Sprayed-on 3D Printed Pattern Filler

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I will see if these videos will upload.
The first one is the finger-scratch test on JasonB's flywheel print before the sanding sponge treatment.

The second one is after sanding with the ceramic sponge.
The grooves on the sanded spoke are gone.

.



 
Last edited:
What are those solvents?
About the only PLA solvent bonder that I'm willing to use is Weldon, #3, #4, or #16 - they seem to be the least toxic. (Weldon #3 and #4 are the water thin versions, while #16 is a thicker tube type glue.) I'm not at home right now, but the SDS sheets are available online. It doesn't seem too toxic, but I'm sure I wouldn't want to be using it regularly for long periods without the proper PPE. The Weldon SDS sheets list Methylene Chloride as the primary ingredient of those.

Found this discussion on reddit:
PLA solvents
 
Not really a worry for me as we use a decent spray filler and no amount of chemical smoothing will equal a sanded and faired finish on PLA.

Quite true. But some people, and I not gonna name fingers or point names - because I'd be among them, just hate post-processing.
 
To some extent it will depend on what you are doing with the pattern, Barry's moulds need to be finished to a much higher degree than something used for sand casting. I've CNC cut wood patterns and they have produced good castings without even sanding or sealing, MDF rim did get some sanding but not the beech spokes

casting.jpg


Even printed ones don't need to be shiny smooth, I only did the CAD for this but you can still see the printer marks on the edge of the rim and hub but the sand was not so fine that they reproduced on the casting, might be different with Petrobond

layers 2.JPG


This was the print straight of the printer, it did need handwork, polishing mops won't get into small areas and can remove detail and crispness, no real substitute for putting in the time and effort and manually refinishing. Why try and rush the pattern surfaces when its been 10years or more that you have taken to get to this stage.

layers 3.JPG
 
As I was polishing the 3D printed pattern for the flywheel you show in your first photo, I was generally trying the sponge on everything, including the curved spokes, and the flat surfaces.

The wide flat rim was more difficult to sand out with the sponge, but it was much more flat to start with.
I think the flat surface was flexing away from the sponge a bit.

The spokes seemed more rigid, and had taller lines, and those came off more easily with the sanding sponge.

With a casting as good as the first photo, I would just remove the flash at the parting line, face the rim top and sides, and then very lightly go over the remaining surfaces with a fine sponge, just a very light touch thing.

Some folks will buff or paint everything to a mirror finish, but if the surface is not machined, I prefer it to have a sand-cast texture to it.

The second photo casting came out with a perfect surface finish, so I am not sure what slight of hand you used on that, but as you mention, it is all about the sand.

With the bound sand I use, it is going to mirror things the thickness of a hair, and smaller.
The sprayed on ceramic mold coat helps smooth things out.
I am wondering if enough mold coat could actually fill in the grooves in the sand, so that the 3D printed part did not have to have much smoothing work done to it.

The ceramic sponge method appears to be very promising, and while it will not be a substitute for a full surface fill and sand finish, I do think it will be an excellent surface prep prior to a fill/sand that will greatly reduce the work required as compared to just a fill and sand.

I think 75% of the heavy work can be done by the sanding sponge, and I think many surfaces can be completed after sanding with the sponge just by using a few coats of shellac, instead of putty/filler.

I hope to start 3D printing some Monitor patterns in January, and then I can give a better report on the sanding sponge/shellac method.
Many of the parts on a Monitor are convex, and that lends itself to a sanding sponge.

I am also going to research how I can improve the surface finish by tweeking the Prusa slicer program.

I think the sanding sponge may work ok with smoothing a filled surface.
I tried it on the auto body putty on the dog, and it seemed to take a light cut without digging or creating flat spots.

And lastly, it occurred to me as I was trying the sanding sponge on JasonB's flywheel pattern, that there was no need to buff off any surface that would be machined, such as the outside and faces of the rim, the faces of the hub, etc.

I may try fogging on shellac, and creating a surface that is semi-rough, and something that would bridge the grooves, sanding lightly, and then applying a final overcoat of shellac.

I will try some things, but I think the sponge is going to save a lot of time; I can already see that from early trials.
For smaller 3D printed patterns, and hard to access recesses in patterns, the sponge will be of no use.

.
 
You don't want any surfaces with draft too rough even if they will be machined unless you want to start pulling sand out as well as the pattern.
 
You don't want any surfaces with draft too rough even if they will be machined unless you want to start pulling sand out as well as the pattern.
Yes, that is a good point.

Edit:
For the 14" BHM flywheel patterns, I am going to go with a net design, and so will do the mylar wrap around the flat outside of the rim (no draft angle).

I really want to start doing net castings, or near-net I guess.
It makes chucking things so easy when you have an accurate and flat surface.

.
 
You are unlikely to be holding a 14" flywheel by the OD, either bolt to a faceplate or if your 4 jaw is big enough hold by the inside of the rim. As it is a 5spoke I expect the faceplate option will be your only one as that number of spokes never works well with 3 or 4 jaws.

Personally from having machined a lot more castings than you I would rather have a bit more machining allowance and put up with draft than only have the thickness of a bit of Mylar to play with.
 
The mylar just allows a 90 degree pull with zero draft angle, it does not affect machining allowance, which can be anything.
The mylar should slip out from between the sand and the pattern. I may have to snag it with a straight pin to get the pull started.

I am not sure if I will have to farm out the flywheels to be machined, but if necessary, I have some folks who can do it that work both in industry and at home at machinist jobs, and have larger lathes.

I have never tried to squeeze anything larger than 12" diameter into my 12x36 Grizzly, but I guess that is worth a try before I farm it out.

I made a plywood backplate that I screwed to my metal backplate, so I could use that, or just use the metal backplate.

Ideally I would have a backplate the same size as the rim, or at least the same size as some of the rim, since the rim should be pretty flat, and thus would give a flat surface to pull up to.

I will go out and do some measuring, and see what I can get away with.
I have seen some install extentions on their lathes; ie: raising the head, but I don't think I want to get into that.

.
Edit:
So the diameter of the 3D printed pattern would be double shrinkage (assuming a permanent metal pattern will be cast), plus maybe 0.1" on either side, for a total of 0.2" machining allowance, plus twice the thickness of thick mylar.
Thick mylar is still relatively thin, so not much added there.

Edit02:
If I really want the flywheel diameter to be exactly 14" after machining, I will have to print the pattern in sections, since the Prusa XL's maximum print is suppose to be 14".
 
Last edited:
I've dealt with disks larger than my lathe capacity by mounting them on a center shoulder bolt or a rotary table and used the mill. The only trick when using a shoulder bolt is how to turn the work while maintaining control. Obviously climb cutting is out, but somehow when a conventional cut goes wrong climb cutting invites itself to the party adding bonus fun. Very shallow clean up passes keep things reasonably sane. If the material to remove is minimal I've done similar stunts with a sanding drum and drill press.

Happy new year to all!
 
I went out and measured my 12x36 Grizzly, and it looks like it would take a disc that is 12.5" (+ -) maximum diameter.

And it would not be easy to raise the centers on the lathe.

My mill is not very rigid, so that could be a problem.

Perhaps rotating it on the rotary table on the mill, and sanding the diameter to size with a belt sander.
The outcome of that may be a bit iffy, and that would not face the rim.

I will check with my sources and see if they would machine them on one of their commercial lathes.

The 14" flywheels are going to be problematic, but no way to be at 1/2 scale without them, and 1/2 scale would really be cool in my opinion.

.
 
I thought the idea was that the pattern slides out as there is no friction with the mylar and then you just remove the mylar once it is not being held against the sand by the pattern.
 
I will see if these videos will upload.
The first one is the finger-scratch test on JasonB's flywheel print before the sanding sponge treatment.

The second one is after sanding with the ceramic sponge.
The grooves on the sanded spoke are gone.

.
View attachment 152358


View attachment 152359
GreenTwin may I ask where did you purchase the ceramic sponge. I am going to need to cast a brass flywheel for the Monitor engine using a pla 3d print pattern also. (First time casting brass this should be fun).
I used to use MEK to smooth PLA 3d prints Despite its good results, I can't find it locally anymore (Box stores only carry MEK substitute now, does not work.) plus its toxicity and volatility not a good thing.
 
I thought the idea was that the pattern slides out as there is no friction with the mylar and then you just remove the mylar once it is not being held against the sand by the pattern.
It is possible that the mylar will function that way, but I am not going to buff off the outside flat surface of the rim, and so it may have some ridges on it.

I guess I will see if the pattern will pull without removing the mylar, and then if it seems stuck, I will slip out the mylar and then pull the pattern.

This is new ground for me, so I can't say exactly how it will function, but for the few items that I have been able to make without draft, I really like near net parts very much.
It is expecially helpful on flywheel rims to have a flat surface to grip when you are trying to chuck the piece in the lathe.

Near-net and even net castings is the hot trend right now in the foundry industry.
Some parts are being cast with all holes in place, and require no machining after casting.
The casting process has come a long way in recent years.
There is really no reason to use obsolete casting methods, in my opinion.
Move up into the 21st Century.
Its not that difficult, even on a hobby level, in my opinion, it just requires a different school of thought.

.
 
Last edited:
GreenTwin may I ask where did you purchase the ceramic sponge. I am going to need to cast a brass flywheel for the Monitor engine using a pla 3d print pattern also. (First time casting brass this should be fun).
I used to use MEK to smooth PLA 3d prints Despite its good results, I can't find it locally anymore (Box stores only carry MEK substitute now, does not work.) plus its toxicity and volatility not a good thing.
I buy mine from random sellers on ebay.

You have to turn the sanding sponge slowly, else it will melt/burn the surface, so be sure you have something very low speed to use it with.

I think the term is 3M "Reloc", which is the trademark name, and they fit into a twist-on holder mandrel with 1/4" shank.

This one has a variety pack.

https://www.ebay.com/itm/1441723386...ZkPMQOoc1C7QDb40bbmw4GRZbA==|tkp:BFBMvrrXu5Zj
Mandrel only:

https://www.ebay.com/itm/1438814602...DpfVboCy8QB8i5W+8N7HOmMGfA==|tkp:BFBMvrrXu5Zj

Here is another package deal.
The blue are fine, the red is medium, and the brown is coarse.
If you have a rough casting, and need to remove a lot of material, start with the coarse, and then move to the medium and/or fine.
The fine will leave a mirror finish on gray iron.

The coarse and medium are too rough for use with 3D printed patterns, but the blue works pretty well if you slow it down enough.
If you start getting smears and black streaks, your speed is too high.

https://www.ebay.com/itm/3351341674...ah8d3JvMElABFRFqYZ5i1kSafQ==|tkp:BFBMwLrXu5Zj
 
I was playing around with various flywheel types/sizes/spoke configurations a few years ago (2019 to be exact), and 3D printed them in sections on my smaller Prusa.
I had forgotten that one of these was an early Ball Hopper Monitor prototype (the one on the right), and so this is how I split it.

I am not sure if these are optimum split line locations, but this is what I came up with in 2019.
This gives a 4-piece print (for one side) for a 5-spoke flywheel design.

If I want a true as-machined 14" diameter flywheel for the 1/2 scale Monitor, I will have to do something similar to this on the Prusa XL.

.
rImg_3124.jpg
 
Back
Top