An unexpected offset in y

[Chuck Keysor]

Hello Host of Experts.....

Using Aspire, I have created a simple Shopbot cut file to make a piece for a railing. In X & Y I carefully zeroed my bit on the lower left corner of my stock, just as it is in my Aspire file. Then I zeroed in Z in the middle of my stock. I then loaded my cut file and with the first cut, it was clear that somehow the start reference point was correct in X, but wrong in Y. It was wrong by 2.1inches! I have no idea why I have this error......

Details: I turned off the machine, and sent the spindle back to 0,0, and the tip of my bit parked exactly over the lower left corner of my stock. So I was zeroed in X & Y as I had expected, and I will call that my actual zero point. But I measured the difference between where my first cut was made and my actual zero, and there was an error/offset of 2.1 inches! The entire cut file was translated by 2.1 inches in the y direction, as my spindle was "air cutting" beyond the far "y" end of my piece.

I went and reloaded the cut file, and looked at the set-up panels I always breeze through, and carefully looked at each entry. I had recalled there was an offset parameter, so I checked that, and it was in fact set for no offset.

Why am I getting this big error (positive 2.1 inches in the Y direction compared to my actual zero point?

For the moment, I moved my spindle to (0,-2.1) then re-zeroed to that point. Now the cutting is happening as it should. But now my zero is 2.1 inches below the bottom left reference point of my stock, and I have no clue what has happened.

Thanks for any suggestions you can offer. Chuck

PS: I'd attach the Aspire file, but it is huge, 10meg. And mysteriously, I have Aspire files for the same part, only of a different length, and those files are tiny in comparison.

[coryatjohn]

What about your toolpath file? Have you looked at it?

[Chuck Keysor]

Hello John. I did glance at it, but don't know :o enough to see anything. The file is about 7megabytes, too large to attach. But, here is the top of the file, where the answer should be. Thanks, Chuck

'Top rail cap short roughing 3-28-13
'File created: Thursday, March 28, 2013 - 08:53 PM
'SHOPBOT FILE IN INCHES
IF %(25)=1 THEN GOTO UNIT_ERROR 'check to see software is set to standard
C#,90 'Lookup offset values
'
'Turning router ON
SO,1,1
PAUSE 2
'
'
'Toolpath Name = 3D Roughing 3
'Tool Name = Ball Nose (0.75 inch)
MS,6.0,0.5
JZ,1.000000
J2,0.000000,0.000000
J3,5.100000,2.225579,0.250000
M3,5.100000,2.225579,-0.125000
M3,3.071414,2.225579,-0.125000
M3,3.071414,2.224576,-0.125000

[Brady Watson]

My guess would be that your part is sitting at 2" up in Y or you are unknowingly using 'origin offset' on the project setup screen when you first open it up.

-B

[Chuck Keysor]

Thanks Brady! I did in fact have the origin offset by 2.1 inches in the Aspire material definition page..... Duh!

It seems as though I messed this up as I was trying to add a safety zone at the ends of my part so that my collet nut would not crash into the vertical ends of my railing.

The part I am cutting is 2.1 inches thick, and my 3/4" core box bit is only two inches long. So with cutting into my spoil board, my collet nut would hit the vertical ends of my railing stock. I stretched out my 3d model (which I imported from Rhino), and thought I had stretched my stock model by two inches. But instead I moved it and didn't notice since it was under my part model! Duh!

Thanks again, Chuck

PS: This is the first project where I have designed my parts in Rhino, and then imported them into Aspire. This part I had problems with was the first such part I had cut. It was quite amazing to think I could design a 3d model of a porch railing system in Rhino, and then import those same parts as OBJ files into Aspire, and cut them on my Shopbot! Wow!

[coryatjohn]

Chuck,

You've completed (or nearly!) the path I have set for myself. I have Rhino already and should get Aspire next month. I also have a NextEngine 3D scanner. Cool stuff!

Great going and I hope you're wildly successful!

[Chuck Keysor]

Thanks John. This is a maddening process though, designing in Rhino, and then working in Aspire, then cutting on the Bot. The single part I made today, before ruining it an hour ago, took me maybe an entire day, with most of it spent at the computer. I could have made it by hand in an hour. (See the model from Aspire of what I want to make in the first attachment.) Well of course I wanted to make the Rhino to Aspire to Shopbot test on something simple.

The last attachment shows the porch railing that I designed in Rhino. In the lower part, you can see the copies of some railing parts I dragged off to export as OBJ files into Aspire. Note the corner posts are blank. I hit so many problems with trying to model quatrafoils and arches in Rhino, that I left the details off of the corner posts. It is a bit maddening, as I have a simpler CAD program, called Design CAD. For simple drawing and trimming of lines, Design CAD is so much easier to run and intuitively obvious than Rhino. But the Rhino does the great 3D modeling on the bigger scale. So I have had to kick the can down the road on filling out the details of my corner posts. So far, it seems easier to do things like quatrafoils in Aspire, so I may do that, and export those back into Rhino, which seems a bit unfortunate, as what about the poor Rhino users who don't have Aspire? Rhino should be top notch for these types of things. (Specifically, doing boolean operations with arches seems to crash and burn on me, and I spend so much time trying to find alternate work arounds, but get nowhere...:mad:)

But back to cutting out the parts.... watching the part get cut out, even though I tried every cutting strategy that Aspire allowed, there is so much motion of the router that simply makes no sense to me. It is hard to explain, but it would seem as though simple motions going the length of the part, getting deeper with each pass would have been very simple and much faster. Instead the spindle goes all over, this way and that way, and then does weird cross cuts at each end of the piece. It would be nice to be able to tell this to cut in the way I would think logical, instead of having to choose from several options, most of which had various obvious problems in the simulation, making them so they were not even choices.

I am more concerned about the cross cutting of the end. The cross-cutting is shown in the roughing simulation in the second attachment. I have one rail system (see the entire porch drawing) that is almost 16 feet long. Unless I can use tiling somehow, I don't see how I can cut my 16 foot long parts on my 12 foot long Shopbot. As I can control Aspire to create cut files, it wants to cross cut the ends of my piece, which I can't allow. If I could just make it cut the top and sides, and not the ends, I could slide the entire piece of stock down in the x direction to cut the entire length as one piece. Well, I am rambling. I guess I like to talk to someone, instead of sitting and scratching my head in front of my computer, wondering if I am stupid, or if these programs are made by people who are too smart that they don't realize that ordinary people must be able to figure them out.......

Thanks John...... Chuck

[adrianm]

What sort of toolpath are you seeing that behaviour on?

If it's one of the 2D ones I'd look at the vectors you're using and make sure they're clean.

[Brady Watson]

Chuck,
What you are describing is 'par for the course'. Each of the programs you mention excel at certain things and are a challenge in others. You have to keep in mind that none of us were born knowing how to work in these programs fluently, so it DOES take some work and practice! Even though I have been using these types of programs for years, I still make time to mess about and try new things. I often discover things I never knew or find faster more efficient ways of doing certain things. It's no different than when I was developing software - you have to constantly be learning and trying new things.

Rhino is a very powerful program for 3D visualization of assemblies and complex 3D model creation. There have been times where I have really hammered on a model in ArtCAM or Aspire for hours - not getting the shape exactly the way I needed it to be (usually precise mathematical 3D parts to customer's specs) - only to spend 15min in Rhino to create the model to perfection. Rhino's drawing tools (just drawing 2D stuff) leave a lot to be desired for me coming from the easy fluid way of drawing designs in Aspire for instance. I most often create all my 2D work in Aspire and just export it as a DXF into Rhino - then use those vectors for sweeps and cross sections etc. It is much easier that way.

Rhino in my experience is pretty clunky when it comes to artistic or free-flowing 3D reliefs, such as those you'd make in Aspire. You could probably get a lot closer if you invested in T-Splines, but it still seems like you'd be 'forcing' Rhino to do something that comes so easily in Aspire. It is much easier to generate these types of parts in Aspire & export them as STL to bring them into Rhino as 'appliques' - Here's a little tip for you - There is absolutely NO need to 'weld' or boolean join your 3D parts before toolpathing. Just get them arranged how you want them to look, select them all and bring them into Aspire like that for toolpathing. You can select a 3D model that is comprised of 10 separate meshes or polysurfaces, select them all, and then export - they will be in the same orientation when you pull them back into Aspire. Don't feel bad...Boolean weld blows up for me on most things too!

In terms of indexing your part - Pull your 3D machining boundary vector in at the ends so that it is IMPOSSIBLE for the tool to go 'low'. If you need a 16' molding, then you should be creating your model at 16'1" or 16'2" or so. This will keep the tool from 'falling off the ends' and give you a little extra meat for trimming with the saw to get nice clean ends. Just make sure your machining vector is shorter in X than the model itself, just enough to keep the tool from falling off the end. I would set your 16' material up so that you machine say 8'1" or so at a time. Toolpath it all - roughing and finishing for that 8' section. Run both toolpaths - then slide your piece down to the left (-x) 8'1" or so...then just run the same TP all over again. If you have good registration, your transition area should be fairly seamless.

Don't be fooled! This stuff takes real work to pull off. Some just make it look easy...

-B

[Chuck Keysor]

John, concerning Rhino, I messed around for a week with Rhino, trying to use their free bits of training and was able to make it more or less do basic things I needed. But this program struck me as being so non-obvious to my brain (not all software is that bad for me!) that I purchased the "Rhino 5 Training Bundle" by Infinite Skills. I think it was $150, and worth every penny of it.

I look forward to you your posting of your first Rhino designed, Aspire toolpathed, Shopbot cut part! Chuck

[coryatjohn]

Chuck,

I found Rhino a bit obtuse as well. I bought the training bundle too but haven't had a chance to try it yet. I'm glad you found it useful. Hopefully, it will be to me also.

[Chuck Keysor]

Brady, I can't thank you enough! Your insight as to the nature of how Aspire and Rhino can best be used together is invaluable.

I spent so long, maybe a day, messing with drawing Gothic arches, and the little associated quasi triangular details called spandrels that go at both sides of the arch tops, and trying to make a solid model out of them in Rhino. These details introduced what I think the error message called, co-plainer elements, which crashed Rhino when I tried to do a Boolean combination. The suggested work-arounds were simply didn't work. When one feels that the problem must be in my head for not being able to figure it out, it is of course discouraging. Knowing that it isn't just me allows me to re-frame my view of this situation, and maintain a positive outlook. Thanks so much for this!

Thanks also for your vital comments on indexing my long railing. My complicating factor is that one end of my Shopbot is 3 feet away from my basement wall! Maybe I will have to do the last foot by hand, instead of trying to turn the entire board around to machine the last foot.

Thanks again Brady! Chuck

[GeneMpls]

Hello Host of Experts.....

Using Aspire, I have created a simple Shopbot cut file to make a piece for a railing. In X & Y I carefully zeroed my bit on the lower left corner of my stock, just as it is in my Aspire file. Then I zeroed in Z in the middle of my stock. I then loaded my cut file and with the first cut, it was clear that somehow the start reference point was correct in X, but wrong in Y. It was wrong by 2.1inches! I have no idea why I have this error......

Details: I turned off the machine, and sent the spindle back to 0,0, and the tip of my bit parked exactly over the lower left corner of my stock. So I was zeroed in X & Y as I had expected, and I will call that my actual zero point. But I measured the difference between where my first cut was made and my actual zero, and there was an error/offset of 2.1 inches! The entire cut file was translated by 2.1 inches in the y direction, as my spindle was "air cutting" beyond the far "y" end of my piece.

I went and reloaded the cut file, and looked at the set-up panels I always breeze through, and carefully looked at each entry. I had recalled there was an offset parameter, so I checked that, and it was in fact set for no offset.

Why am I getting this big error (positive 2.1 inches in the Y direction compared to my actual zero point?

I am having this exact problem (only my cut is off by 6inches) I was doing
a series of test cuts and suddenly when I run a cut file from 0,0 the router
will jog over 6" and then resume the cut from there. I checked the offset
mentioned... it is ok- the only thing I can think of is I rearranged all my bits
and assigned new numbers and one of the tool numbers would not work-
I had to assign a different number to get past the 'wrong tool #' error. I
have a couple hours in this now- anyone have ideas? Thanks Gene

[GeneMpls]

OK- after making a dozen files and trying a lot of tests, it seems to be
related to one specific tool in the database. It is the 90° Vbit from the
starter set- I have deleted the ones I made and tried using the original
tool data from SB but it still retains the goblin that throws my y axis off.
Sometimes it sez 'wrong tool number' and if I change the number it goes
past that and ends with a 'Location of an Axis may be Incorrect' error
box that also says 'Could not correct the Location of an Axis returned to
starting value'. Going slowly crazy.........

[Chuck Keysor]

Gene: Good luck. I hope someone will be able to help you if you don't figure it out first.

Brady:
a) I moved the vector outline as you suggested, to pull it in so that the ends of my stock are not included, and the simulation seems to indicate this will work, though the simulation is rather sketchy, as I can't really see where it is moving at all times...... But when I cut the actual short part, I can study that and be further assured I will be able to cut my longer parts. But again, I wish to express my gratitude, because I would have NEVER thought of pulling in the boundary vector.
b) When I did pull in the boundary vector at the ends of my model, I tried widening the vector so that the vertical sides of my rail cap would be cut. I tried this without using the boundary vector offset command, and it surprised me that the vertical sides of the rail cap were not cut by such a change. So I re-introduced the boundary vector offset, and suddenly, the narrow ends of my railing cap, that I did NOT want to cut, were suddenly being cut.... Well, I fiddled and adjusted and finally seemed to find a combination where my long vertical sides were cut, but not the narrow ends....... I feel like I dodged a bullet......
c) I was thankful when I measured the clearance I have at the restricted end of my Shopbot, and I have over 4 feet!, so I will be able to cut the entire railing by simply sliding the stock as you suggested. Another bullet dodged!

Thanks again, Chuck

[Brady Watson]

Hey Chuck,
No problem - glad to assist. Yes - the simulation can get a bit wonky when you have a long thin model & zoom in too far. I've got pretty good eyesight, but it can still be hard to see what is going on in fine detail on certain models when you simulate. In addition to seeing the subtractive preview, I sift through the SBP code an see what the toolpath is doing - mainly focusing on what the Z is doing, so that I can anticipate any nosedives etc. I also check out the log file that the SBP creates so that I can see what the min & max XYZ values because they can be hard to pick out on a 100,000+ lines of code file.

The automatic boundary vector tool in Aspire is pretty good...however, if I know that the part is supposed to be perfectly rectangular in footprint, then I will just draw a rectangle - It only has 4 nodes, and sometimes the auto boundary vector can be a bit 'dirty' because of stray voxels (3D pixels) that may be wrapped resulting in a less than ideal path. There is certainly nothing wrong with using the auto wrap tool & editing it using node editing. For instance, the ends - you can just go into node editing mode, lasso around the end nodes and nudge them in with the arrow keys.

If you find that you are not completely machining the walls of your part, offset it 1/2 the tool diameter all the way around (I go a little more than 1/2) to give the tool room to go all the way down. Remember, the vector is the boundary - if it isn't large enough to get your part + 1/2 the tool diameter, it isn't going to get everything. The 2D profile pass should cleanup the outer edge nicely unless you do that part 1st and keep the 3D tool on the block the whole time by adjusting the boundary vector accordingly.

If you had a 5x16' machine, you'd be able to get it in one go :D I've only used that length twice for some molding...and some huge boat rudders. Everything else has fit in a 12' length. Good luck with your project!

-B