Code missing from program on shared drive

I've heard many bad things about not using local programs.  Things like, multiple stations trying to use the same program is very bad.

I've heard of people that use coding to copy the wanted program from the shared drive to the local drive, running it, then deleting the program from the local drive when done.  Not sure how they do that, but it should remove the issue of program FUBAR due to server issues, you either get the full copy or nothing at all.

I've never even tried to use a server system for programs, don't even want to try it.

hard drive space is cheap, my programs are all on the machine PC.

So how would we execute your idea when we have multiple (five) CMM's?  I want to make sure that they're not running the latest revision of the program.  I also want to be able to update/make revisions to a program and not have to do it at five different workstations.

That is what I do. I make a program offline. Either email it to my self or put it on a jump drive and open it on the physical CMM. 9 machines. I've never tried to use a server for storing programs. I back up all CMMs once a month. If there is a revision to a drawing and I have to update a program, depending on how much needs to be revised, I may just update it at the CMM or spend some time writing it offline and then going out to the CMM.

It also doesn't help that we do not have any "CMM Operators" our machinists know enough to set up fixtures and run my programs in protected mode. No manual or readpoint alignments. Its all dedicated, 3D printed fixturing and really good setup notes with pictures.

So if its a new program or new part, I have to go to the CMM any way to prove it out and set everything up.

I would guess that each CMM would needs its own individual folder on a server with access to only the proven programs and not the un proven ones.

found this:

nexus.hexagon.com/.../save-measurement-routine-locally-from-server-drive

You should NEVER store a routine, probe file or external alignment file on a shared drive that can be accessed by multiple machines simultaneously.  PC-DMIS constantly reads and writes date whilst executing so, you can imagine what might happen if two (or more) machines were trying to do that to the same routine - at best, you'd get useless results, at worst, you'd have crashed machines.

The Hexagon recommendation is to run local copies.  By all means, store your master programs, probe files etc on a server / sharepoint but ALWAYS copy locally before executing.  That is exactly what Inspect does - there is an option to create local copies when programs are stored on a server.

Guess we are the exception to this rule. We have been running programs off a network drive since 2008.

Maybe we've had problems connected to the programs being on a network drive, but nothing that we have been able to pinpoint the network drive being the culprit.

It's not running from a network that's the problem, it's having multiple machines accessing the same program at the same time that's bad.  If you only have one machine, then you're generally OK (baring any problem with bandwidth / the connection dropping out).

RoboCopy... just execute and have a master then distribute copies to the other machines. Machines operate locally on its own drive.

Its nice when you have a rev change and just change 1 routine and execute, not 5 times if you have 5 machines

I think that's the route we'll end up going.  I was talking about this with our IT guy and he mentioned the same idea using RoboCopy.  Thanks for the feedback everyone!