CNC Root Cause, Corrective Actions & Lessons Learned

cutter said:
So you are saying that .003 vs .002 clean up pass is going to deflect the
carbide cutter. Sorry, but I don't agree.
Click to expand...

I was trying to say that I like my clean-up pass to be a specific, calculated depth for certain reasons. However, in thinking through a few scenarios, it seems like putting the compensation in the rough cut or the finish pass is six of one or half-dozen of another. Are there some advantages to putting it in the finish pass?
 
Last edited:
Cnc

Like everything in cuemaking, different strokes for different folks. I didn't mean to be short in my answer. As you said repeatability is the key.
I learned my way from the machinest 2 doors down. He has a high end shop, Haas VF2, 2 Haas TM1p, and a Tl1 lathe.
He makes parts for companies that make cnc machines. When he's proofing a part, he tweaks
his tool offsets in his clean up pass to hit what he needs. And he's working +.000 to - .0005. So no room for error.
Most cuemakers machines won't hold those kind of numbers. One other thing that he told me was about conventional and climb cutting.
If your machine is rigid with no slop, climb cutting seems to give a better finish cut. But the machine must be rigid and no play at all.
I conventional cut on my bridgeport manual mill, but climb on the Techno with linear ways, and ball screws.
I hope some of the people who say cnc is cheating have been reading these posts. It's not plug and play by any means.
 
Mr Hoppe said:
I was trying to say that I like my clean-up pass to be a specific, calculated depth for certain reasons. However, in thinking through a few scenarios, it seems like putting the compensation in the rough cut or the finish pass is six of one or half-dozen of another. Are there some advantages to putting it in the finish pass?
Click to expand...

The advantage of the finish pass in the last 2 cuts is to try and make both passes of similar amount.Then if the size is not correct, cutter wear etc, it can be measured and then the correct cut taken on the last pass.
In some materials and tooling it is not possible to take .01mm cuts.
 
Hi,

I just tested my "Fin" pockets (12 on the A) and guess what, Jim and Royce were correct. They did the spiral dance on the X again.

My corrective actions and root cause evaluation just blew up in my face.

I have no idea why my last Atooth pocket did not skip on the X.

It feels like I am in the Twightlight Zone.

I noticed that the output rotary section in Bobcad has an area for off set and all 12 pockets were set at #1. I just went ahead and set them 1 thru 12 and am running another test routine with my fingers crossed.

Desperate people do desperate things LOL.

Rick
 
Last edited:
feed and speeds

RBC said:
I'm sorry Rick, but I have to agree with Jim. At least this time I do! LOL

Your machine should more than enough power to snap .03125" bits all day long. Not to mention that at 5" a minute any good bit will easily cut through any wood you'd be cutting. What kind of spindle are you using? What's the RPM?

As for the 36 or 72 features in your cam tree, there are much easier ways to get what you are building than creating a feature for each inlay. You could just post the code for one inlay, use it as a subroutine and hand code the rotational moves in between. There are lots of ways to do this that are much easier than creating 36 separate features, all for the same geometry.
Click to expand...

i agree 5 in per min
at that speed i think the wood would be smoking. cnc zone has a training video feed and speeds on different kinds of wood.

normally if you have somthing mech loose it stays loose
i alway cut a circle to see if there is somthing loose on the drive system
that covers your x and y axis

i had a electrical noise problem one time i couldnt calibrate my steps
it also made my limits trip even when whey were disconnected.

my internet wireless router was next to my control box.
it was all bad ,

power supply cord to you wood router can cause electrical noise too.

mike
 
Last edited:
scdiveteam said:
Hi,

I just tested my "Fin" pockets (12 on the A) and guess what, Jim and Royce were correct. They did the spiral dance on the X again.

My corrective actions and root cause evaluation just blew up in my face.

I have no idea why my last Atooth pocket did not skip on the X.

It feels like I am in the Twightlight Zone.

I noticed that the output rotary section in Bobcad has an area for off set and all 12 pockets were set at #1. I just went ahead and set them 1 thru 12 and am running another test routine with my fingers crossed.

Desperate people do desperate things LOL.

Rick
Click to expand...

Ok Rick, actually though your program is too long I doubt it is there. You have some sort of mechanical or electrical issue. I would assume mechanical. Let me ask this, is your spindle or mounting bracket hitting anything on your rotary indexer? I noticed from your pictures that you where cutting the pockets close to the butt. The tenon seems rather short here. I use a maintenance pin on my back end for a tenon. I just choose a thread size smaller than my weight bolt size.

Jim.
 
Mc2 said:
Ok Rick, actually though your program is too long I doubt it is there. You have some sort of mechanical or electrical issue. I would assume mechanical. Let me ask this, is your spindle or mounting bracket hitting anything on your rotary indexer? I noticed from your pictures that you where cutting the pockets close to the butt. The tenon seems rather short here. I use a maintenance pin on my back end for a tenon. I just choose a thread size smaller than my weight bolt size.

Jim.
Click to expand...

Jim,

I think that makes sense because when I cut one pocket at a time, "go to zero" then "G0A 30" I am getting the same thing with all programs even when I "Saved As". with one feature.

My friend Alan is a CNC instructor and he looked at the code last week and said he did not see anything that would throw a wrench in the works. He also told me as you and others did that the program could be shortened because there were better ways to run it.

You mentioned electrical problem. Would my air compressor start, stop, then re start over and over during the long running routine have any effects on the controller or steppers? If AC amp draws can effect the DC or the computer transfer of pulses or whatever then are there filters or things that can be installed for this purpose. I have a battery back up with surge protection on my computer but I think my Controller is just on a surge protected power strip circuit. Maybe I ought to re think that strategy?

If so I can run with the compressor off and put the set point on the unloader down to only 5 psi. Right now my compressor is set for 110 psi and my low set point on the unloader is 100 psi. It cycles quite a bit even though my regulator on the CNC is low under 5 psi.

Also do you think I should interchange my X & Y stepper motors to test the X function using the Y motor. The lead screws are clean as all I hear is smooth transitions sounds with no groans. I keep the works very clean and all of the fine particles are being drawn away with a high volume vac. system in the CNC enclosure. One of the other things I was thinking about is those anti backlash things on the on the nuts? They seems to be working fine and don't look bad but what do I know of those kind of details. At some point in my troubleshooting I may have to change them out. who knows.

What really spooks me is that my parts and pocket fit is very very good throughout this whole process on every radial.

Thanks,

Rick
 
Last edited:
Travis Niklich said:
Maybe you have a screw loose literally
Click to expand...

Thanks Travis,

I checked and re checked all couplings and mechanics. That kind of stuff I understand.

I think Jim is right about electrical being my problem. I did not know these things were so sensitive. Mike said an internet router being close to his stuff made a problem.

Thanks Again for your thoughts,

Rick
 
have you put a dial indicator on it to check for backlash on the x and y worn bearings, couples and sometimes the nuts on the end of the screw can cause some problems like that to. I don't know much about the electric end of it.
 
Travis Niklich said:
have you put a dial indicator on it to check for backlash on the x and y worn bearings, couples and sometimes the nuts on the end of the screw can cause some problems like that to. I don't know much about the electric end of it.
Click to expand...

Travis,

Mr. Hoppe and myself did look at that last week and it seemed ok but that anti backlash nut is a dynamic thing and I don't have the experience to understand how it reacts relative to rapid moves when it is moving in a out on the LS.

I am hoping that this is something very simple like I need a new computer or something because I was going to get one anyway. I am using V23 with XP and want to upgrade to 24 with win7. It is weird to watch myself write these last words as I have been a MAC user my entire life and don't like Micro Soft. :help: I guess Joe Kennedy was right when he said "eventually one does business with everyone at some time".

Rick
 
I think the electrical issues that Mike was talking about are related to cross-talk interferance, rather than surges. For example, power cables &/or appliances near your control cables. IIRC, you are using shielded co-axial cables. If the issue is repeatible, this probably is not the cause.
 
cnc

I would definitely do what Travis said and really wring out the
machine and check that you don't have an mechanical issues.
Use a test indicator. Not a dial indicator. Jog the suspect axis to a
zero point on the test indicator. Then jog in the other direction
.001. That's what you should read on the indicator. If that is good
write a program that runs the axis away and back to the indicator,
I use 1 foot runs. Do that 10 times. You should be back to zero.
If not it's a mechanical issue. Since you mentioned anti back lash nuts,
I assume you are not using any backlash compensation.
Hope I didn't over explain a pretty simple test.
 
scdiveteam said:
Travis,

Mr. Hoppe and myself did look at that last week and it seemed ok but that anti backlash nut is a dynamic thing and I don't have the experience to understand how it reacts relative to rapid moves when it is moving in a out on the LS.

I am hoping that this is something very simple like I need a new computer or something because I was going to get one anyway. I am using V23 with XP and want to upgrade to 24 with win7. It is weird to watch myself write these last words as I have been a MAC user my entire life and don't like Micro Soft. :help: I guess Joe Kennedy was right when he said "eventually one does business with everyone at some time".

Rick
Click to expand...

Rick, not sure a new computer will do the trick. In fact, I think that Mach3 runs best on an XP setup. You may find that the new computer doesn't even have a printer port. Personally, I use an XP Dell found on E-bay from Garland Tx. This is an off lease computer for $100 or so. It works great. Actually I have purchased several of them. I also use a Smooth stepper. This is a USB controller that outputs like a printer port does. Now then, they are now making a network ver. of the Smooth stepper. This is what I will be getting from this point forward. As far as checking repeat-ability, this is what Travis is suggesting, is a good idea. It will check more than just backlash in you lead nut but bearings or anything else connecting your stepper motor to you moving axis. Since your particular issue is growing over time, I would conclude lost steps and not backlash. Backlash will cause issues like flat spots when X axis changes direction. This is why others have suggested cutting circles. Now then repeatably, setup an indicator where x is applying pressure against it. Then command X to rapid away from the indicator "G00 X 12," then stop and revers back to say "G00 X 0.5" then feed into your indicator at say "G01 X0 F5.0" and stop. Check your indicator and repeat program several times. See if the X axis is moving or not. Check this at several spots along X axis to see if the area that your indicator is in is suspect.

Jim.
 
Mc2 said:
Rick, not sure a new computer will do the trick. In fact, I think that Mach3 runs best on an XP setup. You may find that the new computer doesn't even have a printer port. Personally, I use an XP Dell found on E-bay from Garland Tx. This is an off lease computer for $100 or so. It works great. Actually I have purchased several of them. I also use a Smooth stepper. This is a USB controller that outputs like a printer port does. Now then, they are now making a network ver. of the Smooth stepper. This is what I will be getting from this point forward. As far as checking repeat-ability, this is what Travis is suggesting, is a good idea. It will check more than just backlash in you lead nut but bearings or anything else connecting your stepper motor to you moving axis. Since your particular issue is growing over time, I would conclude lost steps and not backlash. Backlash will cause issues like flat spots when X axis changes direction. This is why others have suggested cutting circles. Now then repeatably, setup an indicator where x is applying pressure against it. Then command X to rapid away from the indicator "G00 X 12," then stop and revers back to say "G00 X 0.5" then feed into your indicator at say "G01 X0 F5.0" and stop. Check your indicator and repeat program several times. See if the X axis is moving or not. Check this at several spots along X axis to see if the area that your indicator is in is suspect.

Jim.
Click to expand...

Thanks Jim & Steve,

I am all over that this evening and see what happens.

Rick
 
Last edited:
Hi,

Thanks to all you guys that have tried to help me with my problem.

It has turned out to be a worn backlash nut on my X. When I ran a test program with 1/4 inch profile circles on the X, the circles were not round and the x was off.

I ordered some nuts from Jim Sickles and he instructed how to measure and put some backlash compensation in using a MACH 3 edit function for the time being until my parts are delivered.

When I first got my Cue Monster I did a lot of initial shaft and butt tapers before I got my saw machines. It only makes sense that my X would have a minor maintenance issues first. This is truly my first maintenance issue with this Unique's Product in the six years I have own it.

Almost back in business.

Again, thanks for all the help and suggestions. Time is our most precious commodity and anything I can ever do in return, I am always at your service also.

The one thing that is good about this whole situation is that I have gone aroud the block with my software, hardware and electrical cords and such and it has made me more tuned into the whole 9 yards.

Rick
 
Last edited:
I told you. you had a screw loose now you just had to find what was wrong with the machine.

scdiveteam said:
Hi,

Thanks to all you guys that have tried to help me with my problem.

It has turned out to be a worn backlash nut on my X. When I ran a test program with 1/4 inch profile circles on the X, the circles were not round and the x was off.

I ordered some nuts from Jim Sickles and he instructed how to measure and put some backlash compensation in using a MACH 3 edit function for the time being until my parts are delivered.

When I first got my Cue Monster I did a lot of initial shaft and butt tapers before I got my saw machines. It only makes sense that my X would have a minor maintenance issues first. This is truly my first maintenance issue with this Unique's Product in the six years I have own it.

Almost back in business.

Again, thanks for all the help and suggestions. Time is our most precious commodity and anything I can ever do in return, I am always at your service also.

The one thing that is good about this whole situation is that I have gone aroud the bock with my software, hardware and electrical cords and such and it has made me more tuned into the whole 9 yards.

Rick
Click to expand...
 
You must log in or register to reply here.
Back
Top