Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Spinnetti on March 25, 2008, 11:29:38 PM
-
Hi all, sorry if this is a repost somewhere...
I'm using the latest release of Mach 3, and on one part, it sometimes just wanders off. Its a simple 2d part consisting of mostly straight lines and simple 1/2 circles. Some times, but only sometimes! it just wanders off in the middle of the part like it lost 1/2 of a line segment. The weird thing is that it does it both on the Mach tool path simulation and the real part, but doesn't do it every time. Any ideas? Kinda scary for me! (broken tools to attest). I took screen shots if it helps.
-
Hi,
The same thing happens to me.
Over the weekend 1 made 200 parts, and sometimes the program just does not follow the the programmed path.
It just moves in a total different way in the middle of the part.
On the toolpath display on the PC also the wrong direction is indicated, which means it must be a software problem
Jan
-
Will ask Brian to have a look and see if he can find anything.
Hood
-
Hello,
Please post the code and I will have a look... And is this with Rev 3.039??
Thanks
Brian
-
I will post the code tonight when I get home.
But in my case it always happened between 2 G01 moves. It does not always happen on the same line of code, but randomly (but sofar always between two G01 moves)
Jan
-
Hi,
Here is my code, with in parenthesis what happened on several occasions
N100 T6 M06
N110 G00 Z2.000
N111 G00 Y70.000
N112 G00 X148.000
N120 G00 X113.000 Y43.000
N130 G01 Z-4.500 F100
N140 G03 X107.000 Y43.000 I-3.000 J0.000 F150
N150 G01 Y41.000
N160 G03 X113.000 Y41.000 I3.000 J0.000
N170 G01 Y42.950
N180 G00 Z2.000
N190 G00 X38.000 Y43.000
N200 G01 Z-4.500 F100
N210 G03 X32.000 Y43.000 I-3.000 J0.000 F150
N220 G01 Y41.000
N230 G03 X38.000 Y41.000 I3.000 J0.000
N240 G01 Y42.950
N250 G00 Z2.000
N1020 G00 X-3.000
N1025 G00 Y-3.000
N1030 G01 Z-4.500 F100
N1040 G01 Y30.000 F350
N1050 G01 X15.000
N1060 G03 X17.000 Y32.000 I0.000 J2.000
N1070 G01 Y65.000 F200
(*After completing this line no move to X53,)
( but a move started to the -Y direction, until I hit Esc)
N1080 G01 X53.000
N1090 G01 Y22.000
N1100 G03 X55.000 Y20.000 I2.000 J0.000 F350
N1110 G01 X90.000
N1120 G03 X92.000 Y22.000 I0.000 J2.000
N1130 G01 Y65.000 F150
(*In the middle of this move to Y65, the mill started to)
( move in the +X direction suddenly, until I hit Esc)
(*After completing this line no move to X128,)
( but a move started to the +Y direction, until I hit Esc)
N1140 G01 X128.000 F200
N1150 G01 Y32.000 F200
N1160 G03 X130.000 Y30.000 I2.000 J0.000 F350
N1170 G01 X148.000
N1180 G01 Y-3.000
N1190 G00 Z2.000
N1200 G00 Y70.0000
N1210 G00 X148.0000
M30
Jan
-
Hi,
I forgot to mention that I use 3.039
Jan
-
I posted my code and pictures here:
http://www.machsupport.com/forum/index.php/topic,6380.0.html
I wish there weren't so many dang forums.. I feel like I have to chase everywhere.
I'm on the latest version, and get this on multiple computers. This is a show stopper for me, and I've crashed tools twice now as I'm working close to my fixtures.
-
I will test today ... I am headed to LA for Westec tomorrow and will be back working on the 3rd... So if I can not find the problem today I will get on it after I get back.. Thank for the code and any info that you can give me :)
Brian
-
I certainly don't want to tread on Brians toes - but I notice that the line BEFORE the line that causes trouble is a G3 move. This is followed by an instruction where ( and I know this is allowable ) only one parameter is specified.
Could it be that Mach is not identifying the second parameter correctly.
I would try inserting the missing X parameter in line 1070 and 1130 and see if that solves the problem for the time being. It certainly can't do any harm.
-
Also could you tell me where you are running the file from?? are you running from the local hard drive ? Over the network ??
Could I please have your XML as well :)
Thanks
Brian
-
Brian,
I run from a local hard drive.
I attached my XML file
As I told before I made 200 parts, and 4 or 5 times the system failed (so 98% is OK).
I think this will make it hard to find the problem.
Jan
-
Thanks for the replies. - Is this the right place for "official" support questions?
Here's my issue detail with code and pictures: http://www.machsupport.com/forum/index.php/topic,6380.0.html
Its really odd though that it only goofs up sometimes, and never in the same spot. I'm really gun shy now though because I have broken two tools - even though my hand is on the stop, I never know where its going to happen. Some additional info that should make troubleshooting easier:
I downloaded the latest version of mach3, and installed totally default - no mods of any kind on two different computers (one Mac, one PC, both on Windows XP SP2 with all current updates)
I also re-did the part in my cam software and still get the same result
The part shows correctly in the display, and as the pictures show in the link above, also shows where it wanders off in two of the cases.
Running just the software (i.e. not running my actual machine), it seems to fail about 30% of the time, but I haven't run it a ton of times.
I run the code from the local hard drive. The computer is not network attached, and has a clean install on it, so very little virus potential.
Any help will be greatly appreciated as I'm dead in the water and can't make parts :(
-
Hello,
I have only 2 reports of this and so it is very hard to find the problem... I will work to get this fixed as soon as i get back on the 3rd.
Thank you for the XML and I hope this will help to show the trouble!
thanks
Brian
-
Brian,
Any chance to look at this yet? From the info I provided, this is all within Mach3, which should help narrow it down. If I can't fix this, I'm going to have to buy other software which doesn't make me very happy.... I need to cut some parts.....
Thanks.
Hello,
I have only 2 reports of this and so it is very hard to find the problem... I will work to get this fixed as soon as i get back on the 3rd.
Thank you for the XML and I hope this will help to show the trouble!
thanks
Brian
-
I have run this file 40 times and have had no trouble... I see that you are not using the 1024.set screen.. Could you please test with the 1024 screen and tell me if you have the same trouble? Also I see that you are not running the files on the C drive... Could you please put the file on the C: drive and tell me if this fixes the problem? The Problem is that it is a memory mapped file :( This is needed to run the VERY large files...
Thanks and all told I have run your file about 60 times with no problem :( I don't think I am going to see the problem but I have a few more machines that I can test on.
Brian
-
Brian,
Thanks for looking at it. I certainly didn't expect you to run it that much. I get it within a few runs. What you are seeing in the pictures I posted is me running it on my Mac (more convenient that going down in the shop). It is on the C drive, though my screen res is 1680x1050. Issue is that when I put the Gcode on my PC which is running the 1024 set, and is certainly on the C:drive (it doesn't even have a floppy or CD drive in it, and no network connection), the result is the same on both my new Mac and my old AMD 1.2gig PC. Any issues running AMD cpus? I've been using it for years without problems until now. What I do is generate the G-code in Mastercam, put on a USB stick then put it on the shop computer. I put it on the Mac to see if I could replicate the problem with a fresh mach 3 download and sure enough, same issue. I still have the crashed part on the mill if you want a pic, and can screen shot on the target computer, but the results are the same on both computers. Can I get older version of Mach to try on? I only need basic functionality anyway. Maybe I should run on Mach 2 or even Master 5?
I have run this file 40 times and have had no trouble... I see that you are not using the 1024.set screen.. Could you please test with the 1024 screen and tell me if you have the same trouble? Also I see that you are not running the files on the C drive... Could you please put the file on the C: drive and tell me if this fixes the problem? The Problem is that it is a memory mapped file :( This is needed to run the VERY large files...
Thanks and all told I have run your file about 60 times with no problem :( I don't think I am going to see the problem but I have a few more machines that I can test on.
Brian
-
I am okay with testing the code for you.. I understand that not having a running machine is of no use to you so that is why I ran it so many times :) I just want to fix the problem... I have an idea where you may be getting this problem from and I will try and have a test rev for you soon... But I am flying blind and this is never a good thing :o
I should have a new rev out in a day or so... Please set your look ahead to 1 and tell me if you have the same problem..Thanks
Brian
-
Ok, Where do I set the look ahead? I took a look, but don't see where that is.
Thanks.
I am okay with testing the code for you.. I understand that not having a running machine is of no use to you so that is why I ran it so many times :) I just want to fix the problem... I have an idea where you may be getting this problem from and I will try and have a test rev for you soon... But I am flying blind and this is never a good thing :o
I should have a new rev out in a day or so... Please set your look ahead to 1 and tell me if you have the same problem..Thanks
Brian
-
Hi, Spinnetti
Look Ahead is in Config, General Configuration.
Chip
-
Thanks.. breezed right over it..
Haven't re-run yet, but did change the config. I hope to try tonight sometime. Thanks.
Hi, Spinnetti
Look Ahead is in Config, General Configuration.
Chip
-
HEY SPINNETTI
i was reading your post and thought maybee we had the same problem, we don't. took your program and ran it - could see where it missed x53 move. -went through the code and made sure that every line had a carriage return.
here is a copy for you,take a look at it. moving on to my ghost hunt now.
bill
-
Thanks for lookin..
I did run a few parts with lookahead set to 1 and it worked ok, but thats not a "for sure".. could have just been luck of the draw. I'm out of town for the rest of the week, but hope to run some more this weekend.
Thanks!
HEY SPINNETTI
i was reading your post and thought maybee we had the same problem, we don't. took your program and ran it - could see where it missed x53 move. -went through the code and made sure that every line had a carriage return.
here is a copy for you,take a look at it. moving on to my ghost hunt now.
bill
-
Brian,
Running the file from my C drive did not fix the problem. The first time I ran the file it messed up
After I changed the "look ahead" to 1, the problems seems to dissappear (running on my home PC, not the mill)
I ran the file approx 20 times without problems.
This does not mean I am 100% sure the problem is fixed, because during my last production run of 200 pieces, only 4 pieces went wrong
Anyway, it looks promising, I hope it works as well for Spinnetti
Thanks
Jan
-
Sounds to me like you have a problem with the lookahead.. That is just odd.. I have run your file here many times and have not seen that problem! I will keep an eye out for it!
Thanks
Brian
-
Hi,
I've had the same problem while working with versions 2.62, 3.00 and 3.01. After updating to 3.039 the problem with lookahead seems to be cured.
But now I have problem with feed rates :
http://www.machsupport.com/forum/index.php/topic,6640.0.html
Rgds
-
I set the look ahead to 1 and am not having problems, but I've only run a dozen parts or so. Problem is, I don't have full confidence that the problem is done, or just less likely. The machine crashes not only scare me, but will cost me more money in damaged equipment.
Any more thoughts on root cause resolution?
Thanks
Brian,
Running the file from my C drive did not fix the problem. The first time I ran the file it messed up
After I changed the "look ahead" to 1, the problems seems to dissappear (running on my home PC, not the mill)
I ran the file approx 20 times without problems.
This does not mean I am 100% sure the problem is fixed, because during my last production run of 200 pieces, only 4 pieces went wrong
Anyway, it looks promising, I hope it works as well for Spinnetti
Thanks
Jan
-
I have not seen this problem other then the few that have posted on the forum... I have only seen this problem with old Revs of mach3... IF you have this happen with the Dev version and can send me your data I will do what I can.
Thanks
Brian