MechMate CNC Router Forum

MechMate CNC Router Forum (http://www.mechmate.com/forums/index.php)
-   Troubleshooting (http://www.mechmate.com/forums/forumdisplay.php?f=83)
-   -   Cutting errors - slop in motors? (http://www.mechmate.com/forums/showthread.php?t=4978)

Rocky Mountain Teardrops Fri 15 May 2015 11:19

Cutting errors - slop in motors?
 
Hello All,
Having some trouble with cutting errors in my Mechmate.

Specs: Standard welded base

carcan Fri 15 May 2015 11:21

Can you explain better ? The only time i had problems is because the small screw on gear was unscrewed

Rocky Mountain Teardrops Fri 15 May 2015 11:56

ARG! Can't get the edit to submit


Specs:
Standard welded table base (per Gerald's plans)
Oriental motors 7.2 geared
Motors have two flats ground on shafts with two setscrews each
Porter Cable 7518 router with K2CNC mount

There is absolutely no slipping of the gear on the motor shaft.
However there is a significant mechanical slop in the gearboxes.
With a dial indicator clamped to the X rails, indicating off of the gantry.
Motors energized, I can push/pull approx .030" on one side, and .025" on the other side.
When I release the gantry, the indicator will go back to 0 (no noticeable backlash)
Motors are very new, this is a hobby machine and it doesn't get run very much.

I can't feel or see any slopping or looseness with the Z axis. (grabbing the collet and giving a hard tug)

Out of round and inaccurate holes:
Tooling: .250" carbide end mill, two flute, upcut
Material: .750" MDF

Step one: pocket larger hole, 100 ipm, 21,000 rpm, .755" outside diameter, .080" deep, one pass
Step two: peck drill small hole, 300 ipm, 21,000 rpm, .760" deep, three pecks
Step three: pocket deep hole, 100 ipm. 21,000 rpm, .310 outside diameter, .750" deep, one pass
http://i222.photobucket.com/albums/d...pssgiz72qq.jpg

Here you can see the step / mismatch from the drill (.760" deep) and the pocket (.750" deep)
http://i222.photobucket.com/albums/d...ps91btdaly.jpg


Angled corners:
Tooling: .375" carbide end mill, three flute, upcut
Material: .750" cabinet plywood (midgrade, not Baltic Birch)

Step one: Rough cut blanks, 120 ipm, 21,000 rpm, .375" depth of pass, conventional milling direction
Step two: Finish cut, 100 ipm, 21,000 rpm, .010" finishing cut, .750" pass depth, climb milling direction

Please note: only the corner on the left is angled, the one on the right is fine.
This was typical of all of the pieces cut on this run. Two corners angled (diagonally to each other) two corners clipped (also diagonally to each other)
Piece came out nice and square, except for the clipped corners.


http://i222.photobucket.com/albums/d...pspzpubyuu.jpg


At this point, I'm leaning toward a motor gearbox issue, but I am hoping for other ideas.

Thanks,
Lee

sailfl Fri 15 May 2015 17:16

Are the motors new? If you know you have slop in the gearbox, contact Oriental and talk to them about the slop.

Rocky Mountain Teardrops Fri 15 May 2015 17:48

Nils,
The motors are very new. I would estimate less than 100 hrs.
I don't know what is normal for these motors. But .030" in the gantry seems like a lot, and I certainly cant get an accurate pocket bore.

NOTE: the peck speed should be 30 ipm not 300
I can't seem to edit my post.

Thanks,
Lee

Regnar Fri 15 May 2015 21:28

If you are using Mach 3 and you speed and feeds are set to high for the machine it will get to approximate locations and shave off corners to maintain speed. Look for CV and try turning it off.

But to be honest something looks loose because your circles have flat spots. With the machine not moving and power turned on start shaking and pushing everything. You will be surprised what you will find loose.

Rocky Mountain Teardrops Sat 16 May 2015 11:13

Quote:
Originally Posted by Regnar View Post
If you are using Mach 3 and you speed and feeds are set to high for the machine it will get to approximate locations and shave off corners to maintain speed. Look for CV and try turning it off.

But to be honest something looks loose because your circles have flat spots. With the machine not moving and power turned on start shaking and pushing everything. You will be surprised what you will find loose.
CV is off (has been always)
From what reading I have done, 100ipm does not seem too fast...
Calculations show I'm making lots of dust instead of chips.


You are exactly right, flattened circles and rounded corners.
Spent the better part of two hours this morning shaking, twisting, and torquing and I just can't find anything except for the gearbox slop.

http://i222.photobucket.com/albums/d...psocuzrobl.jpg


This is a circle cut where the movement was transitioning from X to Y.
Location of this step is approx 11:00 on a clock face.
There is a mirror step at approx 1:00
Conventional cut (left to right in the picture / clockwise)
1/4" carbide end mill, .250" deep, 100ipm, 21,000 rpm
http://i222.photobucket.com/albums/d...ps9sdk72ap.jpg


Thanks,
Lee

darren salyer Sat 16 May 2015 11:56

I regularly cut at 100-140 IPM, so I agree you're not running too fast, if you have a similar setup. I'd look at pinion gears being loose...that can sometimes be an erratic problem.

Rocky Mountain Teardrops Sat 16 May 2015 14:01

Here's another test from today.
Eight total circles, two showed signs of irregularity.

One in conventional (clockwise) and one in climb.
Ignore the inner circle, there was a groove in the spoilboard that might have caused the distortion.
http://i222.photobucket.com/albums/d...pstmdhgsmc.jpg
http://i222.photobucket.com/albums/d...psxy2slgsa.jpg

Rocky Mountain Teardrops Sat 16 May 2015 16:17

Okay, I think this shows *something* better.
Pocket program with Sharpie marker.
100 ipm

Pay special attention to the line of shudders at 6:30 and 12:30
(ignore the line at 3:00, that is the "step" to get to another ring in the pocket program)
Notice the uneven pattern of the rings after every shudder mark.
Also note the flatness of the rings. Almost like an octagon instead of a circle.

Every place that you see a funny mark, the gantry on the Mechmate gave a slight shudder.

http://i222.photobucket.com/albums/d...psri9jtf89.jpg

Here's another run at 200 ipm
Uneven rings very noticeable, along with the massive shudder on last ring at 7:00 position.
http://i222.photobucket.com/albums/d...psaedn7r3b.jpg

Lee

ger21 Sat 16 May 2015 18:51

Quote:
CV is off (has been always)
Are you sure? Is there a G61 or G64 in the status line in Mach3?

The rounded corner certainly looks like it's from CV mode.
What are your accelerations set to?


Do you have backlash compensation enabled?

Rocky Mountain Teardrops Sat 16 May 2015 19:45

Quote:
Originally Posted by ger21 View Post
Are you sure? Is there a G61 or G64 in the status line in Mach3?

The rounded corner certainly looks like it's from CV mode.
What are your accelerations set to?


Do you have backlash compensation enabled?
Well, now I'm not sure. :o
The green "LED" box on the main screen tab "Settings" is not active.
But the check box on "Config>General Config" is checked.
So I guess I do have constant velocity active.
But that shouldn't be causing the problems when I'm making circles or round pockets.....right? :confused:

Acceleration at 6 ips

No backlash compensation enabled. I haven't been able to measure any when I have tried.
(clamp dial indicator to X rail, set zero, run gantry down a ways, record Mach3 setting, run gantry back to dial indicator zero, compare to Mach3 reading)

I'm wondering if I'm having a gear to rack mesh issue. I've never had good, even contact with the gears and rails (noticed today that the gears are wearing / beveled /rounded on the "inboard/motor" side)
Going through and shimming the motors to get a more even mesh to the racks right now, but I stripped the heads on a couple of the grub screws, so that will have to wait until tomorrow when I have more screws in hand.

I think whatever is causing the shudder is my problem. I'm going over everything with a magnifying glass looking for issues.
(and no, it's not the gears/grub screws being loose.....I pulled everything apart today and they are locked down SOLID just where I put them months ago when I ground the second flat on the shafts.)

Thanks for the ideas!!

Lee

ger21 Sat 16 May 2015 20:35

OK, then CV mode is causing the rounded corners. Your accel is much too slow, and should be at least 10-15, if not higher. Also, in General Config, set "Stop CV on ANgles" to 89 and enable it.

The other stuff is probably mechanical, as you suggest.

danilom Sun 17 May 2015 16:01

Also check if your kernel speed is at 25khz as some computers can act strange on values above that. And yes get that accel up to at least 20-25ips (around 600mm/s)

If all fail try another computer. To check the computer with your ear set jog percent to some low value and jog the whole length, the sound should have even tone without any changes in pitch (nice regular step pulses from LPT) or hiccups that indicate interrupt and steppers slowdown and pickup speed again

p.s. slow down the rpm on 1/4" cutter that 21,000 rpm seems too much for that feedrate one example is onsrud 1/4" double flute 200-250 ipm at 18.000 rpm for 1xD cut depth

ger21 Sun 17 May 2015 17:01

At 100ipm, you want to be between 10,000-12,000 rpm. At 21,000, you're drastically reducing the life of the tool.

darren salyer Sun 17 May 2015 21:06

Gerry has forgotten more than I'll ever know about this stuff, but I'll second the advice above. Once I hooked up my SuperPID and reduced RPMs, tooling is lasting substantially longer.

Rocky Mountain Teardrops Tue 19 May 2015 14:40

Problem solved!
Thank you all.

Steps taken.
1. Shimmed motors to make sure the gears have perfect contact with rack.
2. Adjusted motor pivot bolts (mine were too tight, not allowing the motors to keep perfect contact with racks throughout the entire length of travel).
3. Changed motor acceleration to 20. (don't know if this helped much, but the motors are much quieter and "crisper" in their movements now)
4. Turned off CV for angles greater than 89.

The combination of these four items have made a world of difference.
90 corners are also nice and square without the funny overcut that would sometimes show.
Pocket programs are very smooth and symmetrical.


You can see the difference with these two pictures. (same pocket program, same speeds)
before adjustments
http://i222.photobucket.com/albums/d...psg7elr73r.jpg


After adjustments
http://i222.photobucket.com/albums/d...psqkf1m964.jpg


THANK YOU ALL VERY MUCH!

Lee

darren salyer Tue 19 May 2015 16:29

Glad to hear it's sorted out. If I could make a suggestion, only post in one area with a problem in the future. It was confusing toggling back and forth between threads trying to keep up.

Rocky Mountain Teardrops Sat 23 May 2015 19:51

Spoke too soon...
Can't get repeatability.
Drilling a through hole (.250", .755" deep)
Then coming back over, using the same geometry, drilling a .302" hole .650" deep.
Holes do not align..(guessing upwards of .100" misalignment sometimes).

Not consistent to location either.
Sheet is to used as a fixture jig, so the approx 200 holes are randomly spread over the sheet (48"x96" sheet of 3/4" MDF).
Some holes are good, some are okay, 10% are way off.
Bad holes are randomly scattered throughout the sheet.

Must be a backlash issue...but I can't measure it with my 1" dial indicator.

Looking for some more ideas.

Thanks,
Lee

lonestaral Sat 23 May 2015 19:58

Lost steps ?
Check your motor settings.(Acceleration et al)

darren salyer Sat 23 May 2015 20:48

only two things I know would be lost steps or slop or loose gears in the drivetrain. Guess that's 3 things, but I'm lumping the last two in the same category..

Rocky Mountain Teardrops Sat 23 May 2015 21:22

Holes only seem to be off in the X axis, and consistently the same direction.

Mach3 says my(steps per inch) at 3055.75 & 200 ipm (can be higher)
Kernel speed at 25K (24,998)
Mach3 driver test looks go up through 100K

(edit: just noticed that the motor tuning needs the frequency to be at least 36.7K. I'll change Mach3 tomorrow)


For the life of me I cannot see ANY evidence of the gear slipping on the shafts.
Might weld the darn things tomorrow to take that option off of the table.

Lee

lonestaral Sat 23 May 2015 21:30

Might be worth doing the Squaring set up cut test.
Just in case something has moved or had a bump.

ger21 Sun 24 May 2015 07:34

Cut your x axis acceleration in half, and see if it helps.

pblackburn Sun 24 May 2015 13:35

Spur geared motors will have play also. They have to or they will rip themselves apart. It will be less than 10 thousandths but will increases over time. Just know it is there. To find out how much yours has. Place in jog mode and set to single step, use an dial indicator. Move one direction about an inch and stop. Zero inductor then move back by single step until the needle moves. The error from the dro start to the point the indicator moves is your backlash.

Rocky Mountain Teardrops Sun 24 May 2015 15:51

Quote:
Originally Posted by pblackburn View Post
Spur geared motors will have play also. They have to or they will rip themselves apart. It will be less than 10 thousandths but will increases over time. Just know it is there. To find out how much yours has. Place in jog mode and set to single step, use an dial indicator. Move one direction about an inch and stop. Zero inductor then move back by single step until the needle moves. The error from the dro start to the point the indicator moves is your backlash.
Ahh.
That makes sense. I was measuring repeatability.

Backlash now measures:
One side of X axis measures .008". The other X is .009"
Y axis is .007"

This is way less than what my parts are showing.


Funny thing... jog set to step .001"
Both sides of X will take a small step (.0005") then another, then another, then another, then take a .002" step.
DRO and dial indicator will match (.004") but the machine got there in uneven steps.
This is after backlash has been taken out (gantry moved in one direction until DRO & dial indicator are both moving together).

The Y axis does not do this. One jog = .001" on the DRO & dial indicator both.

Working on a program that will drill a bunch of holes with a lot of movement in between (angles, side to side, forward and back) we'll see what that will do.
Then I'll run it again with acceleration turned down.

Lee

pblackburn Sun 24 May 2015 15:54

If it is a 7.2 gearbox it will be around 0.0003 per step. You can calculate it out.

pblackburn Sun 24 May 2015 16:11

Lee,
I have a 7.2 reduction and my velocity and acceleration are 500 and 50. I think you have something else going on. Kernal needs to 35k. Other thing to try is turn off the toolpath preview. I had problems with it causing missed steps. CV mode with angles can cause hesitation at quadpoints and tabs in some setups although this would not do what you are seeing. Just more info for you.

Rocky Mountain Teardrops Sun 24 May 2015 17:48

Ran my "hole test" pattern.
Rows of holes in patterns to run the machine all directions from hole to hole.

Results are not good.

Vast majority of holes are offset in the X axis with some slight Y axis off set.
Most holes show a consistent "left" or X- offset....but not all.

A single row of holes will either show left or right, not both in the same row.
Machine motion does not seem to matter. (even running diagonal back and forth across workpiece to drill holes)

1st hole, 1st row,
rapid to
last hole, 2nd row,
rapid to
2nd hole, 1st row
rapid to
second to last hole, 2nd row
etc

o o o o o o o o o o o o o o this row offset to >


o o o o o o o o o o o o o o this row offset to >




-----------------------------------
first hole,
rapid to
last hole,
rapid to
second hole
rapid to
second to last hole
etc

o o o o o o o o o o o o o o o this row offset <



Can't see how I could possibly be loosing steps with the diagonal motions not showing equal errors.


Guess I really am going to weld the pinion gears, just to eliminate that.
Still cannot see or feel any movement in the pinions or rack.


Lee

hennie Mon 25 May 2015 00:24

Lee, Check your cutting speed If you cut to fast on small holes it loses steps, I had that also, my 5 cents


All times are GMT -6. The time now is 05:36.

Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.