upgrade from manual to dcc

[rdbowman]

I am currently running a manual 765 micro xcel machine with pdcmis ver. 3.2063. My company is looking at buying a new 775 ONE machine. I have around 200 programs that were written on my manual machine using ver 3.2063. Will it be possible to convert these programs to be executable on ver 4.1 with a dcc machine and if so what kind of headaches will I run into. Any advise will be greatly appreciated. Thanks in advance.

[Matthew D. Hoedeman]

I would try to make them run using V3.2063 on the new machine, then 'migrate' them up to 4.1. At the very least, I would use SOMETHING other than V4.1 to try to make them runable. Mostly, it should just be a matter of inserting MOVES (clearance plane or point, or increment) as well as turning on DCC after your alignment, I think. I have never run a manual machine, but I do know that a 'manual' point will turning into a DCC point just by inserting the DCC command in the program.

[#2#]

I would try to make them run using V3.2063 on the new machine, then 'migrate' them up to 4.1. At the very least, I would use SOMETHING other than V4.1 to try to make them runable. Mostly, it should just be a matter of inserting MOVES (clearance plane or point, or increment) as well as turning on DCC after your alignment, I think. I have never run a manual machine, but I do know that a 'manual' point will turning into a DCC point just by inserting the DCC command in the program.

While I have no experience doing this, I agree with Matt. Take it one step at a time. Get them running DCC on V3.2063 (what they are running manual on now) then maybe worry about upgrading to some other version once you are comfortable with them running DCC. The new machine would doubtless be set up with V4.1 and you would have to install V3.2063 on the machine yourself.

I see no reason why inserting the DCC command and a few movepoints/clearplanes won't work to make a program DCC. This assumes of course that you have an alignment very early in the program. If your program was written with all the features being measured and then an alignment at the end (why? I don't know but it's concievable) then you will need to do a little more work.

You can of course continue to run a program as a manual program on a DCC machine. You just use the jog box to drive around and take hits. That can be a little tricky though.

[craiger_ny]

I used to run PCDMIS on a manual. Like Matt said movepoints, movepoints, movepoints. Otherwise throwing a DCC command in there will change nothing. Keep in mind that alignments determine movepoint locations so if you have a goof ball alignment your probe will move relative to it not the machine. My sugestion would be to migrate each program only if you have a part and can run it and insert moves as needed. Run it by Ctr + E each command and watching how each command one at a time goes off. Then you can insert neccesarry commands as you go. Simply putting a DCC command in will not effect your program. It will effect your machine if it smacks something though.

[kass_b]

Move points also be carfull of wrong direction vectors

[Tested]

I would not put the old version of PC DMIS on the new machine if you don't have to. If your going to convert your manual programs to DCC you will want to...
1. rename programs
2. create a manual alignment
3. go to each feature and use F9 to check vectoring and hit locations
4. add DCC comand
5. clear all, mark all, yes for mark manual alignment,
6. turn speed way down
7. step thru as craiger said CTR + E
8. add moves and any further hit adjustments

[#2#]

One thing I think we can all agree on is that you should back up all your programs to a safe location (cd maybe) prior to beginning any work on any of them. That way, if things go all wonky on you, you still have the original un-f-ed program and can try again.

[jjewell]

All the guys brought up good points. I would put the old 3.2 version on first, for many reasons. Some stuff might not work correctly in ver 4.1. Ver 4.1 has a TOTALLY different menu structure and interface. Going to a dcc machine will be daunting enough in the beginning. Oh ya, theres this thing with lots of bugs in ver 4.1 ( I still use ver 3.25MR3 to make a living quite nicely, thank you)

Just my thoughts...

[Lee Johnson]

Stay with what you know on the new machine. Load your 3.2063 in a seperate file. This is not putting down 4.1 it is just totally different and is going to take a while to get used to it. Plus you now have the DCC part of the operation which wil take some getting used too.

There are a lot of changes in the way your data is output. This may give you some headaches on the programs that have been written. Most of your toolbars at the top of the screen have changed along with the drop down menus. You will be searching for a while to figure out where things are.

DO go for the training class it will help you keep some of the hair you might still have left. I think you pull out the last strand around program 597....


I run 3.2063, 3.73 and 4.1 on the same machine.

[T Miller]

I would also think you would want to redo your manual alignment points with dcc points and do another alignment.

[John Kingston]

I agree with the 3.2063 option. As Craig said, CTRL-E each feature, then move the probe to where you want it to move and add a movepoint. Continue through the program until you are done. It will take some time, but would be easier that rewriting everything. Also, backup, backup, backup.

[jthompson]

If your program was written with all the features being measured and then an alignment at the end (why? I don't know but it's concievable)

Although not an answer to the posted question, I can say that we add alignments at the end of programs all the time. The reason is when we do first piece part approvals off of a 5 axis laser, the location of the features are at the mercy of how well our operator aligned and set his fixture up before he cut it. Theoretically (not always) the features should be good to themselves and generally are, BUT 99 % of the time the laser cut is off to the form, sometimes 2 to 3 mm. We would start the program with a basic 3-2-1 using the datum scheme but almost always have to add a form alignment at the end because datums B and or C could be off quite a bit. If we just start with a form alignment, then we would have to add read positions on all of the features and as you probably know, that gets really tedious. So, once we get the datum scheme looking good to the form we just recall the original alignment just before our dimensioning.

[craiger_ny]

My typical part program has many, many alignments.

[Matthew D. Hoedeman]

My typical part program has many, many alignments.

My typical part program has NO alignments, they are done in seperate programs that only align the part/fixture.

[Tested]

My typical part program has NO alignments, they are done in seperate programs that only align the part/fixture.

THEY SOUND LIKE THEY MUST BE VERY SIMPLE PARTS IF YOU CAN DO THEM WITH ONE ALIGNMENT. DON'T REMEMBER THE LAST TIME I GOT ONE LIKE THAT.