Hi Michael,
I ran into an issue I could not quite understand.. I will send you the model to look at. I couldnt find anything skewed in the model and dont understand well the numbers I am seeing.

I have a block sized out with a parting line down the center in X... If I cut it, it is equal.. If I perform a boolean diff of some objects, then cut it, the sides are displayed as not equal.



In this model, there are named objects to direct you in which should do what.. The model is also scaled up by 10 (with a result scaled back down by .1 in style blue) the scaled down model shows the proper size, but with more decimal places, can see different numbers..

The parts come into my cam package as different sized halves. I got a working model by continually performing a shrink trimmed srf through the process, then using a sat export (to get both top edges of the parted model to display as "Z = 0")

Is this just me showing too many decimals past a tolerance point?? I did have to jump through a few hoops to get a working model (2 halves that appear equal in Z).

Not sure if this model iteration has had a few of the shrinks performed, but a simple boolean diff of the objects, then a half part of the block, will show 2 different sized halves in the object properties..

Email sent.

Hi Burr, thanks I received the file. But I'm not entirely sure if I'm following - which particular display are you looking at when you say stuff like:

"If I perform a boolean diff of some objects, then cut it, the sides are displayed as not equal."

Are you talking about a display in MoI, or a display in your CAM package?



A couple of things to be aware of about the object properties size display in MoI - the one that is displayed at the top level of the object properties panel has a reduced number of decimal places in its display in order to make it possible to it all on one narrow line in the UI. Also on top of that the bounding size there is a quick calculation just based off of the display mesh of the object. Since that line is updated every single time you do any kind of selection it has to produce an answer quickly and can't do the full detailed analytic calculation there, otherwise your machine would be doing tons of number crunching every time you did any kind of selection.

But when you click on that line, it will pop open the edit size dropdown - that one not only has more decimal places displayed but it also goes through the higher accuracy calculation as well for the data displayed on the little dropdown menu. So pop open that menu if you want to see a higher accuracy size display.

If you're seeing some stuff in your CAM program, I seem to remember that your program had a tendency to add in its fitting tolerance value to the size of the bounds of the object, if that helps to explain anything you're seeing on that side.

- Michael

Hi Burr, over here when I do the boolean and then cut it in half, I seem to get the same size readout for each half in MoI, here's with the top half selected:



And here's with the bottom half selected:



Seems to be showing the bounding sizes as equal as you would expect - where is it that you're seeing the sizes as different?

If it's in your other program maybe you're running into that kind of odd UI issue where the particular diagnostic readout adds some +/- tolerance value to the actual object size, I think you've run into that before in your CAM program if I remember right?

- Michael

Maybe I have totally not understood what kind of size difference you're talking about though...

- Michael

HI Burr, also I don't really understand what you mean by this part:

"(to get both top edges of the parted model to display as "Z = 0")"

How is it that you're expecting the z size to be 0 units? You seem to have the full block at 19 units tall, so shouldn't you have each half block to show a z size of 9.5 units, exactly as shown above?

- Michael

Hi Michael,

"""""""""A couple of things to be aware of about the object properties size display in MoI - the one that is displayed at the top level of the object properties panel has a reduced number of decimal places in its display in order to make it possible to it all on one narrow line in the UI. Also on top of that the bounding size there is a quick calculation just based off of the display mesh of the object. Since that line is updated every single time you do any kind of selection it has to produce an answer quickly and can't do the full detailed analytic calculation there, otherwise your machine would be doing tons of number crunching every time you did any kind of selection.""""""""""""


I have mine setup to show all my decimal places in that upper right corner. but didnt know it was using the display mesh calculation there..



""""""""""""But when you click on that line, it will pop open the edit size dropdown - that one not only has more decimal places displayed but it also goes through the higher accuracy calculation as well for the data displayed on the little dropdown menu. So pop open that menu if you want to see a higher accuracy size display.""""""""""

That was it... Popping out the edit size shows an accurate split... The top level properties display was showing numbers out of sync...

Yes, my CAM package reads all the raw "bounds data" and I have to extract the models actual "edges" to get a good read of where surfaces actually lay... I was being confused on my results and going back and forth and got lost.

Thanks for looking, this should get me going.

Hi Michael,
Just to be clear, I have to have the "tops" of the 2 halves be placed at "Z 0" to go into the CAM package.. (See the little scaled down blue halves) over at origin..

You answered it well with the edit size precision.. The issue was I was reading the upper right display properties.. If you set that to 4 digits, you will see the 2 different numbers...

I'll make a short video to show some of my process and be clear.. There may be a couple areas to comment on with regard to my method.

Hi Burr,

> I'll make a short video to show some of my process and be
> clear.. There may be a couple areas to comment on with regard
> to my method.

Do you mean you're worried about your current process for rotating and positioning the blocks after the boolean is done?

Object snaps on to the corners of the box using the Move command should be highly accurate for the positioning part, and it should be fine to just use the rotation edit handle to flip the one piece by 180 degrees in the Front view.

- Michael

Hi Michael,

""""""""""Do you mean you're worried about your current process for rotating and positioning the blocks after the boolean is done?"""""

I've always used those methods and had great results.

It was just this particular job where I started to go back and forth and had some confusion.. I dont want to go to much into my CAM package and its methods because that is not the purpose of this forum.. Let's just say that if I was in front view and drew a curve with pick points "above Z zero" though the resulting curve is well below Z zero, the CAm package will read the top most "control point" as the height in Z... But you can rebuild it to show it's true Z value.. This is the same for surfaces and bounds data. The edges of the surfaces will show at the proper height, though the entire surface will show a bit more.. It's usually .002 in all directions, but with this model, IGES wouldnt produce proper results, but SAT did.. (A translation issue)

This video can show what we were discussing before, that you straightend out with the edit properties dialogue, but I wanted to show my use of the boundingbox command for placing an object, and what I saw here, just to get clear for myself.



I am always sure of MoI's accuracy. I was just thrown a little here...

Thanks for looking at the file Michael...

Just got back from my brothers where he finished the mould.








Shapeways demo part seen in it:







The Rapanui man made this so easy!!!!!!!

What are these little pieces ? (if it's not a secret)
Arraw darts metal end for this ?

Looks like it turned out great Burr, thanks for posting the machined part, it's always cool to see that!

- Michael

Hi Frenchy,
"""""""What are these little pieces ? (if it's not a secret)""""""""""

That is a mold for inserts for fishing lures... Lead will be poured into them, then the piece will be encased in a larger resin object with a skirt attached... Their will be a brass tube placed in the mold before pouring..... :o

This sort of thing? (of course more elaborated :)

Great !
i rememeber you posted lure rendering in the gallery
Is it aluminium the mould you are machining ?(i don't see liquid refrigeration for the tools )

Hi Frenchy,
Yes. It will look more like these when done





A guy will paint them up, then they are cast in resins or epoxies.. You would be great at it.. The fishy's like colorful, apparently!!!! :o

""""""i rememeber you posted lure rendering in the gallery
Is it aluminium the mould you are machining ?(i don't see liquid refrigeration for the tools ) """"""""

Thanks M....Yeah, thats it... The mold is aluminum. There is an air-mist coolant system (see the nozzle in the pics). You probably dont see any residuals because he blasts air to clean it just before a picture....

Seems a domain very colorful and diversified!

Nice one Frenchy...

I suppose it is the result of Man imagining what a fish would enjoy!!! lol