Jump to content
Maestronet Forums

Dennis J

  • Posts

  • Joined

  • Last visited

Everything posted by Dennis J

  1. Yes, I think having a flat, hard abrasive surface is the key. As I remember gluing a sanding sheet to mdf or whatever it was didn't seem to work that well. Mind you 80 grit seems very coarse to me. I only finish the blocks for a height taper from the lower block to the upper, 32/30 approx. So I'm faced with a step down arrangement which has to be flattened. Anyway for preliminary sanding what you suggest is probably worth looking at because loose sheets on the saw table round over the corner blocks' pointy bits very aggressively. Not hard to go too far.
  2. Yesterday I finished tapering and levelling blocks on the front shown in the pic I posted previously. I had to take off about 3 mm on the upper and 2.5 on the lower. The lower corner blocks were about .5 mm below the rest. I decided to get them close to finished length before gluing the ribs. I couldn't use a plane without the ribs attached of course so I pared some off the highest blocks along with using sanding sheet on the saw table. When I got them levelled I tested for gaps. One thing quite noticeable was how the apex of the corner blocks were rounded over as were the lower and upper blocks. I expected that because the sanding sheet wasn't glued down. However I've tried glueing sanding sheets to a board in the past with that same sort of problem, especially with the corner blocks. It took about 10 min. on the granite slab to get all the blocks level and flat because the end block was close to 1 mm higher than the others. It tests absolutely dead flat with no gaps on the glass slab. The blocks are now about .4 mm too long so when the ribs are glued on I'll plane them close to level with the blocks and then take any excess off on the granite slab. So it doesn't matter how you reduce or flatten the block/rib assembly it's how you finish that counts. Before I came up with this granite slab and pumice idea I could never get a perfectly flat gap free result.
  3. Here's a pic of my current work back-down on the glass. I took the time to make sure the block ends were square and glued to the form vertically on this glass slab. I used spacers under the form to get the right height up the blocks. It does rock a bit but there's no point worrying about that until the ribs are glued on. So the back will need only a couple of minutes on the granite slab after the ribs/linings are planed down to the level of the blocks. The top will need a lot more work to establish the taper.
  4. There is, but it is very slight. The rib assembly just skates across the slab. I concentrate on the end blocks first, one at a time making sure they are not higher than the corner blocks. I then concentrate on those using both hands to apply pressure.
  5. Not so. The friction involved in sanding will tilt the rib assembly one way or another not matter how you apply pressure. That will result in the curved sections between the blocks being flattened. The granites slab/pumice method I've outlined leaves no measurable gaps.
  6. Well I hesitate to contradict what you say. I think your working methods are masterful. I can only go by my experience using abrasive sheets. I've tried heavy duty stuff stretched on a saw table but only about 240 grit, worn and unworn, exerting pressure only on the blocks. And that levels them pretty well. If the assembly rocks when I test it I mark the block ends with a pencil and sand again to see which ones are high. I then pare them back with a chisel and sand again. I repeat that until all the blocks are level. I then use a .2 mm palette knife to check the whole thing. If any gaps show up along the rib/lining surfaces and they're not too big I go to the granite slab. That .2 mm is not much but I always find gaps at least that wide after that process. That's why I say it is inevitable. It might not occur using lower density block material, I don't know.
  7. I think levelling blocks/ribs comes down to how the job is finished. Using pumice on a granite slab as I do produces a perfectly flat result. In my experience using a sanding board just does not work as a final step. The ribs and linings will always wear away more than the blocks, no matter how careful you are. So a plane is the only viable way to go, but it can be taken a step further with the pumice. I got the granite slab I use from a recycling place. It's probably the sort of fascia slab used on buildings. It was perfectly flat and polished when I got it. I use it a lot for dressing water stones with silicon carbide grit. I just cleaned it up and checked it with a straight edge and it is still very flat edge to edge, which surprises me a bit. It seems to hold the grit quite well and it does have small vugs or holes on the surface which probably help. Pumice is cheap, a lot cheaper than sanding sheets so cost is not much of a factor. It does break down quite fast so a fair amount is needed to remove a relatively small amount of wood. I trim the blocks to reduce the surface area and try to get the blocks close to level, perhaps to within .5 mm or less. I do that with sandpaper (very carefully) or chisels on the end grain of the blocks and test for flatness on a glass slab. I do use a plane as well especially for the top because I cut the upper blocks shorter. I checked this rib assembly back-down on my saw table and it is still perfectly flat with no gaps. I put it aside a few years ago, so the whole thing, form and ribs, haven't moved. You can see how flat it sits on the granite slab.
  8. I'm not against initial levelling of the blocks with sandpaper. If the blocks are cut near to finished length and glued on a flat surface there should not be too much to take off. I've always cut the blocks longer at the lower bout to allow for top plate rib taper. The problematic area I find is levelling the ribs down to block level with sandpaper. Too much of that will inevitably wear away the rib/lining between the blocks rather than the end grain of the blocks. And a misguided plane can easily take away too much in the wrong place also. I've simply found using a slab and pumice grit a safe and very accurate way to do the job. It still needs pressure to be applied to the blocks carefully but does not take very long.
  9. Well it would if the results were there. I test my results with a very thin, flexible palette knife with the rib/mould structure on a slab of glass or saw table. If I can slip it anywhere under the ribs or rock it I take it back to the granite slab, so it must be a pretty effective method.
  10. I find a granite slab useful for flattening ribs attached to the mould, but I wouldn't use it to attach sandpaper to or any other flat surface for that matter. Coarse pumice grit on the granite slab is ideal to level rib structures while it is attached to a mould provided you don't have too much to remove to do the job. The slab itself doesn't have to be dead flat. I must say I don't know how makers use planes to achieve this. It might be one way to lower the end and corner blocks if they are too high, but I don't know how anyone can get a flat surface that way.
  11. No, it is intended to be for the top. I've always had doubts about low, flat arching, but I thought it was probably the way to go. Now I'm not too sure, so I thought I'd try something else. I've made sure that the upper and lower bout arches aren't too flat and the convexity extends the full length of the plate. As far as I know, it complies with a usual top long arch.
  12. I've just drawn up an arching plan using a new long arch. It is 15.5 mm high without any pronounced flattening. The red triangles in the pic have baselines the depth of the arc at each cross-arch position. Each right-angle triangle uses a 5 deg. angle to calculate the height of the inflection point above edge level. I've set that height at 4.5 mm. That is a little above finished edge height, but is an average of 4 at the upper and lower bout and 5 mm in the waist area. So templates made using this layout will allow a bit of extra thickness for final scooping around the edges. The figures in the right column represent the height of the inflection point above the base of the plate. I've managed to use the one french curve to complete all of the convex and concave sections of the arching profiles without any difficulty. So the radius of the arc and angle of elevation and other settings work very well, probably better than any other layout I've made. The long arch I've used here is for the top. I haven't bothered to draw a recurve and I don't think making a template with one is necessary. I have tried top long arch designs to make templates which have flatter profiles before. But this has a full smooth curve. So I classify top long arches as flat or rounded, high or low. I see the back arching as more complicated with a completely different profile. That's my impression of a lot of early makers' work anyway. A complicating factor of the back is the big variation of thickness between the waist area and upper and lower bouts. Is the extra thickness added to the outer surface or inner? I've included a pic that I find interesting. It's one of Muratov's showing the different geometry of Stradivarius top and back arches.
  13. If you take into account the angle of the bevel there is not a great deal of radius there. Sharpening with a side to side or slightly oblique action will cut the corners back to some degree. If a gouge is badly out of shape I square off the edge by holding the gouge vertically and drawing it back towards me on a diamond plate making sure to tilt it so that the edge is square to the angle of the bevel. The further you tilt it the more you will cut back the corners.
  14. I'm not trying to convince anyone, I think it speaks for itself. There is nothing theoretical about it. What excites me about it is that I can design any sort of arching I want especially regarding the ratio of convex and concave components. That is the main determining factor on the overall arching shape. I can draw up a fairly accurate set of templates for the back and front in about four or five hours starting with the long arch templates I've decided on. Totally predictable and repeatable.
  15. Making templates like I have posted is time consuming. But I think that the concept behind it is important. I think exploring design possibilities on paper, especially regarding arching, is something new makers could benefit from.
  16. I'm only interested in concrete working methods which might lead to better outcomes. I'm too skeptical to believe that makers who listen to what the wood is telling them have any idea of what they are talking about.
  17. The Cremona makers lived post the Medieval period. And they undoubtedly had all of the skills and mathematical knowledge necessary to carry out their trades. There are complexities in what I have posted, mainly around edge height, but the underlying concept is quite simple. The inflection point location is calculated by a simple right-angle triangle calculation as defined by its position along the arc. Having arrived at a basic workable radius and orientation of the arc by trial and error (which can be varied depending on what sort of outcome you want) I have no trouble drawing cross-arch templates specified by that geometry.
  18. I think what I have posted about arching fits in the category of violin geometry just like the 4-circle method used to design a pattern shape. There is clearly an underlying geometry to arching however you go about carving plates. And that is that all inflection points lie on a straight line, looking from a 3 dimensional point of view. Deviating from that won't help the aesthetic or functionality of overall arching. The long arches are purely functional. The back of many early instruments seem to have their own peculiarities, particularly approaching the button. I'm sure the subtleties there aren't noticed by many.
  19. If you are prepared to spend the time you might be able to make your own. Cross-arch profiles just about make themselves once you have measurements for heights and inflection point positions. Of course the heights are determined by the long arches' shapes. Once you have that, figures can be set for the cross arches. Working out the general geometry that early makers used for that requires a bit of research and guesswork. The pic shows what is my latest version.
  20. Of course toothed blades track to some extent. But having unevenly spaced teeth won't change that. All that you can do is plane in slightly different directions. I've just used a Veritas block plane fitted with one to plane rough sawn ribs and there will always be high sections left preventing further progress. The solution is to follow up with a plane fitted with a plain blade. In the case of ribs both blades should be sharpened at a cutting angle of at least 50 deg. to prevent tearout.
  21. Well some makers do a very good job replicating examples of past makers' work. And as you say the way tools are used will play a large part in the results. Arching guides can predetermine an arching outcome. Mainly how full or otherwise it is once an inflection point is determined. And they simplify the whole process. Arching does have a definable geometry and I'm absolutely sure early makers knew of it and how to use it. Makers are always asking questions about arching and how important it is because they see very few practical, predictable ways to approach it.
  22. The only thing necessary to construct an arching profile is to locate the position of the inflection point between the upper convex and lower concave curves at any cross-arch location along an arc. And an arc drawn between the widest part of the upper and lower bouts can be used to calculate both its height vertically above the level of the edge and its distance from the edge horizontally. The inflection point is at edge level at both the upper and lower end of the arc, all of the recurve section of the arching is below edge level. So the three-dimensional position of the inflection point can be ascertained from an arc drawn on a sheet of paper. That arc can be drawn from a vertical between points where a double-string-length line crosses the uppermost and lowest cross-arch lines. How ISO lines can be used to draw arching profiles is beyond my understanding. As far as I can see they are a misleading irrelevance as far as trying to draw or make arching guides is concerned. Only geometry that can be used to make arching templates or patterns is of any use.
  23. Dennis J


    I would advise anyone planning to make violins that they study the geometry first. Soundhole placement is covered by that.
  24. Dennis J


    I've tried making inside moulds with basic measurements. And while the results are ok important things like the disposition of and the distance between corners can be a bit problematic. The 4-circles method works on proportionality so everything is balanced and practical whatever the starting measurements, such as the width of the lower bout, are.
  25. Dennis J


    Knowing the geometry of the violin is the only way you can resolve design issues if you want to make your own form/mould. And the 4-circles method described by Kevin Kelly is the best way to go. I have posted arching design methods, the last one is currently on page 57 of the Pegbox. It covers how to make long arches for both the front and back plates which can be used to establish heights for each cross arch as well as how to construct each cross arch profile.
  • Create New...