"A Wrinkly in 3D Printing Land", a tale of an OAP Vs 3D printers

Discussion in 'Workshop Benches' started by Jim Freight, Mar 9, 2023.

  1. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Thanks :)
     
    Vinylelpea likes this.
  2. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Typo in episode 4 "Model Preparation and Printing" now corrected, "Cura 5" was incorrectly "Cura 2" :facepalm:
     
  3. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    8. Internal Spool Hook

    XYZ filament products are sold on spools which are loaded into one of their cartridges which fits into this machine, it comes with a micro-chipped card containing details of the filament which the printer can read and also update the estimated length of filament left on the spool.

    However, XYZ Printing provides an open filament system for the Da Vinci Pro 3D printer which means that you can also use 3rd party filament products as well.

    3rd party products supplied on 500g spools can be mounted in the place where a cartridge fits with the aid of a 3rd party spool hook, typically a printed one which I downloaded as a STL file from Thingiverse. Various ones are available, I chose this one :-

    https://www.thingiverse.com/thing:2657764

    This I considered as a good way of printing something larger, useful and trying out Thingiverse.

    Exercising caution whilst finding my feet I will stay with XYZ filament products to remove a variable from the equation as I learn how to create decent prints.

    1st Attempt

    So I downloaded the STL file, sliced it with XYZware and sent it to the printer where the process kicked off, image shows the file loaded prior to slicing.

    1.jpg

    However what I realised after a while, printing it like this without any support for the horizontal arm was going to fail spectacularly, the horizontal hook for the spool had printed fine and the printer was working its way up the vertical arm, whoops :redface:

    Rather than just stopping the print and binning the job so far I let the print continue half way up the vertical arm and then cancelled the job, at least I had a usable bracket for another purpose. The curved base printed well without any supports and a quite reasonable finish, certainly functional.

    At least I can screw it to a wall :giggle:

    2.JPG

    2nd Attempt

    This time I laid the design on its side, so no overhangs, also trying Cura I added the support option to cater for the curved face of the hook next to the print bed.

    Printing it standing upright with supports would raise the print duration considerably due to building an elaborate mesh akin to scaffolding to support the top hook so I decided just to lay it down on one side, so support just for the actual spool holder.

    It was then written out of Cura as a G-code file, followed by using XYZware to transmit the G-code file to the printer, not possible directly from Cura as printer control via Wi-Fi by Cura uses an incompatible to XYZ protocol , seems there is not a common standard yet, ho-hum.

    This printed fine, but I suspect I did not really need the supporting framework on the side of the hook, which although it parts easily from the main body of the print it did not leave as clean a finish as it was before.

    Filing the surface creates dust which is statically charged which sticks to the item and needs a damp tissue or rag to remove it.

    3.JPG

    Lessons to be Learnt Here

    1. Think 'overhangs ?' before you print.

    2. Especially while learning keep an eye on progress to avoid a huge mess of wasted, tangled materials and the time to clean up a nozzle, with this print job I realised in the nick of time.

    Anyway I have printed the part, it seems to fit, I'll check for sure when I next unload the current cartridge.

    Jim :)

    Back to episode list
     
    Last edited: Mar 15, 2023
    Rob Pulham likes this.
  4. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    9. Dublo Coupling Mounting Block K5

    A long time ago, 1977, I was building some white metal wagon kits by Keyser and ABS Models, then I went to uni and never got around to finishing them, but I kept them with the intention of completing them one day.

    Over 45 years later, :eek:, the student, now OAP has a way of completing them which earlier on was going to be rather tedious. Those kits were designed mainly for the fitting of 3 link couplings so unlike many plastic kits of the time the vans have no floors and the open wagons a flimsy floor.

    1.JPG

    This block is named K5 as that is the label on this van, I have 3 of these (K5 to K7) and a 4th of a different type this block will fit.

    So what's the problem, well the problem is that I wanted to fit the large plastic Hornby Dublo couplings and fabricating the necessary mounting was rather laborious. Those couplings were disliked by many Dublo enthusiasts due to their bulky nature compared with the earlier metal design but IMO they operate much better both for coupling and staying coupled in modest length trains than the metal ones ever do.

    It is intended to run these very heavy wagons with my vintage fleet so fitting these couplings is preferred, NEM socket adapters are on my long list of 3D printing projects.

    Development of the shape of these blocks started with measuring up the wagons and an example Dublo wagon to obtain the relationship between the mounting of the coupling and the necessary end stops to limit the radial travel of the coupling.

    Further measurements were required to set the coupling the correct height above the rails, correct pivot distance relative to the buffers and to ensure that the blocks did not foul the wheels. The latest version v8 now meets all the requirements.

    As with all items so far ABS filament is used, currently in green which is very useful for prototyping and marking with a permanent marker, when I print these blocks for actual use I will print them in black ABS.

    The drawing below is of version 8.

    2.jpg

    The couplings are retained by self tapping screws which enter the block via a slightly larger opening diameter of 1.8mm to align the screw than the remaining hole through the block of diameter 1.5mm.

    Image below is v6, 10% infill

    3.JPG

    The nylon top hat bushes are slightly too long and allow a bit too much coupling droop, I did try printing some in the ABS but initial attempts failed because the nozzle diameter of 0.4mm produced a wall that was too thick.

    Cura 5 is apparently able to produce extrusions narrower than the nozzle diameter, I will evaluate that shortly.

    As nylon is difficult to cut I decided to make a circular recess where the bush sits to take up some of the excess height resulting in v7 shown below. Note that the top surface around the coupling mounting boss has gaps but okay where the coupling mounts, good enough for prototyping.

    4.JPG

    Version 8 uses 50% infill as a step towards the 'production' version, the screws are a tighter fit as would be expected with the higher material density, however it is imperative that they are first fitted without the coupling to ensure they are perpendicular so that the top hat washer sits evenly against the face of the coupling. Otherwise coupling droop will occur.

    Shown loosely in place, the final version will be printed in black ABS and glued into place from the inside.

    5.JPG

    Again loosely in place but with couplings fitted and coupled to a Dublo Mica van.

    6.JPG

    7.JPG

    8.JPG

    Lessons to be Learnt Here

    1) Prototyping in low density speeds up printing, reduces energy used and material costs.

    2) Step by step refinement is required when fitting to an already built wagon.

    It is certainly easier to prototype in a light colour, it would have been rather more difficult to mark a black component where it required adjustment.

    Next I will test van K5 on the layout with these blocks tacked in place.

    This has been a successful development and has recovered these wagons from being scrapped, when the weather improves I will give the bodies and roofs a spray coat of primer. Black mounting blocks will be glued in after painting but prior to fitting the roof.

    Jim :)

    Back to episode list
     
    Last edited: Mar 18, 2023
    Rob Pulham, Kimbo and Vinylelpea like this.
  5. Vinylelpea

    Vinylelpea Full Member

    Messages:
    760
    Likes Received:
    512
    Joined:
    Dec 22, 2017
    Great little project. Those vans are well worth saving. :thumbup:
     
    Jim Freight likes this.
  6. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Thanks, I'd got rather attached to them and they provided an interesting exercise in using this 3D tech, next to save is an open wagon from the same batch from '77 using a variant of this mounting block.

    The intention is to use this K5 block as a starting point so I can transfer the geometry of the coupling mounting to a different wagon mounting arrangement, re-use reduces development time in a hobby as well as in industry.

    Jim :)
     
  7. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    10. Dublo Coupling Mounting Plate K12

    This version was to be fitted into a soldered white metal open wagon (K12) part assembled in the 1970s but unfinished due to studies, life and the tricky nature of fitting a plastic Hornby Dublo coupling to white metal rolling stock, mentioned in the previous posting on vans.

    It also provided an exercise in which I took a copy of the K5 block model and created this K12 mounting plate which would be mounted as part of the wagon floor, which would be plain and covered with a thin planked sheet for appearance.

    However the process did not quite go to plan.

    The drawing below shows the design.

    1.jpg

    The mounting for the coupling uses the same geometry as for the K5 block but cut back at the buffer beam end due to internal ribbing. The printed plate would include half of the wagon floor, the join being disguised by a cosmetic floor.

    1st Attempt

    As the floor part was only 1mm thick I chose to use a layer height of 0.1mm and 50% rectilinear infill. As before, printing in ABS with the da Vinci printer using its 0.4mm extruder nozzle.

    This printed as shown below, a screw has been test fitted to this part.

    2.JPG

    Although the plate was printed flat away from the block, at the block it curves up at the front edge, including the block, neither of which affect operation. But the coupling end stops are deformed near the top, not good.

    The curling up and rounding of the block may be due to overheating with multiple 0.1mm layers as opposed to the default 0.3mm layer height used before.

    Similarly for the deformed coupling swing end stops, I have read that printing multiple items at once gives the layers greater time to cool without slowing down the x-y motion of the extruder between passes. The larger surface of the plate is quite flat, only slight bowing present and flexible enough to be glued in place to the wagon chassis.

    As I was fitting this printed plate, oh dear, two of the wagon sides came apart, a joint failed. On close examination the soldered joint was all fillet and did not penetrate through to the mitred edges of the wagon body. The remaining 3 joints were just as weak and broke easily.

    Soldering white metal well requires more skill than I could muster then and have no wish to spend time on learning now. When I bought these kits there was very little variety available r-t-r, but now the variety is enormous and I now have too many, so the incentive to overcome this issue has evaporated.

    Using adhesives was more difficult IMO as the metal is difficult to clean well enough and due to the loose fitting of some parts required an adhesive with body to take up gaps.

    I digress, back to the 3D printing aspect.

    2nd Attempt

    As a comparison the print was repeated with a layer height of 0.25mm (the floor part is 1mm thick) with the 50% rectilinear infill as before.

    This printed as shown below, with the first shown alongside for comparison.

    3.JPG

    This one has printed far better, there is none of the distortion present on the first attempt and the plate part is flatter, albeit the surface a little rougher, coarser texture, however that does not affect its functionality and the surfaces are not visible in use.

    As before the 'production' part will be printed in black.

    Lessons to be Learnt Here

    1) Choice of layer height can make quite a difference to the functional quality of the printed part especially when it comprises of two very different thickness of cross-section.

    2) Frequent passes of the extruder before sufficient cooling has occurred roughens the surface finish and can distort the form of the part. One approach to avoid this is to print multiple items together so that a longer cooling time occurs before applying the next layer to any one printed part.

    The coupling swing end stops were the highest features of the part so the extruder just hopped between them for most of their height, whereas traversing to a neighbouring part would allow the first one to cool off longer.

    I believe it is possible to set different layer thickness at different layers with a slicer that supports that functionality, this needs confirming, the XYZware slicer does not appear to support that. I need to look closer at Cura 5 and IdeaMaker (which is supplied by Raise3D).

    In this case the second attempt generated a part fit for purpose, pity about the wagon!

    Jim :)

    Back to episode list
     
    Last edited: Apr 9, 2023
    Rob Pulham and Kimbo like this.
  8. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    11. Dublo Coupling for NEM Socket

    My railway has quite a mixture of rolling stock from various eras and manufacturers, the two predominate couplings are the Dublo early plastic couplings (EPC), Trix/Liliput, Tri-ang Mk3 and contemporary hook and bar (NEM socket type).

    Although for shunting I can readily mix EPC type and hook and bar it is not suitable for running a train of wagons, I have enough fun with the multitude of hook and bar couplings not co-operating at times.

    I have many converter wagons based on the dual coupling mounts available on Trix/Liliput and Wrenn wagons, however a NEM socket fitting version would be useful too, especially for some contemporary locos.

    To design these NEM fittable Dublo EPC couplings the task needed to be split into three parts.

    1) The clip end which fits into a NEM socket
    2) The coupling end which in this case is a Dublo EPC
    3) Combining the two parts to suit the vehicle, e.g vertical alignment and required arm length.

    In this way I can blend together the appropriate height and length settings of the standard clip and the desired coupling, i.e. the clip has it's functional geometry and the wagon coupling it's functional geometry.

    The possible fly in the ointment is whether the EPC will have enough radial motion available from the flexible connection between the NEM socket and it's dovetail mounting to the wagon or loco chassis for it to operate properly. That issue will be solved later.

    First the NEM clip part.

    11.1 NEM clip

    Dimensional information for the NEM 362 socket is readily available online, this was used in conjunction with an example coupling to obtain the dimensions to create a basic clip with shoulder.

    Although in German the drawing and dimensions are easily read :-

    https://www.morop.eu/downloads/nem/de/nem362_d.pdf

    An issue with SketchUp occurred when designing full size was that the radiused bottom of the vee could not be drawn at the required radius of 0.4mm, (less than 0.5mm). Online there are discussions that this is deliberate by the SketchUp designers to prevent users trying to create unachievable features when 3D printing, while some posts consider it to be more of a bug.

    There are various workarounds apparently but what I did was to create a 2D circle of 0.5 mm radius and then use the offset tool to reduce the radius by 0.1 mm, delete the first circle thus leaving the required circle of radius of 0.4mm.

    The resulting design :-

    1 Clip Beta.jpg

    Printed with a 0.4 mm layer height, 50% rectilinear infill, 0.4mm nozzle in ABS the resulting test pieces printed quite well, finish not great and I suspect a higher level of infill, probably 100% for a whole coupling would be desirable.

    There was slight stress marks on the outer edges when flexing to insert the circular ends into the socket on a couple but they held well and unclip as required a few times without failure so a promising start. Printed part in green.

    2.JPG

    The tee shaped head is there to mark the position of a shoulder required to ensure the clip locks into the NEM socket, this will be modified as necessary when a coupling hook is fitted to it in SketchUp.

    Continued in 11.2 below

    Back to episode list
     
    Last edited: Apr 29, 2024
    Rob Pulham likes this.
  9. Chris M

    Chris M If 2 wrongs don't make it right ... try 3 Full Member

    Messages:
    156
    Likes Received:
    183
    Joined:
    Mar 14, 2019
     
  10. Chris M

    Chris M If 2 wrongs don't make it right ... try 3 Full Member

    Messages:
    156
    Likes Received:
    183
    Joined:
    Mar 14, 2019
    Hi Jim,

    I use SketchUp for all my 3D models. I have tried a number of other CAD programs (FreeCAD, Fusion 360, Blender) but I like SketchUp the best. I used to be a draftsman and SketchUp allows me to create a model in much the way I would draw it by hand.

    That being said it does have its limitations which often need a workaround. One of those, as you have found, is small circles (definitely a problem if you use the 'follow me' tool). In cases where this is an issue I scale the model up 10 fold (or more). This usually allows everything to draw correctly, when finished I scale the model back down to the correct size.

    With parts like your 'clip end' where the fit is critical I usually print them slightly thicker and then sand them to the exact thickness. This also smooths out any printing imperfections.

    This is a great series of detailed posts and I am sure it will help members who are getting into (or thinking of getting into) 3D printing.

    Chris
     
    Rob Pulham likes this.
  11. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Thanks Chris.

    In my first life I was a draughtsman too, back in the days of draughting machines, plastic film and crayon type pencils of the 1970s but I found 3D modelling with SketchUp quite difficult to grasp initially, just a mindset stuck in 19th century techniques despite my later digital systems career path. I am hooked now, the free version easily meets my needs and workaround seems to be the name of the game with software of all sorts.

    The shapes of the clip end and especially the coupling I found easier to form as a two dimensional surface and then use the push-pull tool to create the 3D model.

    Blending the two parts together went well except at one point SketchUp crashed with it's BugSplat pop-up, that always happen when you don't backup up often enough, there wasn't even a recovery option to restore I could see from the last automatic save, oh, well more practise at editing!

    I have found the clip end can print well enough, although my latest prints have become 'hairy' and a bit malformed which is I suspect is dampness creeping into the ABS filament, so now I am trying out a filament dryer, and that was a long search backwards and forwards through YouTube videos to try and find (hopefully) the right device, I will know shortly.

    A brief write up on looking for a dryer will follow as it is another minefield with some dryers being very expensive, some of uncertain effectiveness, and both :whatever:

    Jim :)
     
    Rob Pulham likes this.
  12. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    11.2 EPC Hook

    This was rather more difficult to draw up as it is virtually all curves and was measured up from a Wrenn example, I have omitted the dropper as I do not use uncoupling ramps.

    I started with a basic rectangle for the stem and another with the overall dimensions of the hook, built it up first with the line tool, added some arcs and whittled it to look right, holding the example close to my eye and looking and superimposing against the laptop screen in the background.

    3D printing is an additive production process, machining techniques are subtractive (e.g. drilling, milling), what I tend to do is start with a 2D design, create the plan view, first with rectangles, and then shape them, and save as a 2D model. Then the push-pull tool is used to thicken it to the appropriate thickness, I save that as 3D model and also export it to the .stl file.

    Generally I have used a subtractive process to design and followed by the additive process to print.

    Circles can be created to add cylinders and be the template for creating a circular hole by pushing them perpendicular through a 3D model or creating a pillar by raising them perpendicular to the surface they are applied to.

    On the 7th iteration I managed to get a good match such that when the couplings are in tension or compression the shafts remain inline, well, pretty close!

    The dimensions shown are what resulted from the final shaping of the hook and were added after the design was created, as such they are for reference purposes only.

    Eventual design :-

    3 Hook Beta 7.jpg

    Printed with 0.3mm layer height, 100% infill, 0.4mm nozzle in ABS, result as shown below and compared with the Wrenn coupling.

    4 DSCF1210.JPG

    So this gave me a good geometric shape for the EPC ready for combining with the clip developed earlier.

    This certainly took some time to develop, however when working at this size of component with an average extruder of 0.4mm bore there is no such thing as a sharp corner and most parts have rounded edges and sides have ribs, one per layer.

    You cannot expect to achieve the sharpness of an injection moulded part that we buy in our r-t-r models, certainly when using a filament printer. The injection moulding process consists of extruding the melted plastic into a precision machined steel mould under pressure so that the plastic material is forced into all the corners and against the sides, smooth or textured as required.

    Filament based 3D printing on the other hand extrudes the melted plastic firstly onto the print bed and then on subsequent layers onto the previous layer, the extruded plastic is not constrained. Ideally it will adhere to the previous layer but can still move out of line, it flattens after being extruded from a circular nozzle and shows the layer edges as a rippled vertical surface.

    Next, combining the clip with the EPC hook to create an EPC which can plug into a NEM socket.

    Continued in 11.3 below

    Back to episode list
     
    Last edited: Apr 29, 2024
    Rob Pulham likes this.
  13. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    11.3 EPC Coupling Assembly

    Further tweaking of the hook design was made :-

    1) The hook tip shortened slightly to allow better engagement with another coupling.

    2) The shank was removed entirely such that the clip stem connects directly to the coupling.

    The 'final' result for the hook and many more reference dimensions have been added.

    5 Hook Beta 8.jpg

    The first wagon that I am to fit this design to is a Bachmann shunters truck which I use in the marshalling yard. By fitting it to one end it will allow easier use of e.g. my Dublo class 08 shunters to shunt contemporary rolling stock and likewise contemporary shunters to shunt Dublo/Wrenn/Trix wagons.

    In practice the NEM 362 standard indicates the height of the socket above the rail head and also the distance the socket entry should be from the vehicles buffer head. However as you quickly find out there are multiple lengths of couplings sold and to add to the fun, straight and cranked versions too.

    I have also found other issues, e.g. that the Heljan 'Tango' locos had couplings that did not match Bachmann stock coupling height and had to be swapped out before the locos could be used. So much for standardisation.

    As it turns out the combining of the two is relatively easy, the EPC hook can sit on the stem of the clip as below with a reduced head :-

    6 EPC NEM beta 1of2.jpg 7 EPC NEM beta 2of2.jpg

    However this now creates an overhang whether it is printed operational (upright) way up or inverted.

    The overhang requires supports, a light weight structure that can be provided by the slicer software, in this case I was using the options in XYZware, the problem is the parts are small and even the lightweight option in XYZware was rather cumbersome.

    Used the same slicing and printing parameters as for the hook on its own.

    Printed upright, clip on bed, there was still distortion of the EPC hook tip, and when printed inverted the clip was mis-shapen beyond use.

    As an experiment I tried thickening the hook part so that supports were not required as :-

    8 EPC NEM beta2.jpg

    Although this printed without distortion it makes for a rather clumsy coupling, akin to the ghastly IMO typical couplings fitted to N gauge r-t-r stock.

    A more advanced way to create supports is to use more advanced techniques in Cura 5 or to use a two extruder machine, one extruder lays the required material for the part and the second lays the supporting material such as water dissolvable PVA. This I will try when I fire up my Raise3D E2 printer.

    Next I tried a design which would not need supports, in this case gradually develop the change in height at the base of the combined parts as :-

    9 EPC NEM beta 3.jpg

    Did not try printing this, the transition from connection to the clip to the tip of the hook meant it looked just as bulky as the previous version.

    Back to squeezing the most out of the da Vinci, brain fired up for something a little outrageous, which I was not sure would even print let alone be usable, why not print the clip and hook parts separately and connect with a dovetail joint :faint:

    Continued in 11.3 - Part 2 below

    Back to episode list
     
    Last edited: Apr 29, 2024
  14. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    11.3 EPC Coupling Assembly - Part 2

    Dovetailed Parts

    A new copy of the clip model was opened to use as an assembly and the hook imported into it, making sure they were individually grouped such they would not stick to each other. Remember ungrouped items when butted up to each other stick together and distort each other if you attempt to pull them apart.

    The two parts were then aligned with the required overlap one axis at a time, a useful video on aligning items in a 3D space is :-



    Once aligned as required the female part of the dovetail was drawn in position and the offset tool used to create the male part 0.1mm smaller all round, mind blowing optimism abounds that this could actually work when using a 0.4mm nozzle. Anyway I was keen to see if something workable could be made.

    The clip, 2D dovetail outline and the coupling hook were parted and the relevant parts of the dovetail attached to the coupling components to create the male part on the clip and the female on the hook. Once completed the two dovetailed parts were individually grouped and then brought together to confirm correct alignment. This is shown below.

    10 EPC Dove Assy 1of2.jpg 11 EPC Dove Assy 2of2.jpg

    The two parts were then exported to separate stl files ready for printing, 3 of each part was imported to the slicer and processed as for the hook but I also checked the box for auto speed adjustment for small parts.

    The resulting parts before any fettling :-

    12.JPG

    Well, at least something was printed where the male part of the dovetail was meant to be, but more of a deformed rectangle, not forgetting this is with a 0.4mm nozzle! The female part had the semblance of a dovetail, however with a bit of filing on both parts one set just about held together on their own. It was most likely that they would need gluing together anyway.

    All parts were of usable form and so were assembled and bonded with Plastic Magic, two are shown here :-

    13 DSCF1247.JPG

    The third is shown fitted to the Bachmann shunters truck.

    14 DSCF1249.JPG

    15 DSCF1253.JPG

    Finally the shunters truck coupled to a Dublo Mica van.

    16 DSCF1255.JPG

    Early on there was concern that the coupling may not have enough radial swing to operate, however the flexible joint between the NEM socket and its dovetail joint do have sufficient swing. The printed coupling hook and the Dublo hook slid over each other easily with the printed coupling not needing to swivel, the coupling on the Mica did all the swivelling.

    Lessons to be Learnt Here

    1) Stretching the boundaries can yield usable results.

    2) Overhangs are worth avoiding with small parts as removing supporting material can damage your printed part.

    3) Contouring to avoid the need of supports is very time consuming.

    4) A simpler form of locking was probably more suitable than a dovetail joint.

    Overall a successful exercise in grappling with SketchUp, slicer and the creation of useful parts, I will in the near future try and print them with a 0.2mm nozzle on the Raise3D E2 printer with dovetail and single part using water soluble supports.

    During this work I ran into issues with damp filament which I will describe in the next episode.

    Jim :)

    Back to episode list
     
    Last edited: Apr 12, 2023
  15. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    12. Filament Dryer

    There are many on the market in a variety of formats, temperature ranges, user controls and size and looking at specific and comparative reviews is helpful but can also be confusing.

    I am not going to review them here as many YouTubers have already done that for us, here I consider what was important to me for my intended use and then what I chose and how it worked for me.

    I live in the UK where the typical humidity in my house and shed are nominally 50 - 70% dependent on the weather, so plenty of moisture available, particularly in the spring and autumn, temperature 18 to 25C.

    In the Beginning

    Before I got started with filament based 3D printing there was two questions that I had of which there appeared to be no answers as the topics never seem to come up.

    Q1) Should you empty the machine of filament if you are not going to use it for a few days?

    Q2) How often should the filament be dried?

    My answers to these questions after my first 6 weeks are :-

    Answer 1) Since I started using my printer I have only unloaded the filament when I thought there was a printing issue or the filament had symptoms of dampness, so at least 5 weeks without any issues such as a blocked extruder nozzle.

    Answer 2) This is much like how long is a piece of string, it depends on the material, the working environments humidity and how long the filament has been exposed to the environment.

    Filament State

    Moisture is absorbed by plastics, which we initially believe to be water proof as it is often used to contain liquids or waterproof an item, but is actually only water resistant. This moisture when heated in the extruder will be turned to steam (e.g. extrusion temperature of 216C for ABS) and disrupts the flow from the nozzle.

    The plastic leaving the extruder nozzle will not stick readily to the bed or the previous layer due to any residual moisture, i.e. moisture between layers, resulting in a weak print, also the plastic can start stringing like some glues, e.g. UHU.

    In addition if you have an open printer then dust will add to the equation and affect print quality especially if part used spools are also stored in the open.

    I spent many 'happy' hours viewing many filament dryer reviews and switched back and forth between at least 3 in the final hours before commiting more cash to what seem to be taking off as a hobby in its own right rather than just another tool in my modellers toolbox.

    Some reviewers did not need them, left spools on a shelf for many months or a year, and many used open frame printers.

    At this point I would suggest that when buying your printer consider a closed environment machine or at least one you can get a 3rd party cover which can stay in place when printing, especially if the material you wish to print with requires a heated bed, e.g. ABS. In my book enclosure rates higher than raw speed, ultimately you write your own specification and buy according to your budget.

    An enclosed printing environment also allows better temperature control of the extruder and heated build plate (bed).

    Back to dryers.

    Dryers

    In the end I determined these factors were the most important to me :-

    1) Was able to dry filament at up to 70C

    I required 70C for nylon, lower cost dryers limit at about 50C.

    2) User controls would allow setting of temperature and run time.

    Some just had labels representing the material, personally I want figures even if the calibration and evenness of the temperature is variable across the enclosure.

    3) Display of temperature and relative humidity

    Placing of the sensors is crucial, however they are not necessarily in the best place, humidity displays may be independent of the control unit but a decent unit should be able to display the temperature sensed and used to regulate the temperature inside of the dryer.

    4) Fan assisted air circulation and vents to release moist air.

    Not all have these features, in fact some accumulate the moisture at the inside top of the dryers enclosure, surely if left for a while after drying is complete it would just drip onto the spool again.

    5) Drying safe when the filament is not being drawn into a printer.

    This is one aspect that rarely cropped up, I had already wondered why any dryers I came across were not motorised. However there were the occasional references to overheating the filament to the extent that on the underside of the spool the filament would start to fuse together when the heat source was at the bottom of the enclosure.

    Some dryers have the element wrapped around the entire enclosure, but limited ventilation or circulation.

    The unwritten detail is, they are designed to dry filament while the spool is rotating, so although drying while not rotating is possible, care is required.

    Personally I would have thought that was a pretty basic necessary requirement (rotating), afterall the more filament there is on the spool the further the heat has to penetrate. So if you intend to dry and then load the spool into a printer then manual rotation while drying is necessary, manually seems to be the only off the shelf option.

    My Choice

    I finally committed myself to the FixDry unit which I suspect is like many products these days is branded by a seller and available with other names on it, I bought mine via Amazon.

    https://www.amazon.co.uk/dp/B094Y217KL?psc=1&ref=ppx_yo2ov_dt_b_product_details

    Not cheap but neither are spools of filament and the time and energy wasted when an hour or so worth of printing is thrown away with a poor finish or worse a hairy or malformed part.

    This unit also has a deflector hood for use when not printing which deflects the air away from the underside of the spool which is useful, however I found that when using it the spool would only sit on the hood and not the rollers thus limiting its capacity to one spool in this mode.

    However I do not see that as a problem as it allows more volume in which to circulate the air around the spool.

    The rectangular structure of the case makes the possibility of motorising the rollers much more practical if necessary later on.

    First Use

    Setting up is really simple, insert the hood, supplied separately bagged with the multi-lingual instruction booklet.

    The unit,

    1.JPG

    The base of the dryer showing the fan and the heating element above it.

    2.JPG

    The deflector hood.

    3.JPG

    The base with deflector hood fitted.

    4.JPG

    The display at power up prior to user setting.

    5.JPG

    Pressing the setting button (the hamburger button bottom left) takes you to set the temperature using the + and - buttons, press again to perform the same with the hours, and again for the minutes, and finally to start the drying. The humidity display is read only.

    I set this to run with a spool of ABS sitting centrally on the deflector for the recommended minimum of 2 hours at 60C.

    The desiccant sachet fitted to the da Vinci spool cartridge was placed inside at the back so as to dry that out too.

    I inserted a thermocouple attached to a digital thermometer through one of the top vent holes to obtain an independent measurement of the air temperature within the enclosure. The temperature sensor within the unit must be in the base as there is nothing attached to the enclosure top so it could only approximate the overall temperature. It took about 10 minutes to reach a stable temperature in a coolish room of about 18C, the independent digital thermometer indicated about 55C around the enclosure when the dryers display showed 60C.

    Humidity at the start was about 45% and came down fairly quickly to between 12 - 15% for most of the drying time, temperature was stable throughout.

    I opened the dryer after an hour and rotated the spool by 180 degrees, otherwise I just let it run.

    Result

    When the filament spool was reloaded into the printer all the hairy (stringing) of the filament had ceased.

    So the dryer had done the job.

    In Conclusion

    Returning to the original questions, the answers are actually related, biased to Q2.

    If you are printing often and/or large parts then you will likely be changing spools when different colours or materials are required, or simply using up the spool before dampness is an issue.

    Or in my case I am printing small parts in one colour, so dampness dictates the unloading and reloading of filament.

    So really Q1 is answered by what you print, how often, how big, how many materials, how many colours and how humid the working environment is.

    I hope this review helps you decide whether you need one and what best suits your particular printing activities.

    Jim :)

    Back to episode list
     
    Last edited: Dec 2, 2023
    jakesdad13 likes this.
  16. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    13. 1960s Skip Truck

    This continues from the 3D scanning project of the same name, where I have 3d scanned a Base Toys chassis and cab of a suitable era truck on which to fit a 3D printed skip truck body of the 1960s, the full introduction can be found here.

    Imported into SketchUp it was moved to align with the axes shown by SketchUp and converted to a component that I will use as a base on which to design the body, noting that I do not want to actually modify the chassis or print it.

    Some extra tidying up of a few ragged edges, lumps under a front wheel arch, and rear wheel arches cleared of extraneous mesh, resulting in the component below :-

    1 Image31.jpg


    Hiding Geometry


    Next I wanted to hide the geometry of the model as though it was created with SketchUp.
    A guide on this process can be viewed here :-

    SOFTENED VS SMOOTHED Edges in SketchUp - What's the Difference?



    In particular from about 7:20

    Please Note: I cannot seem to present this stage as well as I wished to.

    Performing this stage and capturing screen shots has proved very tricky due to time lags in processing and there appears to be some ambiguity (to me anyway) about the sequence compared with the above video.

    So here goes.

    Selected the component :-

    2 Image32.jpg

    and with the Soften Edges tool available on my default tray the Soften Edges tool is now enabled, I was then able to select 'Soften Coplanar', the default of 20 degrees gives adequate smoothing. The slider was originally at 20 degrees but moving it around to get an idea of what could be done with it I moved it back as close to 20 degrees as possible. There does not seem to be a numeric entry option.

    Changing a setting appeared to instigate the process, other times using the context menu and selecting Soften/Smooth Edges was required.

    I really could not get a consistent workflow together to document this properly.

    Result as :-

    3 Image33.jpg


    Viewing the geometry that is now hidden shows the mesh lines as dashes, i.e. soft :-

    4 Image34.jpg


    Hiding the geometry again the model is ready to have a skip truck body added.

    Summary of this preparation stage

    This process can be performed on the model whether it is saved as a component or not.

    What made performing this softening a little tricky is that it could take 3 or more seconds for a response on my Win7 laptop which meant that I could have changed another aspect of smoothing before the last one was completed thinking nothing had happened.

    It seems easy in the videos, hopefully you can do it more efficiently than I have.

    By saving the scanned truck chassis as a component it can be imported into any SketchUp model, which is useful as it seems this chassis is used for several Base Toys 4 wheel truck and van models. There are 3 axle positions under the cab and some Base Toys trucks have significant overhang above the rear lights, whereas this tanker model does not overhang the rear lights.

    Their 6 wheel models have a longer chassis and I may scan one of those too, a shorter chassis is available for their articulated vehicle tractor units.

    Okay, time to move on to designing the bodywork, there will be a pause now as I am waiting for delivery of the skips from Goodwood Scenics which can have a 21 day lead time.

    Jim :)

    Continued in 13.1 1960s Skip Truck - Skip Arrives

    Back to episode list
     
    Last edited: May 15, 2023
  17. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    13.1 1960s Skip Truck - Skip Arrives

    The skip has arrived and I have finally got around to measuring it up so I can create a SketchUp component.

    The prototype image references can be seen at :-

    https://www.stilltimecollection.co....scene-scenery-destination-venue-thetford-95CC

    Further images of this truck are available on the Still Time Collection site above.

    The chassis as scanned and the 12 (cubic) yard skip from Goodwood Scenics :-

    5 DSCF1301.JPG


    SketchUp component of the skip :-

    6.jpg


    From these images the approximate positioning of the skip is made from which I can start creating the body of the truck :-

    7.jpg

    8.jpg

    To be continued ...

    Back to episode list


    Jim :)
     
    Last edited: May 15, 2023
  18. Chris M

    Chris M If 2 wrongs don't make it right ... try 3 Full Member

    Messages:
    156
    Likes Received:
    183
    Joined:
    Mar 14, 2019
    Hi Jim,
    You are getting quite good results with your scanner. I have been considering buying one of the cheaper scanners but Youtube videos from owners show mediocre results and they seem to work best with fantasy figures. Objects with clean lines seem to be an issue and I notice in your images of the scanned truck body above the shape is very distorted.

    I see you are using Sketchup to clean it up? Do you also use Meshmixer? I would be interested in more details of your cleanup process if you have the time.

    Thanks

    Chris
     
  19. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Hi Chris, the main problem is processing power, you need quite a high specification machine to add to your shopping list even with the output of a lower price scanner like the Lizard, it should also be noted that the fantasy figures often shown in the videos are generally much larger than our subjects and there are minimum size of targets specified for scanners too, in the case of my Lizard a 15 mm cube, so this truck chassis was borderline.

    Processing the data

    I was using a 13 year old Win 7 notebook with 6 GB or RAM to perform the processing and a recent refurbished Win 11 notebook fitted with 16 GB of RAM to support the scanner which required an OS of Win 10+.

    Both are hopelessly under-powered for the job.

    The STL file of the raw scan of the truck chassis would take 30 minutes or so to load into SketchUp on either machine, although once loaded they could be manipulated on the screen without too much lag, not ideal.

    However processing to reduce the complexity of the mesh especially smoothing within SketchUp was subject to delays which meant I did not know whether SketchUp was processing the data or not, it got very messy.

    From videos I have watched my opinion is that if you are to use these scans to replicate an item then I would place a minimum specification for a machine as :-

    • Windows 11
    • 64GB of RAM
    • 1TB HDD or solid state equivalent

    Some videos suggest using graphics processor cards/units used for gaming, but I did not see evidence of this, it would suggest that a tower case system would be required to accommodate the required hardware specification or a very expensive laptop.

    So as this truck scanning was for me to learn with and to stretch its abilities so I could find out what it really could do for me it was a success, if it was for reproducing the chassis it would be a fail, not because of the scanner but because of the software and hardware tools I have to effectively use the data from the scanner.

    Reproduction

    I did not intend to 3D print the chassis as in 4 mm scale with a filament printer using a standard 0.4 mm nozzle, this would be impossible, and I suspect would also be difficult with a 0.2 mm nozzle.

    The details of the chassis are fairly crude and the casting even slightly distorted along its length, this I did not notice until I checked it after viewing the crudely cleaned up scan.

    From what I have found so far, and suspected before I started, detailed work in 4 mm scale is not achievable with 3D filament printers, 7 mm scale using 3D resin printers, yes, as has been demonstrated here on P1.

    I intend to use the scanner for other 'jobs' at 1:1 scale so it was not a waste of time.

    Perhaps I should have tried scanning the skip which is a very detailed, probably 3D resin printed by Goodwood Scenics, but for the purpose of designing the truck body just measuring with calipers and creating the outline shape in SketchUp was much quicker especially as only an outline was required.

    Scan Processing

    Well I looked at a variety of packages available, they often have steep learning curves and were not originally intended for 3D printing preparation but for elaborate image and animation purposes, being new to all these packages I took the easiest options to achieve what I wanted to do :headbanger:

    I used the Autodesk Meshmixer - 6. Skip Truck Chassis - Scan Post Processing for mesh repairs.

    Following up with - 6. Skip Truck Chassis - Scan Post Processing - Part 2 where I used Meshmixer to reduce the size of the STL file.

    After the significant reduction in the mesh size I was left with a crude distorted result but good enough as all I wanted was the basic shape of the chassis for basing the truck body design on.

    The remaining editing I did with SketchUp, erasing lumps and bumps and smoothing other areas which was just to tidy it up a bit.

    If you look at part 3 that follows on from the link above, the remaining processing is discussed, concluding with what I have learned so far.

    Summary

    Be prepared to add a suitable specification PC to the shopping list when buying a 3D scanner. :eek: £££

    Jim :)
     
    Last edited: May 19, 2023
  20. Jim Freight

    Jim Freight Full Member

    Messages:
    1,177
    Likes Received:
    950
    Joined:
    Sep 9, 2019
    Whoops, links above corrected :facepalm:
     

Share This Page