Conflat P - in 4mm Scale by Filament 3D Printing ?

1 - Introduction

Well, this is my next mission impossible, typically, 3D printing your own wagons is restricted to 7mm scale as the smallest size practical for fine detail, and using resin based 3D printing techniques.

So what on earth can I achieve in 4mm scale with a filament printer?

Defining the Terms of Engagement

1. Firstly I model in 4mm scale, I have seen some simple models designed and printed this way on YouTube, i.e. Sam's Trains, so within tight limitations, like 1960's Tri-ang plastic chassis mouldings it can be done.
2. Secondly, I am not getting into resin printing, the thought of using toxic chemicals in a domestic environment or even in a shed, especially as I get clumsier is a recipe for disaster.

Having read with interest the exploits of those on P1 MRC who have mastered it in 7mm scale and achieved amazing results I admire their skills and patience

However the amount of work, okay, let's be quite frank here, the hassle of preparation and post processing looks tedious and messy, I have not got the patience and that assumes one manages to print something worth post processing.

Then there is the buying and disposal of the chemicals

The first step is to set a realistic target of what I am trying to achieve.

Target

1. Create a model of something that is not commercially available r-t-r
2. Keep it simple to run with my 1960-70s loco fleet

The wagon that comes to mind first and foremost is a Conflat P to run with my Dublo Co-Bo locomotives as a Condor express freight.

There is what appears to be an excellent metal kit available from Judith Edge of which I bought one to examine last year, but whether I am capable of building it now, let alone at least 10 of them is open to question and doubt.

Whichever way the wagon is built it is probably more cost effective in time to buy the containers.

So, what if I can print what is really just a chassis to the same detail standards of a Dublo Co-Bo which is quite low by todays superb injection mouldings?

To be continued ...

Jim

Update

Source of original model is in the next posting.

Key Stages - Active Links

• CAD process has matured
• First rolling chassis achieved

 

2 - Getting Started

First I need a design, well luckily there is one available on Thingiverse, here

https://www.thingiverse.com/thing:5968262

This provides my starting point, I credit Jonny (IronMink) with the original design and STL files he has shared via Thingiverse.

Attribution License

https://creativecommons.org/licenses/by/4.0/

From Thingiverse two STL files are downloaded, the wagon chassis and the separate brake lever.

Buffers are not included although others are available from the originator on Thingiverse.

Importing the files for use in SketchUp

I use the free SketchUp Make 2017, this version and maybe others are inefficient at importing an STL file, it takes a very long time as I found out for myself when I imported a 3D scan of a die cast 4mm lorry chassis as the basis of a skip truck (really must get on with that too).

Admittedly I generally do my CAD on an elderly VAIO laptop c2010 running Win 7 so this time I tried a different route, i.e. a different format.

Spin3D Plus from NCH Software can convert various formats and what was a common one worth trying was .3ds an old format still in use.

https://en.wikipedia.org/wiki/.3ds#:~:text=3ds is a binary file,well as the data itself.

This is an import option for this format available in SketchUp.

After conversion the .3ds files imported into SketchUp on my old laptop quickly, not like an hour plus I had loading the truck chassis scan in STL format*. Good start.

*Units in STL format: An STL file is dimensionless so be careful how you import it to ensure that the STL units are translated correctly. If the units are defined as equal to a mm, when the STL file is written you must import it as mm otherwise it will not load at the desired scale, in this case full size.

Ultimately It is always better to create models of small objects, e.g. 4mm scale parts, larger than full size in SketchUp, otherwise rounding errors occur and the model breaks and can be difficult to fix. SketchUp appears to have originated primarily as a civil engineering tool, long before domestic 3D printing of small parts, so working down to 3 decimal places of millimetres is always one step away from a broken CAD model.

So to reduce problems the .3ds file was imported full size where one STL unit = 1 mm, once cleaned up then re-scale to e.g. 10x the size of the printed object, this makes editing the model easier and keeps you further away from an unstable CAD model.

Wood for the Trees

When such a file is imported into SketchUp it is a straight conversion, by which I mean it is totally composed of the triangles that was encoded into STL format, this is generally unusable to edit. Let me demonstrate.

Overall view as imported into SketchUp




Closer view




Close up, just look at all the triangles around the buffer mounting holes and the dense meshes between the rivet/bolt heads.




For this to be editable by me I then manually clean out all the surplus lines which are only there to triangulate the whole model, there may be software out there to do this but I haven't found one I can understand how to use so I plod through it manually.

Not too difficult when you get going, but save regularly with numerically incremented file names in case some actual detail inadvertently disappeared that becomes obvious later on.

It is fairly easy to spot these triangles on a flat surface and delete in bulk using the eraser tool, short cut key 'e' and sweep across swathes of them, a quick Ctl Z when a surface breaks, when whoops the deleted line actually defined shape.

After a couple of hours, well on the way, I haven't tried removing all the triangles from the holes and rivet/bolt sides, that is way beyond brain numbing!

As can be seen the shape of the wagon chassis stands out much better than in the first image above. This also loads into SketchUp much quicker now.



Next, I move on to de-cluttering below the solebar.

 

Hi Gormo, the CAD model view is in a parallel projection mode, not perspective so it can look a little odd when viewed out of square from an elevation or plan view.

I am just used to it, except when mentioned by somone, then I do notice it, we naturally expect it to taper into the distance.

The framing also adds a slight optical illusion to distort appearance too.

Jim

 

Hi Gormo, here is a perspective view, generated from the next episode, hence clean below the solebars, working in perspective mode is messy, so normally I work in parallel projection mode, forgot to change the view for this thread.



Jim

 

A very good question Andy.

Here is my answer.

There are several reasons why I am not just applying this STL file directly to my printer via my slicer as the originator intended.

1) As I mentioned above the STL file is for printing a wagon using resin printing techniques, a process that I am not going to use.

2) The nature of supports for filament printing are inferior to those you have with resin printing.

Supports by default I have found to be useless during the Arkitex project because they don't actually touch the overhangs, usually a layer or two short, they only catch a droop from an overhang. Okay for large objects but useless when parts are only a few layers thick. If they do touch they become a part of the print and need cutting off, the materials used in filament printing do not lend themselves to snapping off, they vary from flexible to tough.

3) If I cannot provide supports that contact the printed object with my dual extruder filament printer using a water soluble material (which I shall be outlining in the next episode), then I will have to break the model down into a kit of parts. Too complex to do that with the imported STL file.

4) I want to modify the method of mounting couplings, fit brass wheel bearings, and possibly simplify details I cannot print, almost impossible using the imported file.

Summarising.

So I imported the STL file and cleaned it up so I have a CAD model from which to modify the design to suit my preferred printing process, the triangulation lines in holes, bolts and rivets have been left in place as I do not need to change those, besides it would be madness, even for me, to remove those

The much improved clarity of the model can now be seen by comparing the before and after images.

In one sentence, I required a clear editable CAD model to enable efficient implemention of design changes, so I cleaned out most of the model triangulation lines, this I have completed.

Jim

 

Hi Andy, probably about 6 hours, which included running the Inspector tool add-in on SketchUp a few times because I unleashed a Pandoras Box of broken surfaces and openings which took a hour or so to unravel amongst the dense meshes half way through cleaning up.

I have no problem about drawings I started my working life as draughtsman in a design and development office when we used crayons, plastic film and microfiche back in '73.

From what I gather getting a drawing of a Conflat P is not easy, unless you know which of the plate wagons it was derived from.

My only drawing reference source at this time is a Datafile article written by the BRHSG (British Railways Historical Study Group) in the 1983 Model Railway Constructor Annual which even there only shows a BR outline drawing.

Besides I found this STL file on Thingiverse, so it was probably quicker to reverse engineer an STL as the images shown there indicated it was good enough for me and may need to be simplified, and modified for filament printing anyway.

I would not try and do this for much else, especially if a r-t-r model was available, and building a Judith Edge kit x10 would probably take me much longer, once I have worked out how to print it then I can let the printer spend hours creating them.

Parts I shan't be 3D printing and will buy, are wheel sets, bearings and buffers, the latter are much easier bought as whitemetal castings, some from Lancashire Model Supplies look promising, not quite what the source data indicates, but close enough.

Regards supports, it's as I said above, very different to the snap off ones you can create with resin prints, I went into that in depth when developing my Arkitex parts, so I am pretty sure about the limitations with filament printing and slicer settings for supports. If anybody can offer me an alternative approach to expensive water soluable supports I would like to know about it. But then again, on the web I see very little regarding printing with a 0.2mm nozzle let alone using dual extruder machines to print tiny functional parts.

Speaking of which whilst trying to DCC convert a Mainline J72 I think I have found a subject, a geared drive axle variant (not available from Peter's Spares) to try and print in Nylon, mission impossible #3 coming up, I am not to be deterred by no one else doing it

Jim

 

They do need more extensive calibration than a single extruder machine especially when you use both heads on a single job, e.g. printing in two colours of PLA which I succeeded in doing with a 0.4 nozzle, using one for the object material and another with a dedicated support filament using a 0.2 however will be more challenging.

Well, my R3D E2 IDEX is quite solidly built so squeezing (excuse the pun ) out finer prints than what seems be done normally with FDM has been my aim ever since I bought it and it has surprised me what it can do once the multitude of slicer options are tuned accordingly.

Out of curiosity what make/model is your dual machine, there's not many dual extruder machines around?

Jim

 

3 - Further Cleaning Up & Supports

After a few more hours of fun, yawnnnn, the CAD model in SketchUp is now in a usable form, I can now see what is actual detail.






It was not obvious in the original import that the file represents the Conflat P in model form to prototype dimensions including simplified brake rigging and the axle boxes provided with a coned bearing for fitting axle sets directly.

Provision is made for fitting the missing brake gear parts.

Reason to import and not just print the STL file.

The reason for importing it into SketchUp is so that I can edit the model directly to suit my needs including how I am going to print it. The original design appears to be aimed at resin printing so changes will be required. Also I may decide to print the chassis as a kit of parts in which case splitting up the chassis into components requires a clean model to work from.

A major difference between resin and filament printing techniques is that of supports, with resin printing they are used in a much more effective way than when filament printing.

Resin Printing Supports

They are printed to actually touch the printed object and are broken away from it when the job is complete.

The printing process duration is not affected by how much of the build plate is occupied, every square millimetre can be printed simultaneously for each layer, so supports do not add to the print time.

Filament Printing Supports - Basic

They are printed such they are a layer or two below what they are supposed to support, so they only support what droops on to them. That IMO is not a support.

The more of the area covered on the build plate the longer the job takes as the extruder has to pass over every point occupied by the job, and also travelling between parts on the build plate.

These two issues with basic supports make supports something for me to avoid wherever possible especially for small parts where a layer or two droop is significant.

Break off support for filament printing

For filament printing, the PolyMaker - PolySupport you can break off, however it is more expensive and not-biodegradable, although a derivative of PVA.

In my view a bio-degradeable based supporting medium is better as once the part is printed the support material is discarded. PVB can be re-cycled, but that I have found to be from broken car windscreens, in bulk.

For the model anticipated, break off supporting material leaves marks to be removed and could well risk breaking delicate parts such as brake gear assuming the gear can actually be printed in place to start with.

So, how to have actual supports with filament printing?

Filament Printing Supports - IDEX

This is one of the reasons I bought an Independent Dual EXtruder (IDEX) filament printer, it has two extruders which can work separately or together.

• Create two copies of one item at the same time
• Create a pair of handed parts at the same time
• Create a part with material from both extruders, different colours or compatible materials.

This last mode is where a filament printer can actually create actual supports, the trick is to use a compatible water soluble filament in the second extruder while the first one prints the required part. The gap between support and part is set to zero and after printing the whole print is soaked in water to dissolve off the support material.

However, filament printers are still much slower than resin printers for supports because the extruder still has to travel around a lot, but it comes with an extra price. The time taken for the part printing extruder to park to allow the extruder with water soluble material to take over and vice-versa.

This can include warm up time, travelling between parking place and job and priming or wiping the nozzle on a tower or a wall which is built up next to the printed part. Warm up time can be performed shortly before an extruder is needed so it is ready to print when its time comes.

Other downsides include that the water soluble filament is not cheap and machine set up, Z height and offset between the extruders needs accurate setting especially if printing small parts.

To date I haven't used this dual extruder method because I generally design to avoid the need for supports, e.g. the EPC couplings, dove tailed, separate parts are easier to print more accurately.

Which approach I take for the Conflat P depends on how well I can get it to work and bearing in mind that most of the information available is for a 0.4mm nozzle printer, I am diving straight into using 0.2mm nozzles.

Next, I look closely at setting up the printer for small parts.

 

4 - Setting Up the Printer for Small Parts - Part 1

This where the fun really starts and has resulted in quite a turbulent time mentally getting to grips with all the parameters that slicers provide.

Warning: The Tech Speak that follows may damage your health, jibber, jabber!

I didn't have much trouble printing the EPC couplings as most of the (IdeaMaker) settings used for the 0.4mm nozzle worked well, but this was to be a very different task.

The resolution I want to print at is much higher and the details rather delicate, and worse I want the wagon to be tough enough to survive active use in potentially a rake of 10 wagons which would have Dublo Co-Bo at the front and a Dublo 6 wheel stove brake van at the rear, Condor style .

Line Width and Thickness

Typically one thinks of printing at a line width equal to the nozzle diameter, but in actual fact it is possible with other settings adjusted accordingly to print at even half the nozzle diameter which I managed to do with the 0.4.

There is a square law at play here though, half the diameter, quarter the cross sectional area of the nozzle through which the melted plastic is to be pushed.

So here I have set myself two targets.

1. Basic Resolution - 0.2mm line width, 0.1mm layer thickness
2. High Resolution - 0.1mm line width, 0.1mm layer thickness

Naturally I want to work with option 2, but will start proceedings with option 1.

Initially I will use the HatchBox Pro+ PLA for prototyping, it is a strong easy to use PLA, it may also be a suitable material for the actual running version.

Bed Adhesion

This is paramount, if the first layer does not stick properly, then maybe hours into a print breaking free renders the print to scrap.

I quickly found that using the GeeeTech 15 adhesive with the addition of the AJOYIB flexible bed (which is necessary for releasing small parts from an E2 build plate) was effective, and is so for multiple prints in succession. When it needs replacing it washes off easily with warm soap and water.

Temperature for Printing

My sources of information were initilally from the Raise3D forums but later mostly from some established YouTubers as although Raise3D sell these 0.2mm nozzles for the E2 very little support is provided. Certainly no useful material profiles.

Here is where mixed requirements start to emerge, 'normal' here refers to extruding from a 0.4mm nozzle. So for using a smaller nozzle

• Higher temperature than normal to enable good flow through the smaller nozzle bore.
• Lower temperature to avoid overheating the filament in the hot end because the flow rate is significantly less than normal.

In general, with high flow rates, e.g. using 0.4mm nozzles and larger the plastic needs to be heated quicker so it is ready for extruding from the nozzle by the time it gets to the nozzle.

Excessive extrusion temperature also increases the risk of stringing, which is one hell of an issue where multiple factors combine for maximum frustration, more about this later.

Even worse excessive temperature can cook the filament into charred particles which then jams in the nozzle.

A method used to determine the optimum extrusion temperature is a temperature tower.

Temperature Towers

This is a part designed with features including a bridges, overhang, and often raised lettering and other features to determine how much stringing is taking place and the quality of the surfaces.

Typically it consists of a stack of bridge like features in a tower, and crucially the temperature of the extruder is changed by 5 or 10C at each level, usually highest at the base and lowest at the top.

The temperature of each level is marked on one part of the tower for subsequent examination.

Much more detail can be found in this video



The tower from a Thingiverse file by steinhausler that I downloaded for customisation.

https://www.thingiverse.com/thing:2625999.

As I was using this with a 0.2 nozzle I imported the STL file into SketchUp, cleaned it up, and reduced the width to 50% so the bridges were more in keeping with a the narrow line widths and layer thicknesses described above.

Dimensions were added to the edited CAD file and converted to layer numbers which would enable accurate extruder temperatures at each level. The model was then exported as an stl file and loaded into IdeaMaker which has the ability to set the extrusion temperature on a user selectable layer basis.

Front view.




Angled view




With the temperatures assigned as per the layer numbers.




It was then sliced, the next view shows the temperatures at each level as confirmed by the print preview.




Results

Not very good, stringing at all levels, least worse was at 205C, but this was not the end of it by any means, this was certainly not good enough, subsequent attempts were all poor.

So what was the issue, first was to ensure that the filament was dry enough, so put in the dryer @ 50C for 8 hours rotating it hourly, ... tedious, I really wish these dryers were motorised, wish I had modified this one!

Results, just as bad, filament problem or something else?

Tried a different spool of this PLA+, just as bad.

Okay, back to printer settings.

Continued in 4 part 2

 

4 - Setting Up the Printer for Small Parts - Part 2

Stringing and Retraction Towers

Stringing occurs when the extruder stops laying material on the previous layer prior to moving (aka travelling) to another place to continue, however there is still pressure within the hot end so the plastic continues to seep out of the nozzle, this is also called oozing.

As the nozzle leaves one place the oozing filament is anchored there and stretches as long as it can, and on short distances it then attaches itself to the destinaton where the extruder starts to layer plastic again. These are seen as very thin hairs of plastic much like you get from some polystyrene cements, or UHU. Although they don't mark the surfaces as such their ends get trapped by the next applied layer and end up making two parts appear as blocks of flats with washing lines between them, ugly! Even worse like a plastic spiders web that has seen better days.

This can be time consuming to remove, although some makers (frequent 3D printing hobbyists and pros) use a hot air gun to virtually vaporise them, I don't see that as a viable option when printing small parts, it would be like when I was introduced briefly to welding aluminium, the work piece dissapeared, seems I was never cut out to be welder, not that it troubled me.

Retraction

One setting to reduce stringing is the distance the filament is withdrawn from the hot end to deplete the hot end internal pressure, this is called retraction.

Another type of tower, much similar is printed where the retraction distances are changed in steps, however this must be selected appropriately depending whether your printer is a direct drive, i.e. the extruder is mounted on the same carriage as the hot end so moves when printing.

Or, the extruder is on the frame or case of the printer and is connected via a flexible tube to the extruder, referred to as Bowden tube, much like it was used on bicycle brakes.

Why the difference is important is that in a direct drive machine the filament being in close proximity to the hot end means it flexes little, due to a short, straight and rigid path, so control is precise.

On an indirect or Bowden tube machine the flexible tube, flexes by nature and the filament is a loose fit in the tube so it can also move sideways, so a lot of movement of the filament is absorbed by a situation akin to pushing a piece of rope through a hosepipe when retracting.

The end result is that a retraction distance for a Bowden relative to a direct machine can easily be a 10:1 difference in magnitude, e.g. 6mm vs 0.6mm.

Most of the info. on the web cites 4 to 6mm so remember if your machine is a direct drive it needs to be approximately a tenth of that, what happens if you get it wrong in a direct drive machine?

Well I read about it then later I forgot and what happens is that the soft filament is drawn too far back to the extruder on retraction. When the travel to the start of laying more plastic is completed the extruder tries to push soft filament, it bulges, jams movement into the hot end and the extruder then grinds its way into the the filament within the drive mechanism.

Jammed extruder then needs dismantling, and cleaning out of swollen filament and in a worse case scenario ground filament dust goes everywhere, including onto the hot end as a sticky coating and can pepper the print job, so be warned!

Retraction Tower

A simpler type of tower, but not one I could easily implement with IdeaMaker, so it was the turn of Cura 5.x which has a useful plug-in to create test towers and other parts.

This can be found in the Ultimaker Market via Cura v5 upwards, not available for v4.x.

https://marketplace.ultimaker.com/app/cura/plugins/Kartchnb/AutoTowersGenerator

On selecting the appropriate tower Cura generates the GCode (language of the printer) to step through a range of retraction distances at different heights of the tower.

But once again nothing seem to work even being generous with the distance range, so much so the extruder got jammed as above, whoops, I had been warned!

Ho hum ...

Then I went to try other settings to reduce or elimnate the stringing which were fruitless so I will spare you, poor reader, from all that and move on.

Cura

When I printed a replacement elevator belts for my TTR Elevator/Conveyors with TPU I had significant stringing between the tips of the elevator belt teeth as the IdeaMaker slicer would insist on travelling between teeth often, and TPU just loves to string.

The answer then was provided by the Cura 4.x slicer which has a mode called combing, there is nothing like it in IdeaMaker*

*IdeaMaker & Combing: As far back as 2016 I have seen posts requesting this on Raise3D forums for IdeaMaker but it still has not been addessed even now in 2025, so I think over 8 years later they never will implement it, not in time for me, anyway.​

This can be set so any travelling is made within the part, typically an inner shell or the infill so the stringing is buried within the model, this worked very well even with such a thin belt. What stringing there was occured on the inner surface of the belt and was easily removable.

So, would this work here.

Continued in 4 Part 3

 

4 - Setting Up the Printer for Small Parts - Part 3

Customising the G-Code - IdeaMaker Temperature Tower + Cura Combing.

At this point I had a custom temperature tower with the temperature at each level of the tower easily defined by IdeaMaker, however a way to define the temperatures in Cura seems to be limited to intial layers and then all the others, if any one can point out how to set it up within Cura, pleeease, let me know!

Workaround time, an often needed process in software development and it seems now, 3D printing.

Note: my Raise3D E2 uses the RepRap flavour of G-Code, there are variations between flavours, different machines may use different flavours, so be aware that the RepRap flavour codes I present below may not apply to your printer.​

There is some quite readable information about G-Code on this site :-

https://www.simplify3d.com/resources/articles/3d-printing-gcode-tutorial/


Create a G-Code file containing the important components from IdeaMaker and Cura.

1. Create the test tower in SketchUp, dimensions re-labelled as layer numbers.
2. Export the CAD model as an stl file
3. Generate a G-Code file using IdeaMaker from this stl file
4. Generate a G-Code file using Cura from the same stl file

Read both G-Code files into an editor such a NotePad++ where they can be viewed and edited side by side, you can download a custom language for NotePad++ such as G-Code which then helps you by colour highlighting of the G-Code.

Re-name the Cura file to e.g. "Custom .."

Within the G-Code files comments are usefully inserted by the slicers, in this case the ones starting off ";Layer..." are of interest.

Note: that in a G-Code file, layer numbers are numbered from zero, and in the slicers from one!

Copy and paste the temperature entries from IdeaMaker into the new custom file.


Result of Combining Temp Tower from IdeaMaker with Cura Combing.

No improvement, unacceptable string.

Combing does rely on the nozzle travelling over what will be the inside of a part, however models such as the Conflat P will have very little bulk for this to be effective.

What makes this difficult is that very little is published on activities with 0.2mm nozzles, usually what is available online is for standard 0.4mm and larger bore nozzles.

So, a pause for thought while I look deeper, I will get to the bottom of this!

Jim

 

Meanwhile ...

Cura G-Code for Raise3D E2 - Reducing startup heating times by sequence change

Near the start of the file, the inital startup sequence of heating the build plate (aka bed) and the nozzle (aka confusingly called the extruder) can be usefully modified, Cura heats the bed first and then the nozzle, by default IdeaMaker starts them off together, which is often quicker.

The code to do this revolves around the heating commands for extruders and the bed which can take one of two forms,

1. Start a heating process and carry on to the next code statement, or
2. Start a heating process and wait for the required temperature to be reached before continuing with the next code statement

The annotated G-Code (a ';' signifies comment text) for both to heat together is :-

; Start bed heating to required temp e.g 40C, but let next statement(s) run
M140 S40

; Start nozzle heating to set temp e.g.230C, but wait here until required temperature is reached.
M109 S230

; Issue start bed heating command again but, this time with wait for bed to reach required temp (just in case its late)
M190 S40

These statements are inserted into the Cura generated G-Code file


Further detail on heating G-Code statements can be found here.

https://www.simplify3d.com/resources/articles/3d-printing-gcode-tutorial/#:~:text=Use these commands to set,any other commands to run.

Jim

 

Fun on two fronts

1. Print quality
2. Slicing

On both of these fronts I have been using a small section of the chassis, e.g. a part with two W irons so I can also check the positioning for Peco top hat bearings to the axle boxes, as given the bearings are just the plastic the chassis is moulded in.

Print quality

This is still proving a headache, just when I think I have cracked it, I haven't.

Slicing

Now this has become an issue as whatever STL is exported from SketchUp needs repair by IdeaMaker or is just plain rejected by Cura.

When SketchUp Make imports the original STL file it spends a lot of time 'repairing' it during loading, minutes on my workstation best part of an hour on my old laptop, so needless to say this is a workstation activity. However, if it is checked by the SketchUp Inspector plug-in it fails straight away, it was better after substantial manual cleanout of much of the triangulation as I showed earlier in this thread.

Subsequent scaling had to be in stages to bring the model down from full-size to 4mm scale, so the Inspector could progressively repair the damage done by SketchUp. But the 'clean' file would still have issues when trying to load into IdeaMaker, even its repair facility needed multiple passes, but, okay it was eventually done.

But, all was still not right, when I sliced in IdeaMaker not all parts would be supported as they should be, even for a segment of the chassis, goodness knows what mess the whole chassis would have been like.

So, the file corruption is not identifiable at SketchUp or even in IdeaMaker until it goes wrong during slicing. Cura rejects the STL file but cannot repair it.

It Gets Worse

Same problem with a simple test component created from scratch in SketchUp working at 10:1, so it is little wonder a much larger model fails.

The primary cause IMO is twofold, (certainly for older SketchUp versions)

1. SketchUp was known to be poor (laboriously slow) at reading in a file in the STL format.
2. SketchUp being a civil engineering CAD tool was never designed to detail such small parts as those we try and 3D print, so numerical rounding errors creep in so that surfaces and vertices break almost without it knowing until you try and edit the model again. This did crop up occasionally in my Arkitex Project.

Therefore I cannot use SketchUp Make 2017 for this project.

Time for a Different Approach.

FreeCAD 1.0 is the next contender, not an easy transition from SketchUp but at least that is more of general engineering CAD tool and should be suitable for 4mm scale modeling work, I gather it is popular with forum members working in 7mm scale.

So far trying to scale the imported model within FreeCAD has defeated me, but I believe it is possible from a YouTube video for an earlier FreeCAD version, just need to gather enough brain cells in one place to do it.

Meantime I loaded the original STL file into Cura and was able to scale it down in three steps, would not do it in one step. So, scaled, saved the scaled down file, exit Cura and reloading and scaling repeating twice more reduced what was presented on the build plate to being 4mm scale and not 12 inch scale!

I then exported this 4mm scale version of the STL file and presented it to IdeaMaker without any warnings at all and it sliced properly without missing out supports.

So this is promising, as I need to edit the model to change wheel bearings and add mountings for the couplings that are going to be fitted, certainly not any form of chain in 4mm scale, so FreeCAD is the likely modelling tool for this project, with IdeaMaker retained for slicing.

Jim

 

Couple of Issues when using Cura 5 with a Raise3D E2 Printer

1 - G-Code Transmission to the Printer

Cura (version at time of writing is V5.10.0 for Win 10/11) does not work too well with this printer, like other 3rd party slicers it cannot directly connect to the E2 via a network, Ethernet or Wi-Fi which seems to be typical of machine manufacturers and generally leaves you with one of three alternatives.

1. Connect PC directly to the printer with a USB cable and export G-Code file to local printer - not convenient to trail a 5m USB cable between rooms.
2. Export the G-Code file to a USB drive and plug this into a USB port on the printer, where it can be read as a local drive - cumbersome and slow.
3. Export the G-Code file to disc, import into IdeaMaker and pass to the printer via Ethernet or Wi-Fi network. This also lacks the info added when IdeaMaker does the slicing - tedious but tolerable.

(3) is my preferred option when using Cura.

It must be noted that IdeaMaker passes more than just the G-Code data to the E2, plenty of eye-candy for the display and expected run time data.

It can also exert more control over the machine, if only they would implement combing as discussed earlier.

The eye-candy and run time estimate I can live without, remaining runtime is estimated as the print progresses whether the run time data is supplied to the E2 or not.

2 - Poop and Run

When using two extruders e.g. extruder #1 for the print and extruder #2 for the raft and supports I have found that when #2 completed the raft it purged itself over the last point of the 'printed' plastic, a disc approximately 4mm diameter and 2 mm high before leaving the scene to park and wait for its next turn of action.

This happened on a few occasions, different prints, result, a great wart on the print. If either nozzle subsequently hit it the colliding nozzle would rip the wart off the bed destroying the print, either with a hefty thud if it had cooled or if soft drag it across the print attached to a now blocked moving nozzle, resulting in a gooey mess and ruined print.

It had no reason that I could find for it to perform such a purge regardless of where it was over the bed, even parked, so I could not prevent it.

Jim

 

I give up, too many issues ...

CAD

Meshmixer can clean up and scale the original STL file better than most and export in Collada or STL formats both of which SketchUp can import, the Collada format very much quicker than STL.

SketchUp cannot process a file with so much detail in such a small volume, it specifies errors but it is virtually impossible to fix them within SketchUp.

FreeCAD is not an option after all as an STL file cannot be converted to a parametric part for editing.

FreeCAD (when I found the feature) actually reports many errors with the original STL file.

The fit for printing data information in IdeaMaker is only applicable to resin printers, did not realise that before.

Printing

There are continuing inconsistencies when printing, sometimes no stringing, but poor base layers and vice-versa.

From initial partial test prints I also suspect that the nature of filament printing objects of this size will be too fragile for active use on my layout.

Summary

Well, I have had a go at it and spent quite a bit of time and effort on it over the last two months but one has to know when the odds are stacked against you and call it a day.

What daft idea will I pursue next ?

Jim