Walter O'Bannon

  • Content Count

  • Joined

  • Last visited

About Walter O'Bannon

  • Rank
    Senior Member
  1. Wow, I can't imagine what a job it would be to refinish either of those - just the initial stripping. Nevertheless, the total effect of the finish itself is not particularly violin-like. The one exception could be the trim work on the front panel of the confessional where hands have eroded through the top layers. The Bolognese cabinet from post #5 however seems to have an attractive underglow and the patina around the angels in particular is convincingly original. I might of course be swayed by the presence of all of the compass and scribe marks to just want to see violin varnish on it. The decorative style is also quite a bit like the surviving leather cases by Strad.
  2. A 17th century Bolognese cabinet...
  3. None of these are specifically Cremonese, but I find them to be interesting. The first is an early 18th century table with bone and a purfling like freehand inlay. It is made of fruitwood. The second one is a late 18th century pine credenza. You can see a golden undertone to the red-umber varnish. It has a greenish cast in places, but the varnishing on some panels would not be unpleasant on a fiddle.
  4. This appears to be from Sicily. I would think in hue alone, it is headed in the right direction.
  5. I have always wondered if there are similarities in finishing methods between the classical cremonese makers and furniture makers of the same period. Many people have theorized that regional varnishes are owed to makers purchasing from local chemists. If this is true, would the same varnish not wind up on end tables, picture frames, and the like? Does anyone have any insights on this?
  6. *more of a role in initiating the slip that is
  7. Watching the traveling wave reflect is indeed mesmerizing. I find it interesting that the stick-slip at the hair is occurring at the same frequency as the transverse wave. It's like the dynamic effects of the vibrating string play more of a role than just pressure and bow speed.
  8. An interesting and enlightening video
  9. I definitely like the boiled linseed oil version on the right. The other two look burned in and you see some color reversal. There has to be another medium for it to be in than LO though.
  10. I have not yet done a driven system either. In fact, this was the only modal analysis that I have ever done. Most of the analyses that I do are linear static and nonlinear quasi-static (not on violins). So did you use these cycloid equations to generate a node list that was then imported?
  11. That is an interesting point. I did have to choose the location for the tap tone fixity somewhat arbitrarily before running the model. After all, when tapping a plate you usually tap and adjust your finger location a few times before you find the point it really resonates. Nevertheless, when tapping a plate a fixity does exist and while it may minimally affect one mode because you are on a nodal line it definitely affects the other ones. Truthfully, I know very little about typical frequency response for plates and instruments. There is so much "noise" among the good information that I chose to get through school and form my understandings of mech vibrations from the academic world first. Do you have any examples of typical frequencies of a violin back?
  12. The modes of each test are completely different in shape and frequency. This illustrates the importance of correct boundary conditions in a model or a test. For example, one cannot expect that a top plate tuned to radiate an F will be contributing that same frequency when it is glued to the rib garland. One thing to remember when looking at the mode shapes is that they are all active when vibration is induced into the plate. Their relative contributions will change based on the driving frequency and location from which they are driven. With regards to impact, driving frequency is no longer a part of that but location is. If you maintain the same location but strike with your hand vs finger I would imagine that the pitch you hear should be the same but the timber would not, much like changing mallet hardness on a marimba. With regards to the numbers, the column that you are referencing is frequency in radians per second. If you look at the next column over it is converted to Hz. Also, as I mentioned before I feel that this model is a little thick. If the cuts for the edgework had been made I think that you would see less stiffness - especially in the "Fixed at Glue Surface" model.
  13. Can you take a picture of it in different light - perhaps outside?
  14. I've been scouring my computer to find that image. I no-longer have the model files but I may have a screen shot of the model itself. No guarantees though. Update: I could not find any pictures of the model
  15. Prior to graduating this December, I had a student copy of Solidworks and was working on building a solid model of a violin. The pattern is loosely based on the purfling of a 1710 Ruggeri that Tarisio had in a past auction. I don't remember the exact value, but I think I scaled the image to where the back was around 356 mm. A few deviations from the pattern also occur because I made the assumption that the back was symmetric (*ducks and covers*). All of the components were defined "in context" using the sketch that defined the mold, and changes to that sketch would rapidly propagate to the assembly. Unfortunately, I only got as far as finishing the rib garland and back (without a button). I did however run a modal analysis on the back using various fixity for which you can find a summary graphic below. The model for the back was created with 2 features: An extrude defining the outline with edge offset from the mold's sketch plus rib thickness, and a lofted cut defining the arching and graduation simultaneously. The lofted cut for arching and graduation was created using 5 sketches similar to the arching templates commonly used by makers plus 2 end sketches. I do not however think they were as accurate as they could have been. I have the Messiah strad poster framed in my living room that could have been used as a rough guide, but didn't want to take it apart. The thickness at center is around 4mm. I had plans for adding swept cuts to define the edgework, but didn't get to it. The model run in the analysis has edges of full 5mm thickness. With regards to the analysis itself, three different fixities were applied: fixity at the glued surface, fixed as if suspended between index finger and thumb (similar to a maker testing plate tap tones during graduation), and supported at 4 points along edge (similar to loudspeaker tests). Material properties were obtained from, and the software applied a solid tetramesh. The first 5 mode shapes with their associated frequencies are shown in the image. The goal of this was to develop some experience with the software and illustrate the observer effect more than anything. The same plate tested in with two different methods will produce substantially different results than those of the final assembly. Does a correlation exist? Surely. What is it? We don't know yet. As far as violin makers using FEA software to optimise their personal patterns, we all know that's not very likely. For anyone interested, Solidworks Premium is the most attainable at around $10k for an initial seat and $3k for yearly subscription. Its main advantage is that it is firstly a robust CAD system and has an add on Simulation package. Creating geometry (as John has found out) in a stand alone FEA software is usually more difficult and less adaptable. The main disadvantage of SW Premium for modal analysis is that many companies do not recognize it as a professional FEA platform. Many companies will create model geometry in a CAD system such as Solidworks and import it into an FEA software such as Ansys, Abaqus, Nastran etc for analysis. Some will even have an intermediate software such as Hypermesh that they do all of their preprocessing in before exporting the fem file to be solved by a more advanced nonlinear solver. This becomes prohibitively expensive of course, as yearly costs for commercial FEA software can by around $20k for even a modest number of tokens (job quantity and size that can be run at any given time). With expensive subscription costs come expensive penalties for software piracy. John's recommendation to obtain a student copy of Abaqus should ONLY be undertaken if one is legitimately a student pursuing an engineering degree at an accredited university. If someone is a student however, I highly encourage obtaining a CAD and/or FEA student license! -Walter