machining help needed

[stout1]

Everything that I have been taught and had proven to me shows that using the offsets in a machine is bad. Here's the way I have come to look at it. If it has been proven that the process is capable than there should be no reason to edit with the offset for production purposes. If parts are coming out of the machine out of spec than there must be something else wrong (ie, tool wear, chip build up on fixture, bad parts going in, etc.) What are others opinions about this? My current position is that if parts are not coming out to spec we need to find the root cause as to why this is, not to just adjust the offset in the machine to bring the dimension back. I am currently in disagreement with a production sup. with about 20 years with the company. Thank You for your opinion and letting me vent.

Michael

 

[monkeydog]

Have you checked the process capability of the machines being used to make the part?

 

[stout1]

The machines capability is not really in question, parts should not be set to run in a machine that cannot hold the parts tolerances. Also adjusting the offset to "compensate" for a machine that cannot hold the parts tolerance is just asking for a lot of scrap. If the machine was once capable and now is not than the machine needs to be repaired not the offset adjusted. That's my opinion, aprantly is does differ from some.

Michael

 

[Cockroach]

You need to check your floats, in addition to tooling wear, chip flow and coolant contact near the cut. Typically the machinist is periodically checking his production pieces for various dimensions, and making minor adjustments.

If you are getting large variations between measurements during a production run, then obviously something is wrong. You will get extremely minor variations, not every piece is ideally the EXACT replica of the former! On a mirco scale, machining depends on both the physical process and chemistry of the piece being cut, which are random in nature statistically speaking. The machinist must be diligent is checking a few pieces during the production run!

I use a random number table and have the machinist check those pieces for a few dimensions, again at random. For example, you need 20 pieces of a particular item. Then check the first two for ALL dimensions, then 5 for diameter and threads, 9 for length and thread pitch, check 11 for thread pitch with a master gauge, 16 for all dimensions. Then statistically speaking you have 6 of 20 pieces that have been proven out of the batch. You can work your magic to find the average, standard deviation of dimension and accuracy of the machining.

Clearly this is part of the ISO standard. I just call it smart production. Your machinist must show more due diligence and attention to his work! Check your float, it sounds like this is the problem in this case, in addition to not random sampling a batch run!

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[BobM3]

New sharp tools will wear quickly and then stabilize. It may be only .0001s of geometry but the cutting forces will then increase. This causes the tool and part to deflect more. Also most machines are a bit sensitive to temperature. If you are trying to hold .0001s then you may need to adjust the offsets until the machine and tool reach a steady state temperature.

 

[Fabrico]

If it has been proven that the process is capable than there should be no reason to edit...

That also does not prove a reason not to edit. Even a terrible setup can sometimes prove capable. It is possible that something else is causing the offset to be a problem. Has something else been changed?

As is often the case, if you can't find the problem, it's time to start eliminating assumptions.

 

[stout1]

I know that the machine is more than capable of holding things in tolerance. I want to know why there would be any other need to change the offset other than part variation (than we need to look into why that is occuring) improper install into the fixture, etc. The tolerance is pretty liberal so why do they insist on needing to be able to adjust the offset. What I am trying to say is that if they are getting bad or borderline parts shouldn't they look other places before they just blame the machine and change the settings? That would make the most sense to me.

Michael

 

[berkshire]

stout1 (Mechanical)
One reason for wanting to be able to use offsets is so you can use reground cutters.
you might also want to use the offset command to compensate for normal tool wear as the job progresses.
Depending of course on what your tolerances are.
I know you said in your original post " If parts are coming out of the machine out of spec than there must be something else wrong (ie, tool wear, chip build up on fixture, bad parts going in, etc.)". However compensating for tool wear is a normal part of machining ( move it or change it). The idea is to give the machinist the ability to give you good parts.

 

[monkeydog]

I am a little bit confused here.
If you have parts that have liberal tolerances, you have a machine that is capable of fabricating the parts, and you have parts that are built out of tolerance; And you have a machine shop supervisor with 20 years experience that has a fix for you. Why are you fighting him?

 

[stout1]

We just had a large (in $)ooops and the Sup. thinks that the answer is more inspection. The parts would have been fine if the offset had been left alone. Also not the first time this has happened in the last couple of months. We also have very little tool wear (we only machine AL), most of the problems come from improper mounting in the fixture or people messing around with the program. This is why I feel they should be locked out of the machines. I just don't see the need for them to be changing things every time a part doesn't check out at nominal.

Michael

 

[ornerynorsk]

Offsets are a necessary part of real-world machining. With 24 years in the shop, I can tell you that you will have more headaches without them than with them. You stated the root of the problem in your own post, it's a people issue. Your people need more training/accountability, or you need an overlord of the shop to approve offsets and then lock up the control again.

 

[JJPellin]

There will always be some statistical variability in the dimensions of parts. As long as the statisitical variations are within the specifications, little or no adjustment should be needed. It is possible to over-adjust a system like this. If a part needs to be 2.000" with a tolerance of +/-5 and a part comes off that tis 2.001" no adjustment should be made. With no adjustment, the next part might be 1.999". Unless you understand how the variability is occuring, making checks and adjustments too often could get you further from the ideal dimension rather than closer to it. Years ago we starting applying advanced statistically processes to our process controls to avoid this problem. Operators who were too diligent to try and tweak everthing to be exactly perfect, were actually increaing the total variation from ideal. Some things are better left alone until you have a large enough sample to calculate min, max, average and standard deviation.

 

[stout1]

JJ,

That's exactly where I was trying to get to. That's how I feel a shop should be run but some of the "old schoolers" are fighting it. This afternoon I need to go see a customer that we just made many bad parts for. An ispector, that has checked this part several times in the past and always good, for some reason checked a part with a different methode. Needless to say the dimension didn't come out right so he had the operator change the program and run the rest of the parts (of course they were all bad then). This is getting very frustrating.

Michael

 

[berkshire]

stout1 (Mechanical)
Do you have an SPC data collection system in place in the shop?
My comments about offsets are to get the parts right.
Once the part is being produced to the tolerances, then the machine should be run without adjustment, and the results plotted on a data collection chart. One of the things this should show you is the machine noise ( Inherent errors and slop in the machine.) from this point on the only adjustments should be to get back to the center of the bell curve. You have stated that you know the machine is more than capable of holding things in tolerance. Have you run a backlash test pattern or are you relying on the supervisors word?

 

[Cockroach]

I feel your frustration, Stout1.

On one hand, you have a proven NC program that a machinist "feels" in need of updating. Then you have personality issues between "old school" verses "new school", employees, a resistance to change in operations.

What you really need is a strong manager with a heavy hand. I guess floats and tolerancing due to machining the piece aren't the problem afterall.

Good luck with it.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[berkshire]

stout1 (Mechanical)
My apologies, I did not realise your guys had been tweaking the actual program, Changing a dimension in the program is an entirely different thing to messing with the offset's. As far as I am concerned that is a totally different issue, and is, or should be a big no no.

 

[JDana]

I also work in a shop that according to myself and ornerynorsk, "needs more training/accountability, or you need an overlord of the shop to approve offsets and then lock up the control again".

unfortunately this will probably never happen so I add offsets (specific and well defined by man-readable code) where ever possible. I feel this is much preferrable than trying to figure out how the midnight shift hacked up the original program to get the (hopefully) good parts comming off the machine when I return the next day. Or a week/month/year later if I don't happen to catch it sooner on my own.

I have 14 years in CNC, offsets are not the enemy.