one wavy line of text on 3 line carve

[Wood Art 1]

The last line to carve on these plaques wanders above and below the line. Centerline. plaque is 8" tall x 5.5" wide. Using a Bergerud fixture so all that pieces are a precut size. The "May 3" were the last two to come off the CW. The other 8 came off last week. I have about ten examples with the same font and layout that came out OK inbetween last week and the May 3 cuts. Driving me totally nuts - it's only a short drive.
Is this a font style issue?

[bergerud]

That is curious. They are off in the same way. I wonder what the cutting order of the letters was. I this 1.187?

[Wood Art 1]

version 3.1. This is the last line to carve. The 0 is first, M is 2nd. 2 is third...walked away frustrated at that point.

[myshop1044]

Jerry it looks like the wood is moving, is it tight in the jig?

[bergerud]

Jerry it looks like the wood is moving, is it tight in the jig?

Perry may have a point. The board is only under one roller when the text is cut.

Jerry, do you have sand paper grip strips on the rails?

You might also check the head pressure.

[Wood Art 1]

I have sandpaper on the rails, and I grab the piece to lift the jig to test that it is in tight before putting it into the CW. I will check head pressure this morning and post the result. The jig is approx. 24" long and the piece is always loaded somewhere near the center, wherever I happen to place it (at a random location). The text is set .5 inch from the top of the board. The font is Arial Rounded MT Bold - bold. The text area is 0.5" tall by up to 4.75" long.

[SteveNelson46]

My next step would be to try it on a piece of scrap without the carrier board. If that isn't the cause at least it rules out one possibility.

[bergerud]

I wonder if the teeth on the brass roller could have chewed up the bottom of the carrier.

[Wood Art 1]

Thanks to you guys for helping me focus.
Sunday Morning test. Machine temp 46 deg F. Head pressure at 65#. Lubed posts and screws - Now 75-77#. Rebuilt bottom rails on the fixture with UHMW yesterday before last test.
1) Put same board in fixture- tested for tight by lifting fixture by holding board. Pic "fixture hold".
2) fixture into CW. tested rail clearance - slides completely thru without binding.
3) Used same MPC as yesterdays test, Carving order 0M,23y1a6. same result Pic "Test 1"
4) Used same MPC - changed attachment at top from 0.5" to 0.7". Result Pic "test 2".
Questions???
1) at what point between .5 and .7 does the wavy line get corrected? more tests required!
2) Is this a software problem, hardware problem, or a font problem?
3) Should I just shut up and accept this solution?

[bergerud]

I tried it right on the edge. (Designer 3 and its firmware.)

[bergerud]

T
1) Put same board in fixture- tested for tight by lifting fixture by holding board. Pic "fixture hold".


Do you mean that you tighten the wingnuts until the board is pinched between the sides of the carrier? If so, that is not the way I intended it.

The sides would not stay parallel as the bottom would squeeze in. The sides are supposed to stay parallel; it is the sandpaper and roller pressure which supposed to hold the board. I do not know if this could have anything to do with it, but, check that the bottom of the carrier is flat on the belt near the brass roller.

[Wood Art 1]

Dan, my apologies for hijacking your fixture. Have used this fixture for about two years. Replaced the bottom rails once and sanded them smooth to eliminate tracking roller groove a couple of times. Always use masking tape. ALWAYS tighten side to prevent slippage. Only use the end stop when board length gets close to 19" max fixture length. If the 8" plaques are all cut from the same board - I set the board into both ends of the fixture to set the base fixture width to match board width. Then I set the top of the board slightly less than 1/8" proud into the fixture, using a handmade gauge - setting the board at both ends to tighten the rails. Move the board to the center - approx - and squeeze the board into the fixture.
The wingnuts at the bottom of the fixture prevent the bottom from squeezing in. Only the wing nut in the rails is used to tighten the piece into the fixture. Approximately one turn tightens the board.
Therefore every board that goes into the CW gets measured by the CW.

[bergerud]

Oh sorry, I did not know you added the extra tightening clamp in the middle. Ok, so... still no reason for the weird tracking.

[DickB]

I wonder if the carrier is slipping. The drift in the letters seems to coincide with the order of carving, suggesting that something is slipping. Because the board is proud of the carrier, one roller is going to be contacting the board and one contacting only the sides of the carrier. The latter roller will be exerting less pressure because the carrier is lower. I wonder if this is causing the carrier to slip. If you measure 65 lbs with a board contacting both rollers, the actual pressure with one roller contacting the board and one contacting the carrier is going to be less, is it not?

[Wood Art 1]

Dick, I lubed the guides and got the pressure to 75-77 lbs. the fixture will not move in the CW until I crank up the head when removing the fixture after carving. I will do another test and hold my hands on the fixture and belts to determine if there is any slippage. The Plaque does not come out of the fixture until I loosen the side clamp. I will try some more tests with the attachment @ .51, .52 etc to see if it continues. I will try another font at .5 attachment and etc. to see if it continues. I have about ten examples of the same font at .7 attachment that all came out straight, between the United States and the May 3 examples.

[bergerud]

I do not think the board is moving in the carrier. With the sandpaper and the side clamping, it does not seem likely to me. Easy to test though. Just mark the carrier rail and see if the board moves.

[DickB]

I will do another test and hold my hands on the fixture and belts to determine if there is any slippage.Even if the carrier is not slipping on the belts, it could be slipping on the roller. That is what matters.

With a board under both rollers, pressure should be exerted equally onto each belt. If one roller drops off the board onto the carrier, there will be less pressure of the carrier against that belt. It seems to me that the carrier could lift slightly on that belt, also lifting slightly from the roller. The carrier may still move in sync with both belts but not the roller.

I have had roller contact issues with slightly warped boards when the board contacts the outfeed roller, which were properly adjusted in height for a perfectly straight board, but lifted the warped board ever so slightly. It doesn't take much moment to cause the roller to lose contact and slip.

[bergerud]

If the side of the carrier is straight, I do not think it is slipping on the belts or roller. If the carrier was going to slip, it would do it while measuring the board. There is very little head pressure during board measurement stage. After that, the head is cranked down. I am thinking more like something is messing with the brass roller. (Press down the brass roller to see how close it comes to the front belt.)

[Wood Art 1]

Dan, I just checked the roller clearance to the belt. Pushed down to move the roller to max and with a mirror inside the CW to observe, still have about 1/16" clearance.
Ran a bunch of tests with various attachment figures. Using the same MPC but adjusting the attachment from .5 to top of board to ? Marked the fixture and the belt to see if there was slippage.
Test 3) Result OK, .5 Attach , no slippage
4) OK, .2 attach, (M36y120a,), no slippage
5) Wavy, .52 attach, (M36y210a,), no slip
6) Wavy, .55 attach, (M36y120a,), no slip
7) OK, .57 attach, (M36y120a,), no slip
8) OK, .6 attach, (M36y120a,), no slip
9) OK, .65 attach,

[bergerud]

Could it be that the backlash in the x gear system is getting too large? With vector stuff, where the x drive has to go back and forth, the x backlash can be problem. Could the belt system have too much drag? Has a belt shifted?

Sorry, I am out of ideas!

[Wood Art 1]

Thanks for all your help. I think the problem is related to the proximity to the top edge of the board. No proof, and running out of things to look for. Have to get back to work. About thirty of these plaques to do in the next 10 days. Instead of having the machine determine the order of carving which meant that this line was the last line to carve in projects that use 3 separate carvings and takes approx an hour to complete. I am going to separate out this line and carve it first - a 3 min carve, then if it screws up, I waste a board and 3 minutes, not a board and 1 hour. I will probably try a different font. It's almost like the text function is bumping individual characters up and down.
However, if you guys get any more ideas, please send them. I will give it a shot and report back progress later this week.

[Wood Art 1]

Dan. no evidence of belt shift. and since in most cases prior to these tests, I had a line 2 at the top and two more lines of text (same font) at the bottom of the plaques that ALWAYS turn out OK. Hard to believe that It would only screw up on the last line. Top of board may still be a factor.
Still love the UCB concept, even if I don't use it right. I haven't put a board on the belts in months. When you were first working on the design a couple of years ago, I picked up the concept. As I mentioned I've replaced the bottoms with Ultra-High Molecular Weight Polyethylene UHMW. I still use masking tape. Be interesting to see if these pieces survive the brass roller better. I've seen UHMW pieces 8"x8"x1" used in conveyers get folded to 90 deg and when they return to the inspection port, they are straight again. Tuff stuff!
Thanks Jerry.

[bergerud]

I was thinking of embedding a strip of hard rubber under the base rail for the tracking roller.

You know, with your repeated runs of the same project, and since you already have the clamping, you should consider the procedure:

http://forum.carvewright.com/showthread.php?27235-Experimental-Board-Carrier&p=249253#post249253

[Wood Art 1]

Dan, I will let you know how the UHMW works out, should have an indication in a few weeks. It is available in strips and sheets in various thicknesses at many plastic supply houses. I happened to have 3/4 inch. Had to replace both rails since I had sanded them to get the brass roller grooves out. But if I was starting with a new fixture, only the roller rail would need to be special, the other only needs to be the same thickness.

[bergerud]

You know, my theory of poor tracking has to do with vibration confusing the tracking encoder. It is very hard to test as it would only happen for certain resonant frequencies. I bet the UHMW transmits vibration to the tracking roller more efficiently that wood does. I wonder, as an experiment, if you could try a few more layers of masking tape.

[Wood Art 1]

I've only had the 3/4" UHMW bottom rails on the carrier since Sat nite. One layer of masking tape. The carves that I did over the weekend reflected almost the same tendencies as the previous plywood rails. The plywood was replaced because I had seen tracking grooves in it and had sanded them down a couple of times. Saturday I was not sure that I would get the two rails equal or straight so I made new rails.
Two carves today both came out ok, carve of whole plaque with May 3 set at .6 was OK, carve of whole plaque, same font "U.S. Air Force" set at .7 came out ok. Before I replaced the plywood bottom rails with UHMW I used 3-4 layers of masking tape.
I will watch for faults that might be caused by vibration. Would a heavier fixture help dampen the vibration?
I keep thinking the short distance from the end of the board is a factor, but can't prove anything. I was hoping someone would say, "O Yeah, don't ever use an 8 inch long board with text close to the top end, it causes erratic carving"!!!

[Digitalwoodshop]

This is an interesting repeatable problem... It's not random it is a fixed problem every time... So....

First thing I would do is place a 4 foot level in the machine and crank down lightly. Then check the in feed and out feed trays especially at the metal roller. I am suspecting that the ends of the jig is hitting the HIGH metal roller causing it to slip and off setting the jig causing the text problem.

SO if I was troubleshooting this, I would switch jigs to a board type... This would be a piece of 1/8th inch plywood or 3/8 cut to board width. A Scrap Board the same thickness and width would be taped to the board with the good BLANK taped in the middle. I would measure from the X end of the board to the joint of the Blank. In designer I would draw a rectangle the same distance and place it on the right side of the screen at the board edge. Remember the top of the screen is the brass roller. I would place the text to the left of the rectangle. I would use PLACE ON END rather than center on board. Cut the project and see if this is actually a Software Problem with the FONT.... Place Masking Tape on the bottom brass roller edge of the plywood.

What's different.... for one, the board is being held along the Y or width of the board. Using that thin wood rail jig could cause premature wearing of a THIN section of Rubber Rollers causing LESS SUPPORT...

I draw rectangles on the board to use as tape measures.

AL

[Wood Art 1]

Al, thanks for your input. The rollers on all four corners are approx. 1/4" below the bottom of the fixture. The bottom of the fixture is flat, and 23 1/2" long. The bottom rails on the fixtures are 1 3/16" wide x 3/4 thick. At this point I have moved the text to at least .6 attachment and have had no issues. Will have to test the board type sled later this week. If I have more issues I will post here. Yes the rubber rollers could become indented by the the top of the fixture. Will look into that.

[mtylerfl]

Hi Jerry,

You have had a lot of great suggestions on things to check and test.

I agree it's most likely a Brass Roller/position related problem (tipping, slipping, head pressure, rubber roller wear, etc. as already mentioned). I seriously doubt it's a font problem or an encoder problem. I am interested to hear the end of the story, once the problem is positively identified.

There is something else that could possibly be a factor...workshop temperature. You mentioned the CW was 46° in one of your postings. That's too cold for 100% performance reliability 100% of the time. I am curious if the "good" ones were done when the shop was actually a bit warmer? I think the manual says 50° minimum operating temperature, however, I discovered on my own that the shop needs to be 55° or warmer to have 100% complete success. Wacky things can and do happen in a "too cold" environment. I know, I know, some folks run their machines in the cold and have had good luck. But, sooner or later...oopsy!

If you do not nail down the problem after you have exhausted all the other suggestions, try moving the machine to a warmer area (let it warm up for several hours - overnight maybe) and run the project again to see if that makes any difference.

[Wood Art 1]

Michael, thanks for the input. My shop is only heated during the day when I am using it. Typically I turn the heater on an hour before I go out to the shop. The temperature in the shop will typically drop to about 35-40 on a cold nite (20 deg outside). I have a temperature gauge on the z casting to observe the carving temperatures. It was at 46 deg as noted. the shop temp was approx. 55 at that time. I use a couple of fans to move the warm air around the shop (approx. 650 sq ft with 8'6" ceiling) and the general temp will run about 65-68 deg (sometimes to 70+). My experience has been the wavy text has come off at all times of the day and temperatures and will occur in between good carves. I will observe temperatures at the machine to see how it is affected. Al's point about trying the flat fixture may provide a clue if the wavy is not repeatable in the flat - then it would appear to be a fixture/tracking issue.

[bergerud]

I've only had the 3/4" UHMW bottom rails on the carrier since Sat nite. One layer of masking tape. The carves that I did over the weekend reflected almost the same tendencies as the previous plywood rails.

Ok, nothing to do with the UHMW bottom rails. It must be x drive problems. Maybe, as MT says, because of cold.

[Wood Art 1]

Thanks guys. I've got lots of stuff to check and watch. No evidence of deformed rubber rollers. Temp at machine is now 70.
I'm not going to say it can't be temperature, BUT why do I have about a dozen examples where the last line on a two-sided carve is the ONLY line to ever screw up? Right now I am not carving any text with a .5 attachment. We'll see if it screws up again. 2 done OK.
Just a thought...I will try a plaque in the modified UCB with a filler block at the top and instead of going to the center, start at the corner - the block will prevent the rollers from coming off the plaque onto the fixture.
Back to the shop, got plaques to do.

[DickB]

I was going to suggest trying a jig such as Al suggested, but that is really changing the nature of the problem. But if it works reliably then at least you have a means to complete your projects.

[Wood Art 1]

Thanks Dick. Used a flat fixture like Al suggested. 3 plaques OK. None of those were .5 attached.
I'm now thinking it may have something to do with the roller coming off the plaque onto the fixture. Will add 3 1/2" extension to the top of the plaque in the modified UCB to eliminate the .5 attachment issue. Back to the shop, got work to do.

[mtylerfl]

...I'm now thinking it may have something to do with the roller coming off the plaque onto the fixture...

Yes, that could certainly be it alright. Both rollers should remain in firm contact with the material (or jig) during the full run - - never coming out from the captivity of either roller, and the continuous contact should be without any "speed bumps" (no matter how tiny) due to an uneven surface. Not following those "rules" always invites tracking issues. Only "luck" prevents tracking problems if a project is run without constant captivity of both rollers at all times.

[bergerud]

I think it is clear that this is a tracking issue. From my understanding, the cause of such issues, for whatever reason, is a disagreement between the tracking roller and the x motor encoder. The controller goes into a brass roller fail state and abandons the correction routine it uses the brass roller for. (The brass roller data is also ignored when a roller drops and switches.)

Rolling off of the board onto the rails of the carrier does not rock the carrier off of the brass roller or switch the roller switch. The only way I can see that this could be connected to a tracking issue is if the tracking is affected by the switching of a little head pressure from both belts to the front belt. If this is the case, it means that the x belt system has a problem. Too much drag or backlash.

[mtylerfl]

Yes, we have identified the symptoms, now to deal with the disease. I'll bet giving the pressure rollers a continuous, smooth ride whilst both remain in constant contact with a flat surface might be the cure. I sure hope it's that simple, anyway.

[bergerud]

Michael, you are implying that the fault is in the design of the UCB. I do not believe this is the case. The main point of the carrier is to stay under the rollers and solve the tracking problem. A great many of these have been made and used and this is the first I have heard of any tracking problems. I recall you saying that you have never had tracking problems. Well, many of us have had tracking problems. I was plagued by them until I started using the carrier bases.

Having the rollers roll up and down an eighth inch bump is nothing compared to rolling on and off the end of a board down on the belts. If the CW was as sensitive to bumps as you say, the CW would not even be able to measure a board.

I am worried the carrier base may have damaged the underside of the rubber belt. Jerry has been using the thing for years and it seems possible that the narrow rail pressure on the edge of the belt may have taken a toll. Maybe the underside of the belt is worn ragged and is causing extra friction.

[Wood Art 1]

Mike and Dan. Okay, I carved 8 plaques today. All of them the same size 8"x5.5". A variety of graphics and text. All of them in font Arial Rounded MT Bold. Carving on both sides, approximately 30-40 minutes of carving time on the front. 3 lines of text - 1) at top .6 or .7 attach below the top. 2) bottom line .5 attach above the bottom. 3) a line just above the bottom line no attach. All of them in my modified UCB. Plaque is approx. 1/8" proud of top of fixture. No additional spacers or fillers. Machine Z temp this morning when I opened shop was 55 deg. When I closed tonite the thermometer beside the CW was at 75 deg. All plaques turned out fine, no wavy text.
I did not try to replicate the problem today nor did I change any processes (other than the .5 to .6 or .7 attach).

[DickB]

Having the rollers roll up and down an eighth inch bump is nothing compared to rolling on and off the end of a board down on the belts. If the CW was as sensitive to bumps as you say, the CW would not even be able to measure a board.This is one of the cases where I do at times see tracking errors. If I am using a long board or a long sled and the outfeed rollers are not adjusted just so, I will get a tracking error when the one roller drops while measuring the board. I'm not saying this is the same thing going on here, just observing that when a roller drops the dynamics change and it can lead to tracking issues.

The type of sled that I use is not as flexible as the UCB but it does have some of good characteristics. Once the project is measured the rollers never drop even a small amount so consistent pressure is maintained during machining. The base of the sled is full width so full contact is maintained on the belts with no narrower pressure points. The spindle is kept away from the roller while machining and I believe the mass of the sled also reduces vibration to avoid vibration-induced tracking errors. I too have suffered from tracking issues, even fairly recently when I was cutting out a sign without a carrier (due to the width of the board) http://forum.carvewright.com/showthread.php?28480-More-accurate-machining-when-using-a-sled. I eliminated the problem when I used a sled.

Wood Art 1 stated that in his latest run his board was 1/8" proud of the fixture. Hard to tell but in the photo posted earlier on it looked like it was more than that.

No criticism of the UCB intended. It's a clever design and I agree many people are using it with success. We seem to agree that this is some kind of tracking issue, but the exact root cause appears yet to be in question.

[bergerud]

I agree that having the rollers drop even 1/8" is not ideal. It is essential for the measurement stage but it is not essential for the carving stage. Initial measurements on my machine showed that the carrier did not move at all as the rollers dropped the 1/8". On some machines, however, it has caused a small "snipe" line in carvings. (This is a very small movement which is far from causing a tracking problem.) To fix this, I introduced 1/8" thick shims to fit on the rails. With the shims, the rollers do not drop during carving.

My later ACB design does not have this problem at all. The board drops down into the carrier after measurement and there are no roller drops during carving.

[luckettg]

I had this happen on a 11.5 x 33" board while doing a carving with centerline text in the pocket, centered on the board, always under the rollers. Last line is wavy. My guess is something to do with the X axis feed. I then carved another one, different board, which came out perfectly. The only variances was the boards. There was not any error msgs to check the X axis, but I have seen some strange things since converting to the rubber rollers on the X axis. I have double checked the machine's assembly but found nothing amiss yet. This could be several things, from a software glitch not reading the board positions to the board slipping in the X axis. I will keep observing.
83260

[SteveNelson46]

I had this happen on a 11.5 x 33" board while doing a carving with centerline text in the pocket, centered on the board, always under the rollers. Last line is wavy. My guess is something to do with the X axis feed. I then carved another one, different board, which came out perfectly. The only variances was the boards. There was not any error msgs to check the X axis, but I have seen some strange things since converting to the rubber rollers on the X axis. I have double checked the machine's assembly but found nothing amiss yet. This could be several things, from a software glitch not reading the board positions to the board slipping in the X axis. I will keep observing.
83260

Are you sure the belts are not slightly (or occasionally) rubbing on the o-ring of the x tracking sensor? It is a common problem if the o-ring wasn't replaced with the thinner one in the belt upgrade kit. A high spot on the belt would cause the problem to appear random.

[luckettg]

I cannot say for absolutely certain what is happening when this occurs, as it is intermittent and I do not discover the error until the carving is finished. I can say that I have found nothing operating incorrectly on the machine itself. I have spent the time to run tests and to observe, and everything is working as it should be. Just every now and then, a carving goes whacky (Engineering term to explain things to marketing people?) is the best way I could describe it. I will, or someone else will, eventually ferret out the cause.


Are you sure the belts are not slightly (or occasionally) rubbing on the o-ring of the x tracking sensor? It is a common problem if the o-ring wasn't replaced with the thinner one in the belt upgrade kit. A high spot on the belt would cause the problem to appear random.

[Wood Art 1]

Thanks to all you guys, who have thrown in on my wavy line issue. Since my last post two weeks ago I have carved about 25 more of the same group of plaques. I had one instance of wavy text in those 25. The text was placed at 0.5" from the end of the board. I no longer place text that close to the end of the board. It appears that I have a tracking problem that only occurs when the text is that close to the end. Probably not a machine problem. My apologies to Dan for getting him sucked into this - - Probably because I am not using the UCB as it is intended.
In the future, when I want text close to the end. I will insert into the fixture, (a 4" piece at each end of the board, approx. equal in thickness and width) after it has completed the measuring sequence, so that the roller does not drop onto the fixture. Thereby convincing the machine that the text is really 4 1/2" from the end.
I have heard of lots of reasons, why I have the problem. Is there a solution?
PS> to Dan - so far the UHMW strips on the bottom of the fixture are not showing any wear, even the masking tape seems to last longer.