So I am cutting a new top for my fishtank. After several false starts due to user incompetence, I got everything straightened out, or so I thought. Here are the particulars.
Designed in Easel and exported Gcode to Ground Control 1.23
Cuts started out fine, it got through the primary lid and then did the sides with tabs. The issues started when it went to cut the rectangles out in the lid. It replicated the errors twice it seems around the tabs and then the lower left corners it “shifted” the cut and dropped the lower right corner all together, it got even worse when it did the square near the arch. I figured no problem, I can fix that. Then when it went to cut the outline of the bottom it veered way way way off course. There was no chain binding, nothing physical I could see that would cause this. Fishtank.nc (101.7 KB)
Ugh… I didn’t. That could definatley be a culprit I’m going to do a new gcode with just the squares in the piece thats already torn up and see if it happens again after checking the screws like you suggested. Thanks, if thats the case I see loctite in the future. I will check out CAMotics. I really want to get a hd camera so I can record this thing in action for later review.
Can confirm your gcode looks good so don’t think its related to that. Its all G1 movement so I doubt there is an issue with the Arduino executing it.
Just throwing out another possible issue to check; as its difficult to tell from the pictures.
Did you notice the how deep the router bit was when the sled went off course? I say this because I think I see z-axis issues in the first picture with the rectangular pieces. Also notice that the z-axis ‘height arm’ is at top extent of movement in the last pic which could be a warning sign.
I have the same router and had the ‘height arm’ pop out from the recess in the router body. This resulted in similar wild cut as the router dropped too deep, and could not cut enough in time, to follow the gcode path. This would have been evident, at the time, by marked increased in router noise, and possible sled tilting. You can check the ‘woah’ cut and see if it is deeper than -0.16500 (first pass depth in your gcode). If its more than -0.25" deep, or changes in depth (getting deeper along cut), then z-axis is certainly one of the issues.
Perhaps you could better describe the “several false starts due to user incompetence” so we can rule out issues that are not related?
Just checked the old wiki, because the new one has not been updated with all the information.
The recommendations for rounded corners and squiggly vertical lines (particularly in the lower corners of the work area) are:
reduce feed rate
sand/wax sled
make work area more vertical
This isn’t related to the problem at hand, but I notice that the code is rather inefficient. The first part of the code, for example, cuts only the first bottom horizontal line, not the complete shape. It goes left to right, raising the bit, returning to the start (which is really inefficient), and repeating the process until that bottom edge is fully cut. It would be much quicker if it just followed the entire outline of the shape rather than cutting each line segment. I don’t use Easel, so I don’t know how drawings are done there, but you might want to look at making the pieces into fully paths (shapes) instead of individual line segments.
Good idea about chain guides. Also I see that you don’t have dust collection and dust on chains almost guarantees chain wrapping. It’s best to either get dust collection or frequently clean chains or both.
chain wrap is probably caused by the chains not being parallel to the workpiece.
a chain guide can help limit the damage, but fixing the root cause is better.
