Slippery problem with full circumference rotary lithopanes

[200k]

First I must give an apology because this is such a long entry. But if you can wade through it you will see that I am at wits end in resolving the problem because I’ve already tried everything reasonable.

I am a bit OCD but it helps with such a complex machine. I have a C machine and I am using the LHR rotary jig – the fancy metal one. I have a Ringneck Blues DC plus I keep the machine scrupulously clean by vacuuming out any and all remaining chips after any carve. I frequently vacuum chips out of the machine with a crevice tool while it is operating to prevent build-up on the belts. I also regularly clean and lubricate the entire machine including the rails IAW the seminars during the CW Conferences of 2012 and 2014 (same talk both years). In short, I take very good care of my machine and except for melting a flex cable by using the wrong lubricant, I have been relatively problem free. (hint: READ THE MANUAL)

I have been making a lithopane pattern of four fairies (FourFariesOPT.jpg) that carves the full circumference of 3” PVC. I have carved numerous copies of this pattern now and I have only once been successful in achieving a near exact 360° of carving. Usually the carving finishes before the machine has moved the PVC a full rotation leaving a thin ridge of uncarved surface that is 1/8” - ¼” in width. (NoCarve.jpg) I have to carve it away with a Dremel tool which I have to admit is not near as precise and the CW and it shows with trans-illumination.

I have tried several methods to correct the problem. First I just extended the pattern over the edge of the flat projection in Designer. This shows up in the Designer rotary projection window as a perfect carve (FourFaries-std.jpg). It does not carve that way. Any part of the pattern that overhangs the flat projection board size doesn’t get carved. It’s like it doesn’t exist. Then I increased the “board size” of the project tube by increasing its diameter. This gives me a little more “board length” to make the machine think the pattern is lengthened. When you load the part you manually tell the machine the part diameter is 3.563” rather than letting it calibrate (which would result in the actual of diameter of 3.5”). This trick resulted in the ONLY successful carve of a full 360°. Unfortunately the next one was worse than ever – over ¼” ridge. So even tricking the machine will not give a consistent result.

I ascertained the problem to be slippage of the project somewhere between the machine traction belts and the carved piece. I first thought the carved piece was slipping in the end plug but then I put a small screw to prevent that (PlugScrew.jpg) and there is absolutely no movement around the plug. Since the plug is held in the jig with two screws already, there is no way it could rotate independently of the jig either. So the slippage must be located at the interface between the rubber “fan belt” on the jig and the rubber traction belts of the CW machine itself. I thought putting a bit more head pressure would solve the problem but with a little down shove on the upper chassis (another trick I learned at the CW Conference) I can get over 140 pounds of head pressure – actual measurement! Since the manual says normal head pressure should be 75 to 85 pounds, I don’t think more head pressure will solve things.

So that’s my problem. Short of putting some sort of sticky goo on the traction belts (and ruining them for anything else), does anyone have an idea on how to achieve an accurate 360° carve in PVC?

Edit #1: I carve this project in "Optimum" setting so it will make very small cuts to minimize the pressure the tool will have on rotation. It takes over 3 hours to carve.

Also: My profuse thanks and Muchos Kudos to Michael Tyler who first successfully carved PVC tubing and showed it at the 2014 CW Conference. I had the idea before I saw how he did it, but he saved me making numerous mistakes because he made them first! (like trying to do it with 4" PVC) I still have more ideas but he did all the groundwork. If you've never met him, he's a real nice guy too. Thanks Mike.

[bergerud]

I believe you are correct when you blame the timing belt-rubber belt interaction for the inconsistent calibration. The rubber O ring may also be involved. I do not think the belts are slipping as much as "creeping". The soft rubber compresses and stretches. In the beginning I experimented with different things to get more consistent tracking. For example, I had the jig working better with sandpaper belts and no O ring. (I had the bearings on the tail end of the jig blocked up off of the belts.)

In the end, I just built my own jig. The next thing I was going to try, however, was thin rails between the rubber belts and the jig. Something like wood, aluminum, or plastic strips. Maybe an inch or so wide and 1/8" or less in thickness. (They would need to be 32" or so long for the calibration stage but could be shorter for carving.) I think rails like this would even out the forces on the rubber belts and provide more consistent tracking. (My wooden jig uses a wooden drive rail and has no tracking issues.)

[200k]

Your use of the word "creeping" is more accurate than my "slipping". I agree that it may be the CW equivalent of tire squirm. I like your idea of a thin rail. I have several long oak shims left from another project and I will give them a try. I will also look up your rotary jig plans again. I discarded the idea of making my own then LHR came out with their very nice version but now that I seem to be coming up against a limitation, I will give them another close look. I can look it up but if you have the page, URL, or posting number I would be grateful.
As usual you give a very cogent and expert viewpoint. We're lucky to have your input. Thanks.

[bergerud]

It would be great if you give the thin rail idea a try. You may have to lower the tabs on the jig so they still make it into the slots.

(I have not posted my rotary jig plans for all on the forum. A few members, however, have made it. I will send you the plans if you really want to make it.)

[blhutchens]

I have built Dan's rotary and was pleased with it.
I really like the dust collection built into it.
I have not used it a lot, as I have been busy with other work.
He just didn't get it developed quick enough, I had already purchased 2 from CW.

When doing these lithos, are you extending your pattern slightly past the edge of the board?
If you stop it at exact edge, rotary will carve 100%.
If you extend past it will carve 105%, so the edges overlap.
Rotary calibration every time you restart the machine is paramount !!!

[SteveNelson46]

I had exactly the same problem and solved it (I think) by making a few modifications to the jig.

1. Replace the belt. Make sure you loosen one of the idler pulleys. Don't try to stretch the belt. Don't ask how I know this.
2. Put a washer between the large gear and the chassis to provide some space and to eliminate the drag.
3. Add a washer between the chassis and each of the idler pulleys to keep the belt aligned with the large gear.
4. Enlarge the countersinks for the wood screws so the head of the screw doesn't rub.
5. Lube everything that rubs.

I also cleaned the rubber traction belts with a paper towel dampened with acetone. I don't know if that is recommended but it made them "squeeky" clean.

Although still not perfect, the jig seems to work much better and I don't have as much trouble with calibration. These are the same modifications Bergerud suggested in an eariler post.

[bergerud]

I was hoping that the rails under the jig idea might actually solve the calibration issue entirely. That is, eliminate the need to calibrate the jig every time it is used.

[200k]

I had exactly the same problem and solved it (I think) by making a few modifications to the jig.

1. Replace the belt. Make sure you loosen one of the idler pulleys. Don't try to stretch the belt. Don't ask how I know this.
2. Put a washer between the large gear and the chassis to provide some space and to eliminate the drag.
3. Add a washer between the chassis and each of the idler pulleys to keep the belt aligned with the large gear.
4. Enlarge the countersinks for the wood screws so the head of the screw doesn't rub.
5. Lube everything that rubs.

I also cleaned the rubber traction belts with a paper towel dampened with acetone. I don't know if that is recommended but it made them "squeeky" clean.

Although still not perfect, the jig seems to work much better and I don't have as much trouble with calibration. These are the same modifications Bergerud suggested in an eariler post.

I will do a couple of your suggestions like clean the belts with acetone and lube axles and such. I've already enlarged the counter sinks because the screw heads were dragging on the left plate (drive side). I will take a look at the pulleys and large gear again to see what putting washers in will do. I've already had the jig entirely apart due to an unscheduled and spontaneous dismantling of the drive axle. I had to put it together again and I heartily suggest that everyone use Blue threadlocker on all nuts. Understandably this jig gets LOTS of vibration which is anathema to the security of nuts.

Thanks for the suggestions. I'll post the success of the various solutions members have offered.

[karossii]

I had a similar problem - my solution was to remove the feathering from the outside edges of the carve. It wasn't slippage at all for me. I notice that you have the top and bottom feathered, or so it would appear in the designer. More than likely, this is a feather all the way around your piece, yes? For me, that was causing a consistent ridge at the edge of the carving, just as you have pictured (the strip you removed with the dremel). By removing the feather completely, and then making a manual 'ramp' at the top and bottom to ease into the carving and increase aesthetic appeal, it fixed it.

All that said, I made no more than a dozen of these rotary lithophanes before I moved my shop, and haven't gotten back around to carving any more in the last 6 months or so; so I could be forgetting some other issue that was related. But I am fairly certain that was how I fixed my problem.

[bergerud]

I went ahead and experimented with the thin rail idea. I used a strip of acrylic with masking tape on each side. I repeated the calibration routine many times. The results were better than without the strip, but still varied as much as 1/32" between runs. (I had to lift the head between runs to re-position the strip.) Then I thought to try it under varying head pressures. I found much larger variations (as much as 3/8") with head pressure changes. So, the strip idea is a failure. Glad I tested my idea before someone else did!

As you can see in the last picture, the strip flexes and does not distribute the force as a board would. This local compression of the rubber belts may be the problem. I will need to try a thicker drive rail next.