No onion skin?

[chunkstyle]

O.K., I realize I'm clogging up the forums with my problems so I hope I don't get on anyone's nerves. In trying to cope with the spindle control problem I had run off a part file that I've been tinkering with all day. It's just two small parts, a cabinet side and a cabinet top. Top gets blind dadoed into the side. I just changed some construction settings and re-saved the file. Had the fit tolerance of the dado changed from 1/80" (gave good results off the Thermwood) to 1/128". Previous efforts were giving me a lot of slop. About 1/32" of slop in the fit. 3/8" bit was used and it mic'ed out .3746.

Trying to figure out where the extra slop was coming from I re-ran the parts and noticed that the onion skin pass was skipped. Checked my settings in the link and their all there. Don't know why it skipped it.

One last problem that has been jacking me up all day. I put a mz1.5 and a J2 90,0 in the footer. When the parts are cut it goes down to 97+,47+ and proceeds to plunge! Into the spoilboard if I let it. Has done this all day. I've rechecked the footer and all appears correct with the notation.

Any insights would be much appreciated.

Thanks,
Tim

[ken_rychlik]

For the footer problem, you could write a custom file for that.

Take C5 for instance and create a small file that does what you want.

Then just put C5 in the footer where you want it and it will run that file.

Kenneth

[Gary Campbell]

Tim...
Lets make sure you have your settings for the onionskin correct.

Go to SETTINGS > SKIN PASS

Set your Skin Thickness to a number > .035 to start

Set your Maximum Part Size to 3500"^2 to start
This means every part UNDER 3500 sq in will receive a skin pass

(under size to leave skin on)
Set your Max part Size to 30"^2
This means parts under 30 sq in will not have the onion skin cut during profile last pass

Set your Max part Width to 2.5"^2
This means any part under 2.5 in wide will also not have the last skin cut


using these settings you should be able to cut virtually any part with out movement, even with minimal to moderate vacuum

If you post the exact entries made in the footer box, I can look and see if there is a problem

[dlcw]

For the dado slop problem, make sure you've mic'd the material to exact thickness. This has burned me a couple of times. I've got my depth clearance set to .0156" and side clearance to .007". These settings have served me well.

Yesterday I was cutting a couple of small cabinets to show potential customers. For some reason the dados were sloppy. The previous 23 cabinet kitchen I did every dado was perfect. It turned out the the 3/4" prefinished maple I used on the sample cabinets was .71". On the kitchen I just finished the plywood was .714" enough to make perfectly fitting dados go sloppy.

[chunkstyle]

Kenneth, great idea! I will do just that.
Gary, how the heck do you know that? I musta missed it in the manual I couldn't figure out what those values mean. Why do I think you know san skrit:) I will compare what I have to what you've suggested and make the changes. Again Thanks.
Don, This is a topic that I have ben battling with for the amount of time I've been using e-cabs and outsourcing the panel processing. As I have ben made to understand it by the Thermwood folks, the dado width is made and impacted by the diameter of the tool cutting the dado and the distance of the end of the tool off the table when cutting the tenon thickness. This makes sense to me since the only point of reference the machine needs to have is how far off the table the bit needs to come up to determine tenon thickness. The thickness of the material would have no impact on this part of the joint (fit tolerance). the only way material variation would have an effect is that your tenon may bottom out in the mortise and the shoulder of the tenon would not abut the mortised members face. Our experience (me and the Thermwood shops') have sen a mixed result with our strategies developed around this notion.
I am not going to argue about what works and go with your suggestions. If it works for you I sure as heck hope that it will work for me. Thanks for helping me get over this hump.

Very much appreciated guys.

Thanks,
Tim

[Gary Campbell]

Tim...
I disagree with yor interpretation of the tenon thickness, for the following reason: The design material thickness is output as a variable called "&ZSHIFT" and the bit zero is shifted up from the spoilboard surface by that amount.

Using the example of designing with .750 material and a .375 depth dado, the bit is cutting at a programmed depth of -.375 and leaving .375 tenon thickness.

Therefore if your material is entered as thinner than actual, say by .040, then your tenon will be .040 undersize. Same happens for pocket dados, but the extra depth doesnt affect the fit. However, the location of the pocket dado will be located .040 too close to the top or bottom of the side panels.

I can assure you that accuracy pays big dividends. Audit your parts and find if or how they differ from the nest diagrams and fine tune your settings.

[chunkstyle]

Hmmm......

I'm trying to really get at the nut of this so bear with me.

Let's say that the material, in your example, was .040" thinner. The tennon will get cut using the .75" material thickness as a reference and come down to the table the desired % of the material thickness, correct? No matter how it does the math it has to finally make that determination that it must be "X" distance from the table surface. So no matter what the material thickness is doing, the bit is staying off of the table surface by our specified % of our incorrect .75" material thickness. So far so good?

All tenon's will be cut, if the tenon thickness was set at 50%, with the bit traveling at .375" above the table. Now we have .375" thick tenons on our parts, however the shoulder widths of the tenons are what will be off if the material thickness is rising and falling from the table surface, correct?

Moving on to the mortise, doesn't the program say "O.K., we know that the tenon thickness is 3/8" so we need a 3/8" mortise and we also need to add the little extra that he wants with the fit tolerance". Correct? (this extra addition of fit tolerance calculation might be getting added in the tenon cutting operation but it wouldn't change the fact that the bit would still come down a predetermined height off of the table surface). Now the bit comes down into the panel and routes out a mortise where the width has been determined and the depth will be calculate as so far off the table + depth clearance. Again, the material thickness will not have an effect, in my hypothesis anyhow, to what the material thickness is doing for the fit of the joint. If the material is less than .75" and the mortise is getting cut, the width of the mortise will be unaffected by this discrepancy but the depth of the mortise will get deeper and shallower since all the material deviation is on the side opposite the table side of the sheet.
So in my thinking, the tenon thickness should be constant regardless of the material thickness deviation (assuming our table is dead flat). It's the tenon shoulders that will vary as a result of the material thickness being inconsistent.
The mortise widths will be unaffected by material thickness variation. The depths will be affected by material thickness variation.

I'm not saying that this is right and would welcome any correction to how I'm thinking about this. I'm posting my logic for just that reason as it has been an ongoing issue and someone could get my mind right about this.

Again, thanks for all the help and advice. It's very much appreciated.
Tim

[chunkstyle]

Gary,

The Sheet Footer had this:

MZ 1.5
J2 90

The stopped position wasn't close to this and there was the added excitement of a plunge. Weird.

[ken_rychlik]

It may have something to do with ecabs running off of g code and your other commands running in sb mode.

Also having the commas and decimal points in there correct location is critical.

J2 90 won't work. You are telling it to jog both the x and y and not giving number for each with the commas.

J2,90,0 will

Kenneth

[chunkstyle]

Thanks Kenneth,

I'll try that.
Still no onion skin after using Gary's skin settings. Have Thermwood looking at the export file now.
I would be keen to know how the new construction settings that Don suggested will work. If only the onion skin pass would work!

Thanks for the help,
Tim

[Gary Campbell]

Tim...
Did you ever find what caused the no skin problem?

[chunkstyle]

Hi Gary,

No, Thermwood had a look at the code generated from the link and we went over it together. There is proper depth in the tool paths to create the onion skin. Why it is not performing the last tool path for the onion skin removal is anyone's guess.

I'll have someone looking at it tomorrow and maybe..., just maybe...., we'll get this straightened out. I'll report back with the results.

Tim

[Gary Campbell]

Tim...
Let me know how it works out. No second pass is something I have not seen or heard of yet.

RE your post of 07-30-2010, 08:54 AM :

You need to be aware that once zeroed to bed, the material "design" thickness is output as a variable (&ZSHIFT), which in effect, rezeroes the tool (only when cutting) to the "design thickness" above the bed. This also adds ability to cut multiple thickness' in a job.

If you add this to your scenario above,(design thk = .75, actual = .71 tenon = 50%, 3/8 deep) I propose the following:
Tenons will be proper thickness, as you say.
Tenon dado will be .040 (less clearance) too shallow resulting in gaps between parts.
Tops and decks will be misaligned from ends by .040 (short)

Lets do the other way. (design thk = .71, actual = .75 tenon = 50%, 3/8 deep)
In this scenario you would get the following:
Tenons will be proper thickness.
Tenon dado will be .040 (less clearance) too deep.
Tops and decks will be misaligned from ends by .040 (overhang)

Now lets throw in one more scenario: Zero is to top of material and the zero location, either due to warp, thickness variation or low vacuum caused by used spoilboard, results in error of .040 too high:

Tenons will be .040 too fat.
Tenon dado will be .040 too shallow.
Tops and decks will not fit into the ends and tenons will need to be trimmed for both length and thickness.

How about .040 too low at the zero point:

Tenons will be .040 too thin and will rattle.
Tenon dado will be .040 too deep.
Tops and decks will be moveable by .040 (overhang) to flush and hard to assemble.

I realize that .040 is an extreme example, but half that is common and .020 too tight or loose is just as undesireable to deal with as .040.

Here is what I recommend for those using files requiring zero to top of material:

Get a very accurate material thickness by averaging from other than edge locations. (after some test cuts)

That first you zero to bed, MZ to design thickness above bed, Zero Z.
do NOT do a manual toolchange and rezero during file.... unless you:

Add an off spoilboard zero block and add code to MZ, &zshift and rezero for bit changes during the SB Link file.

Cut in batches of same material thickness. Repeat zero as above for new material.

Call your Congressman (or ShopBot cuz it makes more sense) and see if these files can be changed to zero to bed.

[chunkstyle]

Hey Gary,

I agree to everything that you are saying about the material thickness as it relates to part alignment.
I've had issues with fit tolerances of tenons to dado's. This was an are where there was some confusion about material thickness having an affect on fit tolerance of the tenon going into the mortise.

I'm just gunna copy and past what I sent Thermwood about the onion skin.

Here tis:

The onion skin not working was a bit of a red herring and here's why:

First attempt to cut an e-cab twd file using the link. Parts were 1-cabinet
side, 1-cabinet top. Blind dadoe joinery. Other parts were omitted from
cabinet. Saved as batch job.

Ran parts and the machine cut them out in three sequential passes. The last
outline pass was the onion skin pass. First two passes were conventional,
last pass was climb cut.

Next was the off fall outline cut. Only I didn't realize that the code had
an outline cut tool pathed. This was mistaken by me as being an incorrect jog
to the location that I specified in the footer (x 90, y 0).

The spindle was moving to an odd location (something like x 94 and y 47).
Thinking back on it I realize that it was making a screwy toolpath for the
off fall cleanup. It would then travel down to the end of the table bed and
actually go into the -X!

from there it would travel down the Y until it was located about an inch
into the area where my parts were located and proceed back up the X and into
the parts.

Made another turn and traveled down the Y axis untill it hit a proximity
switch, having strayed into a negative position on the Y axis.

We ran that file a coupla times and got the same result. Very weird. Checked
the code and the machine was doing what was written from link.

Ryan, from Shopbot, was up messing with it yesterday. Found that if we
changed the bed size from 49x97 to 48x96 the problem went away. We are
leaving it this way and there doesn't seem to be an issue (so far?) with
nesting material that's been set up as 49x97.

Why the first run file went in the sequence it did and the subsequent files
(changed construction settings/ fit tolerances on the original and resaved)
I have no idea. Same with the screwy off fall tool pathing and how it
related to bed size.

Weird......

[Gary Campbell]

Tim....
All of the items you described seem predictable to me. That is to say, that I "know a friend" who had most of those same problems with certian combinations of erroneous settings. Is that why I have the table set to 48x96? :confused:

The obvious.... can you set your max penetration to .8+, your outline cut to climb, and check thru cuts last and run opposite direction in the perimeter cuts section. Then set your move speed for the bit and rpm to get adequate chipload?

Or can I assume, that since Ryan is there, its working now?

Has the49x97 table size "returns erroneous passes" been reported to Thermwood?

[chunkstyle]

Hey Gary,

Yes, I can set the max depth as you described. I've been running conventional on the outline and max depth at .375 with an onion skin thickness at .03". Are you saying set the outline passes in climb and onion to conventional? My thinking was conventional on outline and slightly oversized and climb on the onion and bring it to final dimension for outlines. I'm open to suggestions.
I set the minimum off fall to something like a gazillion inches square. That way it won't bother me with the off fall tool path again.
Seem's like you have to figure out what to throw out or turn off and set manually to get this to run.
Spent this afternoon cutting and trying to get my blind dado fit tolerances tightened up. I really hate ill fitting blind dados. I'm going to focus on just getting it to cut correct and try and ignore all the rest of what's weird or doesn't run right software wise or otherwise. I've got about a dozen sheets to get cut so I have to focus on whatever it will take to get that done. Table saw and biscuit joints, who knows.
Yes, I did write to Thermwood and told them about the issues with respect to table size and the odd tool pathing for the off fall. Maybe they know about it already.

Thanks for your interest and advice, It's appreciated.
Tim

[Gary Campbell]

Todd...
I've been running conventional on the outline and max depth at .375 with an onion skin thickness at .03". Are you saying set the outline passes in climb and onion to conventional? My thinking was conventional on outline and slightly oversized and climb on the onion and bring it to final dimension for outlines. I'm open to suggestions.

If you have your skin pass settings so that Skin pass is enabled (as above) then the perimeter cut direction is for the first pass. If run opposite direction is enabled the second pass will run the opposite direction of the first. Your max depth (per pass) setting will determine how many passes it takes for the "first pass". If set to conventional, your parts will be undersize with a watermark at the skin. (my guess) My advice is to run your first pass in the climb direction. This will be a deep pass and deflection will be away from the part. The conventional last (skin) pass cuts a very small amout (the deflection and skin) and cuts very accurately, as it is under lite loading.

I set the minimum off fall to something like a gazillion inches square. That way it won't bother me with the off fall tool path again.

Just check "No Off Fall Cuts" . No cuts will be made and nothing will be stored in Off fall database.

Ive Just finished up the onscreen recording of the Training Video. Sorry it wasnt ready for you last week! :)

Compare the cut widths and depths against those shown in the eCabs nest diagrams, post here and I will let you know what to reset. Usually, if dados are set to conventional and you are using normal feeds, then the fit problem is most likely depth related. Good Luck

[zeykr]

Tim,

I have also reported the misplaced off fall cuts. In my case it happened when cutting using a previous piece of off fall. My machine is a 6x12, but I have the area in sblink defined as a 4x8.

[chunkstyle]

Zeykr,

Well it's good to know that it's not an isolated problem. Hopefully the fix will be made from Thermwood.

Gary,
I'm getting nice results using the two pass climb and onion skin conventional. 15k on the .375" mortise compression and 2"/second. Minimal chatter and the perimeter dimensions are darn good! Making progress.
Do you set your pocket/ dado link setting as a conventional or climb cut?
Still working out the slop here. Conventional pass at a fit tolerance of .0015" gives me a looser than I like fit. Climb cut gives a too tight fit. Too tight a fit is what I would expect for a .0015" fit tolerance so I'm thinking of leaving it in Climb and adjusting my fit tolerance in my construction settings.
Gaining on it, I think.

[Gary Campbell]

Tim...
I think your rpm is too high for that move speed. try 12500 your bits will last much longer.


Do you set your pocket/ dado link setting as a conventional or climb cut?

I set mine conventional

Still working out the slop here. Conventional pass at a fit tolerance of .0015" gives me a looser than I like fit. Climb cut gives a too tight fit. Too tight a fit is what I would expect for a .0015" fit tolerance so I'm thinking of leaving it in Climb and adjusting my fit tolerance in my construction settings.

Do you really mean one and one half thousandths? My guess is that at those very conservative speeds you will not have any slop. If you do it is mechanical and your machines motion adjustments need to be tightened. I predict that as we audit these geometries that you have a Z cutting depth problem. This may be a result of erroneous thickness or initial zero. Heres how you find out:

After parts are cut, take a couple of your sides and measure the width and depth of the blind dadoes and the distance from the end of the panel. Take a few measurements and write them down.

Now open that cabinet in the eCabs editor and click on the nest button. DBL click on the sheet that the parts are on ans then again on the individual part. When the single part is displayed you can use the numbers of the cut geometry to see the exact cut sizes.

First: check the width of the dado against your measurements.
Second: Check the depth. These are the easy ones and if the width is right on, I am sure the depth will be off. Now check that the distance from the dado rectangle to the end of panel matches your measurement.

Now take your material design thickness and multiply by the dado %. That is how thick your tenon should be. You can also look on the nest diagram for a deck or top and it will tell you how deep the tenon dado cut is. Measure this from the top of material to the top of tenon. While you are at it, measure the protrusion of the tenon from the part. This should match the dado depth.

Report back, you can be good to go in the AM

[chunkstyle]

Hey Gary,

Wow, Didn't realize the nest gave the information it does, with respect to dado's and tenons.

Here's what I'm getting from the nest for the cabinet sides dado:

From edge: 0.32
Results from cutting: 0.3215

Width: 0.404
Results from cutting: 0.396

Depth: 0.4063
Results from cutting: 0.430

Tenon Dimensions:

Width: 0.3751
Results from cutting: 0.3685 (I take this measurement to be the length of the tenon past the edge of the panel?)

Depth: 0.324
Results from cutting: 0.402" (This is what I got measuring the tenon thickness)

Tool used: 3/8" mortise compression spiral.

Outline cut: 2"/second, 15000 rpm, climb cut, max depth .375.

Onion: thickness .03", 1.33"/second, conventional.

Pocket/ dado : 3/8" M.C. spiral, climb cut.

Material thickness: 0.72"

Dado depth: 0.375"
Dado depth clearance: 0.0313"
Dado fit tolerance: .008"

Tenon Thickness: 55%

One thing I'm not to sure of is where the fit and depth clearance value's get factored in.

My guess is that the fit is getting added to the width of the dado.
Depth clearance is added to the depth of the dado.


O.K. that's what I did for the 9th try on these parts.

I was finding that cutting the dado's conventional gave me a loose fit no matter how low I went on the fit tolerance construction setting. 0.0015" was the lowest I tried. Went to a climb cut on the dado at that fit tolerance and it was too tight, which seemed right, so I went with climb cutting dado's.

Thanks for the help. Startin to gain on it some. Feels good!

Tim

[Gary Campbell]

Tim...
You are right. Almost there. Can you take a couple more measurements?

Measure a half dozen places around the parts for thickness.
Measure the depth of the "bit tracks" in a few places around the table. See if any places seem different than others. Post your cut thru.

One last test: on scrap @2"/sec 3/8 deep 12500 rpm AND 15K rpm do a 1 foot cut and measure the slot. Also, do you have the dado cuts from conventional... give numbers of width and depth.

What/how/when is your zzero method?

[chunkstyle]

Hi Gary,

In opposite order:

Zero plate is located remotely off the table bed surface. The tool change file goes down over the remote's location and zero's itself out. From there it continues file cutting. However, just to cloud the waters a bit, while re-running these test part files I would move the spindle over to a fresh patch of the plywood sheet and zero out the Y axis. This would make it necessary to zero to the table bed using the portable zero plate, since zeroing out the Y axis from real 0,0 threw out the location of the remote zero plate.

Part thickness is about .725 on avg. anywhere from .722" to .724". Pretty good cores for D3.

Parts were nested using the Y axis in the link. Each file was run after zeroing out with portable zero plate on the table surface that was exposed from the previous run file's part holes.

That said I found the tracks deeper at the outer edge of the X,Y axis. Anywhere from .026-.029".
Tracks further into the table showed a depth of anywhere from .0135"-.019". Mostly around .016" on avg.

The last dado that was cut using a conventional cut had a width of .4175.
Tenons were .406" avg. thickness. This gave me a difference of .0115". Fit tolerance was set at .0015".

I can see how the tables able to introduce inaccuracies if it has moved out of dead flat. This would change the tenon thickness and effect the fit tolerance of the dado and tenon. Wouldn't take much. My sweet spot on the Thermwood was about .0125" on the fit tolerance for ply.

I looked at the overall tightness machine by chucking a bit in the spindle and loweing it down to the bed as close as possible. Grabbed the bearing housing on the spindle and tried moving it in the X & Y. Found little play in the Y but noticed some slop in the X. Heard a slight chunk chunk sound. Traced it back to lash between the rack and pinion gears on the X rails.
Loosened up the motor mounts and tightened the motors up using a clamp. Tightened just enough to get rid of the lash. Retested the spindle play. Much improved. All this was done before test cutting was performed.

I will do the dado cuts as you asked and report back.

Thanks for the help and advice. It's appreciated.

Tim

[jimboy]

A little off the subject, but are located in Negaunee, Michigan?
I use to live in Ishpeming.

[Gary Campbell]

James...
I live in Marquette, the shop is at the old airport, and yes with a Negaunee address. Man! there are Yoopers all over the place!