A 3D Printing Experience

[JimG]

This all started as a completely different project. I started developing an etched kit for a 16T mineral wagon body kit for the S Scale Society using official drawings which AndyB got for me. But that came to a halt when the cost of new photo tools became a bit high.



I was trying to reproduce the 1/4" thickness of the prototype's steel sheet which meant a lot of half etching of 8 thou nickel silver sheet and it became very finnicky to build and I was envisaging the "Who the hell designed this" comments.

I had a bit of a conscience about not proceeding with this etch. The SSMRS did have a white metal kit for a 16T body but no proper underframe to go with it, most people using an LMS 16' 6"/9' WB etch which was sort of close. So I thought that we should have a proper underframe for the body kit and I commissioned Justin Newitt (Rumney Models) to produce one - which he did and a nice job too. Then we ran out of body kits and I had no luck in getting any more produced. So the body etch development was started to give an alternative which would use Justin's underframe and hopefully reward him with further sales.

So my broken hip and Covid experience intervened and then I started thinking about maybe doing a 3D print of the body and I started doodling in Fusion 360 and produced a first attempt which actually printed fairly well but needed a bit of work. Then the printer broke down and I spent a few months trying to repair it - the fault was in the LCD screen which had become obsolete and it took time to find a replacement which needed a lot of upgrading of the electronics - all of which didn't work!! So I had to bite the bullet and get a new printer. I then opened up Fusion 360 to start work on the print again. This was around the time that AutoDesk was limiting the free version and for some reason I lost my drawing file in the small hassle when adjusting to the restrictions. So I went right off Fusion 360 and started to use Solid Edge Community Edition and had to start at square one with the drawing - and also learn a completely new CAD package.

I was aiming to print the wagon oriented at around 40 degrees with the end door at the top so that the majority of of the overhangs and protrusions would not require supporting. I started off with an 0.6mm side thickness and left the floor out since I wanted to experiment wiht a metal floor to increase the weight of the body. But I ran into problems with peculiar ridging on the sides and tried adding a temporary floor to get rid of the ridging, as well as increasing the thickness of the sides.



The faults are obvious here, going from lower left to upper right on the sides. (The other lines crossing them were caused by dirt on the LCD.) So after umpteen attempts with various types of false floors and varying thicknesses of the sides I gave up on the 40 degree orentation and started printing with the body sitting flat on the build plate. That got rid of the ridging on the sides, but the supporting of all the protrusions and overhands became necessary. So there were another lot of attempts until i started getting success.

I now support the wagon as follows :-



The body is supported on medium weight supports sitting 6mm off the plate.





The details on the side doors are added using light supports with the finest ends which will work with the resin.



All the protrusions on the end doors use light supports.



The top angle and corner brackets are supported all round the body.



And the final result is this - quite a forest of supports.

I have also experimented with increasing the pitch between supports to cut the number down, and I might try going wider pitch than this.

The end results I've eventually got are





And the body with the brass floor - cut on my CNC machine with the floor plating engraved.



I'm pleasantly surprised that there is pretty well no bowing of the sides. The weight of the body and floor is 44 grams, so the complete wagon could be around 60 grams or more.

The final thickness of the sides is 1.2mm tapering to 0.6mm at the top - you can just see the ridge lines on the inner sides. The brass floor is 1.6mm thick. So floor and sides are well over scale thickness but all the extra is on the inside and hardly shows.

Apologies for the poor paint job. I gave it a quick scoosh of Halford's grey primer to make it photograph better and the weather outside wasn't great and the can got a bit close. I also slightly messed up the top girder over the end door when cleaning off the supports. Must do better next time.

This is a quick lash-up to take to the SSMRS AGM in a week's time to get feedback. I've got one of Justin's underframes to build to go under the body as well.

I've been talking about this to Richard (Dikitriki) for what seems to be months now and keep promising to put a picture up on WT to show progress. Well here it is.

Jim.

PS

 

[michael mott]

Wow that is a lot of aggravation to get to where you are at, Glad to know that I am not the only one that feels a bit like a "dog with a bone" sometimes. The final result looks great.

Michael

 

[simond]

Jim

i can see why you’re smiling!

I’ve linked this to the 3D Printer thread so it can be found in the far future…

Thread '3D Printing - whether to buy, what to buy and how to use it?'
3D Printing - whether to buy, what to buy and how to use it?

and vice versa

 

[Pencarrow]

Excellent work and a good effort sticking with it to the end. All those bodies piled up makes me think of a scrapyard scene...

 

[JimG]

Wow that is a lot of aggravation to get to where you are at, Glad to know that I am not the only one that feels a bit like a "dog with a bone" sometimes. The final result looks great.

Michael,

It's also aggravating when it takes about five hours to print an orientated body or two and a half hours to do one flat on the plate and then add in draining time after printing, cutting the body out of the supports, cleaning, then curing. So the orientated body had about an eight hour turn round, and nearly five and a half hours turnround for the flat to plate body. So having to wait those kind of times while trying to work out what alterations were working got quite aggravating. It also takes quite a while to redo all these supports for a new file when an alteration is made. Staggering through the 3D CAD operation was relatively quick.

Jim.

 

[JimG]

I’ve linked this to the 3D Printer thread so it can be found in the far future…

Simon,

Sorry, I swithered where to put the thread since it was as much messing around with parts for the S Scale Society as it was 3D printing. Having now re-read your 3D printing thread, I can see similar stories of suffering.

Jim.

 

[Heather Kay]

This is impressive stuff, and mostly going right over my head.

One thing that struck me, though, was getting the orientation right so the shape printed nicely. Might it have been better to print the body inverted, rather than bottom up? It occurred to me that cleaning up the top rave would be easier than having to fiddle around in the details on the bottom of a body.

As I say, I know nothing!

 

[mickoo]

For a first run of any new part I work on a 5:1 wastage to get the orientation spot on, sometimes higher, you're not alone in your stack of waste.

Heather is also right, the body should be printed the other way up, I would also add a 0.5 mm ridge all around the top and add the supports to that, make sure the ridge is thinner than the top so you have a step or witness line to file the ridge down to.

Upside down will remove all the top flange supports and most of the lower door hinge ones, in fact I'd probably not add any supports to the hinges or straps at all, I think they'll print fine. If you want to check then just carve the 3D model up to give you that small section you're concerned with, say a 25 x 25 mm wall section with the hinges and catches on, then print that in the orientation you desire. If the details all work fine then move back to the full item.

Your lines at 40° are due to changes in FEP peel forces, hard to describe, maybe easier with pictures, it's very common whit hollow objects or objects that change shape or size suddenly.

This is a K3 smoke box saddle I'm working on right now, it's hollow to save resin, increase print speed and solve one or two other potential deformation issues.

Everything is tickety boo at this point as the print has two equal areas printing on the FEP.



The problem arises when the two meet and the surface area becomes one, the tension on the FEP alters and covers a larger area, it changes the peel force and over stretches those few layers until it settles down as one large area.



The end result is a witness mark around the point where the peel force changed.



It's pretty much unavoidable, the Form does a better job with it's more sophisticated peeling mechanism, but hobby machines basically just rip the print off the FEP and there's little control over the peel force.

My load bearing/supporting/sacrificial edge is the base of the flange, you can see the 0.5 mm sacrificial extension added to that flange, all the supports are here so any damage they do when being removed is all on the sacrificial edge, not your final product. A quick waft with a file or on sand paper on a glass sheet will cut that edge back to the final point. Because the edge is thinner than the main flange you have a witness line to guide how far to file down.

Here's a valve cover printed vertically as seen from below, you can just make out the added ridge around the base where all the supports attach too, the step to the main base is your witness line to file down to.



CAD designing is the easy part, CAD designing for 3D printing is a whole new skill set and much more difficult. Once you understand how your machine works and where the issues are, usually deformations, then you can counter that by making the CAD part and inverse of the final item. If you're getting an edge with a constant bow in it, then simply invert the bow in the CAD model and it should print straight.

If you do opt for printing upside down then I'd adjust the side walls. Best practice is to get the mass to go lower the further you get from the build plate (or add more supports) so your top of 0.6 and base of 1.2 might cause issues, you might be better off with 0.8 through out, it'll remove the ridges on the inside as well.

 

[Heather Kay]

Heather is also right

Blimey! There’s a first time for everything!

 

[mickoo]

Blimey! There’s a first time for everything!

Have you got this weekends lottery numbers by any chance

 

[michael080]

My load bearing/supporting/sacrificial edge is the base of the flange, you can see the 0.5 mm sacrificial extension added to that flange, all the supports are here so any damage they do when being removed is all on the sacrificial edge, not your final product. A quick waft with a file or on sand paper on a glass sheet will cut that edge back to the final point. Because the edge is thinner than the main flange you have a witness line to guide how far to file down.

That's a brilliant idea. I will try that at my next print!

Michael

 

[JimG]

Heather is also right, the body should be printed the other way up, I would also add a 0.5 mm ridge all around the top and add the supports to that, make sure the ridge is thinner than the top so you have a step or witness line to file the ridge down to.

Mick,

Many thanks for your comprehensive response. I did consider printing upside down way back when, but persuaded myself not to since I was worried that the wee corner brackets on the top might suffer in the cleaning up process. However you have persuaded me to have a go again, the attraction being that I have a fraction of the support work to do. It's the major bind I have with my present orientation is that any minor change in the model requires all that manual support work again. Chitubox's automatic placing is worse than useless in this particular case and I don't even try it now since I've got to clear off most, if not all, of its attempt.

I might have a problem with the small flange round the bottom of the sides which on the protoype allowed a base for welding. It only projects 0.2mm so It might work with no supports. At the moment it works very well when printed right way up.

Jim.

 

[mickoo]

The wee corner brackets will probably fail and end up flat, is that a compromise people can accept for the rest being a better product, I would.

If they bothered my that much I'd stick some 5 thou Plasticard ones on after I'd cleaned up the top face personally. That would be what I call a planned/sacrificial failure and I'd have a post print process to get them back if required. Attempting to keep them as part of the total print will almost certainly limit your options and print quality.

A 0.2 mm protrusion should work fine with no supports but if you're really worried then angle the top face slightly (if you're printing upside down), you won't see it visually and it may give a few layers for it to 'grow' out of the side face, generally speaking though, I do not add supports to items that protrude less than 0.5 mm if the angle between them is 90°.

This double flange overhangs by 0.25 mm all around, as a safety net I filleted the underside joint and spread the transition over three 50 micron layers, this allows the protrusion to 'grow' out gradually from the main barrel. It's not prototypical but it prints successfully and avoids additional supports, plus, if I hadn't have shown the enlargement, no one would ever have known. If it offends that much then waft a needle file around the fillet and sharpen it up.

The angled cut in to represent the joint also avoids a hard 90° edge and allows the layers to merge in and out without a large change in peel forces. Although it's drawn as a V cut the reality ends up more like a U and being that small, is basically a visual representation of the joint.



A big 3D mind set problem is that 3D prints are perfect right off the build plate, it's very hard to get people to accept they may have to do post processing work, in fact if you design the part with planned post processing then it'll often be more successful. After all, we willingly accept post processing for brass and white metal castings once free of the sprue, but not 3D prints.

 

[Rob R]

With the extra bit of wisdom gleaned from the above I have had another go at the S scale GNR open.

Ridge aded all around the top and inverted on the build plate.
Printing right now (0.025 layer height so ready in about 8 1/2 hours).
Must order some more resin now.....

 

[Rob R]

Hmmm.
Another one to add to the Vic Berry sized pile on the workbench.
Still dripping on the printer but the sides have creases shown in diagram below.
Out of resin now so I won't be having another go tonight.
Proper piccie to follow if I don't throw the whole issue through the window first.



Edit. Maybe heavy supports all round and a couple of disposable cross braces across the width of the body?
Or just open the window............

 

[mickoo]

This is one of the biggest problems with resin printers, it's where the peel forces suddenly change in area volume, the top one I can relate to, the lower one not so sure where the area volume change is to be honest.

Is there a particular reason why you're only rotating it in one (pitch) axis? Convention recommends a two axis rotation, doesn't have to be much but should reduce the top crease quite a lot, it may not eliminate it though truth be told.

Somewhere in the 5°-7° roll axis should be enough, some people go higher into the 20-25° region though. You may need to watch out for any protrusions on the shadow face (that's the face that's rolling toward the build plate) as they might need a small support adding to their lowest point.

 

[Rob R]

Fresh start with fresh resin tomorrow if Amazon turn up before I get in from work.
You are quite right, I should have pitched it a little on t'other axis.
Less haste more speed etc
Will still add a couple of cross braces to the top ridge and beef up the supports to heavy all round - belt and braces.
I'll get back to playing with the 63' Opera Window Centerbeam after the agm....

 

[Arun Sharma]

This is one of the biggest problems with resin printers, it's where the peel forces suddenly change in area volume, the top one I can relate to, the lower one not so sure where the area volume change is to be honest.

Is there a particular reason why you're only rotating it in one (pitch) axis? Convention recommends a two axis rotation, doesn't have to be much but should reduce the top crease quite a lot, it may not eliminate it though truth be told.

Somewhere in the 5°-7° roll axis should be enough, some people go higher into the 20-25° region though. You may need to watch out for any protrusions on the shadow face (that's the face that's rolling toward the build plate) as they might need a small support adding to their lowest point.

I'm with Mickoo there. If dealing with a brick-like shape in Chitubox or any other support/slicer program, a single rotation from horizontal exposes an edge and an underside that need to have supports added. By rotating in another plane as well, the lowermost area becomes a single point and whilst this will need a heavy support, the remainder of the underside won't need very much at all.
The classic demonstation example of this is the YouTube video showing the single support required for printing a sphere because it has a single low point.

 

[mickoo]

I'm with Mickoo there. If dealing with a brick-like shape in Chitubox or any other support/slicer program, a single rotation from horizontal exposes an edge and an underside that need to have supports added. By rotating in another plane as well, the lowermost area becomes a single point and whilst this will need a heavy support, the remainder of the underside won't need very much at all.
The classic demonstation example of this is the YouTube video showing the single support required for printing a sphere because it has a single low point.

Absolutely, there is a point where more supports is counter productive, bear in mind that each support joint will create a small bulge and supports close together can have the effect of trapping resin between them in a meniscus way, some of that will partially cure due to low levels of bleed over from LED type SLA printers. It doesn't cure much but can turn into a sticky mess that's just a pain to clean up.

 

[Arun Sharma]

Absolutely, there is a point where more supports is counter productive, bear in mind that each support joint will create a small bulge and supports close together can have the effect of trapping resin between them in a meniscus way, some of that will partially cure due to low levels of bleed over from LED type SLA printers. It doesn't cure much but can turn into a sticky mess that's just a pain to clean up.

Mickoo - I haven't noticed that problem but as a preventative strategy, I always first wash the print with its supports in situ. I then remove the supports [since they are still soft at this stage] and then wash the model again just in case any uncured resin has trapped itself around the base of the supports. I use the Elegoo Mars 2Pro and Elegoo water-washable resin with the Mercury Plus system.