Shift in tool path

[dwkisker (Dave K)]

I think that I understand this issue, but I wanted to post here just in case someone has a solution that I have not yet found.

I’m about to embark on a project in which a bunch of pieces will be cut from 18mm baltic birch plywood using the LR4. I’ve decided to try to use the “Badger” bit which is a 1/8" diameter rough cutting bit that supposedly allows for more aggressive cutting than is usually possible. Initial testing suggests that this may, in fact, be the case.

However, I’ve run into a problem during the testing. In particular, with DOC of say 3mm, the whole gantry shifts to the left as the bit enters the plywood from the edge. It kinda looks like this:

The right side is what I want to happen, but the left side is what actually happens. The shift is around 5-10mm. I assume that is occurring because of the rotation of the bit is exerting strong forces causing it to drift to the left.

Also, the toolpath for this test is directly along X, so the shift is in the +Y direction.

I’ve tried a couple of things, including reducing the DOC and slowing the feed rate (which was initially about 1500mm/min I think.) I also tightened the Y belts and increased the current on the y motors since they were running at only about 42° C.

I’ve only done “quick and dirty” testing, but so far, nothing has reliably solved the problem, although reducing the DOC would probably work.

And, just to be clear, it’s definitely the gantry that is moving, not the workpiece.

So, my question is whether someone knows a reliable approach to resolving this that won’t slow down the cutting too much. To cut out all the pieces from this 4’x8’ pattern is going to take 6-8 hours even using very aggressive conditions. I’d like to not extend the time too much.

And, for those interested, here’s the scoop on the “Badger”:

The "BADGER" 1/8 High-Speed Roughing CNC Router Bit, 1/8 Shank – IDC Woodcraft

[DougJoseph (Doug Joseph (design8studio))]

Hello! I have heard of that bit before.

Just for clarity with getting help with this, this image illustrates the orientation for a LowRider:

You mentioned “the whole gantry shifts to the left” then later mention the shift as being in Y+ direction, which would be a shift toward the back.

Is the bit entering the material while moving from X-minimum side toward the X-maximum side, thus moving left to right, and the shift is happening toward the back of the LowRider?

At the risk of sounding cliche, loose grub screws could result in this. It sounds like your skill level in all this is such that you’ve already thought of this, but just mentioning.

[Dreyfus]

5-10mm deflection is pretty wild, I’ll be watching with interest to see what I can learn here.

Once you’ve checked the gubscres on the stepper shaft pulleys are all tight next I’d be taking hold of the bit and physically pushing to see if there’s any play in the router mount, core bearings, gantry - try and visibly see what’s moving. You’re not using printed XZ plates are you?

[Bartman]

@dwkisker

the whole gantry shifts to the left as the bit enters the plywood from the edge

Do you mean the gantry, or the core?

Is the shift in the X direction (parallel to the gantry), or the Y direction (perpendicular to the gantry?

[CesarH (cesar)]

Only ramp entrance comes to mind

[dwkisker (Dave K)]

@DougJoseph

You mentioned “the whole gantry shifts to the left” then later mention the shift as being in Y+ direction, which would be a shift toward the back.

My bad, Doug. I was not sufficiently precise on my description.

The “left” and “right” refer to the drawing and not the LR4 orientation.

In the LR4 “coordinates”, the desired movement of the router is from -X to +X. This direction is INTO the screen on my sketch.

The shift happens upon entry into the wood and is the +Y direction, which is from right to left in my sketch.

(Sorry for the imprecise description. As it happens, during operation, I sit at the “-X” end of the gantry as that’s where the LCD is mounted. The gantry shifting to the “left” is what I observe, but it’s actually to +Y.)

[jeffeb3]

Does it happen in foam? If so, then it isn’t load based.

Does it spring back at the end of the cut? If not, then it missed steps. If so, then something is flexing.

You are guessing it is flexing under load. I guess that too. But 5+mm is a lot for flex. You can drive youself crazy doing this. But after you confirm it is flexing under load, can you push the bit with your hand (with the router unplugged) and see where it is flexing?

Obligatory: Check the grub screws on the motor pulleys.

[dwkisker (Dave K)]

@Dreyfus

Once you’ve checked the gubscres on the stepper shaft pulleys are all tight next I’d be taking hold of the bit and physically pushing to see if there’s any play in the router mount, core bearings, gantry - try and visibly see what’s moving. You’re not using printed XZ plates are you?

As far as I can tell, everything is tight. There is no play in the core or the router mount. XZ plates are aluminum from V1.

The whole gantry is shifting.

I observe two things as I watch from the -X end of the machine: first, the gantry jumps to the +Y direction and stays there as the router enters the wood and progresses in the +X direction. And, second, the resulting cut is as shown in the sketch… instead of being a straight line from -X to +X, the initial cut has a shift.

[dwkisker (Dave K)]

@Bartman

Do you mean the gantry, or the core?

It’s the whole gantry, not just the core.

[dwkisker (Dave K)]

@jeffeb3

Does it happen in foam? If so, then it isn’t load based.

Interesting idea. I would be very surprised if it occurs in foam, but I haven’t tried this. I’ll see if I can find a piece of foam or something softer.

Note that with other bits (1/2" bit for flattening a piece of redwood), I didn’t observe anything like this, but the DOC was typically much less, <1mm.

@jeffeb3

Does it spring back at the end of the cut? If not, then it missed steps. If so, then something is flexing.

No. There is no springback.

The possibility of missed steps is why I tried increasing the Y motor currents (voltages?) a little bit. But, I only went from 0.9 to 0.95. (This controller is an SKR1.3 with 2208 drivers, as I remember so the change can be implemented from the LCD display. )

I’ll try to get a video of this a bit later, and check the gantry to see if I can tell what’s moving.

[Tokoloshe (Philipp)]

This should be easily doable with the speeds and feeds you are running. I have observed it with the LR3 in a certain kind of plywood and when I was cutting hardwood with a large bit at extreme speeds (for a LR3). Does it look like that?

Core Flex because of too much speed. :)

[Bartman]

@dwkisker

It’s the whole gantry,

You mention that the gantry (not the core) is shifting (twisting) in the Y+ direction.

I can only think of a few possible reasons for this:

• The gantry itself is twisting under load due to too loose clamp screws (check and tighten all of the tube clamps)
• The YZ plates are lifting off the table (possibly due to excessive feed rate or dull bit)
• Looseness in the linear rail connections to the XZ plate, or in the linear rails themselves (unlikely)

[Cincinnatus]

Hello! New LR4 owner in Seattle here. Made my first big cut and saw this exact same phenomenon on plywood (3/4"). It occurs whenever the tool is plunged at every start point for a part cut. (Using ESTLCam latest version; start points denoted on a part cut as gray arrows.)

Adding this for posterity; will check back if @dwkisker finds a solution.

My depth of cut (DOC) is 3mm and the plunge angle was set to 90 degrees. Its weird, it wants to begin the cut slightly further up in the +Y direction of the machine and then shift back down to the proper y-coordinate for the cut. It does this for every layer of the cut. At the completion of one layer, the tool is retracted and then plunged back to the incorrect start position. Pics attached for reference.

Haven’t checked on tightness of everything as mentioned by the group; will report back if I find a fix at my end. Thanks everyone so far for the suggestions!

[CesarH (cesar)]

I would recommend you to use ramp entrance plunges, bits deflects a bit when the flutes engages with the stock in vectric i use smooth ramps and that solved my issues. (It will slow down your cutting times)

You can do a quick google search about bit deflection at start/end

[Cincinnatus]

Thanks @CesarH ! I’ll give ramp entrance/bit deflection some research.

[dwkisker (Dave K)]

@jeffeb3

Obligatory: Check the grub screws on the motor pulleys.

Well, the LR4 is my 5th V1 machine and I’ve finally been bitten by the loose grub screw.

I’m honestly not sure what happened as this phenomena had not occurred in previous testing, but one of the grub screws on the y- motor at -X was actually gone.

I realized that this was the problem when I pushed on the gantry at the -X end right after turning it on and engaging the motors. It actually moved about 1/2"!

I managed to replace the missing one and tighten the other one without too much disassembly (adding a little Locktite) and now, with the motors engaged, the gantry doesn’t budge. I haven’t tried the do-over yet, but i expect it will now work pretty well.

I guess the lesson is that even if something has been checked multiple times, check again…

Thanks to all who offered ideas.

[jeffeb3]

Dave, your avatar reminds me of John Hammond from Jurrasic Park:

Played by Sir Richard Attenborough.

[dwkisker (Dave K)]

@jeffeb3

Dave, your avatar reminds me of John Hammond from Jurrasic Park:

I can see the resemblance, lol.

Fortunately, you didn’t reference the guy that Facebook once thought I was…

[MakerJim]

@dwkisker

you didn’t reference the guy that Facebook once thought I was…

Oh, man… I’ll never be able to un-see that boot hat.

Can’t wait to see that chair finished.