Trace width inconsistencies

[aiden1015]

Has anyone else had this problem? My vertical traces are pretty much bang on in width, but my horizontal traces seem to have a 50% error. My belts are tight, and everything looks OK in my g-code viewer (using ChilPeppr). The photo used GRBL but after my tiny-G showed up, I got my machine calibrated again and ran it again, but have the same result.

Forum is complaining my photo is too big, so here's a link to my board: http://i.imgur.com/5oWDirz.jpg

[WillAdams]

Runout?

What spindle and endmill?

How is the spindle mounted?

What did you do to ensure that it’s plumb and vertical?

[cvoinescu]

That looks a lot like backlash to me. You say you checked your belts; have you checked your pulleys? The set screws sometimes work themselves loose.

[aiden1015]

Runout?

If it was runout, shouldn't I see the error throughout the board? I'm only seeing the error when Y-axis is stationary and I'm moving along in the X-axis. They are supposed to be 0.5mm wide though.

What spindle and endmill?

It's the quiet cut spindle using a 20deg 0.2mm v-bit

How is the spindle mounted?

I'm using the standard mounting kit for the Shapeoko 2

What did you do to ensure that it’s plumb and vertical?

I'm using ChiliPeppr's Auto-level to keep as consistent as I can in my cuts. I cut a board a couple weeks ago and didn't see anything noticeable.

As a sanity check, I ran another test to cut an L-shape whose dimensions are 10mmx10mm and are 2mm thick, so the shorter lines inside are 8mm. I measured them out and they're spot on. http://i.imgur.com/lSFB7ha.jpg

[WillAdams]

If it was runout, shouldn't I see the error throughout the board? I'm only seeing the error when Y-axis is stationary and I'm moving along in the X-axis. They are supposed to be 0.5mm wide though.

Good point. Go w/ Catalin’s thought on it being backlash.

[aiden1015]

That looks a lot like backlash to me. You say you checked your belts; have you checked your pulleys? The set screws sometimes work themselves loose.

I checked the pulleys and they all seemed tight except for maybe two (each on a separate motor), but since each pulley has two set screws, I couldn't see it being that big of an issue. I'm going to run another test or two and see how it turns out though.

[aiden1015]

After tightening the pulley's these look a lot better. http://i.imgur.com/1X5VKuU.jpg

I made my eagle file a bit smaller, and I still have some z-axis calibration to do, but...
The left three traces should be 0.5mm wide and I'm measuring the vertical traces to be 0.25 and horizontal to be 0.13mm. Those 0.12mm turn out to be about 5mil, so part of me is saying "come on, 5mil tolerance isn't that bad." But there shouldn't be that much error to me between axis.

[WillAdams]

How tight are the belts?

I’ve found that they need to be really tight, pluckable like guitar strings.

[aiden1015]

How tight are the belts?

I’ve found that they need to be really tight, pluckable like guitar strings.

I try to keep them as tight as I possibly can.

[jlauer]

Hate to say it, but this is why I stopped using my Shapeoko for circuit boards and went with the 3040 chinese CNC because the ball screws are super amazing compared to timing belts for getting accurate trace widths. I think Shapeoko's are awesome for large wood projects, but for detailed work like circuit boards I got frustrated enough with backlash that I realized the Shapeoko just wasn't designed for circuit boards.