Dimensions consistently off [not fault of Easel, still searching]

I’ve been using easel now for several years, and one problem that I keep running into when trying to cut specific dimensions is that the carve is exactly off by the same amount in all directions. As a simple test, carve a block that is exactly 1" by 4" using outside outline. The block will measure exactly 1 1/32" by 4 1/32" - basically the cutting path is an extra 1/64" towards the outside all the way around. This happens consistently on multiple machines, and regardless of what bit I use, provided I’ve accurately confirm the exact bit width to easel. Basically, I’ve come to the conclusion that I have to LIE to easel about bit width in order to get an accurate cut – for example, by subtracting 1/32" so that it moves the bit over 1/64" closer to the edge of what I’m cutting. This is the only way I’ve been able to get accurate results. I’m curious if anyone else has encountered the same problem and what workaround you used? Also, is this something that we can get “fixed” in the easel software? And if so, can the change be triggered only by user selection on a per-project basis? I.E. I have a lot of existing projects that are incorrectly dimensioned on purpose to get the correct results - those need to produce the same results until I update them and throw the “higher accuracy”.

bit width may not be the exact size advertised. In fact most are off a hair.
Also runout may be in play.

Add those two together and you have a good candidate for your issue.

Have you checked the calibration over a reasonable distance? (Maybe 80% or more of the maximum dimensions your machine can handle).

I’ve carefully checked the bit size. Bit has imprinted size, which matches micrometer measurements. Issue is consistently repeatable on different bit sizes. At one point I thought it was just something odd about my old machine (WhittleCNC) but now with an X-Carve it’s still doing exactly the same thing, so I’m able to rule out the machine also. The 1/32" offset is consistent across any distance the machine is capable of. I’m down to having to blame Easel for adding a fixed offset to the bit width calculations, which is why after several years of being plauged with this issue and trying to identify and solve it from multiple angles I’m finally posting here.

i doubt that tbh, and I am not a huge fan of Easel so that should mean something :wink:

fyi, the only way to truly measure a bit’s cutting width is by milling a slot in some stock and carefully measuring that. Measuring the bit itself will not take runout into account, which may be your culprit.

Hi Scott,

It is highly unlikely that the issue you are describing is a flaw in easel, but we can easily confirm this. After the carve it will ask you if everything turned out all right. Run your test described above and then click “No” and continue to get help from Inventables and submit your problem. This will submit your carve and all your settings data & g-code to our support team. We’ll be able to inspect the g-code to see exactly where it sent the machine. Or, if you’re comfortable reading g-code, you could download the g-code and inspect it yourself. G-code is very basic text that tells the machine to go to specific X, Y & Z coordinates. You can audit the coordinates and see if there is the extra gap or not.

If there is nothing wrong with the g-code–which is what I expect we’ll find–then the issue must be hardware-related. It is possible that the bit is not the exact width you expect it to be–as others have suggested–or that the bit is being deflected by the force from the material it is carving, for example.


I wish a simply function could be added to the cutter selection with “Cutter Offset applied = X mm/inches” as this can be found on commercial machines.

If you did a trial cut and saw it was running out 1.1mm or 1/32" out in both directions you could go back in and enter this figure into a offset box and then allows you to cut a more accurate job

Okay, so I appreciate the suggestion of examining the G-Code for the source of the issue. I’ve done exactly that, and here is my results (test carve of one pass, 1x4" rectangle on outside, 0.125" bit):

The results show that although there is some small variance, which I suspect may be mostly attributed to conversion between mm and inches, the actual calculations from easel appear to match the bit width.

So: I’m satisfied that my 1/32" offset is not caused by a math computation error, and I’ll look elsewhere for the problem. Yes, I agree it would be most helpful if it was possible to add a separate fudge factor instead of fiddling with bit width.

However: I’m slightly bothered by the fact that the sides of my 1x4" rectangle are not fully square. Granted, the variance (red) in the opposite axis to the large travel distance (yellow) is miniscule (.001" at most) but I would have expected it to actually be zero.

Re: suggestion that material is deflecting the bit: This is entirely possible, however I’m using a fairly soft plastic (cutting board material) and a very slow feed rate for accuracy, so there should be very little deflection. The bit dimensions have also been triple checked. I will probably end up doing some manually generated G code to test the reverse carve direction to see if that affects the measured width to prove deflection one way or another.

I know you said it’s not the machine(s), but have you generated gcode in another way to rule that out?
These machines have to be calibrated, and, unlike a screw drive or rack and pinion system, you’ll have to recalibrate due to belt stretch.

Try cutting a simple straight path in the material in question and measure the slot width. How does that compare to what you measured for the bit.

The only other thing I can think of is the steps/mm are off just enough. If you cut a larger shape and the delta is larger, this could also indicate steps/mm is off.

I’ll try measuring the bit slot to confirm bit carve dimensions and post here later. I have repeatedly tried different dimensions (for example 1 vs 4 inches) but consistently gotten the same exact 1/32" extra width regardless. I’ve also (on both machines) checked bit movement across maximum travel distance against a ruler (and in the case of the X-carve against the mm markings on the board) and everything is absolutely spot on. I’m 100% confident that physical movement vs g-code positions is exactly in sync. Believe me, I’m particularly rectal about this, to the point that I’m planning on setting up limit switches on the far side of the travel to be able to do a probe calibration against and confirm everything’s aligned before starting a carve.

My honest opinion is to just adjust your steps setting $100 - $103 to take off 1/32" if it’s always the same. No need to reinvent the wheel looking for answers. Happy Carving!



If I do that, then holes centered on specific points would be located in the wrong place. What I’m getting is an extra 1/64" on all 4 sides of a rectangular block. The only explanation is a difference between what the bit measures as, and what the width of the slot it leaves in the material is – except that the width of the slot is narrower than the bit itself? This is why I’m confused about what’s going on here.

Have you tested your system for:

  • Backlash
  • System deflection

What happen if you do a 1x4" piece once, then rerun the same code to act as a finishing pass?

A work-around could be to use a slightly different bit diameter value vs measured to take up the extra material. That won’t hurt X/Y precision either.

More “pro” software have cutter compensation, to account for runout/actual tool size and tool wear.

How is that possible??

That’s what I don’t understand. Here’s what I know: If I cut a rectangle outside to make 1" x 4" piece of 0.5" thick plastic, and I use a 1/16" bit (which is both marked at 1/16" and my micrometer confirms is 1/16"), what I measure up is 1-1/32" x 4-1/32" exactly. (BTW, this is repeatable with larger bits as well).
The steps are accurate, since I can tell it to move 12" (or even 30") and it’s spot on to a ruler in both dimensions. It’s not the spin of the bit dragging the router slightly, since that would make the rectangle smaller, not larger. The slot that gets cut appears to measure 1/16", but I don’t have the right tool to do that accurately enough yet to be absolutely certain as yet. And I’ve confirmed the G code dimensions are within 2% worst case of what they should be, given the cut dimensions and the bit size. In every case, where I’ve carved a hole in a specific spot, it’s centered dead on to where it should be (although seems a slight bit small - an M5 bolt thread grabs more than it should for an actual .5mm hole). This is good for anywhere on the entire X-Carve 1000mm usable area (over 30"), and was true of my whittlecnc also, although the slop on it’s head was pretty bad comparatively. I am really at a loss to understand why this is true. At the moment, my only real solution is to take 1/64 to 1/32" off the bit dimension and recarve to “fix” the problem. I think I’ll start a spreadsheet of materials, bits, and test results to see if I can spot any correlation. I’ll need to head over to Costco to get more of those commercial kitchen style cutting boards (exactly 1/2" thick soft plastic - great for making your own laser mount or material guides).

I get the same measurements. I’ve tried that. The dimensions only change when I change the bit size and re-run. There is very little play on the spindle (x-carve w/611), my material is fairly soft plastic (cutting boards), and I purposefully run about 50% slower when trying to get accurate measurements.

And it turns out Costco isn’t carrying it anymore, but i found the same thing on amazon: http://a.co/hooc5Bj

Have you tried this with any other material? I would like to see what the same bit does with hardwood.