Horyuji Pagoda, Canon

  • After turning my attention to the lovely Vespa 150 from Paper-Replika, and then the Chinese Dragon from Canon as light relief, I decided to return to architecture with one of the recent releases from Canon. Far Eastern architecture is quite fascinating, and I have been itching to try one of these. The large structures of Potala and Angkor Wat were just that - rather too large - and didn't look good candidates for scale reduction. I have tried castles before, so my eye was drawn to the temple and shrine buildings. Wanting something relatively simple, I selected the Pagoda from Horyuji Temple.

    Horyuji Temple dates from the early seventh century, but the buildings were destroyed, probably by fire, in the year 670. It was rebuilt over the following forty years, and the present pagoda dates from then, though the massive central pillar has been dendro-dated to 594, so may have been part of the original buildings on the site. The pagoda was built to house a fragment of the Buddha's bone, amongst four relief images of Buddhist teachings. The photograph attached is from Wikimedia Commons, authored by 663highland on ja.wikipedia.

    The model has 195 parts, on 27 A4 sheets (plus a cover sheet with photographs of the finished model and a brief description of the building), and 10 pages of assembly instructions.


  • As usual I am going to try to make the model at half the original size, but otherwise largely as supplied, except for substantial strengthening with cardboard. There will no doubt be other minor changes to make it buildable in the smaller scale.

    My first task is always to produce bitmap versions of all the pages, at 300dpi to be printed at 600dpi which is the best my printer can do. Then I can do any required manipulations using GIMP, and print or reprint as necessary with no worries about colour matching. The first four pages of the model are taken up with the four sides of the base. I chose to combine them into a single piece to avoid unsightly joins - this would not fit on an A4 sheet at full size. Similarly I could combine the two pieces of the plinth on which the pagoda stands. The photograph shows the first few pages printed on two sheets of A4.

    After all this, I have had time to make up the base and plinth, both well supported with thick card.


  • Good morning Alan,

    You begin a very interesting model. Already, your precision is showing in the clean, crisp angled base. A wonderful start.

    The Vespa 150 and the Chinese Dragon are outstanding models!

    Good luck with the Pagoda. I don't think you are going to need it ...


  • Hello John

    Thank you for your kind words. I very much enjoyed those two models - the dragon was a complete change from my usual subjects, and was pleasingly colourful. My daughter gave me one of those wooden "skeleton" models of a dragon last year, and they go well together.

    I'll reserve judgement about this pagoda, as I have been bitten before by the Canon models. Just when everything seemed to be going fine there has been a serious fit problem. I think you found the same with the Tower of Belem. We shall see.


  • Now for the entrance steps. The assembly instructions say "The staircases are minute and difficult to fold". I think "minute" is going a bit far, but as designed they are certainly difficult to assemble. I removed a good number of the tabs attached to the end of the steps, and made card formers to shape the steps before attaching the sides. The result was adequate, though not good. The photographs show up the heavy printed fold lines rather worse than they appear with the naked eye. I should really have toned them down before printing, but by the time it occurred to me I had already made two sets of steps (and discarded a third), and I couldn't face starting all over again.


  • Moving inwards, we have the floor of the first storey of the pagoda. This sits inside the plinth, and in the model forms two square grooves into which the lower walls slot. These are designed to lift away to reveal the inner sanctuary containing images of the Buddha and the relic. There is also the vertical central post which supports the rest of the building, to a height of over 30 metres.

    It seemed to be a good idea to check that the walls would fit smoothly in the grooves. The outer set were fine, but the inner ones were much too tight. They are double-skinned, and the strip separating the inner and outer skins at the base is drawn exactly the same width as the strip in the floor forming the groove. As a result the groove is at least two paper-thicknesses too narrow, even before allowance for an easy sliding fit. For both sets of walls I shall build them around a cardboard core, so for the inner ones I shall cut the inner and outer parts apart, and use much thinner card. In fact, the inner skins will not be visible, so I may not bother with them at all.


  • Having checked the fit of the walls, I have now installed the floor and central post. The hole in the floor for the post was not large enough - once again the designers do not seem to have allowed for the thickness of the paper.

    The piece repreaenting the sanctuary images is roughly a cube, with recessed sides and a slightly stepped top. It has several redundant tabs. The top is intended to be made from four quarters folded down and interleaved, but this seemed much too difficult, so I cut the quarters back to tabs, and drew a plain square to glue on top. Once again the central hole was too small, so I had to cut it back.


  • Not a lot to show for the last couple of days, but I hope to have a good session over the weekend.

    Completing the lower walls. I filled the gap between inner and outer skins with 1.4mm thick photographic mounting board, for stability. The roof parts consist of an overlay for the walls, the roof undersides and the tops which fold back over. Apart from cutting off a few redundant tabs, I did these as recommended, with no trouble.


  • Now for the five pagoda levels proper. [The walls and roof already installed do not appear to count as a storey of the pagoda. In fact, they look rather like a later addition, perhaps to provide additional space for the visitors to the shrine.] In the model, these are hung from a central post which slides over the one already fixed to the base. This is crowned by a hexagonal spire, which is a rather pale imitation of the highly ornate real thing.

    This pagoda has five storeys in all - other pagodas range from a single storey to 13 or more, though the interior arrangements do not always reflect the outward appearance. Traditionally there is always an odd number of storeys, though I don't know what the significance of that is. Each of the upper storeys here has a square central core and a widely overhanging roof with upswept corners. I anticipate that these roofs will not be easy to construct.

    Anyway, first comes the core of the top storey (goju), which has heavy projecting beams at the corners and the middle of the sides to support the roof. I filled these beams with a suitable thickness of card, so could cut away all the tabs and just glue the sides and edges to this filling.


  • The roof consists of eight parts, four identical ones for each of the top and bottom surfaces. The four nearly triangular upper faces are joined by raised rectangular-section ribs. One side of this is attached to each abutting face, and parts of the top surface are attached to each side. The whole thing is intended to be constructed using lots of tiny tabs, leaving the ribs hollow and unsupported. The outer edge of the top faces is folded down to form the fascia, and tabs on the underside glue to the rear of this, forming a hollow section approximately 5mm thick (2.5mm for me).

    The assembly instructions suggest that you should join each underside to its corresponding top piece, then assemble the pieces together in place around the central column, with the central building core and support beams already in place. I couldn't imagine being able to do this and getting a reasonable result, so I took an entirely different approach. Firstly, I removed all the tabs from the corner ribs, and cut cardboard filler pieces to the right shape, extending down to the corners of the under surface to provide some strength. Then I assembled the whole of the upper surface - not too bad, though it looks a bit wonky blown up to about three times full size in the photograph.

    Then the complete underside went on. Getting the edges glued accurately was quite difficult, but the central hole allowed me to get a thin rod between the skins to position and press down the tabs, a few at a time. The result isn't perfect, but will be mostly hidden anyway. Finally, I could slide the whole of the roof up the central pillar, followed by the core building with the roof support beams, and fix the lot together.


  • Now for the fourth storey (shiju). The construction is much the same as for the top one, except that the roof is attached to the outside of the central enclosure. The walls are topped by railings which project at the corners, but I decided to cut these off, and attach them to the outside after the roof is on.

    Now for a bit of a disaster. I made up the upper surface of the roof, as for the top storey, but it came out badly distorted, and additionally the opening for the walls is far too big, by nearly a millimetre all round. The roof can be forced to be level, at the expense of rather buckled fascia boards, but then the corner ribs are just about horizontal, whereas the supporting beams are at an angle of about 15 degrees. In fact, this can be seen in the photograph of the completed model on the front page of the kit, when you know where to look.

    Some major surgery is going to be needed - I must go away and think about it for a while.


  • Well, I have had a think, and have studied photographs on the web, and have found a strange thing. The model roofs sweep boldly upwards at the corners, just as I visualise a Japanese temple to be. But the real pagoda roofs are almost flat, rising only quite subtly at the corners. I am really surprised that the designer of the model left such a glaring discrepancy. I can only guess that he (or she) was blinded by dramatic imagination, and neglected the reality!

    I really think that I shall have to make completely new roofs, so I shall buckle down and get drawing. This could take a little time. So much for making the model as supplied in the kit!


    Original photograph (c)Tomo.Yun (http://www.yunphoto.net/en/).

  • Hallo Alan,
    I think it's a problem of construction such a roof. If you change the angle of this roof only about a half degree, you have on both sides of one triangle part of the roof 1°, multiplied with four roofparts you have 4°! That makes a great difference in the wingparts of the roof.


    By the way: south-chinese roofs have a much more remarkable curvature (the curvature of the roofs expresses the connection with the heavens)

    Ein Leben ohne Kartonmodellbau ist möglich, lohnt aber nicht! (Frei nach Loriot)

  • Hi Alan,
    The Chinese roofs normally sweep quite a lot.
    I start building the pagoda but did not finish it yet. It is a Japanese Model which shows the inside construction of the woodwork:
    4.KBW - Japanische Architektur




  • Hello Hajo

    Thank you for your comments. I did not know that there was a difference between Chinese and Japanese styles of roof. It is curious that a Japanese designer should have chosen a Chinese style for a Japanese model! I am very interested to learn that the raised corner is symbolic of connection with the heavens - perhaps a similar idea to the tall spire on many English churches.

    You are quite right that with roofs of this style a small error in the angle of each edge can have a very large effect, but this is only one of the problems here. The basic design of the roof is wrong, and the error in angle at the corners just makes it worse. In addition, this particular roof has a design problem at its inner corners. The corner rib should be set into the roof by half the thickness of the rib, so that the top edges of the roof meet at the corners (see the detail of the first storey roof; the second and third storeys are similar). On the fourth storey, this offset has been taken along the angled side, rather than at right angles to it. This is a surprising lapse.

    Best wishes

  • Hello Günter

    Thank you for pointing out your thread about building a different version of the pagoda, which I had forgotten. That looks like a very interesting model, showing the internal construction methods which help these structures withstand severe earthquakes.

    Will you ever finish the model and show us the results?

    Best wishes

  • So, I have drawn the outline of a new, more sober roof for the fourth level, and shall now try a test build. Then I should be able to use the old parts to paint the new ones adequately, and get construction going again...

    Test build done, and all is considerably less exciting. The parts fit together without stress (as they should, of course, if I have calculated them correctly) and the roof looks much more like the photographs.

    Now to get it coloured.


  • A session with GIMP has transferred the textures of the model to my new shapes. I use GIMP because it does all I need very adequately, and it is the right price. Some people find it obscure and hard to use, but I have spent many years using UNIX, so I am used to obscurity. The coloured roof is rather easier to make than the plain white one, as the texture tends to hide the minor imperfections.

    The twelve support beams are a little tedious, especially when realising that there are twelve more for each of the other levels still to go. I am not going to assemble these parts together now, however, as I have decided that assembling the levels top down is going to be too difficult, as each roof will get badly in the way of fitting the level below. Instead I shall proceed from the bottom up. The last task will be to demolish the top storey which I have already made, and rebuild it with a new roof. I shall put off that traumatic moment for a while.


  • We are now back on track, I hope. Next for my amended assembly order is the lowest level (shoju). As I decided earlier, I cut apart the inside and outside layers, and laminated them with relatively thin card, so that they would slide into the slot in the floor. The roof support beams for each of the four lower floors are identical, so I used the ones already made up for the fourth storey here.

    Happily the roof went on with no trouble. [The nasty glue splodge at the bottom of the wall will be hidden when the model is assembled finally, so I was not concerned about leaving it.]


  • On goes the second level (niju).

    A new, though relatively minor, problem has made my decision to build bottom up more valuable. Each level of the central core building has a small ledge at the bottom, which fits inside the level below. However, the base of the second level is barely smaller than the bottom level - certainly not enough to accommodate two thicknesses of paper plus another for the corner tabs plus a bit more for a sliding fit. The fit between the third and fourth levels is just the same, while the base of the third level is larger (by more than two paper thicknesses) than the second level enclosure. Building upwards I could easily trim the base of the higher level to fit, but it would have been very tricky if following the recommended top-to-bottom sequence.


  • Onwards and upwards... The fourth storey, deferred from a while back, is now in place.

    Now is the time to dismantle what I had already done at the top. Fortunately I had only glued the pieces together quite lightly, so some careful prising with a sharp knife was enough to separate them. The corner roof support beams will have to be replaced, as the vertical posts are now not long enough to reach the fourth-storey roof. I chose to remove the centre beams as well, just to make things less cumbersome.