Results 1 to 20 of 37
Thread: Qualifying CNC Machines
09-01-2005, 03:09 PM #1
I have a small shop with 3 CNC mills and 3 CNC lathes. To satisfy customer QC requirements, I need to come up with a way of verifying the accuracy of all of them. One option is to hire that job out to a service that shows up with a technician and a ball bar. This seems like it could be expensive in the long run. Another option is to buy my own ball bar and do it in house. This is definately expensive up front - about $9,000 - but more flexible and perhaps cheaper in the end.
Does anyone know of a simple method of doing this that doesn't cost a lot? I can imagine using a gage with a calibrated contour that would be traced in the machine with a matching contour program and a tenths indicator. The program would stop periodically so the operator could read the deviation on the indicator and record it. It would probably have to be a different gage for lathes and mills, but it seams feasible. Or am I crazy?
Does anything simple like that exist?
Advice would be very appreciated!
09-01-2005, 04:27 PM #2
Ball bar is the most common test method at present. If you buy your own, do you have a machine tech that can fix the problems that you're going to find with a ball-bar? You're going to spot stuff like out of level, servo tuning, loose gibs, worn ball screws, etc. If you can't fix the problem, you'd have to bring somebody in anyway. There's independents out there that do Ball-bars with minor adjustments and provide reports for $750/machine.
If there was anything cheaper or just as good, we'd have heard about it. What we've looked into is brand new. 3-D stereo laser comparisons. A known shape is ran on the machine and then the laser compares the part in the machine to a 3-D CAD image. There are systems out there working. Some bugs yet. About $30K to set up one machine.
09-01-2005, 06:03 PM #3
Quick thought would be hire a pro and get the job done now. Take the jobs profit and purchase the equipment and training for the next time your achine needs to be qualified. As I believe this is an ongoing thing not a one time deal.
09-01-2005, 10:32 PM #4
I may be totally out of line here, but are you selling the equipment, or the parts you make on them?
What I mean is what does it matter what machines you made them on, or what hoops you've jumped through to get the parts within spec?
IMHO, customer may control manuf. specs, may have controlled processes, and in some cases may have a certain machine that is tested by them to manufacture the specific part < I'm thinking EDM brand/model >.
They may specify the coolant, shop practices, handling of material or foreign substances in the shop etc.
BUT in no way do they have any control over your machine's QC procedure, accuracy or shape.
The only exception would be inspection equipment.
The fact that you may want to know about the accuracy of your machine and the maintenance schedule is, and should be YOUR concern and not the customer's. What's next, having to qualify cutting tools before entering the shop?
It is only my $.02.
09-01-2005, 11:07 PM #5
WOW, that sounds like a bunch of crap to me. Either your customer has an ISO manager that feels the need to justify his job with yet more forms and procedures, or they are trying to price you out for another supplier. I can understand them wanting you to have calibrated measuring equipment, but this seems like crap.
If it is your customers requirment, let them pay for it since they made up the procedure, or talk to them and ask them if your quality has been a problem. If your quality is good, there should be no problem. The machine shouldn't matter, heck if you can produce good parts with a hammer and chisel at the agreed upon price, it shouldn't matter.
I'm curious, what type of products are you producing, that would require certification of a machine?
09-02-2005, 02:01 AM #6
Bob said it right
" Either your customer has an ISO manager that feels the need to justify his job .... or they are trying to price you out for another supplier."
As a Manufacturing Engineer, I would ask about machine capability when I jobbed out to a new supplier for the "First" time, because he was using MY forgings....but once he made good parts, all the qualifying crap stopped ( Put your money where your mouth is )
let me make it simple..
A machine "operator" produces parts within the variables of his machine.
A "machinist" produces parts beyound the variable limits of the same machine " TOOL"
or another way...
"Rock Crushers" and "Sculptors" both use hammers and chisels.
One works the tools and makes chips , but the latter uses them to make Art, with chips as a byproduct......Same Tools !
Machining is the same thing !
09-02-2005, 02:05 AM #7
I should mention that is is perfectly alright for a customer to ask for Coordinate Measuring Machine (CMM)print outs to varify accuracy of parts provided. But that is a cost reflected in the quotation procedure
09-02-2005, 08:21 AM #8
Rich, very well put, but remember that the CMM data is of the finished goods and not the manuf. equipment.
I have just looked throughthe SAE AS9100 documents, and there ain't a darn thing about equipment certification or accuracy data.
I know you're probably just want to satisfy your customer and comply with their standars, but for that you need to know what standard you need comply with. Therefore they should be able to give you somekind of a document to state the necessary tests. Since the such test, procedure or anything is likely non-existent, put the ISO guy to work and make up one. And when he does make up one, you'll realize he, just like 99% of his fellow ISO managers haven't got the first clue about manufacturing.
09-02-2005, 09:35 AM #9
I think the guys here are onto something about qualifing the machine. Perhaps and most likely no needed.
But Now if you can incorporate the test into the price of the job and GET and KEEP the JOB, its worth it, to me at least. Tell the guy he's an idiot for making you do the test can put a bad taste in their mouth. Besides you "qualifiy" your machine and customers happy, now when he looks for a second vendor it will be all the harder for him to find.
I have a customer who wants me to layer their parts in this sticky foam material and roll the edges just so to keep them from scratching, soughta makes sence till you go to the recieving department and watch them rip the box open and "dump" the pieces into an bin.
Yeah waste of my time packing, but the company pays extra to have them packaged to these specs. Whatever makes em happy.
09-02-2005, 10:21 AM #10
True, but who pays for the cost? In your case the customer, but for "compliance" to a non-existing standard? Yeah, if the job is worth it, sure I'd do it too, but what if the machine does not qualify?
09-02-2005, 10:44 AM #11
I make prototypes for a couple of companies here in Seattle, I was asked last year to jump through the same hoops. Most of these customers are very large corporations with huge numbers of personel. I asked if there had ever been a problem with my quality and was told, No, that's why you get the work...
That was the last I ever heard of having to qualify my machines
09-02-2005, 10:44 AM #12
Hmmm, if I was a QC manager interested in my vendor's machine capabilites, I'd be more interested in a machine's Capability Ratio (CR, CP, and/or Cpk) for a given part tolerance. Accuracy can be adjusted for (via tool offsetts, backlash compensation, etc.) but if the spindle bearings, or the ball screws, etc are worn out, it will show up in a variation (6 sigma) study. That, to me, is the true machine's capability to machine a part.
I don't believe the requesting individual really cares what he's asking from you; probably just going through the motions to satisfy his internal requirements. At the very least he should be able to supply the test criteria.
09-02-2005, 10:46 AM #13
There is a old standard used to qualify machines You cut a triangle on top of a circle on top of a square to given dimensions. Use this as a certifying specification for your vendor for each machine. And cut a new one every quarter to verify the accuracy of the machine. For a lathe I would imagine you could come up with a stepped OD and ID and may a thread you can check with a Go no go gage.
09-02-2005, 11:19 AM #14
What part of that test qualifies the machine?
IMHO, you want to hold +/-.0002 on any feature, you tweak the thing until you get +/-.0002 on that feature. If you can't get that +/-.0002, you won't take the job or get another machine.
So what if the customer says, hey, your machine can only hold .0005, we need parts tighter so we will not even consider you? Should it not be MY concern how to make the part correct?
09-02-2005, 12:38 PM #15
I'd have to agree with Seymour's statement that the great majority of ISO compliance types don't have a clue regarding manufacturing. I'd say the field attracts a whole lot more of the types who have "control" issues than ones with the knowledge of how to make something useful.
I do some work now and then for a 1st tier supplier to the auto industry. However, its all tooling or fixture type work. They're not allowed to sub any production work to guys like me (not that I'd want it, based on the going prices) because I don't have ISO certification, and the auto industry requires certification down at least thru 2nd tier. Every time I talk with their buyer, he asks again if I've gotten ISO yet so he can send me some work. Its sort of a joke, because he knows the answer before he asks the question. We get along well, and he has told me in the past that the 2nd tier requirement is the toughest part of his job because it ties him down to trying to find suppliers from a pre-selected group of shops instead of just simply finding suppliers who make parts to print. He's been in the job long enough to have seen the before and after of this requirement, and says without reservation that he's seen nothing to make him believe an ISO shop has a greater chance of making good parts than a non-ISO one.
There's a lot of evidence to make one believe certification is just another non-productive industry that has evolved and thrived based on our ever increasing percentage of people who are largely incapable of producing, and have no desire to produce, anything more than a daily stack of paper. To imagine this does not have a large impact on our ability as a nation to compete is to be blind to the realities of business.
09-02-2005, 01:45 PM #16True, but who pays for the cost? In your case the customer, but for "compliance" to a non-existing standard? Yeah, if the job is worth it, sure I'd do it too, but what if the machine does not qualify?
If it is something your machine should be able to do, get it done, proved, certified etc. Then let him call out the spec till his hearts content.
Now who pays. Hopefully the customer thru the pending order or orders. I would break even to gain good paying potential work. Consider it tooling, anybody buy some expensive dedicated tooling for a customer or a job?
09-02-2005, 07:57 PM #17There's a lot of evidence to make one believe certification is just another non-productive industry that has evolved and thrived based on our ever increasing percentage of people who are largely incapable of producing, and have no desire to produce, anything more than a daily stack of paper. To imagine this does not have a large impact on our ability as a nation to compete is to be blind to the realities of business.
My career(s) were machinist, CNC programmer, electrical engineer/manager and I'm still active in all three. I've seen it work great and have seen it fail. It all came down to management and labor's commitment.
09-02-2005, 10:28 PM #18
My thought is prior to starting to check things out, find out exactly want the customer needs to see. A certified Ball Bar Test, milled sections whatever. Get what is required in writing, make it his or your specification.
Ok SIM, but what does that all mean? ISO exists so there is traceability throughout the entire manuf. process, but in no way does that specify what, or more specifically which equipment is used, and what accuracy does that equipment have. That is a completely meaningless procedure, as only the finished product and the process is what needs to qualify.
Let's say all your CNC machines fail the ball-bar test with respect to original specs, but you have manual methods to complete the part accurately. What then? You won't qualify even though your finished product is good?
Don't get me wrong, I'm all for pleasing the customer, but this sounds bogus.
09-02-2005, 11:15 PM #19
I agree with all of you find out what the customer wants. A ball bar test will only show how good or bad your lead screws are and whether the axis are perpendicular and or parallel. As far as using a triangle, circle, square to qualify a machine. That's what I used for our last ISO audit. I knew it would pass the scrutiny of the auditor when she ask me how I calibrated our cnc production programs. [img]smile.gif[/img] Sometimes you win a few and then again some you just slide on by.
09-02-2005, 11:25 PM #20
Thanks everybody for the responses!
About 80% of our business is the manufacture of instruments that are used in surgery - mainly the installation of artificial hips and knees. We must comply with Good Manufacturing Practices as specified by the FDA but are not required to be ISO compliant in addition.
As you can imagine, our customers are pretty worried about quality. However this customer (new) is the first to request qualification of CNC machines in addition to inspection equipment. They provided their own document as a guideline , but I don't have to rigidly adhere to it. Their document calls for one of these 3 methods yearly:
1. Ball bar check (no surprise there). Since ballbar calibration by a vendor is around $700 per machine plus commute time, I'm not crazy about the idea of doing this for 6 machines every year.
2. "Specialized" calibration. Not explained, but I'm guessing it applies only to unusual or custom equipment rather than mainstream CNC equipment.
3. Sample part test. This is suggested as a way of testing machines when the other methods are not practical. Wire EDM is mentioned as a candidate for this test and I'm hoping small shops with small budgets are seen as candidates as well. I think that will be acceptable provided that I can come up with test parts that will show any problems with the mills and lathes.
Here is one thought I had for a test of a mill's X and Y axes -
1. Secure a disc to the table.
2. Bore a hole in the center and set that point as G54.
3. Program a light cleanup cut around the perimeter.
4. Measure the distance from the ID to the OD at several points around the part.
A separate test would be needed for the Z axis.
Scojen - You described a test piece, how would that be used? Would you be so kind as to provide more details on that? Your idea of a stepped part for a lathe seems practical (but it would need to be big) although I would omit the thread portion.
Any more ideas out there for test pieces?