Controlling CNC Machine Programs - Your recommendations appreciated

E

emiller1680

Hello,

I work for a company with about 50 CNC machine tools and we're trying to standardize the manner in which we control our CNC programs. I'm basically limited to a 4-digit machine program number when it is on the CNC control at the machine itself, but in some cases, I can assign another filename on our servers.

Does anyone have any advice on how I can proceed?

Considerations:
1. Program numbers will differ from machine to machine. That is, machine program 123 is not the same on one machine as it is on another. At least it doesn't have to be.
2. Manufacturing routers. I would love to give the operator a note on the routing sequence of what program number to use. But this will lock me into using that number going forward.

Any help would be most appreciated. Thanks!
 

CarolX

Trusted Information Resource
Hi emiller1680 and Welcome to the Cove!!

I am all for establishing naming conventions - it pays for itself in the long run!!

How about a 2 part file name -

Part one - Part Number

Part two - Machine Number

By listing the part number first youe would eliminate the need to put the program on the router, all the set-up person needs is to know the machine number.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
I have witnessed this problem, too. My understanding is that there are database programs that will remotely control the CNC programs in an elegant manner. One I have heard of is Predator http://www.mis-group.com/dnc/dnc.php Since running the same part on two different machines may require a different program, picking programs by part number can be dicey. Using a database means you do not have to use "smart numbers" (the approach CarolX is speaking of), so you have 9999 programs available per machine.

Program rev management may also be handier with this type of software.

I am hedging all of these statements because I have only heard of these benefits, I have not implemented or worked in that type of environment. Only one that was trying to get there....:cool:
 
E

emiller1680

you're right, smart file naming might not work in all cases. the file name on the server could be smart named though, in a fashion similar to this:
part_rev_machine_setup. so the first setup on part 123 revision 0 on machine ABC would be 123_0_ABC_01, and so forth. the problem becomes, as you say, the four digit number at the control level. i'll look into the predator software; i'm guessing they might control the files themselves, but the naming would be up to us.

in many cases, four digits would be enough. however, in our company, we're rapidly approaching 10K programs on file. yes, we could retire numbers after a certain period, but that could begin to make things even more confusing. it seems to be a difficult issue.
 

rstocum

Involved In Discussions
My experience with this problem came while working at an FAA certified repair station. We had to be able to show control over the content of the program as well as program name & revision control. The Predator software mentioned above was what was used. It allowed you to set up downloading of the program from a partitioned drive. The operator at the machine would receive the program from a location that only certain people could save the program to. When the operator saved the program from the machine's memory that program would be saved to a different location, and could only be used again if the CNC programmer reviewed it, and saved it back to the download location. This solved all requirements very neatly. I highly recommend it.
 
E

emiller1680

we have a similar system to predator in house as you describe. the problem i'm coming up with is the four-digit number. let me try to describe...

let's say i have a program on our file server called 123-0-ABC-01. but the program number is 8888. when my operator calls it up on the "predator-like" software, he can access the 123-0-ABC-01 file name, but it downloads into the machine control itself as 8888. thus, how do i tie the two together? the four digit number is perhaps meaningless, until i put on the manufacturing instructions, "Machine part on setup one. Use program 8888.NC." now that number has meaning, and i have to control the numbering of the program. so not only do i have to control the content of program itself, but i also have to control the nomenclature of the program name and number. here's an example of how disaster could strike:

operator 1 is running a machine. he makes a new program for part A and calls it 8888.NC, simply at random. he finishes the part and takes it off the machine just as his shift ends. he leaves for the day.

operator 2 arrives and picks up a job for part 123 rev 0, where it says on the routing to run program 8888.NC. he looks in his machine library and BINGO! there's 8888.NC. so he runs the program, which he THINKS is correct, but is actually the new program operator 1 created on the previous shift for an entirely different part.

how do i prevent that from happening?

now, let me say, this doesn't usually happen and would be an extremely happenstance case. but it's the thorn in our sides when trying to standardize our system. ideas?
 

rstocum

Involved In Discussions
Maybe. The shop I worked for had no restrictions on filename size, so this same circumstance was not a problem for them. Other places I have worked for had older machines that only took a filename of a certain size. Some of them had to have a capital "O" as the first character, etc. Those places put notes in parentheses in the first several lines of the program. The parentheses kept the controller from reading them as commands. The notes contained the controlled document number of the customer drawing. The operator was trained to run the program only if it contained the appropriate information. He could check it to both his router, and his controlled coopy of the drawing.
 
E

emiller1680

that's exactly what we do at the moment and the more i think about it, the more i think it's all we can do. the operator is required to validate the first few comments of the program with the information on the part routing, just as you describe. i'm not sure we can get around it any other way.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
I have seen that, too. You need a cultural change as to who can change a program, who can write a program, and the steps they have to take to ensure that doesn't happen. I have seen a program tweaked by one setup operator that may not have been a good tweak after all, only to have overwritten the original program. Rev control has to become top priority, and that includes a single source of program numbers - maybe even a single notebook.

Also, you need to make sure the parts are loaded from the program, not the machine memory. The program has the controlled rev, not the machine memory.

As far as tying them together the database program should have a cross-reference of the program number 8888.nc and the file name 123-0-ABC-01. But, the long name will never show up on the machine screen. That is why the loading process has to be the control.

I am not sure if that is clear...but maybe those with live experience can help clarify it...:cool:
 
E

emiller1680

Yep, i think that's exactly what we're in the process of doing. a culture change of who can make edits and then some kind of control as to whether those edits are approved and saved back to the network or not.

we're already using cross-reference sheets on our fanuc control based machines, and i think we're just going to expand it to encompass everything.
 
Top Bottom