13. Type G Frame - Full Height, 2x Width This frame is a generic frame without a specific purpose at this time, it is modeled on the double width shop front windows and comprises of two type A frames joined as one. The first print is to make sure that the dimensions are correct after cooling shrinkage has taken place. As two connected type A frames it can either mount two type A fascias or what is really aimed at is wide archways, windows or doors that cannot be fitted to a single full width panel. The internal parts of the frame are to be reworked as required as only the outer features of the frame matter for connectivity to the connector beams, base plate and adjacent panels. CAD model of the generic frame Print of the generic frame Summary Simple generic frame to support two type A fascias or the basis of a custom frame. Next Type H Frame - Short, 2x width Jim Back to Article List
14. Type H Frame - Short, 2x width This frame is simply a reduced height version of the type G frame for roof top mounting, created for the same purpose as the type G frame. This generic frame is built from two type B frames. CAD model of the generic frame Print of the generic frame Summary Simple generic frame to support two type B fascias or the basis of a custom frame. Next Frame - Alternative Printing for a Better Formed Top Flange Jim Back to Article List
15. Frame - Alternative Printing for Better Formed Top Flange The top flange has been a tricky part to cleanly print as the panel frames are printed face down to minimise the need for supports and where the support is provided via slicer options it does not appear to be working very well. When the lower outer shell surface of the flange is started it not only droops but the laying of the filament around the corners is poor, the filament is not being supported on the corner. The filament does not stick and hence it cuts the corner and often results in the filament along the flange edge being out of line and not staying where it should. The issue occurs here (cruelly magnified) :- The supporting structure is where it needs to be when looking down on it from above, so what was the problem. Here lies a misconception I had about supports, yes I knew that they required to have minimal contact with what they support otherwise they become one with the supported face but what I did not know and I am only referring to filament based 3D printing here, not resin, is they do not actually reach up to touch the part they are supporting at all. This is supposed to be a support? Well it turns out for filament printing that it intentionally stops one layer short of what it is to support (Ref 15.1), and what it actually supports is a drooping first layer of an overhang. It supports a droop! The DaVInci's slicer XYZware Pro makes no reference to it but the Raise3D slicer IdeaMaker allows this gap to be adjusted using the "vertical offset down layers" parameter which by default is one layer (Ref 15.2). This may all be very well when the droop is not distinguishable on a large curvy part but I am printing small parts with angular faces, so single layer, even of 0.25mm droop is very noticeable. What makes matters worse is that I try to lay filament around a corner, so extruding in mid air the filament does not sick to anything before or after turning the corner. It's no wonder I have had problems, I believed a support actually supported. The HatchBox matte grey I have used so far is very forgiving but another I tried, the FlashForge matte white has proved to be almost unusable in the E2, since then I have ordered some HatchBox matte white, which I wished I had been able to buy that originally. I want a matte white for printing concrete style panels in a similar colour to the Arkitex white. Since drafting this artice I have obtained a spool of HatchBox matte white and it is as good as the grey. Alternatives The available slicer options do not really help and adding a light weight removable support as part of the CAD model has proved difficult. What came to mind next was to remove the corners I try to print in mid air and introduce straight bridges which droop very little in comparison depending on the material and print settings. The basic idea was to print two frames at a time and butt the two overhangs together via a thin single layer bridge which can be severed after printing, as shown below. The bridge is made one layer thick and 0.6mm wide to accept a razor saw to separate the two parts. It can be seen in the IdeaMaker slicer preview images below, the first at 50% printed the second at 100% that the first layer of the overhang is a straight bridge between flanges. This was then printed, the result as follows, the central support however stuck to the BuildTak rather too firmly, and flexing the printing surface enough to release the support was difficult. Once released the flange surface above the support looked promising. The next image shows the two flanges cut apart, the form of the edges are better than printed singly but the surface edge that remains is almost like a fabric where the seam has been cut off leaving strands sticking out waiting to unravel. The trouble is at the point of cutting is unsupported, that 1 layer gap again which tends to detach the sawn outer layers. Possibly a thin plastic shim and using a sharp knife would be better than using a razor saw. The edges were tidied up by removing protruding loose hairs with a scalpel against a cutting mat, the remaining loose edge strands washed with DeLuxe Materials Plastic Magic and pressed down by finger as they dried. This has reinforced the edge. The end result is a much improved edge to the flange which is a functional part of the frame. Once firm a gentle chamfer can be applied to edge with a needle file to finish it off. Summary Here I looked at the fun one can have with supports when printing small angular parts with a filament 3D printer and an alternative way of printing this difficult detail so it can perform its function better. A simpler alternative is to make this flange a layer (0.25mm) or two thinner so the overall thickness is what it needs to be by allowing for a one layer droop. The extra thickness is not just caused by the droop down but also the partial tangling of the shell layers along the length of the flange. However the HatchBox matte PLA appears to be quite forgiving in this application and the FlashForge matte PLA very poor. At present the parts printed with the HatchBox matte PLA mostly function correctly if not a little tight but I do now have a simple alternative in printing joined twins and then separating them. Next Custom Base Plate. References 15.1 3D Printing Supports: 3 Simple Steps to Success https://all3dp.com/2/3d-printing-supports-guide-all-you-need-to-know/#:~:text=Make sure your model is,of the 3D printed overhang 15.2 Support Vertical Offset Down Layers https://www.ideamaker.io/dictionaryDetail.html?name=Support Vertical Offset Down Layers&category_name=Support#SupportAdd Sparse Connection in Support Jim Back to Article List
16. Custom Base Plate Due to the limited space available for the Royal Mail warehouse frontage a custom base plate was required so a single frame base plate was drawn up and printed to ensure that the critical parameters matched those of an Arkitex base plate. Printed in PLA and shown next to a 3 hole base plate. With the geometry of the holes for the connector blocks and the pitch confirmed correct between them the required custom base plate was modeled. As this low profile building does not have a back those slots in the base plate are omitted and allows a lower profile to be achieved. and printed in HatchBox matte grey PLA. The assembly of the Royal Mail warehouse frontage onto this custom base plate confirmed its suitability for the job. The CAD model of this base plate can now be used as starting point to create further custom base plates. Summary This was quite a straight forward part compared with the frames and fascias but still a very useful addition to my library of CAD models, it printed easily and cleanly with little dimensional error using an infill of 10%. Next Evaluating the work so far. Jim Back to Article List
17. Evaluating the work so far Apart from initially making this a one liner by finger trouble , so far so good. I was going to show the RM warehouse frontage completed but I have been distracted by a host of design and development of parts to follow of which an offshoot will be applied to that building. Currently ideas for extra parts for factory and shop buildings are flowing out of my skull faster than I can refine and print them (despite filament printing being a slow process ), including improving some original parts which do not enhance the buildings at all. One particular item, rather two, no three, anyway three to start with, that have always bugged me are the lightweight roof edging, warped floor panels and parts for building convincing overhangs. The first two can almost be ignored when the roof is higher than my eyes, which is certainly the case with the large electrical engineers illustrated earlier. However the RM frontage is looked down on and annoys me, so that will benefit from improved roof parts, hence delaying completion for a week or so yet. Also I have been evaluating limitations in the printing process, printing with filament in 4mm scale requires detail tolerant eyesight, at 3 feet, large features are not that sharp to me, so the surface finish is easily tolerated now, 20 years ago, no, next 20 years As parts are developed they can be used to create further parts, a library of proven CAD models is developing and a significant change is required to filenames and applying reference numbers so I can catalogue them. A current almost finished set of parts numbers a 17 part kit of components with a refined part number system, but more about that in a specific article to follow. I would also like to print reference numbers on the parts but previously like on 3D benchys this has proven tricky on parts of relatively small size, I need to perform some tests ASAP. Reviewing printer settings and setting up custom slicer profies for some parts has also been required, especially for parts like window frames and snap in glazing, you cannot easily filament print clear glazing by nature of the process, but for factory windows which may be wire reinforced and/or just dirty, a translucent finish is fine. Both window frames and glazing benefit from specific slicer settings. So, a lot of beavering going on behind the scenes, I have not given up Next Royal Mail Warehouse Frontage Version 1 Jim Back to Article List
18. Royal Mail Warehouse Frontage Version 1 This was the original target for this low profile warehouse frontage but is now version 1 as I have been refining it's structural appearance to look more like a warehouse than an office building. This has required developing further custom parts, especially for representing thick floors, and improving the roof edging, both will be presented shortly, including the need for special parts for a low profile building such as this, e.g. the bare top right corner. Retention of the lower end of the short 'penthouse' panels on the top floor is required, metal clips were used in the original Arkitex system, I am sure there must be a better way, one shifted out of place just for the photo! So here are images of the building so far which includes some of the frames, custom panel fascias and the custom baseplate developed in this project to date. Next Parts for version 2, starting with warehouse floors. Jim Back to Article List
19. Warehouse Floors - Design 1 The first design I considered was to use as many standard Arkitex parts as possible and so minimise the number of custom parts required. Here I utilise the short beams known as locking blocks (LBs) to support a very short variant of a panel, the idea is that these 'panels' contain the equivalent of an LB integrally to the panel for a vertically compact design. The internal framework of the building would have a layer of these LBs to maintain the integrity of the structure. Conventional panels fit against flat surfaces and utilise cosmetic vertical corner pieces that engage in the fluting of a connector beam to clad framework corners as shown below. This is not practical to replicate with beams as short as an LB, so I decided to have two distinct floor spacers, one straight and one including a corner. They are designed to be part structure and part panel, one end contains the equivalent of an LB and the other end interlocks with its neighbour to retain its free end. Both include a an inward pointing bracket for the lug on the bottom of a panel to engage with. These parts evolved to be as shown:- Straight Spacer Corner Spacer Multiple parts for a 2x1 structure to illustrate how they connect together. Printed parts in use. However when the height of an LB plus a connector block is measured this becomes 21.8 mm which equates to 5.5 ft in 1:76 which is exceptionally thick even for a warehouse floor. Reducing in height any further becomes impractical as the 'panel' needs to extend below the connector beam onto the top of the panel below. This is also limited by the additional connector block. Summary Although straightforward there are issues with this design:- Minimum floor thickness that can be represented is too thick for most applications The addition of another layer of short beams and their connector blocks to the structure does not enhance the squareness of the structure. The springy nature of connecting the blocks and beams in the Arkitex 00 system is one which causes difficulty in achieving a clean square result at the best of times, adding more is not desireable. So a different design is required to address these issues. Next Custom Connector Beams - Basics Jim Back to Article List
20. Custom Connector Beams - Basics This was originally just a 'what if' component as it has fairly tight dimensional requirements and is quite small for filament printing with a 0.4mm nozzle, but then again, nothing ventured, nothing gained. The connector beam and its connector block are the core parts of the Arkitex system but with only the standard beam and the locking block it rather limits the style of the structures to simple offices, shops and bridges. Here I look at the general features of the connector beams and feasibility before developing others for specific tasks such as improved warehouse floor spacers, overhangs, extended reach and high ceiling factory walls. The original connector block is made of a translucent polythene type material, it is flexible and prone to wear and damage when repeatedley fitted into connector beams which are moulded in what I suspect is ABS. Samples of at least 10 of each were measured to ascertain suitable dimensions of the beams to ensure that my 3D printed beams were compatible with original Arkitex connector blocks. Material Which material was the first question, PLA or ABS, PLA is easier to print but not very elastic , ABS although quite elastic within the restraints of interlayer adhesion can be very much affected by warping and distortion, especially with small parts. PETG was not considered appropriate due to its high gloss finish and tendency to easily string. As the Raise printer was already loaded with HatchBox matte grey PLA I tried that first and the results were promising using the actual dimensions of a standard Arkitex connector beam as a starting point. The end holes were based on dimensions found to be a good fit for connector blocks during the design of the custom baseplates, namely 3mm square, this was also printed in PLA but with rather more support around the hole. Key Dimensions Measured connector block pin size is approx 2.8mm square Connector beam socket size, after running tests, which allows for shrinkage and ceiling droop inside a socket when the beam is printed laying down :- PLA - 3mm square, shown below ABS - 2.9mm square Printing A number of issues have been revealed in re-producing and creating custom beams which I will now address, remember this is for printing with filament, not resin. Build Plate Orientation Matters 1) For Strength Orientation of the beams on the build plate must consider the planes in which they are printed. Printed parts are strongest along the x or y axes, weak in the z-axis so they are oriented lying down. The end sockets for insertion of connector blocks print without the need of internal supports, in fact slight droop inside the top surface of the socket adds a little useful give to the socket. 2) For Fluting Definition The fluting is required where a vertical corner is to be attached, at least one face must be fluted, the edge of the flute and the outer edge of the beam is relatively critical to allow the vertical corner to stay attatched. The depth is less critical, it just needs to be deep enough to accomodate the tongue of the corner piece but not too deep as to adversely weaken the beam. To aid flute definition the beams are printed with the flutes facing up and down. The horizontal layers are by default printed at 45 degrees to the x and y axes, therefore orientating the beam such that one fluted face is on the build plate it is easily bridged over. Much longer bridging occurs during the transition from 10% infill to outer skin layers. So although the face down flute becomes slightly shallower due to bridging droop it is functionally satsifactory for attaching the corner parts. Also the flutes do not warrant supports in this orientation, printing with the flutes in the z axis would require supports and as I have found supports do not support the body material completely, they support the droop, as they do not actually make contact with the part they are supposed to support. (Ref 20.1). The result would be ill-defined fluting. However for forthcoming parts some beams of other lengths and finish e.g. plain concrete colour would be needed, so to start with I have kept to standard dimensions for proof of printability before developing custom variations. Which Printer to Use Fume extraction on the daVinci 1 Pro and the Raise3D E2 (R3D) are non-existant and poor respectively, ABS is nasty stuff to print in the open as I have said in my materials article (Ref 20.2). Although I have printed my own fume extractor system for the daVinci (which prints ABS much better than PLA anyway), the build tolerances of the daVinci are too slack to print these beams, even after tweaking the dimensions to account for distortions. The R3D however is a much more heavily built machine for light production work and produced the beams in the required dimensions with little error or distortion. For time economy I prefer to print 10 to 12 at a time so some extra work was required to prevent lifting. The unfluted beams were quite easy to print but the fluted ones have a very small footprint and so were likely to curve away from or eventually detach from the build plate, which results in either an unsable beam or the whole job ruined. Printing Multiple Beams It is envisaged that quite a few non-standard beams will be required so reliably printing 10 or so at a time is important. A downside of printing with the fluted surface on the build plate is the small contact area which, if multiple beams are printed in one job, often resulted in one coming unstuck whereby all the others get ruined by the lifted beam being dragged around by the nozzle. The compromise I take is to link them in sets with 2 to 3 layer webs joining them, this eliminates curling away from the build plate. However, unlike the lilipads they become part of the beams first 2 or 3 layers but are easily pruned off. Firstly with sharp scissors and then followed by a sharp knife to fine trim them, just as though they were sprues attached to injection moulded parts. Any fraying of filaments can easily be repaired with Plastic Magic. These parts are internal to the building so it is not an issue if the cutting shows as a change of surface finish. Results This has worked well with both PLA and ABS on the R3D printer. Connector blocks have fitted all of those printed on the R3D so far with a good push fit, the blocks would normally be a light interference fit into the Arkitex connector beams which are a rigid plastic. So PLA is a good option, as it is the blocks that compress slightly on entry to the rigid beams. Standard Beams - Plain - PLA Two sets printed at a time. CAD model Printed Standard Beams - Plain - Third Length - PLA They are actually shorter than third so that three with two joining connector blocks is equal in length to a standard beam. Four sets printed at a time. CAD model Printed Standard Beams - Plain - Half Length - PLA They are actually shorter than half so that two with a joining connector block is equal in length to a standard beam. Three sets printed at a time. CAD model Printed Continued in part 2
20. Custom Connector Beams - Basics - Part 2 Continued from part 1 above. Standard Beams - Fluted - PLA Two sets printed at a time CAD model Printed Standard Beams - Fluted - ABS CAD model as for the PLA except slightly smaller end sockets as stated in key dimensions above. Two sets printed at a time Printed Summary This has been a successful exercise with consistently good results when printing on the R3D printer with HatchBox PLA or ABS, it is however beyond what the daVinci 1 Pro is capable of printing even with ABS. Printing with HatchBox matte PLA works well using the default 'Speed' settings in the Raise3D IdeaMaker slicer, the matte white is quite a good match to the Arkitex injection moulded parts for which the unfluted beams may be considered as light weight concrete beams. So far the fluted beams have been printed in HatchBox matte grey or white, and HatchBox orange ABS, HatchBox filament as bought at Amazon UK has been a consistently easy material to use at a typically £21 per kg spool. Making orders up to the free delivery breakpoint of £25 at Amazon UK at this time comes with quick free delivery (a few days or less) without the need of their Prime services. Just have some odds and ends in mind to top up to £25 when buying a single spool. Some colours are imported from Amazon US e.g. dark yellow and cost approx 60% more than those shipped direct from Amazon UK, I think I would pay the extra for a colour I really want as I can rely on the brand, a recent alternative matte white from FlashForge was nothing but trouble on both printers, so a waste of money. Next Thick Floor Connector Beams. Jim Back to Article List References 20.1 Support issues - part of article 15 15. Frame - Alternative Printing for Better Formed Top Flange https://platform1mrc.com/p1mrc/inde...anding-a-1960s-system.6522/page-3#post-107916 20.2 PLA, ABS, PETG https://platform1mrc.com/p1mrc/inde...le-of-an-oap-vs-3d-printers.6395/#post-106819
21. Thick Floor Connector Beams These beams are standard length plus spacer depth, they are to be printed in PLA in linked multiples as for standard length beams. The requirements for these beam is tighter for these beams as they pass through one end of the floor spacers and in addition these longer parts suffer a little from contraction after printing. Dimensional Errors Measuring up sample original beams presented differences in length of 0.2mm or more due to flashing, wear or post moulding shrinkage, some of which may be age related, also the plastic doesn't seem to be as solid as ABS. 1) Cross Section To add further fun to the mix the dimensions of the parts relative to the print layer thicknesses defined by the slicer can adversely affect the resulting height of the finished part. Although this would be trivial in a tall decorative part it is significant for a part of only 5.2mm height that needs to be a light push fit into the aperture of another part. As the original CAD model the top layer results in a z-axis height of 5.3mm, as shown in the slicer preview image. Reduction by 1 layer results in a z-axis height of 5.05mm, slicer preview image. This is much better. The floor spacers of design two requires a well defined beam cross-section, for any other beams it is not so critical. To this end the beams for the floor spacers had to be changed from 5.2mm square to the z-height being 5.1mm otherwise 5.2 was rounded up to approximately 5.3mm. At 5.1mm it was rounded down by the slicer to 5mm and the required fit into the spacer was achieved. 2) Length Although standard length beams printed well enough as designed the longer beams for the 8mm floor spacers suffered from shrinkage, these were lengthened by 0.2 mm, this is not much but errors soon stack up to distort an already tricky to assemble squarely system. In cases such as this I maintain CAD models of ideal parts for constructing buildings in the CAD tool in parallel to the CAD models for printed parts. Parts adjusted for printing will mostly break any attempt to construct a building purely in the CAD tool editor, prototyping a building in the CAD tool including colouring parts is fun too. Results CAD Model - Print Version Printed in Eryone matte yellow PLA, as the preferred HatchBox dark yellow costs 60% more than most of that brand as it is imported from the US via Amazon(UK). Summary This beam needed to be more precise than previous ones but now it is available for the 8mm thick design 2 warehouse floors it can be adapted for other planned floor thicknesses. Next Warehouse Floors - Design 2 Jim Back to Article List
22. Warehouse Floors - Design 2 To allow the design of the warehouse floor to be thinner the need for a short beam, e.g. a locking block plus connector block had to be eliminated. This second approach ensures that the floor spacer is independent of those two parts and the robustness (albeit limited) of the building frame work is maintained. The floor spacer now floats above a standard panel which allows it to be a minimum of 2mm thick which is not really of much use as it could be weak and certainly not that visible to be effective. I consider that 4mm or a scale 1ft is what would be the minimum thickness for an industrial or commercial (non-office) building floor. For this to be possible custom length connector beams are required whose height would be standard plus the required floor thickness, 3D printing of beams can be a bit tricky as detailed in article 20. The new pair of parts, straight and corner similar to the first approach are as shown below, but in this version the beams pass through the floor spacer without the need of a connector block. CAD Models Printed Parts How they fit together. CAD Models Printed Cosmetics Printing parts in this orientation leads to rounded vertical corners which can be reduced slightly by cutting knicks into them as circled left below. To mimic the joint between a corner spacer and another spacer a groove approximating to the gaps between adjacent spacers is printed (circled right below) so as to make the corners appear as a separate column of concrete components. Assembly In use the appropriate length beams are fitted to the connector blocks followed by fitting the floor spacer by threading the beams through them as shown. Once the building has a complete layer of these spacers then the next layer of standard horizontal beams can be fitted. This sequence is repeated for each floor as required, e.g. the RM warehouse frontage on a revised custom baseplate. Internal vertical beams must be of the same length as those used for the outside. At the left rear having a full corner part which is unsupported due to not fitting panels to the rear causes the spacer to droop, but a corner is required to lock the end wall spacer into position. To overcome this a low profile corner part is added. CAD Model Printed, with a mating part. Continued in Part 2
22. Warehouse Floors - Design 2 - Part2 Assembly Continued Low Profile Corner Fitted To fit the standard wall panels the spacers are slid up the beams as necessary and the panels fitted as usual, except the top flange bears against the floor spacer and the lug at the base into either the baseplate or the spacer below. Remaining main structure, one floor less than version one due to the added height of the floor spacers. Summary These parts are printed in HatchBox matte white PLA, a brand I find to be a cost effectve filament that prints well with my Raise3D E2 printer. This approach is vertically scaleable from 4mm to 16mm with little issue, the 8mm spacers are used in the example of the Royal Mail warehouse frontage. Next Roof panels. Jim Back to Article List
23. Roof Panels A simple item but one that can really upset roof edging in this scale, the original parts are prone to twisting or warping which does not show within a building when used red side up for internal floors, but grey side up on a roof it can play havoc with the original roof edging which is very light. So here I have tried to print an alternative in PLA which providing it can be released from the build plate without damage can be modified for further variations. Unlike the original parts any bends can be flattened out easily. The Arkitex roof/floor panel is dual colour, grey/red, but as I am aiming these for roof work my initial prints are in Hatchbox matte grey PLA. Drawn as 0.8mm thick they are printed 1mm thick, the slicer rounds the required thickness according to the specified print layer thickness, I could adjust the layer thicknesses or the parts' thickness but I find that the results are satisfactory, any thinner would become quite difficult to release from the build plate. Flexible steel build plates release a rigid part easily but not a highly flexible part. CAD Model Printed A 2x1 roof panel to reduce joins. CAD Model Printed Fitted to the RM warehouse Summary Simple but effective, it expands well to double length which increases stability and improves appearance in one go. Can be bent when released from the build plate but is easily flattened out again. Next Roof Edging Jim Back to Article List
24. Roof Edging - Part 1 - Customised Parts The roof edging of the Arkitex (00) kits is very light weight and fits rather poorly, it does not cap the connector block as the 1/42 system and does not include the pegs which engage into the top of a panel, or the lugs between straights and corners. The catalogue/handbook illustrations, appear to be an incorrect leftover from the 1/42 artwork. My version of the roof edging was to :- Cover the connector block Be more solid Sit better on the building To meet covering the connector block it had to be taller and hence more solid, they should naturally sit better on the building. Clearance is required to avoid being lifted by errant original roof panels, roof panels printed in PLA described in the previous article reduces the number of panels required by creating some larger ones which can be flatter. The cross section is rather more complex than an Arkitex part and I have not attempted to mimic the interlocking lugs which originated in the 1/42 scale sets. The deep outer edge contacts a wall panel, the next inner parts contact a connector beam, it also includes clearance for a connector block, and inboard the lowest edge should be just above the roof panel. After 3 to 4 iterations of some parts the design for the roof edging based on Arkitex part functionality became what I describe next. Straight Edges and Outside Corners Handbook illustration, features circled in red are not present on the 00 parts. They would be too small to print with sufficient accuracy and strength with a 0.4 mm nozzle. Straight Edge CAD model Printed Outside Corner CAD model Printed Inside Corners Handbook illustration, again, features circled in red are a leftover from the 1/42 handbook and are not present on the 00 parts. Inside Corner CAD model Printed Continued below.
24. Roof Edging - Part 1 - Customised Parts - Continued Examples in use Straight edges and outside corners, with a glimpse of the north lights in the background, subject of a future extensive article. Short Straight This has two purposes Inside roof corner with outside wall corner above as in previous handbook image Roof edge joining a flat wall, as shown below CAD model Printed (Visually very slight difference to v4) Using two short straights as an inside corner was not practicable, the required interlocking nature of two short straights was not possible because they could not be printed accurately enough so a single piece was developed for this purpose which is described later in 'New Parts'. This aligned and fitted much better than two seperate parts. Short left and right handed parts From the catalogue/handbook CAD models, left and right handed Printed, left and right handed Examples in Use Showing the left and right handed parts presented so far. In addition it shows the single width double outside corners and the special inside corner replacing a pair of short straights which are described in part 2, 'New Parts'. In addition to these customised parts some new parts were created including parts specific to a low profile building. Next Roof Edging - Part 2 - New Parts Jim Back to Article List
24. Roof Edging - Part 2 - New Parts Further parts required, including those for a low profile building. Straight Edge - Double Length The less joints between edges the better so a double length straight part is useful, lilipads were required to ensure that the ends stayed attached the build plate. CAD model Printed Parts for low profile building The low profile nature of the RM warehouse frontage meant that the top wall panels are attached to a single row of beams, this made it impractical to use standard straight or outside corner edging parts. Low Profile Straight Two extra support webs added to aid seating on beam as there is not a roof panel for the inboard edge to rest on. CAD model Printed (Visually very slight difference to v4) Low Profile Outside Corner or Closed End Truncated version of an outside corner, which provides a neater end to a single beam thickness wall. CAD model Printed Other Parts Single Width Part There are places where the roof may be only one cube deep, for this case using two standard outside corners does not look so good, so for this situation a double outside corner was created. Not just two standard corners connected but a little extra added between them which makes for a better fit as a small gap is normally present between adjacent roof edging parts. Shown in use in part 1. CAD model Printed (Visually very slight difference to v3) Unique Part for the RM Warehouse This part needed lilipads at both ends to prevent lifting during printing. CAD model Printed (Visually very slight difference to v4) Continued ... Jim
24. Roof Edging - Part 2 - New Parts - Continued Inside Corner Vert Corner This part was required to replace the use of two short straights that could not be printed accurately enough to create this type of inside corner. From the handbook/catalogue :- A single piece part was developed to perform this function. CAD Model Printed (Visually very slight difference to v4) Next Roof Edging - Part 3 - RM Warehouse Low Profile Frontage Jim Back to Article List
24. Roof Edging - Part 3 - RM Warehouse Low Profile Frontage Roof edging parts applied to the RM warehouse Overall views of the completed RM warehouse frontage Summary This has proved to be one of the most difficult components to develop so far, reverse engineering the geometry of the Arkitex system is complicated by a number of factors which meant finding the 'average' dimensions iteratively. Up to four design iterations were required to get these edging components to an acceptable level where parts could be interchangeable and adequately solid so as to cap off a building neater than the original parts. Factors affecting determination of geometry included :- Tolerances of the parts Subtle damage and distorted parts, including kiddie teeth marks, I 'kid' you not! Distortion through ageing, minimal but present on a few parts Flash from moulding especially on connector blocks and beams Dimensioning of parts to suit the printing process Many parts make up the edging standard pieces, 13 to date and far more were required for other roof features developed for factory buildings, e.g north light roof windows which is covered in a future article. Some parts required separate 'perfect' parts for CAD models of assemblies because some were dimensioned differently for printing, e.g allowing for shrinkage, expansion and overhang droops when printing or rounding by the slicer according to selected layer thickness. The RM warehouse frontage is shown in place at the Montague Dock ready for bedding in here. Next, Correcting Type C Frames Jim Back to Article List
25. Correcting Type C Frames These frames are full height and 3/4 width. When I first modelled them I did so from a short panel but forgot when I created the full height version to have cut outs in the base of the panel to accept the triangular hooks of the one below in a multi-storey model. As first designed. As it should have been. Test fitting was only done on a single storey building, the missing clearance slots soon became apparent when I fitted them to a two storey building and the vertical points fouled the panel above. The change to the CAD model was easy enough but then I had already printed at least 10 frames without the cutouts, this was remedied by creating a seperate adapter piece with the clearance slots to glue in place of the plain base. Adapter piece. This done, the modified frames are now correct for fit and function Next, Overhang Beams Jim Back to Article List