A bit of 3D Printing.

Discussion in 'Workbenches, including workshop techniques.' started by JimG, 7 December 2018.

  1. JimG

    JimG Western Thunderer

    I've been thinking about doing my own 3D printing for a while but didn't want to get too involved in it while I was getting the switching layout into some reasonable state to exhibit in November. But once the exhibition was over I ordered up the machine I was interested in, the Phrozen Shuffle

    Phrozen Shuffle (XL) Review - Stunning Prints, Stuffy Workflow | All3DP

    I followed various of the groups on Facebook dedicated to the less expemsive 3D resin printers and, although the Phrozen was a bit more expensive than the others, chose it since it seemed to have fewer problems than others and it also had a supplier in the UK who also used it himself in his jewellery business and was providing good support to users.

    The first results were good. I had done a 1:32 scale file for a Caledonian axlebox and that became the guinea pig.

    3DPrinting-001.jpg

    The file was set up in the slicing software at various orientations since adjusting the orientation can often improve the surface finish but in this case the flat orientation proved best. The closer strip has 30 micron layers (approx 1.25 thou) and the far one has 50 micron layers (approx 2 thou layers) and I couldn't see much difference between them. You can just about see the embossed "CR" lettering on the closest samples which got itself mirrored in my messing around in the slicing software. :) The resin printing process requires special base layers so that when the part is held upside down in the printing process it adheres well to the plate on the machine. The parts are held off this layer by the thin stalks (supports) and these can be generated automatically by the slicing software.

    The next test was to half the scale of the axlebox and print it in S scale...

    3DPrinting-002.jpg

    ...which came out quite well but the detail - like the lid on the axlebox - was almost lost since it was a bit too fine for the process. The resin has a minimum wall thickness of 0.4mm and some of the detail on the reduced scale file was finer than this. The "CR" lettering is still there, and readable - you can just about see it on this picture. The main lesson from this exercise was that re-scaling might not work well for fine detail. Detailing for smaller scales will probably have to be a bit overscale to survive in the process and if these files are up-scaled, the detail will probably be too heavy in larger scales.

    I also had some S scale files to hand given to me a while ago by Justin Newitt, so I popped them all in one large sliced file to see how they looked...

    3DPrinting-003.jpg

    ..and they came out really well - a selection of BR axleboxes and springs. These, and my S scale ones above, were printed with 30 micron layers and these looked really good. Justin had drawn these as masters for casting hence the sprue work.

    I'm really pleased with the results. The quality is excellent and will be more than good enough for most S scale detail parts. I will have to draw up some chimneys and domes to see how they fare under the process.

    It is not a fast process and the above prints took between a half and one and a half hours. The main factors are the height and the layer thickness. The process exposes the resin to UV light through an LED screen so the horizontal area of the print has no effect on the process. The printer has a small horizontal area - 120mm x 68mm, but has a print height of 200mm so parts longer than 120mm can be printed by standing them on end or angling them towards the vertical axis.

    Just to show that all did not go well...

    3DPrinting-004.jpg

    ... a print that was sliced at 30 micron layers and printed with the printer set to 50 micron layers. :)

    Jim.
     
  2. Steph Dale

    Steph Dale Western Thunderer

    Hehe, we all have days like that ;)

    But, more seriously, those are impressive prints from what's a relatively inexpensive machine. Well, the good ones are anyway!

    Steph
     
  3. JimG

    JimG Western Thunderer

    The resin doesn't photograph well - like styrene - so I'll try and give the prints a light scoosh of grey primer tomorrow to make them more photogenic and show the quality better.

    Jim.
     
  4. JimG

    JimG Western Thunderer

    I sprayed on a coat of grey primer to highlight the details and apologies for forgetting to clean gthe parts first. :)

    These are the 1:32 scale prints...

    3DPrinting-005.jpg

    ...with the mirrored lettering, :) and a very big closeup of the nearest one...

    3DPrinting-006.jpg

    The main thing to note is the chording on the curved bottom of the box and I might be able to reduce this in 3D CAD if there is a setting to increase the number of chords in a curve. The corner of the lid has also lost detail at the closest corner and I might have to beef up the detail to avoid that. But otherwise there is pretty well no evidence of layering on the sides and the lettering is quite crisp.

    Now the S scale box...

    3DPrinting-007.jpg

    The chording is less evident but the detail on the lid is not perfect and a thicker lid would probably be required. The lettering is still just about readable but it is very small and is only noticeable under magnification so I suspect that I will have to increase its size a bit to make it obvious - a bit of experimentation is required. :) Like the larger scale print, the layering is not obvious. To give some idea of actual size of the S scale box, the curved base is just under 3mm diameter.

    And finally Justin Newitt's BR files...

    3DPrinting-008.jpg

    Axleboxes above and springs below

    3DPrinting-009.jpg

    These have printed very well and there is no evidence of layering.

    All in all, I think the results so far are excellent and would stand close inspection by the naked eye. You really need at least an Optivisor or a macro lens to start showing up the shortcomings and I'm not sure if the finish of my quick scoosh of Halford's grey primer might not also be open to question. :)

    Jim.
     
    Rob Pulham, unklian, TimC and 10 others like this.
  5. Dikitriki

    Dikitriki Flying Squad

    Very impressive Jim. you've really hit the ground running....and I love the word 'scoosh':)

    Richard
     
  6. JimG

    JimG Western Thunderer

    Glaswegian/Clydeside term for a small spray, and also applied to any type of fizzy lemonade since that's what usually happened when you opened a bottle after carting it for miles to something like a picnic. :) Glasgow Fire Brigade also used the term to describe their first hydraulic boom foam disperser - the Big Scoosher - TheGlasgowStory: Glasgow Scoosher. :)

    Jim.
     
  7. JimG

    JimG Western Thunderer

    Onwards and upwards with Fusion 360. The most appropriate word for my experience is probably "exasperation". :) I get the feeling that the software is really still in BETA form and that the users of the free version are their BETA testers. But I'm making some headway and finding out the blind alleys up which I should not go. :)

    I re-visted the Caledonian axlebox and adjusted the curved face of the box to make it more evidents and the lid was thickened to get within the minimum thickness of the 3D process.

    3DPrinting-010.jpg

    Here are a set of S scale axleboxes and the matching springs. The spring next to the axleboxes has the Drummond combined spring mount and single lever pivot point. Three of the 1:32 scale axleboxes are to the right. The prints are still mirrored and I have now found out that the slicing software defaults to mirroring in the X axis for some reason, but does allow you to set it to print normally.

    3DPrinting-011.jpg

    Another print of the axlebox in S scale and 1:32 scale. You can just about detect the layering in the side of the larger axlebox. The print layers were 30 microns, or 0.03mm.

    3DPrinting-012.jpg

    A drawing I had done some time ago was of the distinctive Drummond wagon buffer stock. This one is in 1:32 scale and a wee bit of layering is apparent on the main body.

    3DPrinting-014.jpg

    I also scaled them down to S and that worked really well with no layering obvious. Again 30 micron layer were used.

    3DPrinting-016.jpg

    This was another test print to assess the optimum angle for the best results. My reasons for going for an angled positioning are twofold - a comparatively large surface area parallel to the building plate will tend to retain excess resin on that surface during the printing process and that resin will start to set with UV overspill from the LCD. The printing process involves lifting the work piece then dropping it back down onto the LCD screen just allowing a space for the layer thickness for the resin. When the UV sets this layer, it tends to stick to the protective screen on the LCD. Angling a piece can cut down the area in contact and therefore the force required to unstick it. I suspect that a few of my distortion problems have been due to the piece moving on the supports when unsticking it. I'm now using larger supports as well to get round this. 3DPrinting-017.jpg

    The last picture is hot off the press - or printer. :) These are test parts for printing Gauge 3 Mansell wagon wheels for Jonathan. The two on the right were printed flat with no orientation and there are problems with distortion of the plate - witness the almost non-existant nuts on one side, and the rear of the discs has a lot of stray resin baked on. The ones on the left were done with 15 degree orientation and they have come out very well - the lower one is the inner side with the nuts and the upper the outer side with the domed bolt heads.

    My problem now is getting the seams between the wood blocks to show. I started off with 0.05mm gaps on the right hand ones and th0se barely show - you can just about see them with an Optiviser. On the left I opened up the gap to 0.1mm and it is still pretty well invisible. :( I'm going to do some tests now to find out what I have to do to get an obvious join line. I know that the real thing probaly had hairline joins between the blocks, but we modellers like to see quite obvious lines in our scales. :)

    Jim.
     

    Attached Files:

    cmax, Dave, GrahameH and 21 others like this.
  8. Rob Pulham

    Rob Pulham Western Thunderer

    They look superb Jim.:thumbs:
     
  9. JimG

    JimG Western Thunderer

    I've had another go at the CAD drawing of the Mansell wheel discs to see if I could make the wood joins a bit more obvious. I did it by applying a chamfer to each side of each block - well on one block then placing that block in a circular array of sixteen, The outcome looks not too bad although I was surprised at how much chamfer I had to apply - o.15mm.

    3DPrinting-018.jpg

    The discs have had a quick scoosh of Halford Grey Primer to photograph a bit better. They have also been lit quite obliquely to bring up the detail. With flat frontal lighting, the wood joins almost disappear from view. :) They were printed at a fifteen degree angle to minimise printing distortions.

    Jim.
     
    Ian Smith, chrisb, Dog Star and 12 others like this.
  10. Mike Turner

    Mike Turner Member

    Hi Jim

    The bolt arrangement looked odd to me so I googled Mansell wheels and on all the examples I looked at the outer ring of bolts are centred on the blocks rather than on their edges. There may of course be other types and this isn't a criticisms just an observation. As ever inspiring stuff!

    Regards

    Mike
     
  11. JimG

    JimG Western Thunderer

    Mike,

    You are correct!!. I've just checked back on my source information. I will redo them when I get back from New Year holidays. But these might do for fitting tests for Jonathan in the meantime.

    Jim.