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

Continued from the Workbenches Forum.

The adventures of a wrinkly noobie into 3D printing, my optimism, attempts to learn enough to make something useful, the failures and successes, but actually making progress at last

Having found that there is rather a limited amount of consolidated information on how to approach all this and having done some quite time consuming searches into how, what and why, the jargon, the technologies and just "How do you get all this to work" I decided to share my findings warts and all.

I hope you enjoy this tale, the aim is to present useful information on how to get into this potentially very useful side to building our railways in a relatively light hearted manner but as technically accurate as I can.


This thread should have really started on this forum, to rectify this it continues here with the same name as in the Workshop Benches forum and they are linked together.

Original thread with full list of episodes can be found here Episode list.

My next project (the skip truck project will continue I just like variety) is creating replacement windows for my dockside warehouses.

14. Victorian Style Windows for Warehouses

I have a block of dockside warehouse frontages built from laser cut wood kits reworked from full depth standalone designs to a low relief format, these are covered in another thread (Ref 14.1), here I concentrate on further 3D printed parts to enhance them.

The original windows are very simply done to a budget IMO and I was going to leave them as shown e.g. below



However this sort of building in the 1960s was probably of Victorian build or earlier and certainly would not have had glass panes larger than 6ft high doorways. Typically they would be very much smaller, e.g. may be 12 x 6 inches (30 x 15 cm) at an estimate from old pictures. Now fabricating such things from plastic card and strip is feasible but a long and tedious exercise. If I was scratch building the frontage then I could probably have sourced some window frames from specialist manufacturers and made the window apertures to fit the windows. Here, I have the walls and need window frames to fit them.

Time for finding out what else I can print with a daVinci Pro 3D printer from 6 years ago fitted with its default 0.4mm nozzle.

There are two sizes of window frames required, most ground floor, which I have chosen a format of 9 columns by 5 rows, 8 required, and the smaller first floor (and a few ground floor) of 8 columns by 4 rows, 38 required.

Hence a great deal of work if fabricated from individual parts, a few spares are created to cover for the odd misprint or damage when installing.

Intention is to print in natural PLA which from XYZ Printing is actually clear and uncoloured.

I have documented the larger ground floor windows here, the first floor windows were designed and printed in the same way.


14.1 Initial Design Approach

After attempting to create the window frames manually which is a tedious task I went to YouTube and found a review on a few SketchUp extensions, most are for highly detailed windows for illustrative or architectural CAD drawings.



However for my needs I found a simpler tool used for creating lattice frames across many shapes of objects which fitted my requirements much better :-



This can be used to easily build Victorian style frames for small glass panes much quicker than using basic SketchUp tools.

Downloadable (after signing up for a free account) as a freebie from :-

https://sketchucation.com/plugin/1097-latticemaker

This is what I used in the following design for 3D printing.


14.2 Design Process Utilised

Created 9x5 grid using the rectangle, divide, and line tools.




Select entire grid then right click for Lattice Maker which brings up Lattice Properties, representation of the glass pane is not required.




Lattice created.




Selected shows it is grouped.




Needs to be ungrouped for editing.




Deleted all join lines so that it can be processed further, e.g offset tool on outer edges.





With join lines deleted, offset tool can be applied to the whole frame.




Push pull tool to pull forward and create the mounting flange.




Continued in Part 2

 

Part 2

Solid Inspector issues.




Solid Inspector fixes these issues, debris from deleting joints?




Made group and exported in STL format file without errors.




Imported into XYZware slicer.





Laid flat and positioned in XYZware, I work around the print bed to spread wear.




Loaded up my slow print speed 10mm/s instead of 30mm/s settings for a better result, between component speed left at the faster defaults.




Preview looks good.




Printing on the way.




After the heating phase, the printing has started.




Result is pretty good, nearest edge has not joined up properly and there is a bit of stringing, but tolerable.




Certainly for the intended use and bearing in mind the outer edges are nominally 0.4mm wide, same as the nozzle, If I were to mount these with the the nearest edge to the top it would not be visible where the buildings are to be sited.

14.3 Alternative Process

Deleting all the joints above was rather tedious so I tried another way, create the lattice array as above, then hide it.

Create the flange part separately and position so that unhiding the array they could be joined together. This was much quicker, but, I ran into issues flagged by the SolidInspector over 'nested' parts which would convert to an STL file okay but when sliced the preview showed assorted panes filled in.

The inspector could not fix these errors and so was unprintable, so not sure what was meant by a having a good STL file.

After a couple of hours I gave up and used the process described above, which although had a few errors they were fixed by the Inspector and printed fine. An experienced SketchUp user would probably be okay but not a newbie like myself, well done if you can do it!

14.4 Damp Filament?

Just wonder if the stringing is a dampness issue, but as this was a new out of the sealed manufacturers (XYZprinting) bag today it shouldn't be, or was it poor manufacture, old stock?

When I buy a new inkjet or 3D printer I use the manufacturers own consumables for the first year to cover the 1st year of warranty and minimise any issues.

So I think I ought to see what a stint in the dryer does to the filament before printing a whole batch of these window frames.

Humidity started at 51% and settled to approx 12 to 14% for most of the time, at 1 hour I rotated the reel 180 degrees and stopped at two hours to let it cool, the default recommended setting of 4 hours for PLA seems rather OTT, and besides if it is run that long I won't be able to get in a print run this evening and would have to leave it overnight before using it.

Very little difference to stringing, but tolerable, so try printing 9 at once as sometimes a pause between passes has been beneficial in the past (with ABS) as it allows a previous layer to cool longer and the next to stick better, is that the case here?

Tweaked the model so that the STL file is for a laid flat component which allows easy import of the STL file multiple time allowing the slicer software to distribute the components evenly about the print bed without manual rotating and positioning.

Continued in Part 3

 

Part 3

Layout of nine frames for printing.



14.4 Results

All printed to an acceptable quality, remember that a 3D print will unlikely to ever be of the same standard and finish as an injection moulded part such as those in off the shelf kits and r-t-r models.



The gaps in one outer frame edge is not a fault of the printer it is actually present in the output from the slicer, the gappy nature of one long outer edge can be seen in the print preview when greatly magnified, the printer is printing as per the slicer output.

As the resulting edge will be barely visible I am not going to pursue it further for this job but I will be interested to see what another slicer produces from the same STL file, perhaps those gaps originate from SketchUp when it creates the STL file.

14.5 Summary

This has turned out to be another adventure through the mysteries of design and processing for 3D printing, the learning process still makes up most of the time I spend on these jobs.

On the plus side attempting to fabricate over 40 Victorian style window frames was not something I would even consider doing these days unless someone was paying me a very good hourly rate.

If I want more I just print more, if I want another design I now have a way of doing it witout too much hair or sleep loss.

At the moment the printer is half way through a 3 hour printing of the first batch of 20 1st floor window frames, at least all I have to do is keep an ear out for potentially problematic noises from the printer, other than the general curious noises it makes, such as whirring, electronic squeaks, as well as the clicking or thudding of the filament feeder.

I may well try a sample print of these with the Raise 3D printer fitted with a 0.25mm nozzle but the results so far with the da Vinci and a 0.4mm nozzle are good enough for improving the look of these laser cut wooden kits. There is little to be gained by having a varying level of detail across one building, consistency looks better.

All I have to do now is a little light fettling and spray paint them, priming I do with rattle cans, top coat, oh dear, I must get to grips with the spraying gear I bought a year or so ago, I'm terrible, a one year warranty can expire before I have done anything serious with these things.

The first batch of first floor windows have completed and look okay, painting and fitting will be covered in the warehouse thread.



This was my first exercise with PLA and the parts feel more flexible than I thought they would be, the filament off the reel is quite brittle in comparison, I do wonder if this reel is representative of PLA in general, any advice is welcomed.

Jim

References

14.1 Montague Dock - Warehouses

https://platform1mrc.com/p1mrc/index.php?threads/montague-dock-3-warehouses.6399/#post-104783


Back to episode list

 

15 - daVinci 1 Pro - Printing & Levelling Grief

This was not an intended article a seemingly easy repeat run of 20 more warehouse window frames caused hours of grief to track down the problem and another issue surfaced, print bed levelling aka auto-calibration failure.

After the success of the first 3 hour print run of 20 frames in the morning, ref previous article, I started the 2nd run just before dinner that evening.

Second successful run .... no way!

This is a long episode of grief written as an aid to my memory at the time, so if you just want the positive highlights skip to the summary in part 3, otherwise ...

The Gory Detail

Printing Attempts 1 & 2 - 0.8mm frame bars in PLA

Failed almost straight away with the first layers not sticking getting tangled around the nozzle and dragging neighbours off the bed, so I cleaned the nozzle and sensor pin with the supplied brass wire brush.

Printing Attempts 3 & 4- 0.8mm frame bars in PLA

Replaced the blue tape on the bed.
Failed straight away, clean again, then failed again!


15.1 Auto-Calibration/Levelling

Next decided to check the bed was still level (why it should have changed though? ) using the built in auto-calibration (aka levelling) check.

It will sense the bed at four corners starting with the back right, then front right, front left and lastly back left after which it will display its findings and what to do if not level enough.

At the first calibration point it tried 3 times I think it was and came up with the message failed see FAQ (aka go to web site for not a very useful video).

A silent video by XYZprinting which is captioned, not very good IMO as I can't read and view the details simultaneously can be found at :-



So I wiped the four contact pads with isopropyl alcohol on a tissue, note that it is the right hand side of each metal pad that matters (not clear in the video) and it went through all four points before announcing 'Failed see FAQ', arrgh!

Two more calibration attempts, cleaning the contact points, gave up tried printing attempt 6, fail, it seems that the auto-calibration can be troublesome, whoopee!

e.g. video demonstrating manual calibration/levelling:-



Before trying the manual approach I tried further cleaning and auto-calibration.

Printing Attempts 5 & 6 - 0.8mm frame bars in PLA

Two more fails of the first layer to stick and cleanup again, beginning to look that I would need to do the calibration manually.

As a last resort I tried cleaning the contact points and sensing pin on the hotend with Servisol Super 10 switch cleaner, tried calibration again and it pronounced that the bed was perfectly level, hooray.

Printing Attempt 7 - 0.8mm frame bars in PLA

However the first layer still would not stick, ho hum.

Important observations worth noting so far.

1) the sensing pin does not appear to be spring loaded.
2) it and the pads need to be very clean.

And this is very important to note :-

3) when all is clean enough the duration of touch of the pin to the pads is very short, the pin almost lightly bounces off the pad. Before the last clean with Servisol it made contact for noticeably longer as though it was pressing harder.

Back to printing.

Printing Attempt 8 - 0.8mm frame bars in PLA

The first layer still is not sticking properly but at least the bed is confirmed level, well I hope it is.


15.2 Two further options to consider :-

1) Now could it be the filament needs to be dryer, that can affect how it sticks to itself and most importantly to the print bed.

2) Adjust the Z-axis offset to bring the head nearer to the print bed, up to now this has not been an issue with ABS and is currently 1.66mm which I suspect is specific to an individual machine due to build tolerances.

Apparently when PLA is too moist it becomes brittle, that was quite the case when the end gripped in the notches in the reel broke off easily after unwrapping the new reel and subsequently after the 2 hour dry at 50C. Surplus moisture in the extruded plastic will also affect adhesion to previous layers and the print bed.

Therefore the next action I will try is drying the PLA filament again, this time at 50C for 4 hours, at this rate the energy efficiency of printing with PLA relative to ABS is becoming less well defined (ABS prints at higher extruder and bed temperatures than PLA).

15.3 Next day.

The end of the filament did not feel brittle when I anchored it to the spool prior to drying.

As I was awake at 4:30am already daylight in the UK in June I put the filament spool in the dryer for a 4 hour run at 50C so at least it will finish by 9am. Hopefully I can try and get a 3 hour print run done before lunchtime. As usual I rotated the spool by 180 degrees at about the 2 hour mark and let it continue.

Nearly finished in the dryer as I write this.

Printing Attempt 9 - 0.8mm frame bars in PLA

Reloaded filament, as soon as it started the first frame the first layer was not sticking again, so clean again as PLA around outside of nozzle again.

The inner frame parts of the frame are only 0.8mm wide and so have very little to grip the bed with, but why was this not a problem before, especially when I originally printed with 0.4mm wide inner parts.

15.4 Z-Offset

Okay, so reduce the z-offset so that the first layer is applied harder to the print bed, as the layers are 0.3mm I wll try reducing the z-offset by 0.1mm to 1.56mm and try printing again.

Pressing the down button it jumped straight to 1.56mm.

Printing Attempt 10 - 0.8mm frame bars in PLA

Better but still not sticking well enough.

Reduced to 1.46mm (steps in increments of 0.05mm)

Try printing again.

Printing Attempt 11 - 0.8mm frame bars in PLA

Better but stil a problem, when the extruded filament leaves the nozzle and sticks to the bed or previous layer it breaks away clean from the nozzle.

If it doesn't and a little stays attached then that catches onto other parts it passes over and pulls them away from the print bed too.

As this builds up around the nozzle the effect snowballs and cannot be recovered from, the print has to be cancelled and everything cleaned up again.

So now I try using general purpose Duck masking tape, this is more absorbent and not so smooth.

Printing Attempt 12 - 0.8mm frame bars in PLA

Result even worse, but looks as though the first layer is being applied too hard, i.e. spreading out, so put z-offset back to original setting of 1.66mm and try again (after cleaning up and replacing the masking tape as a precaution as it was marked by the filament).

No better, at this point I am giving up on the PLA.

How come all the earlier prints were so much better, a duff reel of PLA or some step change in the printer is wrong, I do note that the bed temperature according to the LCD screen displays 45C which seems a bit low, so many variables.

So try with the ABS I used for early printing activities.

Continued in part 2

 

Part 2

15.5 Back to ABS

Prepared bed with fresh automotive blue masking tape and reloaded the ABS cartridge, to date I have had no issues with ABS.

Printing Attempt 13 - 0.8mm frame bars in ABS

From the slicer settings, changed material to ABS, using the default print speeds of 30mm/s, but selected auto speed adjustment for small parts.

Promising but contaminated with blobs of clear PLA, does this give a clue to the earlier problems, was there PLA in the extruder that was causing a flow issue after the first 20 frame batch?

Printing Attempt 14 - 0.8mm frame bars in ABS - 20 mm/s

Aborted, re-run but also slow the print speed to 10mm/s as adopted for the PLA prints.

Printing Attempt 15 - 0.8mm frame bars in ABS - 10 mm/s

Failed, but this time the tenuous adhesion failed at an overlapping tape join near the rear of the table, the entire frame lifted and a mess followed, my over optimism.

Printing Attempt 16 - 0.8mm frame bars in ABS

So, this time avoiding the overlapping tape which was usually not an issue, the tape I use is not wide enough to cover the table easily like a 50mm tape on a 200x200mm print bed.

Also printing 9 at a time to reduce scrappage as less frames trashed if one goes wrong on the first layer

Oh dear, what I couldn't see on the clear PLA prints is far more noticeable on the green ABS prints is voids at the cross joints of the glazing bars, in addition to the breaks in the outer frame edges which would not be that visible.



But I could fill these with plastic putty.

Printing Attempt 17 - 0.4mm frame bars in ABS

This time a quick glimpse of using ABS to print with the inner frames at 0.4mm, unlikely to work but worth a look.



As expected the pile of single nozzle width layers start to collapse although would be lessened by printing a batch of 10 or more to give previous layer a chance to cool further.

15.6 Back to PLA

Printing Attempt 18 - 0.8mm frame bars in PLA

Try a repeat of the successful batch of 20 but just nine to see if the problem was due to debris stuck in the extruder that was cleared in the first ABS attempt, i.e. #13.

At last success, the parts even look better than the first run so perhaps the extra drying helped.

Before everything cools down too much I need to perform another run this time of 12 to complete the 38 plus spares I require for the warehouses, fingers crossed.

Printing Attempt 19 - 0.8mm frame bars in PLA

Fail - barely 15 minutes after the last run which was successful, blobby and frame ripping so aborted.

15.7 Further Unblocking

This time I tried unblocking by performing two operations.

1) Performing a cartridge load without removing the one fitted. It extrudes about a metre of filament.

2) Perform a nozzle clean.

Printing Attempt 20 - 0.8mm frame bars in PLA

Repeat attempt at 12 frames.
Fail - even worse, two frames wrapped around the nozzle.

This time different approach.

1) clean nozzle
2) unload PLA cartridge completely
3) load ABS cartridge, plus extra load when it queried whether filament was exiting nozzle.

There did not appear to be any contamination visible so unloaded the ABS and reloaded the PLA.

By now not impressed, after all it was a recently bought XYZprinting PLA spool for their own printer.

Reloaded PLA cartridge and selected try again to push more through.

Printing Attempt 21 - 0.8mm frame bars in PLA

Another attempt at 12 more frames.

Barely got to first frame and stringing started and became a real mess requiring another nozzle clean.

15.8 Print a Simple Test Block

Printing Attempt 22 - a simple block in PLA

I will try printing a block of PLA the overall size of the window frame before another attempt to see if that cleans up the printing process.

The initial trail put down on the bed was messy, stringed to the block and significant poping sounds from the extruder, apparently this can be due to moisture escaping from the filament. But how come I can get decent prints occasionally popping or not.

Printing Attempt 23 - 0.8mm frame bars in PLA

Fail - 7 and 8 out of 9 did not stick, run up to print was messy, aborted, wasn't much popping this time either, this is so hit and miss.

Printing Attempt 24 - 0.8mm frame bars in PLA

Another option is to try raising the extrusion temperature, extrusion temp raised from, 195C to 200C, to print 4 frames.

Printed okay, varying levels of popping.

Printing Attempt 25 - 0.8mm frame bars in PLA

Final 6 to get total quantity of up to 39 printed okay, at last job done!

Oh boy, what fun.

Continued in Part 3

 

Part 3

15.9 Summary

Once again I have pushed the limits on what can be achieved with a 0.4mm nozzle filament 3D printer with mixed results.

Printing with PLA on a daVinci 1 Pro is not easy especially when their own brand filament was so damp from new, disappointing.

The auto calibration feature is easily upset by poor electrical contact between the pin and pads, the pin is just a rigid pin protruding from the brass nozzle block, not a switch of any kind. Material of both parts are hardly reliable contact quality, corner cutting here.

Cleaning the pads and pin end with Servisol Super 10 switch cleaner finally did the job applied via a piece of tissue paper, initial attempt with isopropyl alcohol made no difference.

If when you run auto-calibration the pin doesn't just tap each corner pad on the bed then they need cleaning, if it lingers calibration is likely to fail.

I have achieved usable prints but the time and aggro resulting primarily from sub-standard condition PLA only bought a few months ago was disappointing.

It will be interesting to see how I get on with the Raise 3D printer and their PLA.

Jim

Back to episode list

 

16. Warehouse Glazing Panels

For low profile warehouse frontages the glazing is to disguise they are empty, initial thoughts were to apply translucent film to give a murky glass backed with black card or paper.

It seems that the PLA I have is poor quality despite its branding so I will use some of it to make these murky back panels.

The surface on the print bed is always smoother than the top surface so if the smoothest surface is glued to the painted frames that should give a suitable translucent murky glass effect.

For the blacking, spraying the rough side with dark primer, dark grey or black before glueing to the frames should work out okay.

An alternative would have been to print the windows old school kit form, i.e. glazing and shallow bars all one piece with the paint carefully applied with a cosmetics grade soft pad so only the raised surface of the bars are painted.

This worked to good effect on my mineral dock coal tippler building.

In actual fact if I had known I was to have the long drawn out grief with printing PLA frames without glazing I would have taken that route as it can greatly improve appearance.

Real windows scaled down properly from a full size building would be classed as flush glazing in modelling terms.

Two sizes were required.

Ground floor 30x35mm, 2mm thick, 8 required
First Floor 30x25.5mm x 1mm thick, 38 required​

Default settings for PLA used including 10% rectilinear infill.

Ground floor panels printed okay.

However marks from tape joints meant that smooth sides would have visible lines, no matter, top surface to frames is still acceptable.

Results, fitting and use is shown in the warehouse thread here.

Jim

Back to episode list

 

17. daVinci 1 Pro - More Problems – Hot End Repair

First batch of 1st floor warehouse glazing panels, failed, here we go again.

Before I knew it none had stuck to the bed or had been dragged together from one not sticking on the first layer, all ended up wrapped around the extruder nozzle and its associated wiring.

Time to clean up one of the biggest spaghetti messes I have achieved so far, however I got too enthusiastic in removing the filament and ripped a wire off near the nozzle, oh dear what have I done? It was PTFE covered wire and was the same colour as the natural PLA filament, light grey and translucent.

One thing the daVinci Pro is not good at and that is interchange ability with any industry standard parts associated with the hot end, it has an easy to change expensive cartridge instead, £££ ouch.

Just as well we didn't have a skip outside otherwise it would probably ended up there, figuratively speaking (as I would of course recycled it).

It was time to consult Messer’s Google and YouTube for assistance.

17.1 Hot End Cartridge Removal

Removing the hot end cartridge whilst not able to heat and unload the filament makes life rather difficult, everything has cold filament threaded through it.

First, was to cut the filament where it enters the feeder at the top of the printer.

I then found that to release the PTFE filament feeder tube you push on the rings of the connectors at each end so as to spring the teeth that bite into the tube off of the tube.

Then with a third hand, yes it is tricky carefully ease the tube out of the connectors, mustn't damage the teeth or even more repairs required.

With the feeder tube loose I cut the filament at the spool end where it enters the feeder of the hot end cartridge.

Then disconnected the flying leads from the cartridge.

Next difficulty was releasing the cartridge as the remaining filament prevented release from the carriage frame, firm persuasion broke the filament and allowed me to remove the cartridge without further damage, phew. Luckily this filament is relatively brittle due to its poor quality observed early in printing window frames.

17.2 Repairing

There was the broken wire connection part buried in a silicone like goo, now what, where did it go to in the goo.

Looking further online an option was to buy a 3rd party upgrade so that the daVinci could accept industry standard nozzles, as designed the nozzle is integral to the hot end for I expect reasons of cost reduction.

Anyway looking at a few YouTube videos including the ones showing how to fit the upgrade it turns out that the wire I had broken was for the hotend heat sensor which in the video just lifted out.

It was not easy to take any pictures of all this but near the beginning of this upgrade video shows the parts including the heat sensor the guy just lifts it out, approx at 0:47.



So that looked promising.

But not for me it would not, it was part buried in a silicone type goo, using some strong pointed tweezers to start removing the goo I could at least make out there was a few mm of bare wire available to repair the connection by soldering so long as I could get it out of the brass block which includes the heater and nozzle.

So I carefully dug away the goo until I was able to lift out the glass bead form of sensor with one lead still attached and then perform ever so delicate soldering and sleeving with some more PTFE tube to make the repair.

I did not have any of this goo, but perhaps this was an afterthought, it certainly did not look like heat conducting paste, anyway I was able to route the wires so it would just about stay put for reassembly and test.

17.3 After Re-assembly

More printing was performed without issue so I obtained some suitable silicon to stick the thermistor back in place. The material I chose was a clear RTV (Room Temperature Vulcanisation) silicone adhesive and sealant by ProSeal. This is rated up to to 260C, but apparently on the packaging this is the intermittent rating, whereas the continuous rating is 204C.

Anyway, all I want to do is lightly restrain the thermistor in the hot end, the nominal temperature for ABS extrusion is 215C and PLA 190C so it should be okay as it sits in a well on top of the hot end brass block which incorporates the nozzle.

Squirting a little on a glass plate to see what it set like it seems to be very much the same as what was originally used, as far as look and strength is concerned.

So next squirt some in the 'well' and insert the thermistor.

17.4 Applying the Silicone Adhesive

A week or so later after some more printing which confirmed the hot end was happy again it was time to perform a filament cartridge unload followed by cartridge change so that the hot end cartridge could be removed.

Once removed I was then able to lift out the sensor and apply some silicone adhesive around the top of the well the sensor sits in and left it for the recommended 24 hours to cure.



On subsequent reassembly the daVinci decided to throw a tantrum on the auto levelling again.

17.5 Calibration Issue Again - Different One

Auto Calibration, aka auto levelling of the table is another feature of the daVinci 1 Pro that is potentially problematic, as a quick look on YouTube reveals.

In a previous printing episode the trouble was down to dirty contacts, so prior to refitting the hot end cartridge I gave all the contacts a good clean with Servisol as before.

One would expect that refitting the hot end cartridge in its carrier and locking in place again all would be well, think again.

After a few attempts to run the calibration subsequent runs included making step changes include 12 to 17 steps in opposite directions each time.

I found that the right hand adjuster had a problem, the wheel was running loose, having tightened that up, still I would get alternate step changes to make in almost exactly opposite directions, it was not homing in, now what.

So I resorted to performing the levelling manually as in this video, starting at 9:00 minutes :-



Having completed this I thought let's try the auto calibrate again, this time it required only a few steps and after 2 runs it homed in on being level.

17.6 Why the Levelling Problem?

You would think that taking the hot end cartridge out and replacing it would perhaps have made a little difference to the z-axis offset. Why was so many steps needed to correct the levelling, the left hand adjuster could be between +/-12 and +/-17 steps which is very high step count, normally maybe 2 or 3 at the most.

This was probably due to the right hand adjuster rod coming loose for an unknown reason.

When it became too far adrift due to the loose adjuster its algorithm became confused and generated requests to oscillate backwards and forwards. Even when I thought of just trying half of what was requested it would demand a reverse of that half distance.

Whatever the cause running a manual calibration/levelling got the table in the right ball park and then it was happy to auto calibrate.

17.7 What have I learned

If auto calibration calls for excessively high changes, i.e. greater than 2 or 3 steps and it subsequently reverses those adjustments, stop, check the adjuster screws operate correctly and then manually level, followed by auto to finish off.

Maybe I should not have bought this printer all those years ago, but when is the right time to buy, it was the best around sub £1k at the time.

Finally and most importantly, never lose your temper when printing keeps going wrong and you have a significant mess to remove around the hot end, in my case I probably spent some 10 to 12 hours repairing the damage I inflicted including the calibration aggro that followed, it could also have been very expensive.

Jim

Back to episode list

 

18 Warehouse Roof Mitre Tool

This came about when I was constructing the roofs of the warehouses (Ref 18.1) using 1.25mm mount board covered with slate roof paper.

Initially I just cut the edges square and overlapped them, however despite the apex angle being approx 110 degrees, even cutting at 90 degrees the ridges looked wrong.

In plastic I would have just sanded the required mitres to the top edges of the roof panels, but that does not work well with card.

Cutting card at an angle is difficult because as you tilt the knife blade it will slip under the straight edge and a messy cut results.

To get around this I drew up a holder for a standard knife blade such that the blade was held at 53 degrees, why 53, well it feels right, but the edge of the holder that bears against the straight edge is vertical.

The blade is guided by the channel in the base and a peg in the sliding top piece passe through the slot in the blade which allows the depth of the cut to be adjusted.

The first version was drawn as :-






The printed holder, using PLA and default settings :-





Problem, with the blade perpendicular to the direction of cut and the shape of the blade it tended to rip the surface of the printed paper layer.

Next, angle the blade as you would typically do when holding the knife in your hand, I chose 45 degrees as a potentially good starting point.

However adjusting the part to what is now a compound angle, 53 degrees from horizontal towards the straight edge and 45 degrees perpendicular to the cutting direction along the straight edge was difficult.







Eventually after two attempts to create the slider, some opening out of the channel the blade sits in due to the way the holder is printed, cropping a little off of one of the blocks which bear against the straight edge I have a working tool.

Printed in ABS using default settings, tweaking the first clear PLA print was difficult without very good eyesight as the infill shows through the skin. This makes it difficult to distinguish what I am trying to modify, it is so much easier using an opaque plastic, green ABS is what I have to hand so I have used that.

Despite the angles of the sides the base printed without the need of supports, the tapered tip of the sliding part required light supports.







The CAD model needs some adjustment to correct details but at least it should be relatively easy to create a version for other roof pitches should I want to.

References

18.1 Montague Dock - 3 - Warehouses (Part 9)

https://platform1mrc.com/p1mrc/index.php?threads/montague-dock-3-warehouses.6399/#post-104914

Jim

Back to episode list

 

Thanks Paul.

Well so far the original basic straight blade is lasting the course, I have only about another 30 odd more roof panels to do, (I do have spare blades though) although a convex blade may give a cleaner cut particularly in the first tool, it may have extra drag as it penetrates, but as I do not intend doing much work with card and paper, it was the most appropriate for improving these laser cut ply warehouse kits, I won't be trying other blades at this time.

Me too, I hate it when it drifts off the straight edge and ruins a piece, I may well try a guide for vertical cuts as that could be useful for plastics, which I would normally cut vertical and then chamfer with fine abrasive stuck to a flat piece of board.

Jim

 

19. Gantry Crane Rails

Rails inset into the dockside were required.

The two cranes for the Montague Dock (Ref 19.1) have different types of wheels, the pre-production version has conventional wagon wheels but the production version has wheels that have a centre tread and flanges on both sides of the tread.



As luck wouldn't have it, a common track gauge for both cranes was not possible and reworking or replacing the wheels on either crane was not an option because of the way in which the bogies are glued together.

Having a single groove to run the wheels on their flanges did not appeal either so I adopted the design shown below which lends itself to being suitable for both cranes. The production model can ride on the centre rails and the pre-production model on the outer rails of a pair.



The base of these rails was to be mounted directly on the baseboard, held in place with track pins.

Rail height was to match that of the Peco Code 100 Streamline which sits on the 3 mm closed cell PVC foam as seen in the images of the docks so far.

On top of the 3 mm foam will sit another layer of foam to match the thickness of the Peco sleepers, nominally 2 mm thick which will overlap the crane rail base.

The top layer will be the basis of the concrete surface using another layer of foam, 2 mm light grey EVA which will overlap the Peco sleepers and crane rail base.

Railway and crane track railheads will actually be above the 'concrete' surface level otherwise oversize width wheel treads will ride on the 'concrete' and also make railway track cleaning impossible.

Buffers / End stops

Some form of buffers would likely to have been fitted to avoid derailing the crane if it overrun the rail ends, for this I decided to make them appear to be concrete and faced with wood.

Their shape is designed to match the bogie frames of the production crane.

Their height also prevents them being a trip hazard, back in the 1960s they may have been painted white perhaps, but probably not covered with hazard stripes to stop longshoremen, drivers or vehicles colliding with them.



Length and Build

A nominal 600 mm track length was required for both cranes, although too big for either printer as one piece they could be made as pairs of 310 mm lengths with the Raise3D printer as opposed to 3 sections with the DaVinci, limited to 200 mm lengths.

Simple overlapping lugs were applied to aid alignment on the baseboard and allow all the printed track components to be identical, they were to be longer but then the design would be too long for the Raise3D build plate.



After some initial trial prints on the DaVinci the final design then became as shown.



The material chosen was HatchBox silver grey PLA as a reasonable toned down rail surface without painting. Paint would not help the double flanged wheels rotate and would wear off, generally the wheels are too stiff and light to rotate smoothly so they just skid.

Only the buffers and rail ends where little traffic would occur will need some painting, the wide base will be covered by a two layers of foam.

Printing

Sliced with IdeaMaker, FOC from the Raise3D website, using their standard PLA settings for a 0.4mm extruder nozzle. Temperature of 205C for extrusion and 55C for the build plate.

A few trial printings of short lengths were made before the final versions were printed, the 310 mm rail sections took about 2 hours each, 3D printing is not quick, however the results look good to me.

Rafts were recommended by the slicer but no appreciable differences in the final print quality were noticeable, except leaving them to cool, for an hour or so helped keep them flat.

In fact the raft was so difficult to initially remove I nearly bent the printed component, so coupled with a somewhat extended print time I did not use a raft for the other seven parts.

Post Print Cleanup

A little clean up was required to remove small flares to the base of alignment lugs and to square up the ends of the rails to remove slight bulges.

Printing a railhead width of 1.16 mm with a 0.4 mm nozzle is a challenge, but only the open ends needed a little attention.

Four of eight as printed.



I'm happy with the result, decorating and installing in the dockside here.


References

19.1 Montague Dock cranes

https://platform1mrc.com/p1mrc/index.php?threads/montague-dock-2-cranes.6398/


Jim

Back to episode list

 

20. Threaded Inserts - Brass

There is plenty on the use of threaded inserts applied to 3D prints on YouTube which primarily employ the use of a temperature controlled soldering iron with a custom made tip to push the insert into place. Here I consider cold insertion.

What They Are

A brass threaded insert is typically a short tube with a threaded bore to suit the screw you intend to use to fasten two parts together and a knurled outer surface to anchor it to the material of the component it is inserted into.

These were designed for use with injection moulded parts and would have the outer surface knurled to prevent two actions.

1. rotation when the screw which enters them is tightened or loosened.

2. to prevent them being pulled out when a screw is tightened or from a tensile load on the screw from the attached component.

When to Use a Threaded Insert

If the two parts to be fastened together need to be separated multiple times for maintenance or mating parts are to be interchangeable.

If the parts are never or rarely ever going to be separated then self tapping screws can be used.

There are several designs of brass inserts available but not all are really suitable for use with 3D printed parts as they are mostly designed to have the plastic injected or cast around them. When inserting after moulding or in this case 3D printed they have to be pressed in which means that their resistance to pull out is less than would be the case for one that is moulded or cast in place.

The simple type are straight knurled inserts as below.



Inserts which have opposing angle knurls are better to resist torsion when moulded into place but I can't help believing that inserting into 3D print would not be as effective.

Opposing angle knurled inserts.

One set of knurles will move the plastic one way and the second the other way. So these may be ok when inserted hot but not cold. For that reason I am using straight knurled inserts.



Hot or Cold Insertion into a 3D Print

As mentioned above most videos I have seen use a custom made soldering iron tip, which can be purchased online too, to push the insert into place, strong results can be obtained.

However unless you have very good eyesight and a steady hand pressing them perpendicular into the mating face of the component whilst positioning accurately is difficult.

Being different I have looked into cold insertion of a brass insert. This is for mounting a fan and filter holder onto a fume extraction hood I am making for my DaVinci 1 Pro printer. The smell and other nasties from printing ABS for a hour or so is rather unpleasnt even with the windows open in this drab summer of 2023, so for winter printing fume extraction and filtering is required as on the Raise3D printer.

I also think it is morally better to trap particulates in the special filter than just vent them out of the window.

Anyway, back to the subject, many assorted packs of inserts are available online to suit the occasional user like myself but most are not really suitable for 3D printed items. Ideally the ones I would prefer have a smooth straight shank that goes into the component first followed by a knurled flange at the top end.

An example of press fit types can be found here :-

https://www.insertsdirect.com/shop/inserts-for-plastic/pressfit-inserts

At the time of starting this exercise I did not find this type available economically so I have settled for inserts with straight knurled flanges at both ends (image 1 above) which I can either use as bought or remove the knurled flange at one end to convert it to a press fit type.

Insertion Tool

This was based on a small press tool I bought from an online store back in 2014, these were apparently designed for people assembling their own ammunition, I kid you, not!



It has been invaluable for fitting 4mm wagon wheels to axles both when upgrading old stock using their original axles or using replacement silver steel axles.

The adapter was designed to be a push fit onto the 20mm anvil of the press arbor, this was printed in PLA with 50% infill density on the Raise3D printer, this prints PLA better (more accuracy and cleaner) than the DaVinci printer.

Material shrinkage is noticeable when small parts are printed with both printers, in this case to get a gentle push fit onto the press arbor the CAD model (using SketchUp Make 2017) was drawn with 20.20 mm cup diameter.

The number of circle segments was set to 60, the default of 24 is too coarse for the cup, fine for the non-critical outer vertical surface.



Print or Drill the Hole

Typically I print with an infill of 10 or 20% except for items such as wagon coupling parts which even if they had infill specified very little would occur due to component overall thickness relative to outer skin or shell layer thickness.

If the component was printed nearly solid then drilling could prepare the hole for the self tapping screw.

For light loading with small diameter screws then the shell layer could be used to self tap into (a pilot hole is desirable), for a more secure fitting printing the hole would provide it with skin layers on its inner surfaces into which the screw or an insert can bite into.

Sizing the Hole for the Insert

As for the insertion tool shrinkage of the holes in which the insert is to be fitted as well as getting the size right for cold insertion. It needs to be enough to accept the insert by deformation and firm enough to grip sufficiently to be of use.

Sample blocks were printed using the same material and infill density as the intended job to make sure that it was right first time as printing the component would take nearly 13 hours on the Raise3D.

After a few attempts a suitable hole diameter for the insert (knurled outer diameter 5.9 mm) was coincidentally 5.9 mm in the CAD model which prints as 5.6 mm diameter.

The height of the insert was 6 mm so I set the hole depth rather arbitrarily to 8 mm to catch any swarf or plastic shavings and give extra end clearance to the screw if required. Depth was not restricted by the part in which the insert is to be fitted.

The infill provides only part of the support for the insert as the holes are lined with the shell layers which in this case is 2.5 layers, (alternates between 2 and 3 layers) of 0.25 mm which totals 0.6 mm thick. I must admit I'm not sure how the layer count stacks up, okay, 2.5 x 0.25 = 0.625, but how is that created with fixed layers of 0.25 mm?

Insertion Process

The adapter and the insert are held together by an M4 cheese head screw with a washer which provides a protective face to avoid the insert deforming the adaptor face under load during he act of insertion.

The parts, note that is not a crack in the cup edge but where the laying of the extruded plastic starts and stops on each extruder circuit of the part. I believe it is possible to randomise this with the slicer for better strength.



Adapter and insert assembled.






Adapter fitted to press arbor.





Once aligned the insert was pressed into the test block and the press arbor lifted leaving the adaptor with the insert all attached to the test block.

Unscrewing the M4 screw releases the parts and if it is required to press the insert below the surface a little then refitting the screw without the adaptor and applying the necessary downward push allows final positioning.

Insertion completed.



Effectiveness

Very unscientifically I simply tried pulling it out by hand gripping an M4 screw head but it stayed firm, which is easily good enough to hold the fan and filter block to the fume extractor hood. There will be inserts in four positions, one for each corner of the fan so they should be more than ample for the task in hand.

The next stage was to update the the CAD model to apply the necessary dimension to the holes on the fan and filter block.

Concluded in Part 2

 

Part 2

Pressed in place in the extractor hood fan mounting block it has worked out well but there is scope for further refinement when I next use these inserts.



Refinements

Extra lead in at the entry to the hole, either a countersunk or counter bored entry would have made entry alignment more accurate.

However if the brass insert had the knurling removed at the end that goes in first and the hole printed to suit then a better fit and alignment could be achieved.

The friction fit of the shank would have been satisfactory with the knurling at the top providing essential additional resistance to torsion of the insert against the printed hole of the component when the screw was tightened or loosened.

It must be remembered that the inserts commonly available (shown above) were designed to be moulded in place, not pressed in after manufacture of the plastic part.

Finally with the fan fitted in place with M4 screws, rock solid, job done.




Jim

Back to episode list

 

21. Tina2S Printer - Promising, But Issues & Poor Support

This posting is all about my personal experience, the opinions and statements are mine too.

I was on the look out for a small build plate filament printer that could be connected via WiFi for printing small items with a 0.2mm nozzle in PLA.

Firing up the Raise3D to make small models is like boiling an egg in a kitchen oven, an inefficient overkill.

So the new improved Tina2S seem to fit the bill and for a price of under 200 GBP from Amazon UK and a further 24 GBP discount coupon when purchased was well worth a look.

https://www.amazon.co.uk/dp/B0B8NGNVZG?psc=1&ref=ppx_yo2ov_dt_b_product_details

Ordered one and it arrived quickly.

Printer #1

Out of the box it is compact, semi-enclosed with a heated bed, PEI magnetic surface, inductive levelling sensor and very quiet running. Initially impressive, and I mean 'initially'.

However the seperate power supply and USB cables plug into the top righthand side of the case next to the control knob inline with the processor board, and the spool hangs off the lefthand side. So the elegance and compact nature evaporates.

Thier own slicer WiiBuilder and a modified Cura 4.10 is supplied on a Mico SD card and thoughtfully an adapter to plug it into a USB port of a PC to load them onto the PC. This can also be used as a route to pass sliced files to the printer.

Printing was supposed to be possible from the card, direct USB or WiFi.

Printing via USB cable with the supplied sample of 100g of PLA filament, (nothing like the 250g spool shown in the Amazon listing) it printed quietly, cleanly with good results using the default 0.4mm nozzle.

However, the WiFi setup only works with their Poloprint Cloud service, and only if the WiFi firmware version of the printer is v1.4+, mine as shipped was v1.3, any attempts to connect it failed without any reasons being flagged up.

To make matters worse any attempts to load the firmware updates from the WeeFun site failed with error messages or the printer hanging, nothing like the sequence shown in their YouTube videos.

I pursued this with WeeFun support who on first contact responded within 24 hours, he sent me a link to the same updates. Repeated requests for a solution about why these updates would not load, via the support email address or 3 Facebook groups did not yield any further responses, despite the support person being a moderator on at least two of the Facebook groups.

Initially this was because they are in China, 8 hours ahead of the UK and it was a national holiday, that became long past and stll no response.

Tried using OctoPrint 4a as a substitute WiFi server and although that would connect to the Tina2S and Cura on the PC sending a file to the printer would fail, Cura allegedly announcing that the printer was already printing when it was not.

Okay, so all that was left was to use the SD card or direct USB cable to print, by now I am not a happy bunny.

One last attempt at getting something useful done with this printer, tools and spare nozzles are supplied, including a 0.3 and a 0.2 nozzle.

Swapping out the nozzle in place was not possible, the videos were not up to date for the 2S,the hotend assembly had to be removed from the printer, which is straight forward if not a little fiddly. Typically nozzles seize in the block, just as this had done so holding the block in a plastic jawed vice and using a small socket tool was necessary to free it, the supplied spanner stood no chance, as posters on one Facebook group and an Amazon review also found.

After re-assembly, levelling and adjusting the z-offset it printed a half size 3DBenchy quite well, however after that the z-offset had to be adjusted for the 2 subsequent prints, after a few attempts I gave up, the machine was just not rigid enough for repeatable prints with the 0.2mm nozzle, stability of the z-offset height is critical.

Summing Up

This machine is being heavily promoted as a beginners machine and Christmas present at this time (October), however if it is shipped with v1.3 WiFi as mine was it does not work with their cloud app (WeeFun website) so all a beginner can do is print one of the many examples on the supplied SD card.

Supplying tools and offering easy support is good but neither were up to scratch.

When I printed test items of my own with their filament and two type of HatchBox PLA it print quietly and cleanly.

However buying their own filament is expensive at about 40GBP for 4x250g spools (HatchBox 1kg spools are 20GBP).

To use a 3rd party 1kg spool requires you buy or 3D print a form of support, e.g. a desktop roller type as they are too big to hang from the printer.

Filament 3D printers are relatively easy for beginners to use with PLA but can still have maintenance issues, especially when parts do not adhere to the bed which quickly leaves a beginner at a loss of what to do.

Also downloading items from sites such as ThingiVerse can be problematic if the user knows nothing about the need for supports, brims or rafts.

Result

I gave up in disgust and returned it to Amazon for a full refund, because if I left it much longer I would not be able to return it to Amazon within their 30 days returns window after which I think you are at the mercy of the seller directly.

Many parts of the printer do not have a 1 year guarantee, but nevermind, plenty of spares are available to buy from the WeeFun store on Amazon

I could be cynical and think that the subsequent lack of response from support is such that I miss the Amazon return window, no I won't say that, they would not act like a dodgey eBay seller, as they are so committed to the buyer, afterall their support sign off is :-

"We value your satisfaction and trust in our services, if there is anything else we can assist with or questions we can answer please do not hesitate to contact us."

Continued in Part 2 below

 

21. Tina2S Printer - Promising, But Issues & Poor Support - Part 2

Printer #2

A couple of days after leaving a suitably critical review on Amazon their sales person contacted me via Amazon Messaging very apologetically and offered me a discounted machine plus a 40GBP filament bundle as a free gift.

By this time I really was not in the mood to respond, but a few more days later another message offered me a new replacement machine totally free of charge including delivery. Bearing in mind I had already returned the original machine and received a full refund by Amazon, which they were aware of, I considered this offer.

They asked if I would update or delete my original review, however before accepting it I stated that I would apply any update of the review in a fair to all manner, a number of reviews already have 3, 4 or 5 stars for defective machines that were replaced.

In other words, a free machine will not guarantee a review deletion or an upgraded review. They accepted that and a day or two later a new machine arrived direct from wherever, but not in traditional Amazon packaging and using a 3rd party courier.

On receipt of the 2nd machine, inspection showed it had the later appropriate WiFi firmware version, more subtly there was only one spare nozzle, size 0.4mm and a stronger nozzle spanner.

The first machine had a spare 0.2, 0.3 and a 0.4 nozzle, however the nozzle design had changed back to those used on the previous (non-S) model which is handy because a set of three sizes of those are readily available on Amazon.

Anyway, many attempts were made to connect the machine to my WiFi networks, the secure house one and the less secure to suit my DaVinci and Raise3D. Even trying all sorts of adjustments using a seperate wireless access point the printer would not connect to any WiFi set up I could configure using their cloud based app.

However I can drive the printer via the OctoPrint 4a app on a Motorola G7 from their version of Cura 4.10, but not from their slicer directly via the WiFi. Cura 5.3 is included with this printer but is not compatible with Win7 on my laptop.

I replied to the sales person, and they sent me a link to one of their YouTube videos of how it was supposed to work, same as in the user booklet, of course printer #2 responded, 'So?'

She also gave me a WhatsApp tech support number which is for, as one might guess in China. I now attempted to contact them as a last resort, and as China is 8 hours ahead of UK time it means phoning them from the UK between midnight and 9am. Seeing as the sales person was working UK hours I tried in the afternoon, UK time, but no answer.

Next try, UK breakfast time, stiil no joy, they don't answer, they can't call me but can message me, that was a slip up by me, my wife's phone was in silent mode, whoops

So I fired off specific questions to sales again, but I do really wonder if these people actually understand English or it just gets badly translated each way by some application they run.

How images of the messages like 'Cannot Connect' or 'WiFi Timeout' tell more than I quoted, I do not know.

The only technical information received was that the network required to be DHCP as the printer does not have a default IP address.

Anyway to cut a tedious story which had spread over nearly 5 weeks in all I wrote specific questions about what security protocols the Tina2S could handle. There's nothing secret about that surely, but all they kept asking for was images matching what I said, so either they did not believe me, or may be an image is used to match an image in a look up table so they can send a reply I do not know.

I repeated my questions on all three of their Facebook groups, finally got an answer, send us a picture

At which point I replied curtly, and closed my Facebook account which I had only opened again to view their barely active user groups.

Conclusions

If you buy one of these machines, which are quite advanced for the price, and it can print PLA well, be warned that if anything goes wrong, support, which admittedly is not labelled 'Technical' appears to be anything but technical, you are virtually on your own.

Make sure you give it a good testing so if faulty can be returned easily and free to Amazon within their 30 day window, do not let it drag on.

Both sales and support are all very polite, charming and apologetic, but that does not make a device work.

The longevity of these machines is unknown as they are of light weight construction, but I viewed it as a possible occasional printer of small items using a 0.2mm nozzle.

I will get back to trying this machine with a 0.2mm nozzle and see how well it works via Octoprint 4a.

My original critical review on Amazon UK of the machine and support stands, just look through 1 star reviews, pity there's not a zero star option.

Which version will you receive?

Jim


Back to episode list

 

22. Montague Dock Walls

These dockside walls are to be made to look as though they are stone built, and capped with broader flater stones, the model length is approximately 11 ft long (3.3m). Creating a relatively uniform wall of that length is tricky such that it look like a continuous build as it would be in real life.

Each panel became 245mm long and would interlock with neighbours in an attempt to create a continuous wall.

The model design method used here is for SketchUp (Make 2017 in my case) and uses the component feature.

Design of the basic wall was started as a flat surface of regular rectangular blocks to suit the external appearance of the wall such that each section contained, a single rectangular block which was created as a component and replicated.

Making the brick a component makes for a lightweight SketchUp model and adjustments to one adjusts all at the same time, this is useful when adjusting the size of the bricks and spacing to suit the size of the wall.

Now the tricky part is to get to mortar base in place, a suitable technique is contained in the YouTube video "Modeling a Brick Wall - Skill Builder" created by SketchUp.

In reality the depth of brick and stone mortar lines would be closer to that of a brick paper, i.e. flat, as opposed to moulded grooves, however the exagerated moulded, or in this case printed mortar lines look better than printed to me. Also for custom stone or brickwork it is easier to pickout the grooves wth a mortar colour paint than draw lines in a mortar coloured ink or paint on paper, especially the vertical mortar lines.

Typically printed papers I have used have a satin sheen to them, never matt, and even multiple coats of protective matt spray does not flatten them to matt, I don't want my walls looking as though it is constantly raining.

This video shows how to add depth to an otherwise flat decorated wall in a SketchUp model, from approx 9:18 it applies to 3D printed walls as well, remember that the coloured textures will not be printed, only the physical form the model represents.



It is important to note during the scaling process that you need to scale beyond the surface and let it snap back, this I found to be the trickiest aspect of the process.

The resulting plain wall design :-





With position for a steel ladder inset into the wall, similar to that on a canal lock near to where I live :-



Capping Stone design :-





The walls and capping stones were printed in PLA with 10% infill, which is almost irrelevent as the walls and capping stones are thin.

The Raise3D printer was used as it's bigger build plate means less wall panels per metre.

The walls were printed stone face up.

The capping stones printed face down, by printing face down a smoother finish was achieved by virtue of the BuildTak surface of the build plate and also significant supports were not required.

Next image shows a test fitting of plain wall and capping stones with the wall reduced to the required height (using a miniature table saw) and sprayed with grey primer. The walls will be glued to the hardboard strips.



A wall with ladder position just printed.



Fitting of these will be shown in the Montague Dock thread.

Jim

Back to episode list

 

23. PLA, ABS, PETG - Part 1

April 2023 I started using my printers in earnest and after consuming a few kilograms of filament I thought I would pass on what I have learned as a newbie to 3D printing with plastic filament.

My machines at present comprise a 6 year old daVinci 1 Pro, a Raise3D E2 and a Tina2S, very different beasts even when printing PLA.

There are various forms of PLA which behave differently, no two printers behave the same, one persons settings may be unusable by another depending on subtle mechanical and firmware differences.

The nature and moisture content of the filament, adhesion of the extruded filament to the build plate and to the previous layer, and the thickness of the first layer all add to the variables affecting the quality and success of a print.

This is compounded by which slicer you use and what the default and advanced setting values are, build plate (or bed) temperatures, enclosed or open build plates, so what works for me may not work for you.

The materials I have tried to date are 1.75mm PLA, ABS and PETG, I will discuss each in turn.

More detailed information may be found at the website references below, also there are many material reviews and comparisons available on YouTube.

Material Friendliness

Two aspects to this, to the user and to the environment, specifics are mentioned for each material.

Disclaimer: DO NOT take these notes without question as many people can be allergic or have an adverse reaction to something which is fine for others, if in any doubt consult official sources. I offer this overview in good faith as a user and not as a qualified scientist.

PLA

• Low Odour, no toxicity (Ref 23.1), can be printed in the open.
• Supposedly environmentally friendly by being plant derived and compostable, however it is only industrially compostable, did not know that early on, it can last well in your home composter without decomposing.
• Low on stringing and available in a wide variety of colours
• Can be painted.
• Reasonably tough but can be brittle compared with ABS or PETG.
• Poor elasticity, can deform easily when it clip fits to another part.
• Available in matt, gloss, metal, silk and wood effects.
• Relatively easy to glue with cyanoacrylate adhesives.
• In addition Plastic Magic from Deluxe Materials (DM) has proven very effective for bonding PLA to PLA parts. I contacted DM and asked them which of their products was suitable as they do not currently mention it in their product range, their reply was "We have heard that Plastic Magic AD77 works.". I am still using its predecessor of the same name but marked AD24 and that certainly bonds PLA-PLA well.
• For gluing to larger scenic parts such as wood, urethane glues, e.g. Polyurethane Ref 23.2, can also be used, however they do have disadvantages, slight expansion on curing e.g. with Everbuild Lumberjack, comes in large containers, even 750g is rather large, and less than a year to use it, even before opening it. Requires that one surface is porous.

ABS

• Nasty throat annoying odour, some toxicity (Ref 23.1) and gives off harmful particulates (Ref 23.3), best printed in a well ventilated area or in an enclosed machine and the fumes extracted through specific filter types.
• Oil based, long time to decompose.
• Higher temperatures for extrusion and build plate required when compared with PLA, tends to string when it starts to get damp.
• Can warp and lift from the build plate when on a 3 hour+ print run, this can be alleviated using rounded corners to parts in contact with a heated bed or attaching lilipads*.
• Can be painted.
• Tougher than PLA, almost as tough overall as PETG.
• Much better elasticity than PLA so can be used for clip fit parts.
• Has a range of finishes available.
• Easy to glue with ABS compatible cements, a commonly used material for many moulded products.

PETG

• No odour or toxicity. (Ref 23.1).
• Yet higher temperature than ABS, tends to string easily, and likely to require lilipads* on sharp corners to prevent corners lifting.
• Difficult to paint, generally has a high gloss finish.
• Good elasticity, greater strength than ABS particularly between layers.
• Difficult to glue, which can be useful, RTV silicon can be used to limited effect.

*Lilipads (also referred to as mouse ears) are small thin partial discs that look like lilipads/ears which are attached to corners by just enough to hold onto the part being printed. Some slicers include this as an option, otherwise it is easy to create your own, which is what I do. A set of 4 in a SketchUp file, 10mm diameter and two layers thick, which I import and apply, one is shown below, note the teeth to enable easy detaching from the printed part and prevent it being part of the components base layers.



General

More information on

• materials at reference 23.4.
• gluing 3D printed parts can be found at reference 23.5.

Food containing products should not be 3D printed for multiple reasons, here are a few.

• The layered build of a 3D print is full of grooves and crevices that cannot be cleaned.
• Your machine is not hygienic.
• May leak.

Additional on PLA

I have used four types of standard PLA (as opposed to e.g. PLA+), as sold by most printer manufacturers and 3rd parties.

Basic

XYZprinting PLA which comes as 600g refills, a non-standard size for their cartridges and relies on microchips to supply temperature and material remaining information, the latter can leave 10m still on the spool. However it can be bypassed as the DaVinci 1 Pro can accept 3rd party filament either by a printed hook internally or fed externally from a 1kg spool.

Raise3D Premium, 1kg spools, good quality, red and white used, glossy and expensive.

WEEFUN small 250g spools, sets of four, expensive, only used the supplied with printer 100g sample, printed well with good finish.

HATCHBOX, 1kg spools, various colours, my default supplier via Amazon UK.

HATCHBOX filament prints well on the Raise3D E2 and Tina2S, but my daVinci 1 Pro is not good with PLA, the prints do not look very smooth.

I especially like their matt grey PLA which prints well with a smooth finish, suits me for models I will be painting.

A slight metallic, HATCHBOX Silver, which seems to print coarsely, layer adhesion not as good as standard PLA, okay for larger objects but can show as discontinuities, possible fracture points when joins are made between the start of an extrusion and completing a complete circuit of an object, e.g. a layer of a shape such as a circle or rectangle.

Different settings including randomising layer starts need evaluating.



Wood effect

FLASHFORGE 3D Printer Filament Wood PLA bought by mistake, prints relatively easily albeit a little hairy, not stringy as such, looks and feels good for handles of tools.
.


Clear

ERYONE PLA, tended to be a bit brittle and easy to tangle on the spool at times, do not expect a clear result due to the printing technique, more the translucent effect you get with low obscurity house glazing. Although I believe you can post process to make it clearer , but if the material is thick enough to require infill then below 100% the pattern will show and I suspect if 100% infill then heading to opaque as thickness increases.

I did manage to print an open top small test box about 70mm x 60mm x 20mm one layer thick with a 0.4mm nozzle on the Raise3D but without polishing it was still more translucent than clear.

Additional on ABS

Once again I find the HATCHBOX range are very good value, prints well in the DaVinci (since I modified the built plate surface and added a fume extractor, both to be described in due course) and the Raise3D. In both cases dampness can cause issues and the Z offset was critical in the DaVinci to achieve a sound first layer.

As for PLA XYZprinting ABS filament comes as non-standard 600g refills.

Summary and References in Part 2 below.

 

23. PLA, ABS, PETG - Part 2 (Continued from above)

Summary

There is a lot to learn to create a sound good looking printed object, bed adhesion and filament drying, and keeping it dry are further considerations as all these plastics are hygroscopic to varying degrees.

There are so many brands of filament out there, I have just stayed mainly with HATCHBOX because it performs well enough for me and is available easily via Amazon UK at a middle range price.

Filaments that tend to tangle on the spool will effectively knot, jam and be ground into by the extruder teeth. The result is a loss of feed to the nozzle and a ruined print. In addition with a direct drive (extruder gear in the moving print head) like the Raise3D, plastic dust was peppered onto the print and left a mess on the top of the hot end which fused like a glazing.

Worse still on one occasion the filament carbonised, (became part burnt and solid) inside the nozzle and caused a difficult to remove nozzle clog.

Cheaper brands can be more prone to tangling because they are not wound evenly, but even some better brands can too, listen out for banging noises or a jumping spool, that’s a warning that a mild 'knot' suddenly released. The Raise3D machine is capable of pulling a loaded spool out of an external holder or even snapping filament which is brittle like the transparent material I used above.

The transparent PLA seemed prone to tangling as the surface of the filament seems almost sticky on the spool, however it only jammed twice in most of the 1kg spool, care in routing into the Raise3D printer from outside of the printer was more reliable than when mounted in the printers spool compartment.

To date I tend to find that the HATCHBOX filaments are the most cost effective for my uses which includes modeling, domestic repair jobs, simple tools and fittings, other brands may be better performers but for the moment unless I need a material or colour not made by them, they are my first choice. Consistency reduces waste, time and electrical energy, and less scrap for landfill, which it will all end up as, be it as failed print, skirts, brims, rafts, supports or end of life components.

Recycling one’s own scrap does not appear to be cost effective in terms of energy from the YouTube videos I have viewed.

I hope this was useful, any comments, suggestions and discussion always so that we can all learn from each other.

Jim

Added more about PLA in Part 3 below (16th Dec 2023).

Back to episode list


References

23.1 All3DP - PLA vs. ABS vs. PETG: The Main Differences
https://all3dp.com/2/pla-vs-abs-vs-petg-differences-compared/#google_vignette

23.2 Autodesk Instructables - Testing the Best Glue for PLA 3D Printed Parts
https://www.instructables.com/Testing-the-Best-Glue-for-PLA-3D-Printed-Parts/

23.3 WikiPedia - ABS - Hazards for Humans.
https://en.wikipedia.org/wiki/Acrylonitrile_butadiene_styrene#:~:text=Hazard for humans,ABS is stable&text=Concerns have been raised regarding,by a buildup of UFPs.

23.4 Simplify3D - Materials Guide
https://www.simplify3d.com/resources/materials-guide/

23.5 All3DP - Gluing 3D Prints: The Best Glues for PLA, PETG & ABS
https://all3dp.com/2/gluing-3d-printed-best-ways-bond-3d-prints/

 

23. PLA, ABS, PETG - Part 3

PLA - the environmentally friendly option, or so I was led to believe.

I spoke of this in part 1 and have since found more to show that it is not really that friendly, bad enough pure, but many variants of it have many other additives to aid flow, colour, strength and give varying effects from wood, silk, gloss, glowing and more.

Here are two videos from YouTube by prominent 3D printing guys giving their findings.

Maker's Muse :-

"PLA - 3D Printing's Biggest Lie"



CNC Kitchen :-

"I tried composting "biodegradable" 3D prints!"



Summary

Well so much for me thinking I could compost my scrap, but the idea of melting it down in a small cheap oven is a good idea (Maker's Muse) because instead of having a pile of odds and ends I can at least create small sheets of material that could be used for other purposes and in the mean time can be stored efficiently.

I bought a small cheap oven for dessicant drying, and will post about that in the near future.

Jim

Back to episode list