Kevin-the-lubber

As I say, no first hand experience of sintering at all, but here's an excerpt from one of my go-to sites, All3DP.com; "... engineering-grade parts with excellent mechanical properties, fine resolution, and incredibly fast, SLS (selective laser sintering) 3D printing is what engineers and industrial designers turn to for functional rapid prototypes and end-use parts..... advancements in technology have made SLS printers more affordable and compact... there are now a good number of office-friendly benchtop SLS 3D printers on the market.... SLS also offers a high degree of predictability in material and mechanical properties, so it’s popular in aerospace, medical, and regulated industries. There’s also no need for supports in SLS printing, which expands your ability to design and produce very complex geometries.... This technology can create parts that are finely detailed, strong, durable, heat resistant, and flexible (when needed) all at once. Printer manufacturers often boast that SLS prints rival injection molding products in terms of strength and precision. When compared to injection molding, 3D printing can create parts that have internal channels, lattice structures, and other features not possible with molding...." Resolution down to 0.05mm. Sounds good, doesn't it 😃. And we're talking metal, plastic, all sorts of material choices. Now the bad part - that resolution would currently set you back at least $40,000. If you can settle for 0.1mm, £6,000. If a machine was available that did even 0.1mm for less than $1000 I for one would be sorely tempted, if it was less than $500 I'd be first in the queue. I haven't tried Shapeways or any other print service but I probably ought to at some point, just to see what sintering delivers.

I've printed with a 0.2mm nozzle to the point where the surface finish is really not that bad, though sanding filament is much less easy than resin; it's more that getting FDM dialled in can be a right pain whereas this is super-simple with resin. I think all the manufacturers know what the community wants and this will determine the direction of development: affordable, reliable, easy to operate, no fuss, plug and play, no mess, no post print cleaning etc. Oh, and big. We want parts, not a process! Neither FDM nor SLA entirely fit that bill which is why I think, sooner or later, someone will try budget level sintering or whatever else comes closest to that spec (I have no experience of sintering but it looks like the real thing). The same is true with the software side. My ideal would be Tinkercad with bells and whistles. It took me 15 minutes to learn enough to produce my first part in Tinkercad, more like 15 hours in F360. Obviously Tinkercad is quite limited but with a few additions it could remain very useful indeed, not least because WYSIWYG.

Rick, I'd be a little careful there. FDM certainly has it's place and is currently a good approach for quite large or functional models but my guess is that it will be overtaken by resin within another few years, a bit like Blackberry vs Smartphone. Elegoo will be releasing the Jupiter early next year, with a build volume of 277 x 156 x 300 mm. This is creeping closer and closer to a 'standard' FDM build volume and, while the cost is at present relatively high, it will only ever come down. But in any case, I doubt the majority of hobbyists have room for giant models and the challenge we face is reproducing detail at a very small scale. The main problem with resin - fragility - is less of an issue once the parts have a bit of substance and, conversely (in my view), resin is much, much better for thin or very small model parts than FDM, so long as brittleness isn't an issue. And I expect resin manufacturers are beavering away at improving this aspect of resin for consumer level products. Personally, I wouldn't orientate the software to any particular production method unless that was either the point, or I had no choice. For all we know hobby level sintering or suchlike may be just around the corner.

See, I knew someone would see something I'm blind to now! Thanks, I think I'll try that, it should also solve the dowel hole issue and save me redesigning that aspect. There's some important detail on the outer face but there are ways around that.

I didn't explain this very well and should have used photos; The 'pins' side is the underside of the print and it has LOADS of supports. This face doesn't show on the model so finish doesn't matter much. This one is partially cleaned up and of course in practice this is relatively easy but unscientific. I don't think there's any way around this, this face has to be the underside as I need the A grade finish on the topside. The side piece has to be printed with supports on the face shown for the same reason. However, as the dowel sockets are on the side, their faces are perfect and, in engineering terms, this face comes out as designed, except that the dowel holes lose integrity as the tops are unsupportable. A ledge system (which I'll try this evening) will yield a perfect mating surface on the main body (but a rough top edge, which doesn't matter). But the mating face on the side piece will be an underside so imperfect. I can think of a few convoluted ways of getting two perfect mating surfaces which I doubt I'll even play with, they are that daft, but maybe someone here will point out something obvious or clever - it's easy to miss the obvious when you have your head buried in these things.

Yes, I've been thinking along those lines myself. I should have made clear, I can of course sand down the bumpy resin face very easily and the problem itself is not that problematic - the quest is for the best solution in the context of resin printing. Ledges, which probably is a better approach than pins for all sorts of reasons, still doesn't address that one mating face will be as smooth as the proverbial while the other mating face will still be a spotty teenager. Though easier to smooth accurately than the dowel method. The grail is to somehow get two perfect mating faces.

Can I pick the brains of this community. I need to join two resin printed parts as a T-joint. The main body is printed such that the jointing face is supported, meaning I lose some precision. The edge of the part to be 'T'ed is the side of the print so nice and sharp. I have very little tolerance to play with, 0.2mm max, otherwise it adversely affects the fitting of various dependant parts. As per the screen grabs, currently I'm using a dowel approach but this is unsatisfactory as the dowel holes get distorted during printing and re-boring tends to shift the centre by maybe 0.1 - 0.2mm. I'm scratching my head for a better approach and just thought I'd see if anyone has suggestions. This is the assembly, the two sides get joined to the main plate. The 'webs' are actually floors and are part of the sides. A close-up of the dowel method. The 2mm dowel is part of the main plate, the socket part of the side.

Dowel/kebab sticks will be my fallback for the spars if resin doesn’t work, but I’m fairly sure the masts themselves will go well in resin as there’s more meat on them and lots of room for a stiffener.

I found a solution earlier this week and will explain it with some pictures presently. The shrinkage itself isn’t a problem, in the sense that it doesn’t matter that the overall assembly will be about 0.5mm smaller than I intended; my design dimensions were in any case as much my choice as determined by the original kit and, so long as the whole assembly fits nicely together and to the hull, all will be well. There are about 30 ‘big’ parts to the assembly (and loads of smaller bits) and the shrinkage was preventing these from fitting together properly. I completely agree with what you say about the toolmakers, Daniel, that was part of my job in days gone by (not in modelling though) and I would have been very, very proud of the results if I’d made this kit. Especially the feathered edges of the side galleries and the rebates for the wales and rails, that is exceptionally fine work. I’ve settled for ‘close enough’ on my version and know it’ll need a little bit of putty here and there. Bill, the problem with resin printing small tubes is that you create a ‘suction cup’ in the centre during printing. Air gets trapped between the cured part and liquid resin which plays havoc with the centre hole. And when the tubes are this small, I’ve found that the hole gradually gets closed in anyway. If I had a lathe I could of course just bore them out post printing. But I don’t and, if I did, it would be miles easier just to turn them as solids from wood or steel! So, I tried making the centre of the main lower yard in 14 (!) sections that dowelled together. Nothing magical or scientific about the number 14, I just wanted to see how it went if I made lots of short, stacking tubes that I could bore out back to size by hand. I especially wanted to see if the inherent brittleness of resin meant the yard would crack on bending, bearing in mind the CF core. It didn’t, I could bend it way, way more than I expected without any damage. But it would need a fair bit of filling and smoothing of the joints - not that this would be a problem. Although I’m far from drawing any firm conclusions I think it’s quite possible to go down this route. My next experiment will be to try two halves, exactly like the kit, and see how that comes out. Seams around the ‘barrel’ are much easier to hide than those that run along the length, but I’ll try it anyway. In fact ‘round the barrel’ enhances it in my view, makes it look more handmade than machine part. Then see if either method works well all the way through to the very thin spars, where there is only enough room for a 1mm CF stiffener. The great challenge is that, once again, you’re dealing with super-thin wall thicknesses and the fragility of resin becomes an issue. It would be such a heartsink to be halfway through rigging only to have a spar literally snap off on you and you can bet it would be right in the middle of a complicated web of ropes!

Bill, did you shape the dowels or did they drop straight in?

Impeccable, Daniel. I have yet to tackle the entry port but doubt I will have the patience to do it at this level of detail. What software are you using for the scrollwork?

It is indeed, you can scale separately on the x, y and z planes, which is where the fun begins. As just about everything is curved, when you stretch one plane the relevant radii change. As before, not by much, just enough to be a royal pain. I did some experiments with design changes yesterday to (literally) give more wiggle room, that gave promising results and I’m sure I’ll find a workaround. Two steps forward, one step back.

Oh the frustration! I thought the very slight tweaks I had to make were just a need for a fraction more clearance here and there, that I was being a little too ambitious in my quest for a perfect fit. I’ve learned that it’s actually because resin shrinks during curing; not by much, average about 0.5%, which in 99.9% of instances wouldn’t make the slightest difference. But over a 130mm span that’s about 0.6mm, whereas my design tolerance is 0.15mm. So, I’m now chasing my tail with re-scaling parts, only the shrinkage isn’t even and the amount varies according to the mass of the part in question. I must have done something truly terrible in this or an earlier life to deserve this torture🥵.

Agree, reasonable price and it looks like a lot of content for the cost. Are you doing the ‘Rhino 3D Tutorials, from beginner level to advanced level’ module? It took me quite a while to find the right ‘tutor’ for F360 (Kevin Kennedy), traversing everything from superhero show-offs to one who seemed to know less than even me! I installed Rhino yesterday, will hopefully find time to start learning within the next couple of weeks.

I think Daniel would be more than able to do that himself if he wanted to, and probably make it look a lot easier than any attempt made by me. Bill, this probably isn't hugely helpful but, if you had a mind to do so, you could probably carve replacement parts from evergreen. Whatever you do they'll be an absolute pain to paint.

Thanks Pete. Still fiddling with printing, not too much more to do now.

That’s another good idea, won’t work in this situation though as these are parts that lie flat on others, just helps to have position pins. To answer the original question, they work okay, just need to run the sockets all the way through as some become suction cups and become filled in. I’ll cover it in more detail presently on my Victory log.

I'll be able to tell you later this evening, once the print run has finished. Your system sounds interesting, can you share a sketch or photo? I'm not sure I'd be able to work in clamping pressure, a lot of these parts are not even 1mm thick and are extremely fragile until glued together.

It's there on mine Bill, on sprue 8, the number is on the back side. Very small part. What is it with Heller, did they see no virtue in locating pins!

Yes, that's more or less me as well. It's particularly useful right now, when I'm also needing to experiment a lot with the printing. For instance I've found that 0.5mm locating pins (out of necessity, space is very limited) are perfectly printable but they break off too easily when removing supports. So I've just gone back through every item in the assembly, swapping to rectangular pins, which give them a lot more meat. Although this is fairly straightforward, down the line I'd know I need to design around that from the outset, but right now I appreciate how easy this it is in F360 to make a change like this. It may be that I continue using F360 in parallel to SW or Rhino, just as a tool that's good in some situations but not all.

Afraid I can’t help you there Bill, light years away from that stage. What is making it so difficult with CA?

Having history has allowed me to design on the fly - not good practice, a complete break with my training and not always even sensible, but it is kind of handy when you’re not exactly sure how things are going to need to fit together or intersect. But I’ll be the first to admit you can instead pre-plan. History can be problematic in F360 as well and I know some people don’t use it. I think so long as I’m able to retrace my steps it’ll be fine.

Thanks Richard, sounds good. I guess I'll have to see, when I try it, what it's like not using a design history, how you can go back and tweak things if you need to.

Forgive the questions that would be answered by simply downloading and getting stuck in but I’m waiting until I have a clear desk before starting the time limited free trials. 1. Do both Rhino and SW have the design history function like that of F360? Can you easily re-order the sequence of commands like you can in F360? 2. I’ve read that Rhino can be a PITA for complex filleting - true or false? F360 can be hit or miss there too. 3. My biggest gripe with F360 is how it handles ‘failed’ commands I.e. when it can’t do what you’re asking of it; too often it locks up for a very long time and ultimately crashes. What happens in Rhino and SW? 4. How do Rhino and SW compare to F360 in terms of the graphical representation of the model? I have to say, I like F360 in this respect. 5. How do they compare in terms of processing speed? F360 can become very, very slow as the model becomes complex and the timeline grows. 6. Ive read that you can import / export between e.g. Rhino and F360. Does that mean you can move it in and out of each package for different stages, or does it become a mesh? I thoroughly dislike the import mesh function in F360, for me it’s a last resort. 7. Are there particular functions of either Rhino or SW that would be naturally important to ship modelling, that perform poorly? Things where, like in F360, you learn through experience to do it this way rather than that, even though the latter might be the more obvious path? For instance, projecting to a surface in F360 will often require that you edit the projected lines afterwards, in order to be able to surface them. (Sometimes the projections don’t join properly). Or using pattern - almost fatal to pattern in the sketch, fine on the model, so you learn to make the model of the pattern and project back to a sketch, if the sketch is what’s needed. At least that’s how I do it. 8. The discussion so far implies there’s very little forum-type help for Rhino. With F360, when I’ve got stuck I have usually been able to get free advice fairly easily from other users. What is that like for Rhino and SW?

I was being a bit tongue in cheek there, falsely implying that drawing one line is no more difficult than another. If only it was so. Though perversely, sometimes that's not a bad mindset to have, it sees you having a go at something you might otherwise shy away from. I have a careers worth of succeeding because I knew no better! It's funny how age and experience see you become less (over-)confident, isn't it, you'd think it would be the other way round. I think I'd treat any transposed drawings as no more than starting points, sometimes good, sometimes bad. Back in the day, the drawing office would sometimes give us full size plans and try to persuade us to just dot punch through the drawing. Never happened. I'm not so sure about the scale effect, even at 1:100 the eye picks up all but the smallest anomaly. And if I went on to model the Cutty Sark I'd probably go for slightly larger than 1:100, maybe 1:75, just to gain a few fractions of a millimetre on very small parts and detail and give the printing a better chance. Turns out my little bit of un-fairness is happening in either the slicer or printing. No hint of it in the soft model. Printing is the chief suspect, that it's developing a bit of a warp that gets corrected further up, and needs some strategic support.