Ballast Cleaning Train - Part 16 (SLA Resin 3d Print Version)

There is a lot that could go wrong with the design and print of this generator wagon chassis. For a start there is a lot of thin levers, rods and brackets in the underframe that may not print well. Then, there is the distance between axle boxes; did I get it right taking into account fitting of brass axle bearings and free running wheels.

Well, the print (6hrs 42 mins) was near perfect. The only damage was a few breakages as I cut away the intense support structure. (I think the Lychee auto support facility goes OTT on quantity but, it is a godsend for the novice.) Plasticard infills repaired the breakages.

One mistake was to use a hairdryer to quickly dry off the print after water washing. The heat warped the axle box tie bars badly. The print is quite pliable prior to UV curing so, I used various materials with required thicknesses to prop and force the tie bars as straight as possible before the UV curing process. The UV then hardens the plastic allowing parts to hold their shape. The tie bars were much improved although a little warping is still present that I'll have to live with. I really don't want to reprint the model.

The wheel sets fitted perfectly between the axle boxes. However, they did not turn freely. I had made a big design mistake by setting the brake shoes too close to the wheel treads and flanges. The brakes were locked on! I had to snip off the break shoes, cut the brake rods back a little and re-glue the shoes in place. Did I say I really don't want to reprint the model. (It uses a lot of resin and takes too long.)

Talk about printed detail - the individual leaves of the leaf springs are clearly visible! Each leaf must be no more than a quarter millimeter thick.

Postscript: I had a eureka moment about the warped tie bar. I glued a piece of stiff wire behind the tie bar which pulled the bar straight.

To Part 17.

To the beginning.

Read about the SLA (stereolithography) resin 3D printer here.

Read about the FDM model here.

Ballast Cleaning Train - Part 15 (SLA Resin 3d Print Version)

 Printer worked flawlessly this time.

Here are the cabins for the ballast cleaning machine and generator wagon plus two more components (print time 7hrs 42mins). Once again some fine details visible.

The cabin on the left has two hand rails replaced with copper wire because I knocked the printed ones off whilst removing supports. The auto support generator facility of the Lychee Slicer Application was a bit over ambitious there stuffing the rails with unnecessary supports that were hidden by other supports  (and on the printed model). I could not see them in the Application to manually adjust but, have since found out how and have adjusted the offending ones.

I believe there is just one more model to print and that is the generator wagon chassis.

To Part 16.

To the beginning.

Read about the SLA (stereolithography) resin 3D printer here.

Read about the FDM model here.

Ballast Cleaning Train - Part 14 (SLA Resin 3d Print Version)

Change of plan. 

I like to be around during the print to keep a check on things - and it is a good job I did (more on that later). It is a case of knowing when I am available against time needed for a print run so, the cabin mentioned previously is deferred in favour of others.

First up are the conveyors (print time 6hr 50mins). Two needed a reprint due to inadequate support structure and even after further optimisation the third one down in the photo has a deformed end. Fortunately this will be hidden when assembled so no problem. Removing the support structure is fraught with difficulty as it is easy to break elements of the model. In this case plasticard implants saved the day (e.g. the white part on the right hand end of the nearest conveyor).

On checking the print progress I was alarmed to see the print had paused and an error message on the screen saying the USB stick was disconnected. It seems for each print layer the program reads the file on the USB stick memory rather than importing the file to internal memory for processing. The error message included an instruction to reinsert the USB stick. Having done this the print resumed exactly where it left off without any damage to the finished model. That is impressive! However the same problem arose 4 or 5 times early in the printing cycle before it settled down.

This event has been reported by others. I have not found the cause but it could be due to one of three reasons, programming fault, dodgy contacts or unstable USB stick. I changed the USB stick and the next print run was faultless. But, my confidence was a little shaken and now feel I must monitor print runs periodically.

The next print was the girder frame (print time 9hr 42 mins). This model is extremely delicate and needed a mass of supports, many in awkward places making them difficult to remove. Great care is needed. 

I am pleased to say there were no deformations!

I was really impressed with the quality of this model, especially due to its complexity.

The hand rails (photo below) prove that this technology is ideal to achieve fine details. They are only 0.5mm diameter with a 0.5mm gap to the girder.

To Part 15.

To the beginning.

Read about the SLA (stereolithography) resin 3D printer here.

Read about the FDM model here.

Ballast Cleaning Train - Part 13 (SLA Resin 3d Print Version)

I was not totally satisfied with my FDM 3D printed ballast cleaning wagon models, due to lack of detail, plastic layers evident and design errors. This lead me to buying an SLA 3D printer, which gives injection moulded quality without the high development cost.

This series of posting is as much about my experience as a novice in this technology as the development of the model.

I started with the ballast cleaning machine to be followed later by the generator wagon. Potentially the model could be printed as one piece instead of the 60 odd parts made on the FDM printer. However, I am aware that invariable it is not until the SLA printer has done its work that we can review the finished model for flaws and since the process takes many hours then a serious flaw in the model results in wasted time and materials. In view of this I split the model into about 20 parts. (I do know that tools are available to assess the model for flaws before it is printed)

For efficiency the build plate of the SLA printer should be filled with parts as the printing time is dictated by model height rather than number of models on the build plate. As a novice I expected failures so, to avoid waste I limited the number of parts.

First, I produced a small batch of odd ball parts. Their .stl files were loaded into Lychee slicer software and the automatic support feature activated to create the essential support structure. The output file was loaded into the printer and print started. 3.5 hours later the prints were ready and yes, one of the parts failed. The first oil drum in the photo has a split and is deformed. This was solely down to inadequate supporting structure. This one's support structure was different to the others. I simply copied that of a good model and printed again satisfactorily.

1st attempt with support structure

Next was a more complex model - the ballast cleaning machine chassis. After 5.5hrs the chassis was ready. Yes, there were flaws. They did not look too damaging but as there were quite a few deformations and I accidently broke off some of the steps I decided to strengthen the steps in the design, review the support structure and print again. Quickly I learnt that the support structure is critical and the automatic support generating feature of Lychee is not a panacea for success. The deformations helped me in understanding where more supports are needed and these were manually added. 

2nd attempt

After washing uncured surface resin away the support structure is severed from the model using snips. Great care is needed in doing this because delicate elements are easily damaged - and this did occur. Thankfully a spot of Superglue worked well to stick back damaged parts.

There were still some very minor deformations that I can live with, except one place being more noticeable. I did not want to print again so I filled the gap with a piece of plasticard filed to fit. I also accidently knocked off the edge of the buffer bean (This stuff is very brittle) which was corrected with more plasticard.

I am very pleased with the quality of finish. Surfaces are smooth, very small parts well formed and rivets that I added to the buffer beam in the design are visible.

Next to prepare and print is the cabin.

To Part 14.

Read about the SLA resin 3D printer here.

Read about the FDM model here.

SLA 3D Resin Printer #2

I have nearly gathered all the necessary equipment. Only a UV chamber left to obtain and that is on order. 

Apart from obtaining the necessary extra equipment there is some important preparation to do. First was read/watch as many tutorials that I could find. 

One of these revealed the importance of levelling the machine. I had to make some standoffs for the two rear feet due to a sloping floor. 

Another was to print a calibration model to check optimised settings. Several .stl models are available free of charge from the web for this purpose. I selected 'RESIN XPT VALIDATION' as it is low profile for quick printing.

I had previously tested the two main slicer applications Chitubox and Lychee and decided on Lychee since it fully supported both my printer model and resin (Anycubic water washable resin +) and I liked its automatic support creation feature. However, I did not need supports for this calibration model since it is flat. Having configured the default, optimised printer and resin parameters I imported the .stl file and output the printer file. 

I needed to make a heater stand (see inset picture) to accomodate the heater's power cable that sticks out the bottom of the heater. This was made using my FDM printer. The cable was then bought out of the machine via the gap between machine and its cover.

The heater does not have temperature control, just an overload cutout so,  a separate temperature controller is needed to switch power on and off to it. There is a small access hole at the back of the printer through which a thermocouple probe could pass. The probe was taped down in the gap between cover and vat near the front.

Whilst waiting for the UV chamber to arrive I decided to print the calibration model since it does not really need UV curing. The working temperature of the resin is 20 degrees C +. The chamber ambient was 12 degrees and the resin 15 degrees so heating was needed. The heater raised the resin temperature to about 22 degrees in 1 hour at which point the print was started (by the way I used a temperature measuring gun aimed at the resin for accuracy). 20 minutes later the model emerged. 

To find out whether the result was adequate I watched another tutorial about it. The model details confirmed an adequate parametric configuration, even though most of the holes were covered. I think this may be due to being printed flat to the build plate.

This is the heater I am using that fits the available space in the Creality Halot Mage printer. It is an unbranded Chinese product available from many UK suppliers.

To Part 1.

Read about my experience printing my first model here. (4mm scale Matisa Ballast Cleaning Machine & Generator Wagon).

Basingstoke Model Rail 2024

Here we go again. Our annual visit to this show.

It is encouraging to see such fine layouts and bounteous, fully stocked trade stands. The hobby is not dying! Second hand rolling stock prices were mostly affordable for anyone entering the hobby or wishing to expand their rolling stock.

What was a surprise was the ease of parking on site and quick entry. We rolled up soon after opening on the first day. Usually, we don't attend at this time as invariably the car park is full with a long queue at the entrance to the exhibition itself . Not this time. The show was busy but, without uncomfortable over crowding.

Only one layout had been seen before. My top three favourites were:

ROPLEY. Faithfully modelled from the prototype of today. Liked this as it is a station we have visited several times. Always interesting to recognise the environs of a place well known to us.

BAMFYDEL. An authentic looking, uncluttered scene and yet it is an entirely fictitious location.

GRINDLEY BROOK. On entering the first hall we were confronted with this massive layout. Its size should not be surprising as it is a large station complex constructed in 7mm scale. In fact this layout occupied the entire centre space of the hall whereas in previous years you would expect four or five smaller exhibits. Traders stands were placed around the perimeter with (by contrast) a 009 layout stuffed into the corner of the room.

7mm scale done to this level of expanse and detail  can be easier on the eye than the smaller scales. Because of this it was my personal best in show.

Only a couple of 1mm drill bits were purchased from a trader. I could not find any in the local town shops, unless I bought them as part of a larger set that I don't need.