Edge detection Error

[55president]

I’ve been having an edge detection error that I’ve been trying to eliminate for quite some time and I just don’t know what to do.


To start off I recently did a 500HR overhaul on my machine.I replaced gear and bearings in y motor pack, replaced FCC cable, replaced board sensor, sensor cable , z motor pcb board, y truck bottom roller bearings, (2) ,Y belt Idle pully assembly ,Retaining ring for 5mm external shaft, and oiled the flexible shaft.
checked motor brushes, but did not replace.


2 carvings later -- I tore a y belt. I think my material was too tough and my
bit might have been a little dull. (continued...)


After changing the belt - I’m was getting an Edge Dectection Error as soon as
it started to check the board ( white pine, Lighted with a trouble light) . Checked my sensors board sensor reading 150+, I figured it might be
the wiring for the board sensor. So I rechecked that, and I noticed a point on
one of the wires that has rubbed through.I wrapped this with tape and buttoned
everything back up.


Same Edge Detection Error.


Took it apart again. Replaced the wire to the board sensor with the old wire
( it was still good), and for good measure replaced the new board sensor
with a good one I had. Buttoned it back up....
board sensor IS working - reading up to 155.
Same Edge detection Error, but instead of the error at the edge It goes to the
center of the Board, before the error occurs


I updated the firmware, same story Edge detection error.


Kind of at a loss, here any suggestions?

[Digitalwoodshop]

IF the board is TOO dark the reflected light will not be high enough. Masking Tape on the board will help.

AL

[55president]

It's White pine with a Florescent trouble light shining down on the board inside the cover, I've never had this trouble before.

But haven't tried the Masking tape yet , But I'll give it a shot......

[55president]

Same story with the masking tape -- Edge dectection error, But without the light it stops at the edge and doesn't start
to read the board, but with the light on it's go 1.5 inches or so then give the error.

[TerryT]

Just a wild a** guess but did you try it with the flouresent light off?

[55president]

Terry -- yes without the light it will givethe error right away, with the light
it will measure about 1 1/2 inches the ngive the error message.

[fwharris]

Lets try a test of sorts,

Put your board in and lower the head down to the board. do a board sensor check (0,7) and with the head all of the way down move the truck across the board and see if the reading stays the same or changes. Also move the head up and down.

[Digitalwoodshop]

I was going to Post the very same test.... IF the reading drops out as you move the head manually in Y.... The FSC Cable could be bad... As it flexes it looses connection....

AL

[55president]

Al,Floyd-- I loaded a 12" maple plywood blank and checked with the sensor .After several tests, the sensor starts to read the board around
115 and as I move the head in the y plane , the reading goes up as much as 120 ,but pretty much stays in the 115-110 range and
drops where there is a darker spot down to 85 then back up to 110-115. So Yes it does fluctuate but never goes down to zero.
This is with the florescent trouble light on and off so it's the same either way. I just changed the FSC cable.I made sure the cable was
fully inserted into both boards.

[fwharris]

Al,Floyd-- I loaded a 12" maple plywood blank and checked with the sensor .After several tests, the sensor starts to read the board around
115 and as I move the head in the y plane , the reading goes up as much as 120 ,but pretty much stays in the 115-110 range and
drops where there is a darker spot down to 85 then back up to 110-115. So Yes it does fluctuate but never goes down to zero.
This is with the florescent trouble light on and off so it's the same either way. I just changed the FSC cable.I made sure the cable was
fully inserted into both boards.

Ok so that sounds good on the readings with the plywood blank. Did you try your original board also to see what the reading are? One thought that just occurred to me. Your original board, was it a solid board or one that is in a sled or carrier??

[55president]

The original board was a 1"x4"x36" white pine board that I was using for
measuring calibration. I did not do this test with that
board but will.....

[55president]

no carrier.

[BBrooks]

I read this several times. I don't know anyone who knows the CW better than Al but Is it possible that the loss of the Y belt has taken the truck out of
alignment on the rails? I know when I replaced the truck assembly when upgrading to the Carvetight that some how I had a screw in too deep or in the wrong
spot and it made the head tilt ever so slightly in one spot that caused it to fail to read the board correctly. I had to slide the truck extremely slowly to detect it, without power.
Good luck and let us know.

[Digitalwoodshop]

Since the machine is using the board detector as a flashlight shining light out into the night... Do this at home.... Shine a flashlight into the air at night and you do not see any reflected light.... Then hold you hand out in front of the flashlight 3 feet away... You see the light on your hand... This is how the Board Detector works.... 2 IR LED's like your TV Remote Control shining light toward the board. Between the 2 LED's is a straw like tube and at the end a Photo Detector. It has a Narrow field of view like Horse Blinders.... With NO board in the machine the LED's shine light down into the gap between the belts and since there is nothing close to reflect back the light the Sensor reads ZERO.... Wave your hand under the detector and watch the readings and you will see the numbers get higher as MORE reflected light is seen. With that basic theory out of the way... WHAT can cause a Detection ERROR....?

1. A Weak Reading.... AS in a Dirty Lens, One broken off LED of the 2. To check the health of your board detector I started using a sheet of white copy paper as a Standard... Looking at this post and seeing numbers for plywood don't help me decide if the detector is working correctly. Yes, I know it is working but is a 96 really a good reading? That is where my White Copy Paper comes in... Everyone has White Copy Paper... So the Standard is 156 for White Copy Paper... DO you get a 156 for White Copy Paper with the paper on the board AND the head cranked down? Yes or NO?


SO lets go with the NO first... Say is is 96..... That could be a Dirty lens..... OR one of the LED's has snapped off... Likely due to vibration from using a QC long past it's failure and the BB Marks is letting the Bit Flop Around at High RPM... The High Frequency Vibration causes everything from letting dust work it's way down inside the board detector through the cable way on top of the machine... OR let Dust get into the Board Detector through Slight cracks in the backside cover flap. The Dust will settle inside the Board Sensor at the lowest point... The LENS.... OR snap off a LED or 2.... EARLY Board Detectors were not glued to the black holder, the LED's held them in place... Once the LED Copper Legs Snapped the board fell out once the back flap is cut off...


MY fix for the Board Sensor is to remove it from the machine and using a knife cut down the back flap removing it... When you do this either 1 of the LED's will fall off in your hand and you found the problem... OR you will see the dust covered lens and it is easy to pull out and clean... Blow out the dust then install the lens and I use masking tape to seal it back up catching the edge of the lens with the tape to hold the lens in place. Inspect the tube to the sensor in the middle between the LED's.


So that pretty much covers the NO part of the problem.... Now for the YES.... it reads 156 on white paper..... BUT you are still getting a Edge Detection ERROR....


First thing to do is move the head from one side of the board to the other in Y and watch the readings as the head stays over the white paper. IT should stay at 156 or close... IF as you move it and the FSC CABLE.... That 14 pin cable that brings the drive voltage to the Z Head Servo Motor and the power to the Board Detector and RETURNS the Reading of the Reflective Light that the Center Sensor is seeing.... So a INTERMITTENT reading could be a bad FSC Cable since it is just very thin copper FLAT wire covered with plastic on all sides. AS the Wire Flexes over and over.... Eventually the thin copper wire SNAPS.... Then you get Intermittent Readings as the ends of the copper FLAT wire touch then DON'T Touch... Usually as the cable Flexes.... THAT is why I call the FSC Cable a CONSUMABLE.... And to buy a few and have them in stock for a quick repair.


So that covers the YES 156 BUT Intermittent when the head is moving as in WHEN it is asked to detect the edge of the board....


So back to the backyard and shining the flashlight into the night sky.... You hold out your hand and you now see the reflective light...

Now start with 2 flashlights..... One 3 feet above you shining directly into your eyes... The 2nd shining UP into the night sky and you hold your hand 3 feet away and the 2nd flashlight shines on it.... Before without the 1st flashlight shining in your eyes you could see the hand.... BUT now with the flashlight shining into your eyes... YOU cannot see the reflective light... Your eyes are being overwhelmed by the light in your eyes....

SO with that little Theory.... Back to the CW.... Since the Board Detector relies on a CHANGE from DARK or NO Reflective Light to reflective light as the board edge is detected.... What IF.... The area along side of the board had a layer of sawdust and the Fluorescent Light caused the light to reflect off the sawdust and like a flashlight shining in your eyes.... You are blinded... And don't detect the edge.... OR you are now using a 1/2 inch board and the head is cranked down closer to that layer of sawdust near the brass roller.... STILL Blinded.... You don't detect the board edge as you are getting a bright light all the time.... no change from dark to light... No edge detect....


SO with ALL that THEORY.... Go back and do the tests.... White Paper, Move the head, Even measure the sawdust reflected at the brass roller with the fluorescent light on and OFF.... IF necessary.... CUT the back off of the board detector......


When you buy a board detector and a spare I always RECOMMEND you buy a BOARD DETECTOR CABLE.... It's a CHEAP 4 wire cable with 2 plugs... IF you had a Spare Board Detector AND a extra CABLE I would recommend you turn the power OFF and then UN PLUG the Board Detector Cable on TOP of the Y Truck and plug in the spare cable and a new board detector and turn the power back on and hold the detector in your hand and shine it on white paper and look a the numbers... THIS would test the FSC Cable as Good or BAD.... This is known in Troubleshooting as Half Splitting... Placing a Known GOOD Board Detector and Cable in the middle you then test the FSC Cable and back to the computer..... IF the Cable was bad OR the still installed Board Detector was bad you would get good readings now....


With TWO of my 4 machines, the first in 2007 still under warranty, when changing a board detector I pulled the cable DOWN and it slid along the SHARP Edge if the Y Truck while being PINCHED with the Cable slot plastic cover. The QC Vibration then caused the sharp edge of the Y Truck to CUT into the Insulation and in my case TWICE... Got a Intermittent Problem at first then a ZERO Board Detector Reading, AND the CLOSE COVER ERROR.... AND NO Z Encoder Data..... Of the 4 wires, I bet the 5 volts DC to power the Board Detector was the wire that had Cut insulation and Shorted the 5 volts to Ground... HENCE.... NO Board Detector, NO Close cover, NO Z Data.... That first machine got it's first Frequent Flier Miles with it's first of 2 trips to LHR for a Warranty Repair... The end was a Stripped Square Hole for the Flex using the hand ground Square Allen Wrench sold by LHR and MY NOT Hearting the QC HOT ENOUGH to loosen the FOREVER RED LOCKTITE. I STRIPED the Square Hole Round.... Hence more Frequent flayer Miles to TEXAS.... Ron's SQUARE DRIVE Socket made for a much better method... I did SNAP OFF my first 3 of 4 Ron's Square Tips trying to remove some FOREVER LOCKTITE and a QC.... FINALLY GOT IT... on the 4th... I blame ME for snapping the Case Hardened Tips I bought from RON... I DID NOT Heat the QC Hot Enought.... The tips were FINE... That is why I bought 4 the first time.... I expected trouble with the QC's.


OK... Enough.... go fix your machine and back to carving....


AL

[Digitalwoodshop]

So more pictures to further tell the story of my 2nd machine that got this failure from me pulling the cable down and pinching the wires... and not being under warranty it was time to learn.... YES... I was NEW once and didn't have a clue.... It was all hands on ON the JOB Training.... I had to learn or PAY.....

The rest is history.... I use Masking Tape to close up the Board Detector and the cable path.

The last 2 pictures I don't believe I have ever posted before... I had broken my last Ron's Square Drive tip and my Frustration Level was very high.... I had again not heated the QC and Forever Locktite enough and out of frustration I drilled a hole into the Truck and into the hollow Spindle Shaft and inserted a Allen Wrench.... In an attempt to lock the spindle to remove the bad QC.... Well, the Hollow Shaft Crumbled as did the outer Truck Housing.... Call that a "FAIL".... Tossed it in the parts drawer and installed a whole new truck.... I then started ordering the Trucks without a QC... I had switched to Ron's ROCK shortly after that.... And the Carving Gods were again HAPPY....

AL

[55president]

Al -- Thank you for your detailed message.


OK -- The white paper test -- 1st test -Head cranked down with the white copy paper
on the plywood. On the the very edge along side the board it reads 0, moving the truck along the y-axis it starts to read the paper at 152 as I move the y-truck over the paper the reading does fluctuate down as low as 135, but does not drop out to anything less than this.


2nd test --same as before, but it starts to read at 142 fluctuating up to 152 and
down as low as 135 but no lower.


3rd test -- with a florescent trouble light -- similar readings to the first 2 tests.


4th test-- using card stock (thicker white paper) -- readings start at 155 and fluctuate down to 145.
But I notice that at any one point when I stop the the number will fluctuate
by 1. If it starts to read at 155 and I keep it at spot it will read 155-154-155-154
etc......I hadn’t noticed this on the earlier tests but It might have been doing that
all along. Not sure what is normal.




Holding a the white paper freely up to the sensor I get a reading as high as 158.


Furthur info: The FSC cable is new ,but could be defective I guess.
I ordered 2 ( as per your advice in earlier read posts) and replaced it at my 490hr overhaul, cut a few blocks - maybe another 10 hrs when I tore the y-belt trying to cut some new material I was trying out. At that point I ordered a new belt , and made a jumper for the cut motor as well as the other cover switch.Because I was tired of taking the cover off when I'm trouble shooting. Also at this time I also took off the front roller, trimmed 1/4 off the belt and drilled 2 holes underneath the aluminum roller base to move the plastic retainer over the 1/4 to keep the trimmed rubber belt out of the brass roller. I was making measurement tests when this all went south. (by the way for the first time ever my machine measures pretty accurately: within .07 inch after 8 tests,Is this good?, I’m not sure,I’ve never been able to measure accurately). This is when I started getting the edge detection error. I figured when I replaced the y-belt, I took off the z-motor pack and I might have pinched the board sensor wire. So it all came apart again. I did see the wire bare in one tiny spot so I put a small piece of electrical tape around the wire at this spot. Again Edge detection error. Took it all apart , Changed to my old board sensor wire, and old board sensor which both were still good and for good measure I replaced the FSC cable again as I had a spare.Same story. I guess I could change the FSC cable again.I'm just hpoing I didn't really screw something up badly......

[Digitalwoodshop]

WOW.... I am Impressed as you did all the great things...

Have you replaced the board detector? Could be vibration as the machine auto moves that is causing the edge detect error....

IF you are doing something and get a fault and reset the board you must cycle the power or I have even gotten edge detect error.

I would like you to follow the instruction to format the memory cards and re install the latest or the version you want. Could be a problem with the card? IF you have 2 cards, only do one.... IF all goes back to normal then try the old card and see if it comes back..

Thanks for explaining how the Y Belt broke.... Testing new harder material....

Good Luck,

AL

[bergerud]

Could the edge detection error have to do with the y drive or y encoder instead of the board sensor? There could there be excessive backlash or some other problem with is connected to the breaking of the belt?

[Digitalwoodshop]

WOW.... That is a Excellent point.... Y Backlash....

Since a Y Belt was broken, a inspection of the Y Gearbox is requested... You may very well find the bearings bad... causing Encoder Problems....

[55president]

I have not checked the y-belt gearbox -- but I replaced bearing AND gearbox at the 490 hr overhaul. (10Hrs ago),
so I'm pretty sure it's not slop in the gears.
I also looked at the encoder to see if there was dust ( there was none0 & I resealed the back of the motor too
after checking. In fact my machine is particularly hard on the y gearbox so I'm always checking
these bearing on a regular basis.

[55president]

I do have a new Board detector so I'll try that tomorrow. I also forgot to say I upgraded to the latest designer software
and flashed the machine with the latest firmware ,both after I was getting the edge detection error.

[55president]

I always cycle down between the steps I try ,I have not tried with the other card too ,so another thing to do.

[bergerud]

Maybe the encoder disc is loose. Sounds like you have checked everything else!

[badbert]

At this point I would suggest thinking outside the box. My machine a close cover error that was caused by the 14 pin cable being crooked. How are the two related? They use the same power, along with all of the other sensors. The close cover was just the first error, None of the sensors worked. Maybe something else is causing the problem, and this is just the first error... Does this make sense? Try running sensor tests, and make sure all of the sensors are working? If you over oiled the flexshaft where does the oil go? Could that be causing the problem? Just trying to add to the think tank...

[55president]

I bought a spare flex shaft and soak the spare in a tube made from
small hot water tubing slightly larger and longer than the Flexshaft. It has a
cap glued on the bottom and the top cap remains unglued to keep out
the dust in the shop.
so When I'm ready to change out I'll pull the spare, wipe it off and
let it set for a while b4 I put it back in so I don't over oil anything.
I'll send a pic of the set up.

[55president]

What would be the fix for a loose encoder is there a tightening screw?

[liquidguitars]

also you could check the sandpaper belts for a rolled up end, hard to see but this could be making the table a little out. Al and i rip a 0.50" off the end of the belt on the keyboard sides.

and also check the power conections under the unit and off and on switch for a short....

[bergerud]

What would be the fix for a loose encoder is there a tightening screw?

I am not sure since I have not had a loose disc myself. I think others have used a drop of glue.

[liquidguitars]

2 carvings later -- I tore a y belt. I think my material was too tough and my
bit might have been a little dull. (continued...)

Do you think you damaged the y pack... if you tore a belt that's hard core, however you would get y stall error.

[55president]

Liquidguitars -- I use the rubber belts and I've modified the belt closest to the brass roller to not touch it.
I trimmed 1/4 off the belt and moved the plastic retainer on the bottom of the belt holder over to keep it
out of contact contact with the brass roller. I had to drill 2 new holes in the aluminum to do this.

I will check the A/C power too Thanks....

[liquidguitars]

can you lift the board up a little when loaded? I like to have a some spring in the board.. to do this I load the board off center to get a little fulcrum to lift.

[55president]

LG-- If the rollers are down -on 1/2 plywood there is really no
spring in-between the rollers. Not sure what you mean.Please
clarify. Thanks FXP

[liquidguitars]

Try this, place a .75 thick board in the cw but do not center the board but place the board off center " front of board under Board sensor". Make sure you do not over tighten the table to alow a little spring movement. test by lifting up on the long end you should have a little spring in the roller and not "rock soild" Also some plywood is under .50" so do your test on a thick peice of wood if you can.

[55president]

Oh -- you mean testing the rollers to see if they are stuck and not compressable?

[liquidguitars]

no just don't over compress them as the board sensor can give a error if two close to the wood board, did you try 3/4 thick wood by chance?

[55president]

I think I fixed it -- I will know Tomorrow when I try and carve a small block. I took it all aprt -- put a
new board sensor in, checked the y gear box - which was fine -- no encoder problem,
as i was putting it all back together. I noticed the belt retainter ( on the back of the y-truck)
was sticking out too much causing some contact around where I was getting the error-
a small tap with a ball peen hammer put it back in place. I didn't get the error , but I only tested
once. It seem if was the contact from the retainer there would be a Y-Stall error, or am I wrong?

FXP

[Digitalwoodshop]

Good JOB !!!!

AL