!

 

Thanks for the input; keep it coming!

 

C

[PE @ IHS]

If the spindle orientation doesn't repeat, then it is a hardware fault (encoder??). This has nothing to do with the type of machine or the control.

 

Peter Eigler

[Guest CNC Apps Guy 1]

quote:

---

If the spindle orientation doesn't repeat, then it is a hardware fault (encoder??).

---

quote:

---

The ISO code excuse from Mazak is ridiculous. It is an excuse from people that generally don't understand CNC programs unless they are in Mazatrol.

---

I would concur with Peter.

 

The G-Code issue is one of my biggest problems with Programming Conversational. You have so little control over exactly how things are done. This is why I'm such a fan of a "regular" control. No fancy BS to get in the way. With that said, the Integrex is an OUTSTANDING machine, It would be the $#!+ if it had a Fanuc 16i Control on it. That and full C-Axis contouring in the Sub-Spindle would be the only improvements I'd make to that machine.

 

You may want to check out a Mori Seiki MT style machine if the Integrex can't do the job.

 

JM2C

 

[ 09-03-2002, 11:45 AM: Message edited by: James Meyette ]

[McRae]

Peter E - STOP...

 

The machine as shipped goes thru a rigerous inspection process and at installation, a field service engineer confirms copious settings, switches, mechanisms and parameters. To blame the machine first out of the gate is the wrong thing to do and sounds too much like to operators that I have in my shop. The machine will do what you tell it to do and even in some cases when you tell it to. If the tooling offsets are correct, the fixturing is correct, and the process is sound, the machine will make parts correctly.

 

Lets look a little into this first. My question is if there are features that have a critical tolerance between the front and back sides of the part, then I would locate these with pins. When doing second op work on a sub spindle, does the transfer process introduce variation.

 

To trouble shoot this lets do a little work on the Failure Modes and Effects Analysis (FMEA) before we scrap the million dollar machine...

 

Peter E gets todays "Flame of Shame"... we need to form opinions from emperical evidence, not "Broken Encoders" or "Bad Video Cards".

[Guest CNC Apps Guy 1]

Andrew, OUCH!

 

Should the subspindle repeat within it's advertised spec? It was also mentioned that the Mazak Engineers are there actually setting up this turn-key. I would think it would be a safe guess that they have taken the things you mentioned into consideration and are attempting to asess the situation. If the Integrex is built in Japan, it has been known to be a rough ride on occasion on the boat ride over the pond. Strange things happen to equipment of this nature in rough seas. I worked as an AE for a Japanese MTB and usually a few times a year had machines damaged in transit(visibly and not visibly). Anything can happen in transit, ANYTHING! I saw the remains, yes remains, of a 45,000 Lb. machine that looked like it went through the rinse cycle in a washing machine, creacked, broken castings.

 

One thing that perhaps should be looked at least, because sometimes things are simple, is when zeroing the subspindle make sure that the machine is taken back to Absolute (G90) as opposed to staying in Incremental (G91). Sometimes the small things cause big problems.

 

JM2C

[chris m]

Andrew

 

Thanks for your emails; we will continue to correspond.

 

I agree that it is very unlikely that the machine is !@#@%ing it up; that is typically the last place that I look when I have problems. Many of my operators [and one of our former programmers] immediately say "the machine is screwed up" but that is rarely the case. The Mazak guys will be back in here again today to discuss what is going on and what they intend to do. I'm sure they'll tell us that we're the "only ones who ever had this problem..."

 

Don't you love it?

 

Edited after seeing James' last that came in while I was typing:

 

James, the machine was indeed built in Japan but has been set up and running in this country for about 1 year now; 1st at the Mazak tech ctr in Connecticut and now here. The apps/service guys have not mentioned anything about rigging issues and the machine looks mint. They even had Mitsubishi in here looking at it and loaded some new software in the machine last week; still no dice.

 

C

 

[ 09-03-2002, 12:26 PM: Message edited by: chris m ]

[Guest CNC Apps Guy 1]

I was not necessarily trying to point the finger, just looking at possible casuses. I too have found that when something is wrong it is usually my fault or the setup guy's, but with that said, I try to never point to the machine first (unless it's been a problem in the past), 99% of it is program/setup related. The thing that seemed strange to me is that you can't repeat positioning wiht the Sub-Spindle. That is odd. Maybe a connector has vibrated loose.

 

Is the part possibly slipping in the jaws on the sub-spindle?

[McRae]

James- Now your thinking. I suspect work holding is the culprit.

[Guest CNC Apps Guy 1]

You kno wthe funny thing about holding onto castings is that even the very best Investment grade castings can vary by .010 in. I'm wondering what type of casting it is?

[PE @ IHS]

Andrew

 

I ran a brand new Integrex 300 SY. After the "rigorous testing, blah blah... the toolchanger didn't work on the second day, the tool calibration probe was stuck, and the chip conveyor could not be shut off. All in the first week of operation. Don't judge the mechanical experience of those you have never met.

 

Peter Eigler

[chris m]

James & Andrew

 

We were discussing the possibility of part slippage this morning here and I am realtively sure that it is not occurring because:

 

1) the features that are off seem to be relatively consistent and it is difficult to envision the same amount of slippage each time

 

2) features machined in succession with little stock removal are not right

 

3) the chuck is clamping on a finished surface for the second side and there is no evidence that the part moved

 

If we ever get personally involved with debugging this thing there are a bunch of things we'll try...

 

C

[Craig Madsen from CM3D]

Is there any reason that you stop the main and sub spindles? Once the spindles are stopped, the servo system will have only minimal control over the spindle position.

 

As a experiment, stop the spindle and (carefully) in a area where you can apply pressure manually, without placing any part of you in a position to be hurt by anything if you get a unexpected reaction, try to turn the spindle. I think you will find it is easy to move for a small distance (in excess of your .008") then gets progressively harder to turn. When you release pressure it will return, but maybe not to the exact starting position.

 

If this is the case, I would suggest that the transfer operation might have better repeatability at some low rpm. (maybe 150 - 300 rpm). The spindles will probably have a routine to syncronize speed and orientation.

 

Even then, I think you may also want to consider the postioning accuracy of the spindle in C-axis mode. If you have a small part of 1" dia. and your spindle is accurate to .001 degree, this is a rotary error of .000008" (pretty small). On a 10 inch part this is .00008" (still pretty small). BUT, if the accuracy is .01 degree on a 10 inch part, this is .0009 error. (since this error is usually a +/- error, the best you may expect is almost .002"). Multiply this by 2 spindles and you could be .004" out on a good day.

 

So . . .

1 Look at the stated specs and repeatability for the machine and run the numbers.

 

2 Check the 'static' position strength of the servo system.

 

Sorry about the length of this post.

 

Craig Madsen

[email protected]

[Jack Mitchell]

Chris M,

 

You do realize that the ratio of competent service technicians is actually similar to those employed in our profession. I only wish I could be present to observe this situation.

The fact that the machine tool was built in Japan reinforces that it is indeed built properly.

If you ask for satisfaction – Yamazaki will fly their maintenance people in to your plant. IMTS is running this week – the service technician is there right under your nose, as is the president of Yamazaki. - mention your concern to his translator & watch the sparks fly.

 

If you’re running with soft jaws, c-drill, drill, & ream a ¼” hole into one of each jaw set.

Sweep the jaws from each respective spindle – of course, each is clamped before sweeping. Program the dowel about an inch above centerline.

Rapid the machine tool & home it out, sweep the dowel, etc!

Actually, this reeks of a machine tool smash – sweep the tooling bores please.

 

Andy at Protech Precision in Mississauga Ontario has likely been through every imaginable scenario with such setups with this exact machine tool; A.W. Miller in Toronto might give you his phone number if you ask for it.

 

I believe the culprit here is either clamping pressure or a turret smash. I believe the handoff of a stationary part to another stationary spindle, is also begging for inaccuracy. – The intent of the sub spindle is to eliminate stopping & starting and other wastages of productive time.

 

Although possible, there could be a quality issue here, however; Mazak does make among the worlds finest machine tools and guarantees them to the nines.

 

Regards, Jack

[chris m]

Gentlemen

 

How can stopped spindles be less accurate than turning spindles? When drilling, milling, reaming, off-center boring, etc, the spindles are stopped and locked; are these operations all inaccurate? Please explain. If C-axis positioning is .001 degree accurate; how bad can the position really be?

 

I appreciate all of the suggestions of various methods to evaluate the machine condition; Mazak has tried many of these things and they are saying that the machine is physically fine [although some of the cutting heads that I've seen look like the they've been bumped] but they still can't make it work.

 

Something about the C-axis gain...

 

As far as handoffs to stationary spindles, the run time of the part is about 35 minutes and the first op on the second side is a rough milling op, so:

 

1) I'm not really worried about the "extra" 10 seconds to stop the spindle

 

2) I would hope spindle synch would be easier and faster at rest than trying to match RPM on different sized chucks with different workholding

 

3) the right spindle would have to stop after the handoff if it wasn't stopped first, so what would I gain?

 

If it was a cutoff/pickoff as from a bar-feed operation I would agree that a synch on-the-fly would be quicker but in this case I don't agree

 

Keep the good feedback coming guys

 

[ 09-04-2002, 07:26 AM: Message edited by: chris m ]

[chris m]

Just wanted to bump this back up the ladder; plus the little flame graphic is cool...

 

... it sort of describes our attitude toward our Integrex right now...

 

Okumas rock

 

C

[McRae]

[ 09-10-2002, 12:00 PM: Message edited by: MfgEng ]

[PE @ IHS]

Andrew

 

People who spend $500,000 or more on a machine don't give a damn about "paretto analysis" or what component is assembled where. They expect the entire machine to work as promised when the PO was signed. As much as I am a big fan of Mazaks (I have run them for years), I have seen instances when Mazaks came from the factory and began malfunctioning within days. This can happen with any piece of mechanical equipment. Nothing mechanical is guaranteed to work 100% of the time, since even REAL engineers make mistakes, as do machinists and assemblers.

 

Peter Eigler

[Webby]

Peter and Andrew - Cool it on the personal comments.

[McRae]

[ 09-10-2002, 11:59 AM: Message edited by: MfgEng ]

[PE @ IHS]

And some people are capable enough to do all three with success....!

 

Perhaps Dave is right, and we should start behaving..otherwise he will spank us with his ping-pong paddles!

 

Peter Eigler

[Guest CNC Apps Guy 1]

quote:

---

Those who can, Do

Those who Can't, Teach

---

Ouch, that hurt.

 

I know that's not true for me so I'm not offended in the least.

[McRae]

Correction to earlier post

 

Those who can, Do

Those who can't, Teach

Those who can only talk about it, Sell

Those who can't do anything, Manage

[chris m]

I've gotta meet you some day; they say everybody has an exact double somewhere in the world...

 

Just as an update to our Integrex saga; the machine has been sitting comfortably and resting this week [its tired, I guess] because the guy they sent in to run it left Mazak [last Friday was his last day; which he told us at about 11:00AM Friday, of course] and the guy who worked on it before is wrapping up the show in Chicago; sort of personifies Andrew's statement about management...