Ball Hopper Monitor - Casting Project

[GreenTwin]

You can import sample models from around this web, such as this rocket motor.
.

 

[Jasonb]

Picking up on Peter's suggestion, rather than a cup that exactly matches the old trick of forming a sphere with a tube could be used as the end of one circular item will contact all round a sphere. Both ways would reduce the risk of the sanding disc creating flat spots

Don't know what the joint line might come out like but is it worth considering printing future large items as two parts, The first 2" don't really need any support a scan be seen from the early photos. You then print the "flatter" top 2" separately with support. Overall time saved should be reasonable and there is also the option to use a shallower layer height for the top half as that is where the "contours" start to show.

I always send out STEP files for people that want the CAD for 3D printing.

 

[GreenTwin]

I was moving my Solidworks machine tonight, and inadvertently unplugged the Prusa XL in mid-print, much to my dismay.
As advertised, I plugged the power back in, and the Prusa pickup up where it left off, and never missed a beat
That is an extremely nice 3D printer feature !
.

 

[GreenTwin]

The 2 inch sanding sponge at a very low speed is not prone to digging at all, even on the small round bead.
At higher speeds, you can do a great deal of damage in just seconds.

I really think the sanding sponge is going to take care of most of the lines, inside and out.

I always forget which file type to send out.
The Prusa Slicer seems to be picky about which Solidworks file it will import, and it will open a SW STL, but not a STEP.
I am going to experiment with file types that SW can generate tonight, and find out which ones the slicer will open, and perhaps how to eliminate that facet thing.

I will try to make a short video of the sanding sponge in action, when this pattern half print finishes, which should be soon.
.

 

[GreenTwin]

The PLA surface is somewhat tough to sand compared with a wood pattern, but as I mentioned, the sanding sponge actually melts the top off the ridges, and smears the melted material into the valleys, which is ideal, since no dust is created.
So it is not a sanding process per se, but rather a melting/smearing process.

It sounds odd, and it is an odd finishing process.
I was not sure exactly what I was seeing at first, and I could not figure out why no dust was being generated.
.

 

[GreenTwin]

For some reason, the Prusa Slicer only wants to import Solidworks STL files, which is ok, since that works.

I did a "save-as" in SW, using the STL format, and selected options.
I turned up the resolution as high as possible.
The STL file went from 706 KB to 73,841 KB.
I will see if the Slicer program will open it.
This would solve the facet problem I think.

Edit:
The 3D model comes into the slicer program very small, even though the units are in inches, and the same 3D SW file was used to create this higher resolution STL.
Back to the drawing board.

.

 

[Jasonb]

The bit of the manual you posted in #340 says it will import STEP files.

Very odd if it will only import Solidworks files as that would limit its use to those who use Solidworks for CAD and rule out anyone with another CAD package.

 

[Jasonb]

Out of interest I drew one out myself and tried it in CAM to see what the CNC machining time would be.

If I started with a rectangular block 6hrs, if I started with a cylinder 5 3/4hrs

From other wood patterns I have done it would require minimal sanding so less than 5mins, add a bit of time for cutting and glueing a few boards together to form the blank and 6 1/2 hrs per pattern half. Maybe 15mins more on the window side.

Based on a solid pattern and detail on the corebox to get the outside right.

 

[GreenTwin]

The bit of the manual you posted in #340 says it will import STEP files.

Very odd if it will only import Solidworks files as that would limit its use to those who use Solidworks for CAD and rule out anyone with another CAD package.

The Prusa slicer program will only import Solidworks-generated STL files, not Solidworks-generated STEP files.
Not sure why, but the STL file works fine with the Prusa slicer program.
.

 

[GreenTwin]

My Solidworks 2012 version will export:

.sldprt
.sldfp
.prtdot
.sldftp
.x_t
.x_b
.igs
.step, .stp (AP203)
.step, .stp (AP214)
.ifc
.sat
.vda
.wrl
.stl
.eprt
.pdf
.u3d
.3dxml
.psd
.ai
.saml
.cgr
.prt
.jpg
.hcg
.jsf
.dxf
.dwg
.tif

.

 

[Jasonb]

That is what I meant, surely it should import any standard format such as STL, STEP, etc files created with other cad packages not just solidworks. I doubt it knows what was used to create a STL fine anyway. What was that rocket motor drawn with?

You managed to print those flywheel files I sent you and I think they were STEP as were the helical gears

 

[GreenTwin]

I will send you some samples.
.

 

[GreenTwin]

That is what I meant, surely it should import any standard format such as STL, STEP, etc files created with other cad packages not just solidworks. I doubt it knows what was used to create a STL fine anyway. What was that rocket motor drawn with?

The Prusa slicer will import:

.stl
.3mf
.step
.obj
.amf
.xip.amf
.xml
.stp
.svg

I think all programs probably do not create the same file format exactly the same.
The Prusa slicer program is the latest from the net, and my Solidworks program is 2012, so 12 year differential.

I would guess the Prusa slicer would read pretty much any modern 3D program output, as long as it is one of the above file types.
.

 

[GreenTwin]

test files sent, various formats.
.

 

[GreenTwin]

We got pattern !
Not perfectly smooth, but certainly usable with some buffing.

Someone on ytube mentioned that the support settings needed to be changed to prevent the supports from adhering too well to the print, and I found that to be true with this print.
I need to adjust that before I print the coreboxes, since it is the interior of those that needs to be smooth.

The 10% infill is extremely rigid; there is no way to bend this pattern or deflect any part of it it with your hands.
You can stand on top of it, and it does not deflect.
I will use 5% on the next print.

I will start with the ceramic sponge tomorrow.
This is a good start.
I like this pattern half a lot.
The light reflecting on the grooves and facets makes it look rougher than it really is.
The top is rougher than the sides by a factor of about 3.

Edit:
For molding the interior, it does not have to be perfect since the corebox will actually define the interior shape exactly on the core, I will grind off the bosses, and put two layers of mylar on the flat ends, to allow the pattern to release on the bottom side.
I will fill in those grooves too.
Molding the interior of this pattern keeps me from having to cast a very thick pattern half; I can keep this pattern half shell thickness.
.

 

[GreenTwin]

This is a video of the buffing process.
It does not look like it is doing that much, but it actually makes the surface quite smooth, and smoother than it looks.
2 inch ceramic sponge mounted in a 90 degree variable speed drill.

One idea I had was to knock off the high ridges with the sponge, then make the bound mold, and then sand the inside of the mold.
The art-iron folks do all sorts of post-set resin-bound sand work; grinding, cutting, drilling, sanding, scraping, carving, etc.

I probably need to make a small test mold section, to see how much of the linework is going to get mimicked.
Very find bound sand is bad about picking up the slightest imperfection, as is Petrobond (tm) sand.
Petrobond will copy a fingerprint.

I am pretty sure I could sand the inside of the mold with fine sandpaper though, and that would be quite easy to do, and much easier than trying to fill the outside of this pattern with multiple coats of shellac.
.

 

[petertha]

Your video is quite revealing as its showing the initial 'mountain top' knock down of both 3DP striations & (I think) background facets which is going to be a function of how the STL mesh density was set and/or how the 3DP processor handles it. ie even if you had 2mm wide facets, the CAM side may not be able to resolve that as a function of stepover + nozzle size + facet orientation + .... (Listen to me, I don't even have a printer yet! LOL). But I've seen some results from SW files I've passed to others. Does your SW version have a separate option popup utility where you can further define STL facet specs beyond 'save as' which presumes some default settings? I could find a screen grab but just wondering.

Back to the real world, nothing wrong with your surface prep at this point because the lighter color of sanding is giving you appropriate surface conformance feedback to your eyes. And maybe for casting purposes the end is reasonably near. Most of my experience is with composites work & this would be considered hour 1 of day 1 of a week long job. You should make a visit to an autobody supply shop & ask them about primers that are suitable for their 'plastic' car components - the molded production stuff & also composites. They are a different chemistry than the products used on conventional metal prep. I'm not a chemist but they have different etching properties because plastic is very off-standish to otherwise tenacious paints & glues that readily bond all kinds of other products. I know some guys make rudimentary molds directly from 3D prints & pull epoxy & urethane hard parts them, sometimes with no mold release at all, which indicates that its an inherently slippery material. Also plastics have a bit of flexibility so these products are meant to be sandable but still maintain some flexibility. Old school (typically polyester resin based) primers & putties like the Bondo ilk, may well work find for your purposes because you are only packing sand & not laying on an epoxy coat like a typical composites mold which has the power to suck all this time consuming finish right off the 3DP upon mold/plug release.

Anyways, I'm advocating using some kind of sprayed on a coat as step-1 along with blocking until the high points show just like where you are at. Then stop & spray on another coat, ideally a different tint if possible so you get even more visual feedback of thickness but no biggy for these purposes. All of this build-up and knock sequence down occurs over a small dimensional distance in cross section, but I'm saying its a more controlled & forgivable process than just material removal until it looks or feels right. The problem I've seen with 3DP material is it starts to look like suede leather (or probably more like carpet under a microscope) as more material is sanded away. That's probably what your next step top coat is about. But I'm saying with the right product you may likely get away with a single primer / surface coat product one stop shopping. Duratec was a popular choice back in the day but a) polyester based I recall so stinks & requires catalyst b) typically sprayed but I recall using foam brush on small objects that didn't warrant all the equipment prep. But I'm sure there are more modern equivalents to these which are based on other 'nose friendly' resin systems like epoxy or urethane & easy mix ratios if they are A+B type systems.

Coincidentally I bought a solvent based, air dry (pre-paint) primer sealer from a industrial paint store intended for some tougher duty wood projects that would see some occasional liquid splash. They use it a lot on MDF products, not just cabinets but CNC routed parts where the open core is a bugger to fill & seal. For my purposes it filled & levelled very nicely (we are talking roller & foam brush work here). What amazed me was how it stuck to the polyester drop sheet. Usually most products flake right off that material. Anyways, maybe there is some solution out there because paint chemistry is evolving so fast. Good luck!

 

[a41capt]

After having spent the last 3 years learning and perfecting my 3d printing, I can add that you would be better served to apply Tree Supports to your print, thereby cutting your print time and filament usage by better than half. As far as surface finish is concerned, I learned from my son, who builds costuming elements for several cosplay groups, that the use of 3d printing UV resins as an overcoat, smooths the surface nicely and allows for finish sanding of contours and facets to perfect form.

A simple UV light source sets the resin, which has a slightly thicker than water consistency, and flows nicely over the surface filling voids and print lines, while not building so heavily as to destroy the original part size.

Good luck!
John W

 

[GreenTwin]

Thanks for all the advice and encouragement; much appreciated.
I am going to try and think "outside the box", or in this case "inside the mold", which will be resin-bound sand.
I am going to buff the 3D print a bit to take off the obvious high spots, then ram up the mold (cope and drag), and then I am going to use fine sandpaper to sand the inside of the cope sand mold.

This mold will be for making a permanent aluminum pattern, and this 3D printed pattern is temporary, and will only be used once.
So instead of doing a lot of coating/sanding/filling work on the 3D printed plastic pattern, I am going to try and just smooth the cope sand, since it will be hardened and rigid.

This would sidestep most of the 3D printed pattern surface work.
I am going to try a small test piece and see if this approach will work.

For example, say the mold half below was for the top of the 3D printed pattern; just sand the interior lightly with sandpaper, in lieu of sanding the plastic pattern.



I was in a big hurry making this part, and so I used painter's tape to fill in some spots on the pattern.
The intent was that high spots can be buffed off a casting, while low spots cannot easily be filled.

Looking at this zoomed in picture, you can see that I could have just sanded the inside of the mold to remove the painter's tape lines.


I did not occur to me to sand the inside of the mold, but the art-iron folks do all sorts of post-molding operations to their resin-bound sand.
I missed an easy opportunity to smooth out the lines in this casting.



And I told JasonB that I always have flat ends on the ends of a coreprint, but I notice this one is domed.
LOL.
This was another case of ommitting the corebox, and using the bottom of the pattern, which is common.
.


I notice from looking at the photos that if the mold will be sanded, it would actually be better to have low spots on the 3D printed pattern, which would translate to high spots in the sand mold that could be sanded flush.

The ultimate goal here is to cast a permanent aluminum pattern half that has a very good surface finish, with very minimal prep work on the 3D pattern. Any work on the 3D printed pattern is wasted time, since it will be discarded.
What is important is what the surface finish of the permanent pattern will be.
You can see how this casting came out of the mold, and this is the intent; ie: smooth clean surface on the permanet pattern casting right out of the mold.
This aluminum casting has not had any cleanup work done to it, so a light buffing would be all it would take to have a great surface finish.





.

 

[GreenTwin]

In the case of using the sanding sponge on this pattern to knock off the high spots, this is ideal because it leaves low spots/grooves, which are trasmitted to the sand mold as high spots, which can then be sanded off.

If the high spots are not removed, then they would create a void in the mold that would have to be filled.
.