Project myOX Y gantry plate bent plexiglass dust shield v1b000

Serge E. submitted a new resource:

myOX Y gantry plate bent plexiglass dust shield - This is a plexiglass shield to house the controller board and protect it and stepper from dust/chips

Read more about this resource...

 

Here is a picture with the handmade prototype installed on myOX (back in August 2014) :


There is no vent or anything else. It was cut and grooved using a table saw then bent using a (digital) heat gun. The latter allows for a steady sufficiently high temperature without bubbling... More on that soon enough in upcoming posts, photos and videos.


Ya, the wife really loved the idea of having myOX on the (old) dining table in the house ! Now I'm confined to the near sub-zero garage . We're also getting a new modern dining room table (and chairs). Time to get a heater installed. It will keep her car nice and warm for the winter ... at least while I use the myOX. Although the dust and chips might become an issue if I don't start working on a full 'dome'. I do have a full 4'x8' sheet of PC - polycarbonate, sitting in a far corner for a while now ... Dome might actually also help with the dust/chip control ... creating, what is it call ... negative pressure suction ? The vacuum inlet is anywhere (!?) in the dome and the dust/chips should 'find' it (in theory, right ?)

 

Also note that the handmade prototype is actually longer than 24" (unfolded) by ... oh ... 1.5" or so. This was to get the bottom 'flaps' while this public version is limited to 24" (to optimize material usage ?) I might rethink that part before long.

The other aspect to consider : the bottom is open. The idea is that the air sucked in by the fan needs to go somewhere (obviously !) It could thus escape out of the open bottom which would not be helping the dust/chip cloud (!) So I might close the bottom with a rectangular piece, need the flaps, and add a fan at top blowing out. This could be good since hot air naturally rises anyways ... but it would force all of it pass the stepper.

Having the tendency to overkill things, I was thinking of placing a 140mm fan (which I have left over from an other project) on the side cover... instead of a big myOX 'logo'. So the smaller (80mm) fan could be blowing out as the 140mm blows inward straight on all four driver chips (the entire CNC xPro in my case). This could be the end design ... although it would hide the controller LEDs. So I might use the 4 tiny fans which could replace the huge 140mm. If one or two fail, the others are still moving air.

The side cover could also be bent from a single piece of milled plexiglass ... but that would create 'waste' material. I'll think about it.

One could use aluminum, instead of the plexiglass. Just use thinner stock, still groove 1/2 way down along narrower bend 'lines' ... The whole thing needs to be recalculated in such a case. Not sure how well the aluminum would bend, but heat gun on max. would possibly help (is 450 C or so enough ?) Just a late night (out of the box) thought ... One would have to work both sides to get logo.

 

Well, as mentioned in the myOX build, I had a revelation of sorts earlier in the week. So call me a monkey's uncle if you want, but I'll have to rethink the route (no pun intended) taken so far - everything done the same way.

Here's the thought which hit me : there is basically (oversimplifying) two approaches taken by CAM software.

1) The first, as done by SketchUcam, optimizes the tool path, thus the job run time, the tool wear, etc. It does this by having me, the programmer, hand pick the areas having to be machined and telling it how : do an inside or outside cut' do pockets and drill holes ... which covers all one would want to do in 2D or 2.5D. That is how the gantry plates are likely cut and, guess what, how I should of planned this resource - the plexiglass shield. I could of saved time by doing just a 2D drawing and using SketchUcam for routing just what needs to be routed. It would be a very FAST run - mere minutes. The approach cannot be used for all jobs.

or

2) the CAM software, like MeshCAM, "simply" goes back and forth, offsetting the parallel paths of the tool which covers the entire surface, moving the tool up and down along the way to carve out just about anything or just cruise above the material. It can even simulate 4 axis work, repeating the zig zag parallel 'scanning' with a rotation of the work piece done between each full pass. It is basically a separate job for each 'face'. In the case of my shield, this resource, it is a complete waste of time and tool wear ... On the other hand, if I wanted to engrave a photo, lots of details all over the material's surface, it becomes the only way (right ?) The family tree project is a good example best handled by something like MeshCAM (I have lots of detail for the bark and such, most of it taken from drawing the tree by hand).

MeshCAM actually blends three types of passes. A roughing pass to get rid of the bulk of the waste material. Then it combines the parallel finishing pass (back and forth zig zag over nearly the entire surface, in X, Y or both plans). To finish with the 'pencil' pass, which "cleans" key edges. This gives great results, when configured properly (may not be my case for a while still).


The moral of the story, think about the end result THEN pick the best route (no pun intended). I should of known better, I preach this at work for software projects, web pages, etc. ! That's an other one for the forthcoming Dos & Don'ts ...


Having said all of that, I will be changing the approach taken for the shield, taking advantage of SketchUcam's abilities.

The family tree project will actually end up using both SketchUcam and MeshCAM. Here's my thought. MeshCAM will generate G-code for the tree trunk, branches, overall background and framing, not quite cutting through the latter, leaving the name plates flat at a specific height. SketchUcam would generate the G-code for routing/milling the actual text in each name plate... This text is relatively small, MeshCAM will take for ever to begin to be good, using smallest bit possible while zig zaging over entire surface. By doing the text with SketchUcam, I can use any reasonable bit size and let it trace the centre of each stroke. It will literally write the text ! This should be incredibly faster compared to the alternative. I'll just need to carefully fuse the two or more Gcode files, have a tool change pause where appropriate, without messing up alignment while fiddling with the bits, and let myOX at it.
Somewhat of a 'back to drawingboard' moment. Eh, I'm learning here ...

Video and such of the shield being cut and then bent should now follow over the weekend. The cold spell is not helping, the near -30 celcius (with wind factor) has turned the garage into a freezer !! Do you think the wife will let me bring myOX back on dining table ?

 

Which explains a substantial portion of the budget (not included in the OXcalculator). This can become an expensive hobby ...

myOX, in its limited testing, has done enough so far to see a growing list of projects ... from the wife, surprisingly. I think the cost of negotiation will turn in my favour before the year is out ... eh, wait ! We're just starting the second week of January. Man, my wallet is in trouble now ! I should of built a pocketOX instead... I could of worked peacefully in the warmth of the tiny reading room at the far end of the house. ... What was I thinking of when I got myself into this ?

 

Here's a popular model for wooden "shoes" : (not sure the wife will go for them)

These might be more acceptable : (if link worked, click image to see more from this creator - nice !)

or maybe these :

These for certain, eh ?

There's also plenty of wooden handbags. Here's a few (harder to route/mill ?) to pick from :
or
or

... ad nauseum, as they would say in good old latin.

I didn't find any ready made G-code. But I thought this individual had something ...

Some handbag models could be cut with a PhlatPrinter if your mahogony is in sheet form.

After reading what this individual did, I'm sure someone here can figure out how to machine a pair of shoes with matching handbag using their OX derivative. The Z axis' maximum depth might be the only real constraint.

Should we do contests to see who can mill/route the least expected things with their OX ?

Ain't Google wonderful ?! A few click and the "impossible" pops up like magic. Like they say, never say never ! If someone can think of it, there's bound to be someone out there able to make it ... and an other to improve on the design and method. How many billions of people on this little rock of ours ?

It's mill...ing time. Well, actually, I better catch a few Has I have myOX to tweak and then the shield to route and bend ... capturing it all on film. So stay tune for this weekend's news...

 

Oooops ... the link for the wooden sneakers / clog was a little off. Here is the direct link with details about the design. I like them ... too bad there is no G-code readily available, I think myOX could of managed doing them (maybe not me, but myOX probably).

Although handmade, apparently the "Woodwalk’s design is created by computer". How do we get a copy of the CAD file ?!?

 

Serge E. updated myOX Y gantry plate bent plexiglass dust shield with a new update entry:

2D for SketchUcam / 3D for MeshCam or similar

Read the rest of this update entry...

 

Serge E. updated myOX Y gantry plate bent plexiglass dust shield with a new update entry:

Now partly SketchUcam'ed

Read the rest of this update entry...

 

Well, I flattened the layout and used SketchUcam, tweaking its paths by editing some segments and posted the SketchUp file a bit earlier.

I just could not go to bed without seeing what SketchUcam would do, especially since I tweaked and played a bit of shell game with it... Taking a peek at the resulting G-code with Universal G-code Sender's viewer I get this :



I'd say that is darn near perfect (although it might be hard to tell from the above screen capture). The depths look about right, not certain why there is some red in the fan grill, but it still looks ok. I did not order the groups, so there is a bit of skiping around. The origin seems a bit off, but I left the SketchUp file with both the flattened SketchUcamed AND the 3D version of the shield. So that probably explains the odd placement of the origin.

But most amazing, given all the tweaking done manually to get SketchUcam to do the myOX logo just right ... the logo looks more than AWESOME

Yes, you probably noticed a little flaw, didn't you ? I forgot to do the cutout of the shield itself from the rest of the raw material

It's also hard to see, but there should be some tabs to hold the cutouts (the rectangle for display and button holes).

If it wasn't nearly 3 am, I would run to myOX and give this G-code file a try right now ... but I have to catch some before work - only a few hours away now.

 

Serge E. updated myOX Y gantry plate bent plexiglass dust shield with a new update entry:

SketchUcam magic, with a bit of tweaking to get there

Read the rest of this update entry...

 

After some trial and error, I finally got the SketchUcam to behave as I needed without having to change the design. It basically involves doing some editing of SketchUcam generated elements, using some of its feature in unintended ways, ... I did groups, similar features and allows ordering of the routing/milling while keeping the cutting out at the very end.

I think I already updated the resource file. I'll need to update anyways as I missed the 4 holes to mount the fan with screws (and the screw holes to fix the shield to the gantry plate, which will need at minimum two tiny holes drilled / tapped on at least the top edge.



Looks pretty sharp, no ?

Here's the view from UniversalGcodeSender's viewer, now including the final cut out passes and proper zero reference :



So I should be doing a dry run in scrap wood to see how it turns out later tonight. Then I'll be filming with a plexiglass run, the bending with heat gun, etc. Thanks to SketchUcam, it should only be a few minutes of work for myOX.

I will also do a cutting board layout which is almost ready for myOX.

 

So, here is the final 'run' ... still in a 'simulator' :



Just about everything is there ... I think I missed the tabs for the outside cutout of the entire flattened shield. I can't seem to add them ... maybe due to the 'tweaking' done as SketchUcam seems to have problems when there are cuts within cuts within cuts ... Changing the order of groups, when they overlap is also 'impossible' (for me, so far, as SketchUp doesn't seem to let you grab the groups that are completely under an other group).

Nonetheless, I tried to sort the groups to minimize air travel. The group with the four holes at shield's edge could use optimizing though. It currently does : front, rear and then middle two. I should of kept each hole as its own group rather than a group of 4.

Anyways, the update (b0v005) now has the SketchUp 8 and 2015 with both flattened and 3D version of the shield while thre is a 'run' SketchUp 2015 file with just the very tweaked flattened. Then the CNC is the G-code SketchUcam generated for myOX. You should not use it, except to view in a simulator. SketchUcan does document with as much info as I bothered to include, like group names, so you might be able to adapt to your machine by editing at your own risk and peril.

 

Serge E. updated myOX Y gantry plate bent plexiglass dust shield with a new update entry:

Should be it for right ('near') shield ...

Read the rest of this update entry...

 

Well, in my greenness, it came to me late one night : you do have to design for your machine.

You see, the 3D SketchUp is the part I want to tool. But using SketchUcam, it is nearly impossible to do (for me) without a bunch of tweaking as there are shapes within shapes within shapes. It may be the way I am grouping things, or just the way SketchUcam was designed in the first place.

So I dropped the flatening and trying to use SketchUcam for this one part. I took the 3D version of my creation and fed it through a MeshCAM license (bought a bit earlier to do mostly image to g-code projects). It actually surprised me as to the level of control one can get. With a bit of playing around with some of the permutations, I actually get a not so bad tool path. It still wants to cutout the part before milling/routing all of the inside, but using 'tabs' can help avoid sending your unfinished part flying.

As it turns out, I also have pieces of 10"x24" 1/4" plexiglass. The shield, which needs one per Y gantry, was designed to be ... that is right ... 5"x a bit less than 24". Why not mill/route both left and right shields in one job ?! So I have redesigned to allow for the 1/8 between the two halves, doing some quick cut and paste, and deletes, to 'mirror' what I had so far to make the other shield.

Fed the new dual 3D design from SketchUp Make 2015 to MeshCAM, passing through a conversion of DAE (3D export) with MeshLAB to get an STL (the SU plugin was not showing up for me), then playing with permutations within MeshCAM for the 'best' results and did a run on a piece of scrap OSB. Aside from the OSB not playing nice (being what it is), it turned out pretty good... No movies - the camera had an empty battery ! The piece of OSB, has I had to screw it down to my work surface, turned out a bit short !!! But I just wanted to see what else I could of done wrong. So it is best I don't post images either.

As it turns out, I didn't have a surface between the two pieces ... or they are a tad too close. Maybe they are not really 1/8" apart ... not to mention one needs to allow for a bit more than tool diameter, especially on the outer edges, for rough cutting and the likes.

The one advantage of MeshCAM over SketchUcam, at least in my use (aka lack of experience with both), changing tool and such is a breeze with MeshCAM. The other, well, it seems, with all of my tweaking, you have to delete and redo the SketchUcam elements

I'll gladly stand corrected if I am way off in left field on this. It would not be the first time and there is no shame in admitting being wrong, it's part of learning.

Reordering components (groups) in SketchUcam needs to be easier, at least for folks like me. Some groups, given they are under other groups are simply impossible (for me) to pick. There should just be a table of all groups and one should be able to pick and slide the groups in whatever order without having to pick them from the graphic view of the model. It would take seconds to do a reorder.

So I'll tweak by SketchUp and refresh the file here, sticking to a 3D model. You could always flatten the design and SketchUcam it at your end.

I still find MeshCAM to jump around quite a bit, wasting time with loads of inefficient rapids. But I think it is my fault as I told it to do depth first ... or was it waterline first ? Anyways, it tries to work the tool down layer by layer instead of component wise.

I'll play with it a bit more tomorrow, before posting the files and maybe some pictures and an uncut movie.

The cutting boards will then follow ...

 

Ok, so I tweak the design in 3D for MeshCAM to play nice with it and the first run, in wood, turned out pretty good. So I went to set myself up for doing in plexiglass ... remember, the nice 10" x 24" x 1/4" pieces I got for dirt cheap ? Well, my eye is not very precise - nor are the pieces. Even though the packaging clearly states the above dimensions, the pieces, once unpacked, are only 9"x23" or so!

Of course, newbie and trusting that I am, I had made a nice holding frame in 1/4" MDF and I had even figured how to force MeshCAM not to route the outside edge - just define huge tabs way longer than your parts dimensions and a hair thicker. It will do a 'pencil' run, but that is ok as it does it at height of tab; make it a bit more than a hair taller than worked material so the tool doesn't touch anything.

Unfortunately, I forgot to charge the video camera before starting the job and filming. So I didn't get much to look at.

A few other observations, and tweaks I'm doing right now before filming the aventure ... errr ... actual 'final' run for my new shield, in no particular order :

• for such a job, there is no real purpose in doing a parallel run, especially in both X and Y axis. In fact, the 'rough' in 3D followed by the "watermark" and "pencil" passes will give a cleaner look in plexiglass.
• my design had the vent area sunk 50% into the material. Not a bad idea, except the fan grill gets kind of thin with little support and it wants to curl upward. The tool, especially parallel passes, just messes the finish. So I'll sink that area just a hair, maybe a 1/32" just so it marks the area
• don't plug everything into the same wall outlet ... especially if just a 15 amp breaker. To date, all was fine, but adding a heater is a big no !

I also have a sliding Z - the NEMA 17 is just not enough for my big router. So I will try a trick found in an other forum : use a bungy to add resistance without being in the way. Manually, it seem to do wonders. I'll soon find out.

Off I go tweak the design and give it a run through myOX ...

 

Serge E. updated myOX Y gantry plate bent plexiglass dust shield with a new update entry:

Ooops ! Adjusted a few pockets and such

Read the rest of this update entry...

 

the problem is ... once glued, you're screwed. However, no matter how often it gets screwed, it can always come unglued.