[MrRogue]

I have a STRANGE PATTERN on the Profile Surface of a part and can not figure out WHY this is not a smooth surface.

I have been working on my two 96 & 97 VM16's for the last three winters, this is the fourth.

I have had and have an AWESOME technician that has been timelessly helping me over the phone for hours and hours solving problem after problem... The guy IS A BAD ASS with these machines.

I think we are both a little stumped with WHY this is happening to the surface finish.
I'm thinking there is still something bad or wrong in the mechanical area.........


1996 Milltronics VM16
Repair history:

Y ball screw: New bearings & Seals. Took it out completely and cleaned it off and cleaned the bearing nut that has all the bearings in it.Then I re-installed it with New Pulley and Belt. The Y drive motor also got a New Pulley as well.
I've had the Y ball screw out, cleaned, and on V-Blocks to spin it. It's not warped or bent.

X ball screw: IT"S a New ball screw. With New Bearings & Seals. The Ball screw has a New Pulley, and Belt. The X drive motor has a New Pulley as well.

Back Lash has been CHECKED and Entered into the control

Both X & Y drive control cards have been TUNED with the "Signal Gain" and the "Balance" screws on each card.




Cutting outer profile of part into a 1/4" aluminum plate.
Using a 1/2" 2 Flute Endmill.
RPMs 4000 (have raised and lowered to try different speeds)
Feed have tried 8 imp to 20 imp
Passes, one cut pass and a finish pass (have tried different amounts for finish cut)

I think I have done enough testing within the master cam 9.1 to rule out the CAD being the issue.

maybe bad spindle bearing...... not sure
maybe the bearings are bad on the table ways.... not sure
maybe someone that has also had this issue can recognize the pattern on the part surface and know what the trouble it or what direction i need to look in. I don't want to just start taking the machine apart without a direction of focus.


Attached are some pictures to show the outer profile surface finish.
The first picture orientation is bottom and right side standing at the machine looking in at the vise / fixture. Bottom of the part is closest to the sliding door.
This is one of MANY different tries at feeds & speeds to achieve a better finish surface.
This was around 4000 rpm with a feed of 8 ipm ( ridiculous !!! )

The second picture is a shot of the top right of the part sitting on the fixture in the machine.... you can see the STRANGE PATTERN really good in this picture. It almost looks like precision oscillation / rotating parts of the machine.

see IF you can come up with What's Wrong and Where I should look. I NEED, WANT, HAVE TO HAVE better looking parts coming out of this machine and soon. I need to be working WITH it, NOT ON IT.......



Thanks ahead of time for checking this out and your opinion.
MrRogue

[newman55598]

Is your cam program setup to output arcs or line segments. It looks kind like they are line segments. What is your step over and depth?


Sent from my iPhone using Tapatalk

[MrRogue]

sorry forgot to add that

The MaSterCam is producing I & J for arcs not R\

I've analyzed the curves and they are arcs not segments. I also sent the entire folder for this part to a MasterCam guy. From the SolidWorks drawing to the NC file(s) and all in between. NCI, MasterCam file, Export file from SW to Import file MC.

The part is completely arcs and has no straight lines / surfaces on it.

[newman55598]

They could be short arcs. Depending on cam tolerances. I would try cutting a 1 inch circle or something that is a perfect circle and see how it outputs the code and see how the cut comes out.

[MrRogue]

Yes need to machine a hole using just the conversational on the control.
My entire part is made up of arcs. MasterCam guy verified that I have complete arcs and not segments. but yes, I still need to machine the hole just to get that done for reference.

[MrRogue]

They could be short arcs. Depending on cam tolerances. I would try cutting a 1 inch circle or something that is a perfect circle and see how it outputs the code and see how the cut comes out.

CAM TOLERANCES: .0001

[Brian L]

A truly round surface will help, it needs to be a complete circle with one line of code to tell if it's the code or the machine. Second thing I would check is runout of the tool and type of the tool, I get a lot of the same type of scalloping and just attribute it to a pretty worn out machine, mine is a '92 Partner 4 (knee mill) which is pretty sloppy. You may have your ball screws in good shape, but have you double checked the backlash adjustments on them? Put an indicator in the spindle on a surface in the vice and push/pulled on the table to see if you have any movement in the ways? Are they square box ways or linear? If box ways you should have gibs to adjust, if linear, it might be time to get new.

While you have the indicator in the spindle, push/pull on the head stock and see if you have any movement from the spindle bearings or Z axis ways. All of the things that move could be causing the surface finish you see, so it's just a process of checking and eliminating one at a time.

[allen mullis]

I assume the are working with a DC machine. Have you cleaned your tach feedback armature on the axis giving the trouble? Also, program a movement in the axis giving the trouble to go back and forth at a given federate and check that the following error is consistent between directions. Write a program like this:
N1
G1 F100 Y-5.
G4 P2.
G1 Y5.
G4 P2.
GO TO N1
this will make it loop back and forth so you can compare the two directions. If they differ greatly, this would indicate a drive balance issue. If the machine is shuddering on one axis the first thing to check is the tach feedback though. I have had to clean about all of my DC motors at one time or another due to this problem. It will make you think thrust bearings, ball screws, drive, rails, etc.. Easy to clean though, take the motor in question off, pull the cap on the rear, pull the encoder bracket, pop the phenolic divider plate out, you should see two wires, red and black maybe 18ga , they lead to two small brushes on a small armature. This ids a DC generator that gives velocity feedback to the drive, if the armature is dirty it will give the drive mixed signals, telling it for a thought that it is not moving, so the drive ramps up, then the signal returns to normal and the drive recovers. You have to understand that the drive has no idea what position the axis is actually in, it is not connected to the encoder. The control sends voltage to the drive and the drive translates it to power for the motor based on the level and polarity of the signal voltage. The only way the drive has to identify that the motor is actually moving is the tach feedback. So now that you have it apart, take a pencil with an eraser and use the eraser to clean the little armature. Put it all back together and see if this doesn't solve your problem.

[MrRogue]

A truly round surface will help, it needs to be a complete circle with one line of code to tell if it's the code or the machine. Second thing I would check is runout of the tool and type of the tool, I get a lot of the same type of scalloping and just attribute it to a pretty worn out machine, mine is a '92 Partner 4 (knee mill) which is pretty sloppy. You may have your ball screws in good shape, but have you double checked the backlash adjustments on them? Put an indicator in the spindle on a surface in the vice and push/pulled on the table to see if you have any movement in the ways? Are they square box ways or linear? If box ways you should have gibs to adjust, if linear, it might be time to get new.

While you have the indicator in the spindle, push/pull on the head stock and see if you have any movement from the spindle bearings or Z axis ways. All of the things that move could be causing the surface finish you see, so it's just a process of checking and eliminating one at a time.

Back Lash was tested:
X .001
Y .0005
Entered these numbers into the control

Linear ways I believe is what I have.
I've done the push pull test before but don't remember the numbers. It was last winter or the one before...
I'll do it agin.

[MrRogue]

I assume the are working with a DC machine. Have you cleaned your tach feedback armature on the axis giving the trouble? Also, program a movement in the axis giving the trouble to go back and forth at a given federate and check that the following error is consistent between directions. Write a program like this:
N1
G1 F100 Y-5.
G4 P2.
G1 Y5.
G4 P2.
GO TO N1
this will make it loop back and forth so you can compare the two directions. If they differ greatly, this would indicate a drive balance issue. If the machine is shuddering on one axis the first thing to check is the tach feedback though. I have had to clean about all of my DC motors at one time or another due to this problem. It will make you think thrust bearings, ball screws, drive, rails, etc.. Easy to clean though, take the motor in question off, pull the cap on the rear, pull the encoder bracket, pop the phenolic divider plate out, you should see two wires, red and black maybe 18ga , they lead to two small brushes on a small armature. This ids a DC generator that gives velocity feedback to the drive, if the armature is dirty it will give the drive mixed signals, telling it for a thought that it is not moving, so the drive ramps up, then the signal returns to normal and the drive recovers. You have to understand that the drive has no idea what position the axis is actually in, it is not connected to the encoder. The control sends voltage to the drive and the drive translates it to power for the motor based on the level and polarity of the signal voltage. The only way the drive has to identify that the motor is actually moving is the tach feedback. So now that you have it apart, take a pencil with an eraser and use the eraser to clean the little armature. Put it all back together and see if this doesn't solve your problem.

Did the drive tuning yesterday with Balance & Sig Gain
Set to .05(1/0 ditherimg) both table moving back & forth.
Got my following errors to read .0000 X & Y (static setting)

I have been told about the dust in the encoders and I have seen it in a few before. I changed out the one on the X drive motor last year.
Time to see what's in the Y encoder....

[allen mullis]

hopefully running that little program will indicate whether there is a tach feedback problem or not. If you run the programming different axis and see different results, then you should be able to ID the problem motor. I really have a feeling it is the Tach brushes. It is usually very apparent at about 60 IPM, the axis will shudder as it moves.

[MrRogue]

Hello again...


I just went thru the Y motor to clean out the carbon dust build up in the tac armature. It had a lot of dust in there. I just blew it out. fairly easy.
I have never seen a CLEAN machine from any aspect but I think I can tell when I'm looking at something someone else has messed with. There are trails of CRAP work in both of my machines.
I guess I'm a perfectionist when doing the common basic type work like soldering, wire routing, ext...ext... It's either right or its wrong there's no in between.

SO after going through and cleaning the armature out, I put it all back together again and tried to run another part to SEE IF it would produce an outer couture with a good clean cut surface. It did not......


I'm getting the same strange pattern on some edges of the part. I'll post pictures of the part.
I have been trying to identify the location of the problem by looking at the location of the little vertical lines on the part being machined.

When looking into the machine at the part, the top edge at the back and the bottom of the part closest to the machine door.

The top right and top left seem to be the worst with the most pronounced pattern of vertical lines.

X and Y are both moving at the same time but I'm thinking there is more Y movement in this section of the part while cutting.

Along the bottom of the part there is very little Y movement and mostly X movement. This surface is almost acceptable.

See attached pictures......

Thanks

[MrRogue]

Is it POSSIBLE that the encoder could do this IF it's bad some how?

I have replaced an encoder before and seen the insides components of one.

IF the clear plastic wheel with the little black lines on it that is read by a pickup sensor has a piece of trash in between two of them or a line missing or damaged, IS what I'm seeing the results ?

[allen mullis]

My only 3 suggestions remaining are to program 45 degree movements in all directions at a regular federate and compare following error variations between each two axis combination. This should narrow it down to an axis. Include X/Z and Y/Z. And try it at different feed rates. The next suggestion is to change all your backlash parameters to zero, just a thought, something to try. Have you tried swapping drives, x and Y, Z and Y, Z and X, to see if the problem changes orientation?

[allen mullis]

Anything is a possibility, I had issues with one of my partner 1 machines that had accumulating error in one direction on X axis, it was obviously a problem before I got the machine as the cable to the motor had been changed and re-routed. I swapped encoder, motor, drive, tried everything I knew to do and it ended up, all I had to do was disconnect the shield drain from the encoder and leave the main cable shield drain connected to the ground block. Evidently the two being connected was causing stray signals from the encoder.

Is it POSSIBLE that the encoder could do this IF it's bad some how?

I have replaced an encoder before and seen the insides components of one.

IF the clear plastic wheel with the little black lines on it that is read by a pickup sensor has a piece of trash in between two of them or a line missing or damaged, IS what I'm seeing the results ?

[Brian L]

I would suspicion the encoders being the problem, but.... did you check tool runout like I mentioned before? And how solid are your XYZ axis ways holding? I assume you are climb cutting and if the ways on the machine are sloppy, the tool can "pull" the table along and cause the type of inconsistent surface you are seeing.

[MrRogue]

I replaced the X Cale on this machine years ago. 2007-2008
At first I was chasing a similar problem.
I didn't tidy up all the extra wire length and the green w/yellow tracer (ground) was picking up a ballooned pulse from the power wired every time the motor was sent powered to move.
This power wire created an extra pulse in the system making the encoder receive extra pulses.
Figured it out with a dial indicator on the table moving away one inch and coming back less then an inch. Worse every time I ran it.

1" out 1" back
The control said it was back to zero
The dial indicator said less then 1"

Fixed that problem winter before last.

Last winter it was bearings and X ball screw.

Frustrated !!!!

Anything is a possibility, I had issues with one of my partner 1 machines that had accumulating error in one direction on X axis, it was obviously a problem before I got the machine as the cable to the motor had been changed and re-routed. I swapped encoder, motor, drive, tried everything I knew to do and it ended up, all I had to do was disconnect the shield drain from the encoder and leave the main cable shield drain connected to the ground block. Evidently the two being connected was causing stray signals from the encoder.

[MrRogue]

I did some push pull tests.

I could pull hard and get it out 1/10 but then it would return zero.
Pushing was even less movement.

I would suspicion the encoders being the problem, but.... did you check tool runout like I mentioned before? And how solid are your XYZ axis ways holding? I assume you are climb cutting and if the ways on the machine are sloppy, the tool can "pull" the table along and cause the type of inconsistent surface you are seeing.

[newman55598]

Have you tried cutting a square parallel to the axis? Then if it did it it would show you which axis is at fault.


Sent from my iPhone using Tapatalk

[MrRogue]

Here is a cup of water to help SEE the vibration in the X & Y
Y having a larger Pulse / Vibration

Sorry the video recorded on the wrong plane on the ol iPhone so it's sideways. I do not know how to turn it 90 to be up right. ( OK the video won't load up at all)...... not sure how to post it up.

A truly round surface will help, it needs to be a complete circle with one line of code to tell if it's the code or the machine. Second thing I would check is runout of the tool and type of the tool, I get a lot of the same type of scalloping and just attribute it to a pretty worn out machine, mine is a '92 Partner 4 (knee mill) which is pretty sloppy. You may have your ball screws in good shape, but have you double checked the backlash adjustments on them? Put an indicator in the spindle on a surface in the vice and push/pulled on the table to see if you have any movement in the ways? Are they square box ways or linear? If box ways you should have gibs to adjust, if linear, it might be time to get new.

While you have the indicator in the spindle, push/pull on the head stock and see if you have any movement from the spindle bearings or Z axis ways. All of the things that move could be causing the surface finish you see, so it's just a process of checking and eliminating one at a time.