[panaceabea]

I have run the very simple string below thousands of times on my milltronics and bridgeport with zero problems and somewhere around 150-160 times on the new HAAS but for some reason a few minutes ago the HAAS just decided to skip the tool change from T15 , a center drill to T30 which is a 3/32 drill bit thats aprox 2" longer than the center drill.

Machine didnt change tools and as a result destroyed 10 parts with the tool holder and damaged my fixture. The fixture is repairable and the damaged parts only set me back a little over $500 , my main concern is why would the machine have skipped a tool change ? I want to have confidence that I can leave it to run un attended for long periods . Parts lost today were cheap , but I do on occasion machine injection molding dies that run 40+ hours machine time that I cant so easily afford to scrap.


What might have caused the machine to ignore the tool change ?

G57
(MOVE BACK TO TOP TO DRILL)
(Tool 15 = CENTERDRILL)
N1 T15 M06
M08
M03 S3000
G00 Z1.1 G43 H15
N2 G81 X3.9077 Y-0.331 Z-0.1 R0.1 F12.
N3 X7.1077
N4 X10.3077
N5 X13.5077
N6 X16.7077
G80
N7 G00 Z0.1


N1 T30 M06
M08
M03 S5000
G00 Z1.1 G43 H30
N2 G83 Q0.1 X3.9077 Y-0.331 Z-1.1 R0.1 F18.
N3 X7.1077
N4 X10.3077
N5 X13.5077
N6 X16.7077
G80
N7 G00 Z0.1

[mactec54]

Hi panaceabea

Your Gcode format is not very good I think also your tools may be not set right
is the 1.1 above the part for some reasion

Each tool should be touched off the top of the part

I have attched a txt file that you can run make sure the numbers are the same as yours
you may need to remove the G98 as this will make your retracts go to 1.1 instead of R.1

Also you may need to move the G57 to above the S3000M3 & S5000M3 some controls don't like were it is

[Geof]

This is a new Haas? It will have proximity sensors not mechanical switches.

I don't think we have ever missed a tool change and had the program continue to run but we have sometimes had the toolchanger get lost and trigger an alarm because a chip landed on one the the proximity sensors that detect the position of the carousel.

The big problem is finding the culprit chip still in place so you can wipe it off and have the problem go away to more or less prove that was the cause.

I think this is going to be one of those glitches that you have to grit your teeth about and cross your fingers as you press the green button. Maybe Murphy is now satisfied and it will never happen again.

[vlmarshall]

Wow, I've never seen (in 12 or 13 years of daily use) a Haas skip a tool change...but I've also never seen anyone re-use the same line numbers in a program.
I'm going to guess that the duplicate line numbers confused the Hass's Block Lookahead somehow.

[Geof]

....I'm going to guess that the duplicate line numbers confused the Hass's Block Lookahead somehow.

Not likely. The Haas controller only takes notice of the line numbers when you do a subroutine or line jump. The local subroutine call M97 P22 for instance looks for the first line numbered N22 and goes to that. You can have many lines labelled N22 but it goes to the first and ignores all the rest. The jump command M99 P22 will jump to the first N22 it finds.

[vlmarshall]

Not likely. The Haas controller only takes notice of the line numbers when you do a subroutine or line jump. The local subroutine call M97 P22 for instance looks for the first line numbered N22 and goes to that. You can have many lines labelled N22 but it goes to the first and ignores all the rest. The jump command M99 P22 will jump to the first N22 it finds.

Too true. Ah well, I have no useful ideas as to why his machine skipped the toolchange.

It's 2 AM... why am I on here?? Addicted, I guess.

[panaceabea]

Yes its a 2009

The tools are set correctly ,the 1.1 inch above the part is a point it can rapid to and be clear of clamps, the duplicate line numbers are just a result of piecing together multiple strings of code manualy (cut and paste) in a text editor and not taking the time to delete the line numbers. I have never payed attention to line numbers , never had an issue before that resulted from random N** sequence. The program the above block is part of has run thousands of times with no problem on two other machines and at least 165 times on the new haas over the past few days. It had run 9 other times earlier in the day with no problems.

There were three tool changes and operations that ran before the change to T15 and then the missed tool change. This same program with the same tools, no changes had run at least 165 times over the past few days with no problem . I cant figure how it could be a prox sensor in the tool changer unless it were just after start up and this problem occured after 9+ hours of running the same program. The machine had cycled thru at least 50 or so tool changes at this point in the day.

I will just sit with the machine the next few weeks and make sure it goes thru all the tool changes before leaving it un-attended again in hope Murphy doesnt return

[djr76]

If you are gonna sit and babysit, you can set up M00 after each tool change so if infact it does go wonky, it wont crash into your part. It will suck if you have a alot of tool changes in a relatively short period though since you will have to push the start button to advance each time.

[Caue]

The machine didn't skip only tool change, but al the comands between N1 and N2, because if it has processed G143 H30 no damage would occur. So, it's very hard to accept this situation, but if you are TOTALLY sure that no one has touched the machine while it was running, look after the machine during the first days of operation, after fixed, and imediately turn the setting 36 - Program Restart to ON (the default factory condition is OFF). I'm pretty sure that if it was turned ON before, this damage wouldn't occur.

Keep in touch about this issue, I bet we all want to know about a happy ending for your history.

[Superman]

What might have caused the machine to ignore the tool change ?
G00 Z1.1 G43 H15

I don't like the look of this line,
G43 ( take up tool length ) usually forces rapid, but with a G00 at the front may cause the machine to ignore the length comp , and give the appearance that it missed a toolchange. Query your manuals if these 2 can be on the same line ???

I would write it as to be safe

(MOVE BACK TO TOP TO DRILL) 
(Tool 15 = CENTERDRILL) 
N1 T15 M06 
X3.9077 Y-0.331 
M03 S3000 
G43 H15 Z1.1
M08 
G99 G81 X3.9077 Y-0.331 Z-0.1 R0.1 F12. (G99 = return to the Z1.1 after each hole)
X7.1077 
X10.3077 
X13.5077 
X16.7077
G80 
G0 Z1.1

[mactec54]

Hi superman

A G99 will not return the Z to the Z1.1 it will return to the R value of .1 in the canned cycle

It has to be a G98 to do this

[mactec54]

Hi panaceabea

I agree with Caue that if everything was set up correct as you have said even if it did miss a toolchange it still would not crash your tool holder into your work & damage your fixture so something else happened to do this the worst it could do is break the drill & keep going

[Caue]

Only to reinforce my point: if the tool that is loaded into spindle is the smaller one (T15) and the machine is executing the operation f the longer (T30 and considering lenght offset value H30), the tool wouldn't even touch the part, maybe.
What certainly occur is that T15 was loaded into spindle and H15 was being considered.
I'm not here to disagree or to put doubt on a customer complainment, but the only reasonable way of occuring this is a block search with Setting 36 turned OFF.
That's why I told you to turn this setting ON and take care of the machine during the first days of production. We'll always be here to help you, but this help is tottally useless after the damage already occured.
Good Luck!!

[panaceabea]

I just checked setting 36 and it is on , and has never been accesed before today that I know of .

I have a vid surveillance that covers the entire shop and looked back at the tape and at no point was the machine touched other than parts changes before the skipped tool change .

I was the only person in the shop that day so I am sure no one else could have touched the machine.

I havent ever had a problem with having the G00Z1.1 G43 H** all on the same line , its the way mastercamx2 writes it when set for the haas 3 axis vertical processor, Its also the way the HFO TECH was writing it the day he came out to help with TLO setup mental block I was having I dont believe it has any effect on the which tool length offset is loaded from the tool library wether its on the G43 line or the line right after it .

I use the z1.1 to clear clamps but do not want a retract of 1.1" at any other point than right after a tool change, I like it at that point because I know no matter whats going on it will clear every clamp on the setup.

The retract hieghts for different tools at different points in the operation are all set seperatley to save time, depending on whats going on and how much clearance is needed at a paticular place on the part, quite commonly I rapid at .05" above the surface of the part , if I use g98 set at z1.1" it would just add to much time .

I appreciate the differences in programing styles but I dont think that has anything to do with why the machine skipped a tool change .

Below is a pic of the 5 parts it missed the tool change on , in this pic I have already removed the 5 parts from the fixture closest to the camera, as it cycled thru all the proper tool changes on that side of the jig.

There are 17 tool changes in the run before the point in which this tool change was missed and 1 after , it did all 17 tool changes before the skip and then did the one tool change after the skip. The program has not been modified in any way and has run over 200 cycles since the skip with no incident

[Caue]

About the program you can be pretty sure there's nothing wrong. Just like you told, there are different styles of programming, all of those are correct. The doubt is still floating in the air... unfortunatelly.

[HuFlungDung]

T30....does this mean that this is a side mount toolchanger? How, what and where does the controller look to preposition for the next tool?

[panaceabea]

YES 40 Position side mount T6 M** for tool change

[HuFlungDung]

So are you supposed to call the next tool (without the M6) after the current tool is installed or does the control just read ahead and look for the next tool by itself?

[Matt@RFR]

Hu, you can pre-stage tools if you want, but the control looks ahead 100 blocks (wild guess there) and does it for you.

[HuFlungDung]

Ok, I was wondering. Now of course, when anything complex like prestaging is offered as 'automatic', it immediately comes under suspicion in the sense that a bug lurks within the logic