Building An Instrument

[Chris Ghent]

Things to keep in mind when CADing for laser or waterjet output is that all the lines must be either straight lines or arcs (no Bezier, Cubic, etc. curves), all must meet at vertexes, circles must "pigtail" in to start, and the lesser number of vertexes the less expensive will be the production of your part (the laser/waterjet equipment has to "think" at each vertex which adds an appreciable amount of time to perform the cuts).

Richard,

 

While I have had problems with the shapes of curves in the design phase and have touched on this in dialogue with Dana, I had not realised there were restrictions on the shapes of curves in the cutting process. A bezier curve is one with a shifting radius, right? Isn't it a total limitation that a curve like this cannot be cut?

 

And a cubic one sounds like the same only with a logarithmic based lengthening or shortening of the radius?

 

"All must meet at vertexes" means when a geometric shape, say an arc with fixed radius, meets a similar, you can't round off the corner, it must come to a point?

 

And "circles must pigtail in", well I'm stumped there..?

 

I just checked a pic of a Jeffries end and of course the shapes rarely follow

these rules. In fact the most pleasing curves (in every part of life!) are those

described as flowing!

 

I don't have expertise in AutoCad, I am just used to sitting down and learning new programmes. Your warning is heard. I suspect R2V progs have come along recently, one I tried yesterday (AlgoLab R2V) did create arcs, but could not decide how to handle a vertex consistently. Also, as it was drawing a line without physical width (steady down the back) it wandered around inside the width of the rasterised line it was converting enough to be disconcerting.

 

I found it much easier to import a scanned image into my CAD program and draw (CAD) right over it. That way you create what you want without having to "fix" stuff you don't want.

 

When you say draw, if you are meaning you can freehand over the top in vectorspeak, then that would be great, and I'll do whatever you are doing. However, I suspect you mean arrange arcs and lines by means of inserting points and influencing the line produced between them between into a curve?

 

I entered this particular area of concertina building (cad) believing I would be able to show a computer a line drawing and it would accurately replace the lines with some sort of "software guidance tracks". I now know that idea is so divorced from the way it all actually works it is risible. Just another compliment to the human brain I guess.

 

As regards .4mm offsets, would the "jobber" (great word, I usually try to de-Australianise the speech in my posts, maybe I shouldn't) be able to act on an instruction like, "cut .4mm to the inside of all closed shapes?"

 

Hope you might have a minute to continue my education, no hurry...

 

regards

 

Chris

Edited a couple of times to remove evidence of carelessness and stupidity...

Edited February 17, 2004 by Chris Ghent

[Richard Morse]

I had not realized there were restrictions on the shapes of curves in the cutting process.

Not in the cutting process itself, but the restriction is in the capabilities of the code that instructs the cutting process. Most laser/waterjet/CNC milling machines, etc. use G-code which doesn't support the type of curves we'd like to see.

Isn't it a total limitation that a curve like this [bezier, cubic] cannot be cut?

It is a design limitation AND a challenge for us to create a design than so closely approximates a fluid curve (from sequential arcs) that it's near impossible to tell the difference. Not a *total* limitation....

"All must meet at vertexes" means when a geometric shape, say an arc with fixed radius, meets a similar, you can't round off the corner, it must come to  a point?

I'm not sure what you mean by "round off the corner". An arc has 3 vertexes (or points) - a center point and two endpoints. Those endpoints MUST meet the next line (or arc) section or the cut path MAY jump to the nearest vertex (maybe to the "other side" of the offset), or it may stop and start again somewhere else (in which case you'll not have a cut through there and it'd be like a perforated puzzle piece which is broken out).

 

If those two arcs are tangent at the common vertex, the transition will be imperceptible. If the arcs meet at an acute angle (even the most least acute angle), there will always be that "abrupt" change in direction. The only way to "round off" that transition would be to create an arc between the two which is tangent to both of them.

And "circles must pigtail in", well I'm stumped there..?

Whenever a laser/waterjet "cuts in" (the first stab at penetrating the material), its cut ablates more of the material than if it were already cut through and following a path. This isn't very noticeable when the start is at a corner as the slight variation is not very noticeable. But when you start in the middle of a line or curve it is very noticeable, and particularly with very regular curves like circles. The cut has to start somewhere on it, and that point will look like a planet (oid) in an orbit.

 

To make the circle a very smooth and uniform cut, a small arc is inserted that tangents out to the CAD start (not just anywhere) on the circle. It looks like a little pig's tail. You tell the cutter to start "on the pig's tail" and when the cutter gets to the circle (the part you want smooth), it will already be at the right depth and in motion so it doesn't blimp out there.

When you say draw, if you are meaning you can freehand over the top in vectorspeak, then that would be great, and I'll do whatever you are doing. However, I suspect you mean arrange arcs and lines by means of inserting points and influencing the line produced between them between into a curve?

The program I uses does a function that lets me "trace" a bitmap as well as "freehand draw" over one (I use a graphics tablet and pen rather than the mouse or trackball for stuff like that) and subsequently "smooth" the jaggies and convert the line into any type of curve including arcs. Still, this needs to be tweaked as the resolution is so fine that I want to simplify things to make the cut faster (and less expensive).

 

When I CAD by "arranging arcs and lines" it actually goes quite quickly as my program is flexible enough to "affix" vertexes together while manipulating a segment (like altering the radius of an arc or rotating it). I can also make a new line (or arc) be tangent to any other no matter what size I make the next segment or where I put it - the program creates a moving vertex that's tangental to both. Once I set the arc I'm working on it creates the new common vertex.

As regards .4mm offsets, would the "jobber" [would] be able to act on an instruction like, "cut .4mm to the inside of all closed shapes?"

Yup. Though your instruction to him would usually be to "cut on the path" or on one or the other side. Depending on the beam focus/power/cut rate/material/etc. or pressure/aggregate/orafice/rate/material/etc. that offset could vary enormously. Best to let the jobber do the interpreting.

[Chris Ghent]

Richard,

 

thanks for a big jump in understanding. A supplementary question; could you define the exact meaning of the word "tangent" as you are using it here,

 

If those two arcs are tangent at the common vertex, the transition will be imperceptible. If the arcs meet at an acute angle (even the most least acute angle), there will always be that "abrupt" change in direction. The only way to "round off" that transition would be to create an arc between the two which is tangent to both of them.

 

If I use the popular meaning, "an abrupt change of course" I can't understand the paragraph, as the lines would have to meet at an acute angle to be tangential.

 

Dictionary.com offers a clue, but uses the word itself in the explanation, destroying the chance of an objective explanation ie.

 

"A line, curve, or surface meeting another line, curve, or surface at a common point and sharing a common tangent line or tangent plane at that point"

 

The only way I can make sense of this is if the meaning of the word as you use it here is that there is no change of direction at the vertex.

 

Anyway, hope you can see my dilemma here.

 

regs

 

Chris

[Stephen Chambers]

You will probably consider this to be cheating but, when I went to the Greenwich International Festival of Early Music last year, I was shown a prototype anglo concertina kit !

 

It was being exhibited by the Renaissance Workshop Company Ltd. (must be a copy of the one Oliver Cromwell played ?), of Bradford, and made for them by Andrew Norman, but it doesn't seem to have gone into production ? (At least, there is no sign of it on their website : www.renwks.com)

 

Interesting idea though ...

[JimLucas]

...when I went to the Greenwich International Festival of Early Music last year, I was shown a prototype anglo concertina kit !

 

It was being exhibited by the Renaissance Workshop Company Ltd. (must be a copy of the one Oliver Cromwell played ?), of Bradford, and made for them by Andrew Norman, but it doesn't seem to have gone into production ?

 

Interesting idea though ...

Made by Norman? Then that would seem to make two different anglo kits about to come on the market.

 

A few months ago there was an announcement in Concertina World (the ICA Newsletter) of a soon-to-be-available kit, but I'm sure the maker wasn't Norman. There was even a competition to name the model (which I entered, of course), with one of the kits as a prize to the winning entrant.

 

I suspect it's about time for the results of that contest to be announced. I'll see if I can find out more and report back.

[inventor]

Latest news on the Anglo-concertina kits from "Concertina World" is that "the maker and developer of the Anglo Concertina Kit, - - - - - , is no longer in business at the Bradford premises." . Well I am not surprised; making concertinas -even parts for concertinas is a lot more involved and time consuming than people (and that includes concertina makers themselves) think ! Some years ago I had a 46 button concertina delivered to me from Wheatstones it cost £625 (say $1000) it had around 2000 different parts; in wood metal leather, and felt; some of which have to be made with an accuracy of a tenth of a thou(santh of an inch). At the same time a musician friend took delivery of a hand made Recorder this cost her £600. It consisted of just 4 bits of wood with a few holes in it !!!

Inventor

[Theo]

Mark

 

Just a thought for building your CAD skills and solving another problem.

 

You could try using CAD to design the levers and posts for a rivited action. The shapes will be much simpler, allowing you to build your CAD skills gradually. There will be many similar parts enabling you to take advantage of the ability of automated systems to do repetitive work.

 

Theo

[Richard Morse]

CAD can be used to draft the action parts, but I don't know of any system for automating this type of work. Except for the posts and general shape of the arms, there are no repetitive action parts - mainly due to the different arm lengths and post positions.

 

I had tried to regularize groupings of similar arms to save design, production, stocking time, but it didn't make sense as the time to deal with regularizing had to be accounted for, and it turned out that the production time was identical, and the stocking was only about 20% less though would actually be more cumbersome for record keeping.

 

I've found CAD great for Drafting and very, very poor for Ddesign.

Edited July 22, 2004 by Richard Morse

[d.elliott]

I visited the kit workshop in Bradford, met the chap doing the development work, examined the reeds, pans, plates, arms, keys, bellows kits etc.

 

The chap concerned made a wide range of 'old' instruments, this anglo was to be just another in the overall range of woodwind, plucked, strummed string and key board instruments.

 

I thought that he had overcome all the easy problems, many of the mass production problems, and was well into prototype production. I am only sorry to hear of the demise of the business, and the loss of such a concentration of very good craftsmen. I wonder what happened to all the lazercut end plates, the tens of sets of reeds, etc etc etc

 

Dave

[mike byrne]

It was being exhibited by the Renaissance Workshop Company Ltd. (must be a copy of the one Oliver Cromwell played ?), of Bradford, and made for them by Andrew Norman, but it doesn't seem to have gone into production ?

I spoke to Andrew Norman at lunchtime and he has never been approached by or made any Concertina kits for The Renaissance Workshop. He does supply spare parts, but that would be an expensive way to make one.

[Stephen Chambers]

I spoke to Andrew Norman at lunchtime and he has never been approached by or made any Concertina kits for The Renaissance Workshop. He does supply spare parts, but that would be an expensive way to make one.

Apologies to Andrew for any confusion. I'm sure his name came up in the conversation, but perhaps there was some misunderstanding. It was the best part of a year ago and I can't recall quite what was said.

 

Anyway, I hope to put some kits on the market myself, when I get a chance.

Edited July 24, 2004 by Stephen Chambers

[Chris Timson]

Latest news on the Anglo-concertina kits from "Concertina World" is that  "the maker and developer of the Anglo Concertina Kit, - - - - - , is no longer in business at the Bradford premises."   

It seems the story from their website is that they are moving lock, stock and barrel to Toledo in Spain, which might slow things up a bit. The story is here.

 

Chris