Wierd things going on with Alignments v4.2

**rangerboat72** · 04-05-2007, 06:10 AM

Come on man, don't scare me with "potential bugs" before I've even got a chance to install it...

**MrComment** · 04-05-2007, 10:07 AM

Im hoping that if I post them now, maybe there will be a chance that they get squished before you install it.

Besides it makes me feel good to get it off my chest, but then again I dont expect it to be bug free at this point anyway.

**MrComment** · 04-05-2007, 05:22 PM

Proof and the smoking gun.

For my unofficial fellow v4.2 Beta Testers.

I attached a simple fixture program that I started from Scratch in v4.2 (Build Date: Mar 22 2007).

The problem which you can see from the Excel Spreadsheet is that the Alignments seem to only execute during Run-Time every other time. At first I thought that this was only a issue with Recall Alignments, but I proved that wasnt the case when I replaced all of my Recall Alignments with unique ones.

In either case, it makes using v4.2 almost impossible unless you want to Step through your programs line by line.

I still find it odd that it would skip every other alignment though. Maybe its a new feature?

**JamesMannes** · 04-05-2007, 06:20 PM

Originally posted by MrComment View Post

For my unofficial fellow v4.2 Beta Testers.

I attached a simple fixture program that I started from Scratch in v4.2 (Build Date: Mar 22 2007).

The problem which you can see from the Excel Spreadsheet is that the Alignments seem to only execute during Run-Time every other time. At first I thought that this was only a issue with Recall Alignments, but I proved that wasnt the case when I replaced all of my Recall Alignments with unique ones.

In either case, it makes using v4.2 almost impossible unless you want to Step through your programs line by line.

I still find it odd that it would skip every other alignment though. Maybe its a new feature?

I haven't looked at your data, but what is the date of your build is it 03-22-07(sorry didn't see your build date b4)? I have been using that build since 03-23-07 and have no problem like what you describe. Perhaps try an un-install and re-install?

**MrComment** · 04-09-2007, 04:58 PM

Just got done uninstalling and reinstalling the software. Tried it but unfortunately it still does the same thing.

**ARIES** · 04-12-2007, 05:31 AM

MrComment, please, can you describe (step by step) and more carefully the problem of the crash?
Attaching the complete PRG program related to the part of code that you have written.
You also add the PROBE file and eventually the DEBUG files produced from both the PC-DMIS versions (4.1 and 4.2).
Add also the DEBUG files when the program run fine and when there are the error.
You produce the separate debug files, one for every test.
Please, also attaches the register file (produced with the settings editor).
I will examine asap this your problem on our machine with SP600 and to try to reproduce it.
Thanks, regards.

Originally posted by MrComment View Post

Code:

A15        =ALIGNMENT/START,RECALL:A10,LIST=YES
            ALIGNMENT/LEVEL,ZPLUS,PLN4
            ALIGNMENT/TRANS,ZAXIS,PLN4
            ALIGNMENT/END
A16        =ALIGNMENT/START,RECALL:A15,LIST=YES
            ALIGNMENT/BF2D,ZPLUS,LEASTSQR,CREATE WEIGHTS=NO,ROTANDTRANS,USE SCALING=YES,-0.044,0.343,0,0.233,0.9988976
               SHOWALLINPUTS=NO,SHOWALLPARAMS=NO
            ALIGNMENT/END
            PREHIT/ 1
            RETRACT/ 1
SCN27      =FEAT/SCAN,LINEAROPEN,SHOWHITS=NO,SHOWALLPARAMS=YES
            EXEC MODE=DEFINED, NOMS MODE=FIND NOMS,CLEARPLANE=NO,SINGLE POINT=NO,THICKNESS=0
            FINDNOMS=2,SELECTEDONLY=YES,USEBESTFIT=NO,PROBECOMP=YES,AVOIDANCE MOVE=NO,DISTANCE=0,CAD Compensation=NO
            DIR1=NULLFILTER,
            HITTYPE=VECTOR
            INITVEC=-0.2249569,-0.9743687,0
            DIRVEC=0.9743687,-0.224957,0
            CUTVEC=0,0,1
            ENDVEC=0.2249569,0.9743687,0
            PLANEVEC=0.9743687,-0.2249569,0
            POINT1=-82.205,40.898,-2
            POINT2=-120.026,-42.922,-2
            MEAS/SCAN
            BASICSCAN/LINE,SHOWHITS=NO,SHOWALLPARAMS=YES
            <-82.205,40.898,-2>,<-120.026,-42.922,-2>,CutVec=0,0,1,DirVec=0.9743687,-0.2249569,0
            InitVec=-0.2249569,-0.9743687,0,EndVec=0.2249569,0.9743687,0,THICKNESS=0
            FILTER/NULLFILTER,
            EXEC MODE=DEFINED
            BOUNDARY/PLANE,<-120.026,-42.922,-2>,PlaneVec=0.9743687,-0.2249569,0,Crossings=1
            HITTYPE/VECTOR
            NOMS MODE=FINDNOMS,2
            ENDSCAN
            ENDMEAS/
            PREHIT/ 0.5
            RETRACT/ 0.5

Maybe a Potential Bug?

Not sure exactly what is going wrong here. Ctrl-E to the move point just before the alignments, then Ctrl-E on SCN27 and the machine will execute both commands just fine, no errors.

If I execute the entire program though (and without moving the part) then the program crashes into the part on this scan.

I'd almost have to guess that PCDMIS possibly is ignoring the alignment shown in this code segment. (But only during Run-Time)

I think I also noticed some small degree of transformation error between recalled external alignments between two programs also?

One more oddness is that all of my Operator comments keep popping up after any 2DBF alignment executes anywhere in my program. But for this particular probe crash the Comment bug doesnt pop up before it trys to execute SCN27.

I got enough part deviation from the cad to notice when alignments dont work right before measuring the features.

**Lee Johnson** · 04-12-2007, 12:53 PM

Is this only a problem with scanned features ?
I was hoping to see a new version I could use come out.
I ocassionaly use scan features so this would be a problem for me on some programs.

**MrComment** · 04-12-2007, 03:03 PM

Debugged sort of. (At least not exclusively a 4.2 bug)

Actually as it turns out this isnt a v4.2 problem.

After fiddle farting around with this and that today, I found the same bug in v4.1.

Which got me looking at what was different between this program and others that I have made, which I know work.

I noticed that if I recreated the CIR and a move point to get out of the hole while in the Scan Alignment it now appears to work ok without stepping through the code. Well at least so far so good anyway.

Maybe somewhere along the line as I continue to rewrite this part of the program I'll be able to identify exactly what I did wrong.

But for now I guess my bug was something about the way I wrote the last program.

So Sorry everyone, The sky isnt falling after all.

**MrComment** · 04-12-2007, 03:54 PM

Debug Files

"This one works v42.zip" include the probe files and Program in v42 that works ok. No issues on the machine.

"This one dont work v42.zip" includes the probe, debug and Part Program in v42 that doesnt work.

Specifically the bug is, that on the Machine you have to Step through the code for it to work correctly. "Correctly", meaning like the previous program runs.

"Debug2.zip" is the debug file when you step through the program which doesnt work correctly.

As it turns out the issue is also repeatable on v4.1

Ive looked through the code but I cant seem to figure out what Im doing to make the Program behave different when Stepping though, versus Executing normally.

Attached Files

**MrComment** · 04-12-2007, 04:18 PM

One more thing I almost forgot.

And this is a bug in v4.2

The Operator comments in my program keep popping up when I execute a BF alignment. Apparently once for every input in the BF alignment. This doesnt happen in v4.1.

**ARIES** · 04-19-2007, 03:08 PM

MrComment, be patient, I am really reproducing your programs on our Global machine. Thanks.

**ARIES** · 04-20-2007, 07:50 AM

I have succeeded in reproducing this problem. We are investigating on possible other problems.

Originally posted by MrComment View Post

One more thing I almost forgot.

And this is a bug in v4.2

The Operator comments in my program keep popping up when I execute a BF alignment. Apparently once for every input in the BF alignment. This doesnt happen in v4.1.

**Cumofo** · 04-20-2007, 08:00 PM

It happens to me too!!!
Arrrghhh!!!
I think if you write a comment in a block before and alignment it bugs hard. Some type of read ahead script gets stuck in a loop.

_G

**T-Rex** · 04-23-2007, 05:51 PM

The Operator comments (multiple pop ups) has been addressed.

Wierd things going on with Alignments v4.2

Wierd things going on with Alignments v4.2

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Related Topics