[Kingjamez]

I've read the suggestions of many on this board to lap the ways on my X3. I'm hoping to help the "jumpy" behavior of my machine. It has a resolution of about 0.002". When I command it to move 0.001, it moves ~0.0002", then I command another 0.001" and it moves ~0.0018". The pattern repeats.

When I read about lapping I hear people talk about the physical exertion of moving the table back and forth, over and over to smooth the ways.

Why can't I just make a NC program to do that for me 500 times? I'll stand there and monitor, while tightening the gibs as the polishing takes effect.

Is there an advantage to doing it by hand?

-Jim

[Fastest1]

I cant imagine there would be a difference as long as you were going to the extremes of travel.

[arizonavideo]

Lapping can remove a little or alot of material. You don't really need to make 500 passes. 20 passes with 220 grit should be all you need. You are only to remove the high spots not go down farther. 330 grit should be OK too.

500 grit compound will go into the pores in the cast iron and be a bare to get out. For finishing on my ways (IH mill) I cleaned them with solvent and ran them dry by hand for 20 to 50 or so passes and they shined right up. Clean off the black iron with solvent them rub in molly grease. Wipe of the excess and apply way oil or 30 wt oil or your favorite lube.


.... then SMOOTHNESS....

[TXFred]

I recommend that when you write the program for this, put a G04 P30. after every couple of passes. This will insert a 30 second pause, and allow your stepper motor to "rest" in its low power state.

The program will take longer this way, but it will keep your stepper from frying. The program that you're talking about is 100% duty cycle, and the stepper will get very hot.

I know because I've done it.

[Kingjamez]

Thanks for the tips. I can't decided if I should use the standard grit method, or go straight to the polish.... more research I guess.


Thanks for the tip on the pausing. I'm using servos that are way under utilized, but a break for them certainly can't hurt.

-Jim

[SCzEngrgGroup]

I recommend that when you write the program for this, put a G04 P30. after every couple of passes. This will insert a 30 second pause, and allow your stepper motor to "rest" in its low power state.

The program will take longer this way, but it will keep your stepper from frying. The program that you're talking about is 100% duty cycle, and the stepper will get very hot.

I know because I've done it.

That really should not be a problem on a properly setup system. I did this on my X2 to get it worn in after I CNC'd it. I just wrote a program that ran all three axes end to end non-stop. I fired it up and walked away for about 4 hours.

Regards,
Ray L.

[mwood3]

That really should not be a problem on a properly setup system. I did this on my X2 to get it worn in after I CNC'd it. I just wrote a program that ran all three axes end to end non-stop. I fired it up and walked away for about 4 hours.

Regards,
Ray L.

Ray,

you used lapping compound in the ways when you did this?

[SCzEngrgGroup]

Ray,

you used lapping compound in the ways when you did this?

Nope. Just a little oil.

Regards,
Ray L.

[mwood3]

Nope. Just a little oil.

Regards,
Ray L.

Ok, well i imagine it will be at least x2 the resistance(if not more) with the lapping compound method that is being discussed here...no oil...just adding lapping compound and tightening the gibs as you go...I second the opinion of letting the motors rest between cycles if it takes very long to lap it in. I'd imagine this to be effectively the same as having a cutter engaged the entire time the program is in motion...

[Johann ohnesorg]

If your amount of travel is not consistent with the amount of infeed on the handwheel/motor shaft there´s something different that´s wrong. Measure the screw, maybe it´s metric, 25.4mm (an inch) and 25mm are not too far apart.
If you still want to lap it in, measure what is wrong in concern of geometry. You have no idea if the ways are not parallel, if you have a high spot on your slide, if your gibs are straight (test them against a flat, it surely only bears on the ends) and o.k. or if your feedscrew is pinching is simply put in under a slight angle because they milled the seat not rectangular to the ways. if you can´t mesure it, you can at least locate it. Put the slide on without feedscrews and tighten down the gibs snug, now push it around to see if it binds in a spot or if it binds, lets say up front it slides and back there it locks up.

Measure first, then correct.

[TXFred]

If your amount of travel is not consistent with the amount of infeed on the handwheel/motor shaft there´s something different that´s wrong. Measure the screw, maybe it´s metric, 25.4mm (an inch) and 25mm are not too far apart.

It's not a ballscrew ratio issue.

From his description, what's happening is that the table is binding and then releasing. Bad ways can cause this, especially if they have been set so tight as to remove all play from the table.

The stepper moves 0.001". The table refuses to move because it's too tight.
The stepper moves another 0.001". The table now moves. Sliding friction, which is less than static friction, takes over and the table moves enough to release the tension from the initial 0.001" that the stepper attempted to move.

Note that he is moving 0.0002" followed by 0.0018", which is 0.002", the commanded value. So it's moving accurately, it's just not doing it smoothly.

Fred

[HawkJET]

Ok, well i imagine it will be at least x2 the resistance(if not more) with the lapping compound method that is being discussed here...no oil...just adding lapping compound and tightening the gibs as you go...I second the opinion of letting the motors rest between cycles if it takes very long to lap it in.

As Ray says, a properly set up system doesn't need to "rest" the stepper motors. And the argument that there may be more friction is also moot because the steppers are ALWAYS running at FULL TORQUE. The motor has no idea how much friction there is until its torque requirement exceeds what the motor will produce. If it turns the screw, it will be as hot as it will get no matter the load.

Also, stepper motors are typically designed to operate with a 40*C temperature rise (which is 72*F). So if the ambient air temp is 75*, the motor will run about 150* with no problems. This is hotter than you can comfortably touch the motor. Besides, why would you design the system for anything LESS than 100% duty cycle?????

[neilw20]

Imagine pulling against a magnet with a gap.
It is like having a spring attached.
The more torque, the more it lags behind, until eventually the stepper can lose a step.

Lapping the ways, with the screws disconnected will soon let you feel the tight spots.
Just pulling it backwards and forwards by hand on the X3 works OK.

I originally dismantled my X and Y and scraped the high spots, and got it all parallel. Lots of blue, but chalk works quite nicely to see the high spots.

You need some precision ground rods to check for parallel, but I only had to improve the flatness on the Y slides on the main base. It had a slight warp.
Just sliding by hand gives a good feel, and you need to do this to make sure there are no high spots.
It took a few goes to bed the gibs in, and after 200 hours work, mostly full travel on X, it was getting really sloppy again.

And another thing. Those crappy wire cut couplers will eventually break. I just made some nice spiral ones to replace them.
Reground the gibs and a bit of lapping and it is bedded in quite well.
The more you use it, and adjust every so often, and it seems to bed in better every time.

I am now better than .001" on all axis. Z head nod is about about 0.0005". You go 0.0005" down past where you want, then come up 0.0005" before the tool tip changes direction.
This is a combination of stiction, Z gibs, and the gas strut causing this. Live with it. Too hard to fix. Regrind gibs when the y get worn.

I just program in the machine error and can work in increments of better than 0.0001" in Z after the spindle has warmed up for 30 minutes.
Spindle growth is about 0.001" and temperature variations in the spindle are noticeable.
If you stop for 2 minutes, it starts to shrink by 0.0002" every 5 minutes. - Just don't stop for coffee, then it is OK.
It is not so much the spindle cooling, but the heat equalizes from the bearings to the shaft.

Yes I can measure all this. My DTI is in 1 micron steps. That's 0.00004"
After doing 2000 hours of precision work I have got my little machine well tuned. 3600RPM all the time. Original bearings. Still a noisy sh1t.

Eventually the crappy wire cut couplings will fail. I made new spiral ones. Loose grub screws on couplings come factory fitted.
I have the column attached to the brick wall, and without this, Z repeatability is at least 20 times worse.
http://www.cnczone.com/forums/828672-post5.html

GIBS GIBS GIBS .. The better they are, the better it works.

[Kingjamez]

It's not a ballscrew ratio issue.

From his description, what's happening is that the table is binding and then releasing. Bad ways can cause this, especially if they have been set so tight as to remove all play from the table.

The stepper moves 0.001". The table refuses to move because it's too tight.
The stepper moves another 0.001". The table now moves. Sliding friction, which is less than static friction, takes over and the table moves enough to release the tension from the initial 0.001" that the stepper attempted to move.

Note that he is moving 0.0002" followed by 0.0018", which is 0.002", the commanded value. So it's moving accurately, it's just not doing it smoothly.

Fred

Ding Ding Ding... you hit the nail on the head, that is exactly what is going on. I haven't had time to lap the ways yet, but I think I'll just do it by hand. It seems this is something I need to feel, in order to determine when I should next tighten the gibs.


Neil,
I like the concrete wall idea... I've got a big one right behind my X3. I get no where near the accuracy your talking about, but perhaps with a little effort I can get closer.

-Jim