Strange fadal program error

I've been doing this long enough to know that the machine or control almost never makes an error. In fact, I can't ever remember seeing it happen, unless it's a breakdown situation.

Until now.

Here is the situation:

Simple G82 cycle for counter boring parts. Fixture holds 14 parts. First 2 parts good, to correct depth. Next part bad, about .250 shallow. Following 6 parts good, then 2 more bad ones. Final 3 parts good. Bearing race cutter rapids into part #2 with predictable results, one destroyed race cutter.

Same program, thousands of parts for over 10 years. No program changes. Ran fine all morning. The 3 bad parts appear to be the same depth. Machine is a Fadal VMC15, about 15 years old. It's been pretty reliable, just normal maintenance.

Anyone have ideas? Corrupted program? EMI? I'm at a loss on this one. Need to think about it while I make a new race cutter.

First, if you have a backup copy of the program check it line for line against the one on the machine. Somebody may have messed with it?!

Otherwise, I would first post the program here so we can check it for errors. I don't know how many times in the past I've had an error in a program stareing me right in the face and could not see it.

If the program is good, we can move on from there for other suggestions.

Steve

I have ran Fadals for years and have seen only one problem also. Different problem than yours however. I HAVE seen a program change itself somewhere along the way from a PC to a very old lathe control. It loaded the same program fine a hundred times before. I am sure this AIN'T YOUR situation. Wish I could help. Good luck.

Jeff

Doubtful it's the machine.

I might be convinced it's the machine IF you've run other part programs and experienced similar issues. Otherwise, I'm guessing operator error.

Here are some questions, food for thought:
1. Is the program in question installed on an as needed basis? Drip fed? Or constantly resident in the machine's memory or storage device? If the latter, any chance there's an older version, pre-de-bugging, that you've forgotten on the harddrive (assuming you have one -- my vmc 15 only has the factory 28k ram)? If it's drip fed, the same question applies to the connected computer. Also, if drip fed, communications can lead to vulnerabilities. I'm no expert on this, but seems Fadal is temperamental (from what I've read) when it comes to communication. It likes the blocks to be delivered nice and slow. I never drip above 9600 baud for that reason. Anything could be causing a "hiccup" to momentarily increase or decrease baud or snyc between the machine and computer and possibly lead to an error (possibly emi, as you suggest, a computer virus, software running on the communicating computer, corroded serial cable connections, heat, loose connection withing the comm cable sheathing etc). You'd only need a momentary interruption to cause a problem.

2. Assuming the program gets reinstalled as needed, have you been using the same computer for communications in every batch of parts? Mine's on a network and the program can come from just about anywhere... So, more chance for human error to be inadvertently introduced.

3. Is it consistently the same part in the sequence with the problems? That would imply a program issue, not machine. Raise your tools (or remove them altogether) and re-run the program. I'm betting the machine will make the same "mistake" every time.

4. Got "#"? I like using macros. Sometimes I use them dynamically to change tool diameters and lengths as well as part home locations. Any chance the latter two are integral to the program in question? Also,

5. With regard to the most recent program you ran prior to the program in question... Is it possible you've used a rarely used modal g-code? If so, maybe there's a missing gcode in the program's initialization sequence that in most situations wouldn't be a poblem, but for that rarely used modal setting?

6. You don't mention whether the program uses sub-programs or subroutines. If you are using them, then the number of chances for the program being responsible for the error, shrinks. Any chance an "R" parameter is being inadvertently changed by a macro?

7. Not sure if this could cause the problem or not, but... I only use Format 2 (assuming you have the cnc 88 control). Any chance the format has been changed?

8. Assuming you're using different part homes, have you checked the offsets to ensure all of your part homes are set to the same z-level (which should all be zero in my book).

And when you get right down to it, none of those may have anything to do with your problem. The point is, it could be any of myriad issues where human fingers are involved. Intermittent problems with your Fadal, while possible, are highly unlikely in this situation. Hopefully, something will click and you'll identify the problem. I'd be curious what you find.

Good luck.

girlpower said:

8. Assuming you're using different part homes, have you checked the offsets to ensure all of your part homes are set to the same z-level (which should all be zero in my book).

Click to expand...

That's where I would start, I've had it bite me before.

My 15 XT just did the same thing on the X axis last week. It was a bad resolver on the servo. Pull the resolver and spin it by hand, it should spin smoothly.

Mark Hockett

Have not had time to address this problem... have another more important crisis. Here are some answers to questions.

No macros. The program does contain subprograms, and does match the one we have been using for years exactly. The machine is hard wired to the shop computer where we store all the programs. We run lots of different jobs, and the program is downloaded as needed, always from the same computer. Format 2 used, with only 1 part home. It's starting to point to a machine problem... The operators say it's been acting up on the z axis recently.

Thanks for the suggestions, I will post when I find the solution.

steel hand said:

Have not had time to address this problem... have another more important crisis. Here are some answers to questions.

No macros. The program does contain subprograms, and does match the one .....Thanks for the suggestions, I will post when I find the solution.

Click to expand...

You didn't quite say these z move errors repeated in the same locations, just inferred it. So, if you run the program again will it repeat the same errors in the same places?

Mechanical:

Check your Z axis ball screw to servo motor coupling…It may be loose thus moving back and forth. You would expect a misalignment during a tool changing but maybe not. It may just be ‘loaded’ against the set screw just so… to still work

Check to see that your counter weight sprocket and chain is working smoothly throughout the whole travel.

Both of these are long shots for various reasons but possible causes.

I have a friend with a shop who has changed his G00 to about an inch above his normal feed starting point, then feeding at F50.0 to the final G00 stopping point because he was getting about a .06” z error and this fixed it. The problems not solved yet but from the sounds coming from the Z axis when it’s moving I suspect his z axis thrust bearing is bad. I assume your z axis is not noisy?

I’d check these out before buying any electronics.

Steve

steel hand said:

Have not had time to address this problem... have another more important crisis. Here are some answers to questions.

No macros. The program does contain subprograms, and does match the one we have been using for years exactly. The machine is hard wired to the shop computer where we store all the programs. We run lots of different jobs, and the program is downloaded as needed, always from the same computer. Format 2 used, with only 1 part home. It's starting to point to a machine problem... The operators say it's been acting up on the z axis recently.Thanks for the suggestions, I will post when I find the solution.

Click to expand...

I would have to believe the resolver on your Z axis is flaking out. I have replaced several on Fadal machining centers.

Update: I have not looked over the z axis bearings, resolver etc. I did edit the program to test the machine before I ran parts again. Have a travel indicator on a mag base in the machine to verify it repeats on the z axis. Worst it can do is damage a cheap travel indicator if it malfunctions. I go run it when I have time... and so far it seems to work perfectly. I hate these kind of problems. I'm sure it wants another expensive custom tool in the spindle before it repeats the problem...

Update #2: We ran 3 different jobs on the machine. No malfunctions. Went back to the job that caused the problem... 3RD PART CRASHED! Start looking closely at what is happening. Counterbore before the crashed tool is pretty dull, but spindle load never exceeds 60% Load 4 more parts, run first 3 tools, 3 bad, 1 correct depth. Z position readout says it is going to the same place every time. Easy to see because there is a P200 dwell in the G82 cycle. I know that can't be happening because of the parts. Check tool offset, it's good. At this point I'm very suspicious of some kind of mechanical problem. I set 2 steel blocks on the vise top, carefully bring bottom of spindle housing against it. Set up travel indicator, zero it. Start to bring z axis down... it stops according to the indicator, but the position readout keeps going as long as I turn the pulse generator. Pull access cover on Z servo motor to inspect motor/ball screw coupling. Looks ok to me... but I don't work on these much. The crazy thing is, it does not lose position. I can go .100 below where z axis is physically stopped, then reverse back, over and over. It touches the block at the same point. I'm wondering if this is some feature built in to prevent machine damage, but the parameter is set too low. Or is it simply because of the dull tool? I have had these tools get dull before and it will cause an alarm, exceeding 100% spindle load. It's not like we are pushing the machine here, it's a .781 diameter counterbore cutting aluminum. I will put a new tool in and watch it carefully, but can anyone explain this to me?

steel hand said:

Update #2: We ran 3 different jobs on the machine. No malfunctions. Went back to the job that caused the problem... 3RD PART CRASHED! Start looking closely at what is happening. Counterbore before the crashed tool is pretty dull, but spindle load never exceeds 60% Load 4 more parts, run first 3 tools, 3 bad, 1 correct depth. Z position readout says it is going to the same place every time. Easy to see because there is a P200 dwell in the G82 cycle. I know that can't be happening because of the parts. Check tool offset, it's good. At this point I'm very suspicious of some kind of mechanical problem. I set 2 steel blocks on the vise top, carefully bring bottom of spindle housing against it. Set up travel indicator, zero it. Start to bring z axis down... it stops according to the indicator, but the position readout keeps going as long as I turn the pulse generator. Pull access cover on Z servo motor to inspect motor/ball screw coupling. Looks ok to me... but I don't work on these much. The crazy thing is, it does not lose position. I can go .100 below where z axis is physically stopped, then reverse back, over and over. It touches the block at the same point. I'm wondering if this is some feature built in to prevent machine damage, but the parameter is set too low. Or is it simply because of the dull tool? I have had these tools get dull before and it will cause an alarm, exceeding 100% spindle load. It's not like we are pushing the machine here, it's a .781 diameter counterbore cutting aluminum. I will put a new tool in and watch it carefully, but can anyone explain this to me?

Click to expand...

Sorry I didn't see this thread before, and I honestly don't know much about Fadals, other than they seem like good, reliable machines.

From your first post, I suspected a mechanical, rather than program problem.
My first suspect would be the z axis encoder. The .250 error was a clue and it being inconsistent (some good parts, some bad).
It could be just a slipped coupling, but given the age and use of the machine, a worn out encoder wouldn't surprise me.

Try running at a lower rapid rate than the usual 100% balls-to-the-wall and see if it does anything different. It may or may not, but it might help you reduce the crashes for now.

If they're like most brands, it shouldn't be that expensive or hard to replace yourself.

If it was a loose bearing/ballscrew nut on the z axis, a backlash check and some hard listening should tell you if that's where your trouble is.

Again, as you indicated at first post, the likelihood of a good program suddenly going bad is pretty slim. And erring on the side of not offending an experienced machinist, I'll assume you already ruled out some other possibilities, like a slipped tool in the holder, or a fixture problem.

The handwheel thing is hard to figure out, I think I remember a machine I ran years ago that would do that - an Okuma.....or maybe Cincinnati mill ? I dunno.


But just a hunch from the symptoms.....encoder.

Strange fadal program error...

Steel Hand.....
Me thinks tis time to get out the ole 2 by 4 about 4 feet long...
Set up your indicators and your blocks like you mentioned above...
Use your hand wheel and move the machine into position.
Bring the Z axis down on the blocks like you mentioned and continue to crank (slowly) the handwheel and see that your problem is still there. Then crank the Z axis back up off the blocks and then down a little until you see the axis move. Readjust the indicator so it is on the spindle housing. Set a zero on the indicator and note the position on the readout on the control.
Now with your handy precision 2by4 by 4 or 5 feet, try to pry up on the spindle housing and watch your indicator and the value on the control. See if either of them change. If you indicator moves and the control remains the same, you got a lost motion problem (mechanical) If they both move, you got a drive problem. Maybe not enough gain.....
I love to use ye old pry bar 2by4 to look for loose gibs, loose ball screws, loose ball screw nuts etc etc.
Good luck.... I'm still thinking....
Regards Walt....

I'm thinking you should try to pry the spindle housing down as well as up. Then you might even try to pry on the spindle itself. Mount an indicator from the face of the housing to the face of the spindle itself. Press tool unclamp and then release the unclamp button and note the movement of the spindle and then without moving anything, pry up on the spindle itself and note any movement. This would indicate movement of the bearing pack on the spindle or movement within the spindle housing itself.
Still thinking....
Walt..

We are pretty sure it's an electrical problem somewhere. Have been working with a tech over the phone to troubleshoot it since yesterday afternoon. First thing we did was place a piece of wood underneath it. Barely put a dent in the wood. Now working on swapping some amplifiers to try to isolate the problem.

There is a Tech in Fremont named Aijia, I think, who is the FADAL expert. He is now buying older machines and rebuilding them. I will look around for his contact info.

Strange fadal program errror...

steel hand,
Barely put a dent in it.... OK
What was the numbers doing on the control at that time.
What was the lag? What was the distance to go? And then what was the position numbers doing...
If you have a scope or a recording volt meter, try looking at the tach feed back. You know, minus command, plus tach, plus command, minus tach. Watch tach feed back and look for overshoot.

Had a machine once that shut down on tool change....
It would go to tool change position, overshoot, realize its error and try to come back down but it hit the overtravel switch at the same time. It would start down and shut down (E stop) all at the same time and coast to a different position each time.
Increased the gain, tightened up the drive and lived happily ever after....
Regards Walt....

Still thinkin

Strangefadal prgm error

steel hand,
Just kinda bumping this back toward the top.

You fix? You no fix? Run out of 2x4s? Smash indicator?

We just gotta know..

Regards Walt...

Solution to the problem

We swapped the Y and Z amplifiers, and the Z axis behaved normally. Ordered a new amp, installed it this morning. Everything works as it should now.

Thanks for everyone's help. I learned a lot about these machines through this experience.

Thanks to the guys at FadalCNC.com. They talked me through it, and had a new amp here fast at a reasonable price.

Strange Program Error

Steel Hand..

Thanx for the update.....

Most people want to know what actually fixed the problem.
Also, it is nice to know if we helped or what you had to do to fix it.
We may have the same problem someday....

Regards Walt...