Cut path keeps wandering after first pass

[oscarl48]

Total frustration today. I attempted to do a cut out of a pattern with a .2 depth per path cut to minimize pressure on the bit. After the first path pass it just seems to wander. I have never been able to get a clean cut through on any pattern so far. The second pass is always a little bit off and the third is off a little bit more from the second. My earlier attempts were on smaller boards without being under the rollers so this time around I gave myself plenty of extra board so it would remain on the rollers. Exact same issue.

This is attempt number 5 for a cut out with different patterns and they have all done the same. Is there something I need to tweak, tighten, or do differently?

I understand possible slipping on roller but this is attempt number 5 with different patterns and lumber (same issue). I am 0 of 5 attempting to do a cut out. I do use painters tape to keep the roller in contact with the lumber.

The frustration is the carving came out exactly like I wanted which this then ruined. Not to mention a beautiful piece of black walnut.

The pattern I am developing is for a mantel clock that will have 5 cut patterns glued together to build up the body of the clock. I was hoping for precision to ensure all the boards have the same cut out dimensions.

Any advice/help is appreciated.

[fwharris]

Just a WAG, have you checked your head pressure? To low and the board could be loosing tracking.

[Digitalwoodshop]

1. Try Masking Tape.... Painter's tape is much too slippery....

2. Leave extra wood on either end to better stay under the rollers holding the board to the brass roller. When only one roller is pushing the board down the brass roller looses contact with the board EACH PASS by 1/4 inch...

Been there done that...

AL

[oscarl48]

AL,

Thank you. Looks surprisingly a lot like mine. Or not so surprisingly. I'll try masking tape and see if there is any difference.

fwharris,

Checked roller pressure and right at 80lbs. Second test at 78lbs. Third test something clanked loud and 120lbs. The clutch did not engage. Hopefully I did not break something. I'll tear apart tomorrow morning and see if its just gunk in the gears (i hope and not something that snapped). Thanks for the advice as well.

Oscar

[oscarl48]

After reading several threads on clutch not engaging ran over to walmart and picked up a can of white lithium grease. It works. Cheap fix. I like it.

[bergerud]

As for the stair stepping problem, how close was that long cut to the brass roller? I seem to be the only one who blames cutting too close to the brass roller for tracking problems. Try taping some rails onto the sides of the next cutout you try. The wider the better.

[blhutchens]

I have had the same issues. There are those who say its the wood, too close to the brass roller, etc..... I can't seem to isolate the problem. I have re-calibrated both machines and tried every suggested remedy and still only get acceptable results on 1 in three cut outs on multiple passes. The main one that I have trouble with circular bases and caps for rotary lithos. I cut a full depth daddo for the pvc to sit in first then do the cut path.

I have quit doing full cut outs on more precise cut outs. I will do an outline with the carving bit at a shallow depth (1/32) and cut with band saw or scroll saw.

[blhutchens]

It looks like the first cutout for the clock face worked properly. That is my theory, that it has something to do with circles.............I don't know why it will work fine 3 or 4 in a row and the fifth one, (usually when i step out of the shop for a minute and can't stop it) it will mess up.

[oscarl48]

Dan, Brent, thanks for the replies.

Brent I am at a point that I am going to start doing that. I think I have read most of the threads on the topic so understand some of the issues but I am upset at losing all that beautiful hardwood. This machine is an amazing carver but as a cnc cutter it has performed poorly for me.

Dan the long cut was away from the roller but the circular cut was near the roller. I left an inch of material between the edge and the cut. I will continue experimenting with your jig on smaller projects. If it is vibration it should help in dampening some of it.

Thanks again to all for the advice. I am over being angry at the machine and will not throw it out a third floor window.

[bergerud]

Don't give up! This is a problem for many. Very discouraging to wait for a long carve on expensive wood only to have the project ruined in the last operation. I too had resorted to single shallow pass - band saw - router table for cut outs. Even then, they are sometimes not right.

I do not buy the brass roller slipping theory which is most widely believed. Not on a well behaved board under the rollers. If brass roller slipping was really a problem, the presence of the O ring would make a drastic difference in behavior, but, it does not. I see deep tracking marks in my hard boards, and it takes virtually no torque to roll the brass roller; I cannot believe the roller slips.

I have noticed the problem worst when the bit cuts near the brass roller and I have had 100% success with cutouts on the carrier bases I have made. I think it is clear that the problem has to do with the brass roller encoder data. The machine does bad cutouts without the encoder data. Start a cutout, stop it, lift the head, move the board off of the brass roller, clamp the head back down, and continue. The machine will do the cutout using only the x drive motor encoder data and it will be way off. You will see extreme stair stepping.

The brass roller data is not used as the absolute x reference when cutting, the x drive data is the reference. The brass roller data is only used to correct and update the x encoder data. Why is this? Why would the programmers not use the brass roller data as the absolute reference? Answer: it must be unreliable when cutting.

I noticed early on that a small slip of the board on the belt (for example when pushing up a roller) shifted the cutout. Why, I thought, can the machine using the brass roller data not be able to correct for this? It does correct for this while carving, but then, carving is different. The movement of the board when carving occurs when there is no cutting.

My conclusion is that the vibration of cutting can, at times, confuse the brass encoder. The disk must get into vibrations and the rulings move too fast to count or keep track of. The firmware has to then fall back on the x drive encoder data. Each time it has to do this, the tracking is a little more off.

This is a very hard theory to test as vibration transmission and resonance are fickle. The vibration of the disk would change depending on head pressure, type of board, cutting rate, whether or not there is tape under the board edge, and so on...

I do have a possible work around and I hope those of you who experience this problem will give it a try. I figure that if one loosely tapes a "tracking" rail to the edge of the board, the transmission of vibration from cutting to the encoder disk will be greatly reduced.

[lynnfrwd]

There seems to be 5 passes assigned to that bit cut path. You increased your chances of error by 5. One swipe with a 3/16" would have been sufficient and correct based on the photo. Good chance if following the 7" rule one swipe of the 1/8" would have done the job. I disagree with this multiple pass cut outs for every project.

[lynnfrwd]

Ok 3 passes. I was looking at the example instead of the OP.

[oscarl48]

lol. I'll split the difference with you. It was 4 passes.

I had the same idea since the first cut seems to be close to true. I'll have to order a 3/16 bit to test. Cherry and black walnut are both fairly soft. I would not try that on padok or purpleheart. I think I would be hesitant with maple as well. Have you had better success with a specific species of hardwood?

[DickB]

I do a lot of cutouts in my clockmaking, both single pass in 1/4" Baltic birch plywood and three passes in 3/4" hardwood. I typically have multiple parts on a board. The wheels also include drill operations, which are done before cutouts, so the head is moved allover back and forth on the board with these multiple parts and movement must be accurate. I also often have carving and additional machining with the 1/8" straight bit along with the cutout parts. I have on rare occasions had stair stepping but most of the time I get good results.

I almost always use a simple sled with 1" rails (keeping the cutting bit at least that far away from the brass roller) and 4" ends guaranteeing the 7 inch rule. I did not see info in your post regarding your board size and how close you cut to the edges of your board (maybe I missed it), but the sleds have served me well. If you are not using one, or Dan's jig, you should consider it. Dan, I'm not sure about your vibration theory, but I suspect the type of sleds that I use like your jig do reduce vibration into the roller due to the extra mass and distance from cutting bit to edge.

7660976610

[oscarl48]

Completely jealous. Good looking gears. If you don't mind sharing how did you make the spiral spokes on the second screen. I have the gear generator program but on that program I've always thought the spokes were utilitarian looking and not very graceful.

I think the message I am getting from everyone is for cut outs use a jig. So from now on I'm going to try using jigs for the cutouts. :)

The board design is 10 in wide by 20 in long. My board was around 10.5 in by 26 in.

I was planning on having this as one of my patterns for the pattern depot but with all these issues I think instead I will simply share it. Share the headaches. ;)

I haven't quite worked out the back yet. I want a panel access to get to the clock mechanism but I haven't put into designer yet.

[Digitalwoodshop]

Like posted above, it all comes down to the brass roller loosing contact with the physical board by a fraction of an inch on every stroke. In my pictures above, for frugAL use of wood, simply taping blocks of wood the same thickness and width on each end of the good wood keeps wood under the rubber rollers at all times. When you cut path up close to the end of the board edge in length less than 3.5 inches from the end. The board can LIFT while only be supported or being pushed down on the belt by one roller. This leads to the step cut. It's simply properly understanding the 7 inch rule of wood. Doing a project that ends up only being supported by ONE rubber roller during the cut path process is like riding a bike with one hand.... Normally not a problem... but then one slip and you crash...