Dennis J

  • Content Count

  • Joined

  • Last visited

About Dennis J

  • Rank

Profile Information

  • Gender
  • Location
    Adelaide, South Australia
  • Interests
    Anything and everything.

Recent Profile Visitors

483 profile views
  1. I totally agree. If Joaquin's spirals can reproduce the outlines of the upper and lower bouts given a form waist measurement the design problem is largely solved. I'm just not sure this is the case, I haven't tried it. The same figures can be made using splines, as I have done myself. But although they can comply with all the relevant points they still can be varied in shape quite a bit. It depends on how the carbon fibre splines are manipulated. Which is not a bad thing because it means different outlines can be created. Doing it myself I've noticed that you can not get too far away from the near-circular at the widest sections of the upper and lower bouts before things go wrong. But it is self evident that there are no sections of a violin's outline that are circular. I suppose that those advocating using circle geometry are are not saying that, but as you say, curves produced by carbon fibre are very attractive, it is just a matter of controlling the process.
  2. This topic began with Joaquin Fonollosa's post suggesting that Joachim Durer's spiral and associated rectangle could be the basic element of violin design. And to myself it looks pretty convincing, although I'm not sure how it might best be applied. I get the impression that using such a rectangular template the same width of the waist of the proposed violin's inner form, placed in the right way, is all that is required to get the relevant position of corners, length of body, width of bouts, &c. If that approach was invented/used by Amati senior, or any other geometric method achieving the same outcome, then apart from corner shapes, there would originally have been a standard violin design or shape. I'm sure that would have been considered desirable by makers at the time, and many makers now for that matter. After all the outline shape is visually important and can be tricky to get right, but it probably has little to do with the sound qualities of the instrument.
  3. Hello Francois. Your analytical skills are a bit beyond my understanding but I'm sure employing splines in a controlled way will be a design asset. If you intend to physically use splines I would recommend carbon fibre. I've used .5 mm and 1 mm thick rod for violin drawing, but you would probably need something thicker for the broader sections shown. Happy bending.
  4. I'm watching this subject with increasing confusion. If it is about design methodology which might have been used in the past I find it hard to imagine that makers were using giant compasses on large sheets of paper to construct the necessary circles for a guitar. Or that they might have scaled the designs down and then scaled them up for working drawings. On the other hand, having the desired ratios and dimensions in mind, it would be quite possible to use bent splines to achieve what was required as far as shape was concerned to fit those measurements. I'm sure that having used this method to make outlines of the guitar shape they then could have been used in conjunction with some sort of form to bend the guitar sides. Using splines with the necessary thickness, spring and anchoring positions and angles is just a matter of experimentation which, when learned, can be repeated again and again.
  5. I have spent a lot of time experimentally using Muratov's suggested "old" method of using splines to create my "own" mould based on the N. Amati MB example, increased by 1% in size. It can be done using only the lower bout width and length measurements of the proposed mould. However getting the right shape for the upper, lower and waist bouts at the same time as complying with the required ratios of measurement is no simple matter. All of the bouts require a fairly specific start and end point in conjunction with a suitable thickness spline ( I've found that .5 mm for the Cs or 1 mm for the upper and lower bouts serves the purpose). All of the specific design points designated such as position and widest part of the upper/waist/lower bout can be linked and complied with by various shaped curves using different start and end positions and manipulation of the bent splines. So there are a lot of variables and practicalities involved and it ultimately becomes a choice of what shape is the most suitable, albeit within a fairly close range. The bottom line is that it is all based on geometry, however it is applied, and getting too far away from it will result in a violin with a wonky shape. There's not much room for error or variation.
  6. It seems to me that the guitar is as clear an example as possible of using bent forms to make and also design an instrument. If you can design the outline shape of a guitar with bent lengths of wood the chances are that the wider sides of a guitar can also be bent in the same way. All that is needed to start is the length and waist measurements required and for the force to be applied in the right place to achieve the right outcome. Why would luthiers of the past have bothered with complex, manipulated geometry to achieve something that can be realised by using dynamic forces producing fantastic natural shapes automatically?
  7. I agree that Joachim's spirals cover just about everything needed to complete a violin form outline with the location of corners, blocks &c. and I can't see how anyone could disagree with that. Having gone through the process of Muratov's suggested method of using splines to create spirals in association with ratios of width and length I can see a lot of drawbacks associated with the method as far as design is concerned. But it is based on the MB pattern attributed to Nicolo Amati and it does work with the right manipulation of the splines. To draft the pattern I taped a sheet of paper to a piece of MDF which would both serve as a base for the holes needed for turning pegs and a drawing surface needed for a centre line &c. It has occurred to me that the early makers may have used the same sort of method to design and make new forms/moulds. They could have used a board a little wider than the pattern with a sheet of paper attached to the surface with a few dabs of glue. They then could have used whatever drafting aids/patterns they needed to design the necessary shape ready to be cut out. This would explain the absence of scribed lines on the existing moulds of that time. But the holes for pegs which may have been made to hold splines are evident on many patterns.
  8. No problem, Kae. I should say for anyone who might be a bit confused about this is that the 660 mm figure (double vibrating string length) is the distance from the bridge line to the point at which the vibrating strings converge. I think that one of my posts has a photo showing the entire length of that line. It can be moved towards or away from the centre line to change the angle of the vertical which is used to make the arc. Something else that I think that I mentioned before but is worth repeating for clarity's sake is that the edge or crest height varies from 4.5 mm to 5 mm in the plan. But the calculated height of the inflection point is added only to the 4.5 mm edge height for each of the four central cross arch positions to mark the template blank. The 4.5 edge height is included because I've made all the templates full depth. With the uppermost and lowest cross arches at each end of the arc the inflection point height is 4.5 mm, the same as the edge height. And the vertical lines marked from the base of the arc to the curve shown on the arc are used in conjunction with a right angle triangle calculation, of 5 degrees for the top and 4.5 degrees for the back, to arrive at the height of the inflection point at each cross arch position.
  9. Thanks for your nice comments. I'm not sure that I can take it much further, but I hope it helps those who are willing to try it. I know it is a flexible system to some extent but I'm happy with it as it stands. There is a limit to the number of arching figures that are feasible or practical on any violin. Maybe it could also be applied to violas, or even cellos. The double-string-length line refers to the point at which the vibrating strings extended come to a point. I've used a vibrating string length figure of 330 mm which is doubled to 660 mm. The DSL line can be drawn to cross the lower bout cross arch at any practical position. I've used 52 mm from the centre. Vertical lines are then drawn between the upper and lower cross arch positions where the DSL crosses them. From that vertical an arc can be drawn between the upper and lower cross arch lines. The base of the arc I've used ends at the lower cross arch line 82 mm from the centre on the last example I've posted. The 40 mm radius circle centred on where the verticals cross the centre line establishes the height of the arc. Any of the figures except of course the 660 mm DSL length can be changed. But, as I said, I've just concentrated on the practicalities at this stage and I'm not sure that there would be any point in doing so.
  10. I'm not in a position to make comparisons between early instrument profiles and mine. But as I said earlier I'm sure that early makers would have had a full understanding of the geometry of arching. One thing that has crossed my mind is that you could make a physical arc/curve as a template to be used during the shaping process. It would have to be a crescent shape to bridge over edges, &c. and pivot at each end. And of course it would have to be positioned at the edge level, correspond with a double-string-length line and be used at the right angle of elevation in the final stages of arching. I don't know if that is practical, but who knows?
  11. In relation to David's comments I'd like to provide some of my thoughts. Each cross-arch is generated from the same arc-based curve and its specifications at each point. So together they form an homogenous element which helps form the overall arching shape. And that curve is an integral part of the violin's geometry. As for context I see no problem there. The two widest arches at the upper and lower bouts have inflection points closest to the edge of the plate. This allows for a wide, convex, low arch which is entirely appropriate. The narrowest arch at the waist also has an inflection point close to the plate edge. Which works perfectly for the highest arch on the plan. The arches at the upper and lower corners, as well as the one at the bridge position, have inflection points further from the plate edge allowing for a long, graceful scoop necessary at those positions. I've thought of a couple of analogies about a violin's topography. Regarding shape: If you were to place a sheet of silicone on a flat surface and squeeze two opposite edges together with a thumb and finger it would rise up to a maximum height at that point and fall smoothly away in all directions. Much like the central area of a violin. A natural shape. Regarding profiles: Much like the glide slope of landing plane descending from a fixed altitude in a convex trajectory, passing through a fixed inflection point enabling a smooth landing. I think the drawings of the arching template profiles I've created using the method I've described speak for themselves. You can't make them up without a plan.
  12. I can understand Jackson Maberry's concerns about using arching templates, but I think he might be complicating a fairly simple issue. I like certainty about what I'm doing. What I have learned making these templates is that once the cardinal points of all the archings are locked in there is only one arching figure possible for each cross arch, give or take no more than about 1 mm from that line in the upper section. Any more deviation from the line will make a smooth transition from the convex upper section of the arch through the inflection point into the concave scoop at the bottom impossible. I'm not sure that fluting or shaping around the ffs should be a problem because all of that work involves taking wood away, so in the final stages of shaping the template will bridge across that area, or at least in my opinion, should. What I like about making these templates is that I can develop the profile dictated by the inflection point and bottom of the scoop to a high degree of accuracy, and If I go too far I can make another template. It takes time but is not difficult, most of the time is taken checking for symmetry, &c. So I eliminate mistakes making the templates. Mistakes made in shaping plates without templates are irreversible and can only compromise the final outcome.
  13. Here is another version of my template layout with some slightly different parameters. Such as the arc's base at the lower bout measuring 82 mm from the centre instead of 84 mm. This allows for a slightly longer scoop at the lower bout. It also shortens the arc's height, so I've compensated for this by increasing the arc's angle of elevation from 4.5 degrees to 5 degrees. The curve based on the arc is similar to the previous spiral form. It is used to calculate the horizontal and vertical position of the inflection point for both the front and back plates. The edge heights I've specified are probably a bit on the thick side, but allow for variation at the centre bouts and corners during final shaping and edge work.
  14. Here's another pic of the arching profiles. I've drawn a straight line on each from the top centre through the inflection point. It shows how little deviation there is from a straight line. I haven't bothered drawing a straight line like this when making the templates, but it probably wouldn't be a bad idea to use as a guide when shaping them. But the main guide should be the inflection point intersection.
  15. Thanks for your comments Mike. Yes, the location of the inflection points along an arc or a curve of changing radius is fundamental. I'm sure that from experience many makers know how to handle this tricky area of arching, perhaps some better than others. As I say you have a surface curved outward on one axis and inward on the other axis. As for blending in at each end of the arc, the arc represents the location of the inflection point (which is neither positive nor negative) not a shape, and at the ends of the arc the inflection point is at the level of the edge so no physical blending-in is needed. The scoop or negative part of the arching at that point carries on around the ends of the plate without any sort of adjustment necessary. As for the f-hole area my firsthand knowledge of Cremona instruments in this part of the world is very limited, so my judgement is based almost solely on photographs. But your comments about the lower f-hole area are interesting, if I know what you mean. What I have found making these templates is that two arch positions need to "fit". The most important, obviously, is the narrowest at the waist because room is needed for a scoop, so that sets in place all the other archings along the arc. The other is the lower corner arch which has a very strong recurve. So much so that I have tried to widen the arching figure there by changing the the arc on the plan to a curve with the changing radius so that I could move the inflection point there outward a bit to spread or widen the curve and also reduce the length of the scoop. So the concavity in that area is quite pronounced. And I agree that is how the early Cremona makers routinely shaped their instruments. Michael Darnton's laser line image of an Amati clearly shows the strong recurve I'm talking about in that area. Here again, based on photographs, I don't see much evidence of deviation from a smoothly curved arching around the f-holes on many early instruments. I know that an example in David Beard's book shows some sort of hollowing around the f-holes, but other examples show none. The only thing I see there is the usual, quite subtle fluting around the f-holes.