What's going on here? #1

I wanted to cut a t-track slot into a piece of CLS timber and designed it in CC and cut it in CM (files attached).

OK, I could have used a router table (10 minutes) or a table saw (15 minutes), but decided on using my XXL as I could be repeating the task in the future and thought using CC and CM would do just as well (but maybe take a little longer).

The CLS is longer than the XXL can cut, so I designed the toolpath to cut some of the slot, with the intention of sliding the stock along a guide and running the job a second time. Based on this premise, I (initially) only re-zeroed the Z axis, but the edge of the second cut was 2mm out along the length of the stock - along it’s whole length, as shown in the photo:

Strangely, this is parallel with the original cut along it’s whole length, so I have no idea what went wrong, but I know this:

The stock is pine and isn’t known to be particularly straight and parallel (here in the UK, at least) but this piece is fine, is clamped flat along the guide (just visible to the left in the photo) and the guide is parallel to the waste board. All I did was release the clamps, slide the stock forward and re-run the cut.

Why would this happen?

Bench Front T-Track v2.c2d (6.0 KB)
Bench Front T-Track v2.nc (1020 Bytes)

First things first, Peter. Once you have set up the cut, the zero point is persistent so there is no need to re-zero anything. The persistence lasts even if you switch the machine off and come back to it days later.

On the issue of where a particular spindle and cutter combination may be zeroed for a specific job, I find it really helpful to click on the legend Position and note down the Machine Position values because if necessary, you can jog back to the absolute position displayed. Useful if you have run another job in between the first job and the desire to repeat any aspect of it.

Looking at the workpiece image rather than the actual file. There appears to have been a displacement of exactly the amount of the defect. Either your workpiece moved, or you had an additional obstruction behind the workpiece against the track, possibly you misplaced a digit or two when re-zeroing the Z axis ̶o̶r̶ ̶y̶o̶u̶ ̶p̶r̶o̶g̶r̶a̶m̶m̶e̶d̶ ̶t̶h̶e̶ ̶d̶e̶f̶e̶c̶t̶ ̶i̶n̶t̶o̶ ̶y̶o̶u̶r̶ ̶t̶o̶o̶l̶p̶a̶t̶h̶. I discounted the warped pine hypothesis because it would not warp that accurately. :grin: Could the guide have moved when you released the clamps?

Now the files: Nothing untoward can be seen in the .c2d file and the .nc looks as expected. The conclusion has to be you have unintentionally adjusted something mechanical.

1 Like

I moved the workpiece, but only moved it forward and re-clamped it into position.

The clamps are separate and the guide didn’t move. I was careful to ensure the stock was held against it, along it’s whole length when I put it back together, and I think it would have been obvious if any detritus had fallen between the stock and the guide when I clamped it.

It’s a puzzle!

So that act should not have been responsible for the defect.

Right, so no obvious mechanical setting issue. When re-zeroing the Z axis, did you do this by jogging to known values or by mechanically re-zeroing the spindle and tool bit?

Indeed! I will just wait with popcorn and a beer and see who else rocks up. :wink:

I rapid-positioned to X and Y, then Z+6mm, and zeroed Z from there.

I’m jealous!

Hey Peter,

Could you add a little more detail to what we are seeing in the picture? Is the deep part of the cut from a previous cut or from the second cut?

If the stock is the same size as in your .c2d file, then it looks like the bit you’ve circled actually needs to be cut out so the slot is central and equidistant from the edges of the wood.

Is the additional cut 3mm by any chance?

1 Like

The deep part is the original cut, before I moved the stock towards the front, to cut the rear part. It was during this second cut that the issue became apparent, so I kept tweaking the X zero until the cut lined up with the first one.

The bit I’ve highlighted is a third cut, over the original to see how far out it was - essentially, the project is trash now, so I did this to see what the difference was.

Bizarrely, the depth of cut was different too. I measured them and found the first cut had only been 11.5mm deep and the second 15mm, so I’m even more puzzled (sorry about the focus):

The design shows this being central and the stock is correctly sized, but this is coincidental. The main purpose of the cut was to be parallel with the left hand side of the stock (as viewed from the front). This photo might show it better:

Yes it is, @Luke, and it wouldn’t be a stretch to think the depth of cut differs by 3mm too.

This is a long shot, but the pull off distance from the sensors is 3mm. Is there a chance you homed the machine but didn’t zero everything on one of the passes/times ran?

With this project, I secured the stock and zeroed the X, Y and Z axes for the first cut, assuming this would be all I needed to do, but I ended up having to incrementally adjust the X axis to take account of this ‘shift’.

Thanks for the extra details - they are helpful in that they make everything more confusing!

It looks like you slid the stock in the other way around for the second cut. As in, the guide was against the slotted edge for cut 1, and the unslotted edge for cut 2. This is probably unlikely but the sort of mistake I would make.

Another possibility is that your sweepy caught the top of the ujk hold-down clamp during some pass and you lost a 3mm of travel.

Other than that… pixies.

I’m really, really good at doing that!

I definitely didn’t do that! What on earth do you take me for :rofl: :rofl:

I was conscious of the close proximity of the end mill to the clamp, so I was sat on a stool watching it from the front, feed-hold button in reach, but it cleared it well enough.

Darned pixies! They get just about everywhere, these days :roll_eyes:

If you home and go back to zero, does it go back to the last zero you set?

1 Like

In your #2 post, did you mention that you disabled the BitSetter to run the first operation then enable the BitSetter to run the second operation? Does that require you to re-initialize the machine?

1 Like

At the end of each run the machine would move to the rear and wait, I’d rapid move it to X/Y/Z+6, and jog the Z axis down to the new surface, reset Z zero and run the program again, so I don’t know the answer to that question.

I didn’t think it would be necessary to do that, but I’m cutting a project today, so I’ll give that a go, but by ‘home’ do you mean jogging the machine to home position (NE) or re-initialising it?


It does and I did, yes. The projects were days apart and the machine shut down, so I needed to initialise it, anyway :+1:

This has all the makings of “simple” lost steps. Once you re-initialize (Home) the machine, all of that is reset, your absolute reference is regained.
My thought was that if you ran the cuts, homed, went back to zero and noticed it was off that would mean that position was lost at some point during the cut (or in between the two). You may never know why, but it is probably not a setup issue or software bug.


This topic was automatically closed after 30 days. New replies are no longer allowed.