How Many Parts?

[JimLucas]

I've decided that this subject belongs in this subForum, not just as some frayed bit of another thread under General Discussion. So I'm moving my post here, and I'll put a pointer to this post where I had originally put its content. I hope some others will provide parts counts and related information for other types of construction, both current and proposed.

 

I seem to recall Harry Minting (Wheatstone), at an ICA meeting back in the 1980's, quoting a figure of 1500 parts in a concertina.

I have to admit that the figure that came to mind when I quoted Geoff Crabb was actually 1,600 bits - but another part of my mind refused to allow me to quote it - what ever the real figure is - it's very much larger than most people realise...

I just wrote a little program to add up the various parts in different configurations. It's crude. So far it only attempts standard vintage English-style construction (what I think I'm suffiently familiar with), but it can select between things like rivetted and non-rivetted action, veneered or non-veneered ends, bushed or non-bushed holes. I haven't yet added wire valve limiters, name plate, or a few other things like that, but they and correction of any errors should only increase the number of parts, and then only slightly.

 

There is at least one matter of interpretation: Pads are composed of three separate layers, but these aren't separately assembled for each individual pad. So I counted each pad once, but added three parts for the assembly of the separate layers before the pads are punched out of the composite stock. I have not counted glue of any sort as a "part". If I did, should I simply count the different kinds of glue used? I feel that counting each separate dot of glue applied would be unreasonable.

 

Aside from the pads, I have not assumed pre-assembly of any of the parts. These are counts down to the last detail. However, it is only a count of the parts, and completely ignores the simplicity or difficulty of producing or assembling the indivual parts (or selecting and obtaining the materials). E.g., it says nothing about cutting fretwork, routing reedpans, drilling holes for buttons, shaping finger plates, plating metal ends, assembling the ends and pads, etc.

 

Though I'm not reporting the details here, I have divided the materials used into 6 classes, with subtypes:

wood (various kinds, even in the simplest construction)

metal (various different kinds of brass, probably steel, and maybe dural or nickel)

leather (different kinds for different purposes)

cloth (both felts and woven types)

plastic (not necessarily, but in some construction)

paper (including carboard for spacers and bellows)

Oops. I left out the possibility of bone/ivory. For now, consider them to be "wood". Ditto for glass buttons.

 

A few results:

standard 20-button anglo (assumes air valve); no veneer, bushings, capped buttons, or rivetted action; 6-sided ends; 5-fold bellows

... 945 separate parts, of at least 51 different kinds

standard 30-button anglo (assumes air valve); no veneer, no rivetted action, buttons bushed, but not capped; 6-sided ends; 6-fold bellows

... 1275 separate parts, of at least 52 different kinds

30-button anglo; veneered ends, rivetted action, buttons bushed and capped; 6-sided ends; 6-fold bellows

... 1363 separate parts, of at least 55 different kinds

40-button anglo; veneered ends, rivetted action, buttons bushed and capped, brass inserts for strap thumbscrews; 8-sided ends; 7-fold bellows

... 1870 separate parts, of at least 57 different kinds

 

simple 48-button English; no veneer, button bushings, veneer, etc.; 6-sided ends; 4-fold bellows

... 1601 separate parts, of at least 57 different kinds

same as the above, but with 5-fold bellows and air button

... 1659 separate parts, of at least 57 different kinds

48 button English; air button, buttons bushed & capped, veneer, rivetted action; 6-sided ends; 6-fold bellows

... 1881 separate parts, of at least 61 different kinds

same as the above, but w/ 8-sided ends and gauze under the fretwork

... 2055 separate parts, of at least 63 different kinds

same as the above, but w/ 56 buttons, 7-fold bellows, and added wrist straps

... 2352 separate parts, of at least 67 different kinds

 

all the features of the above, but an 80-button duet w/ 8-fold bellows

... 3067 separate parts, of at least 57 different kinds

 

Since the Gold Tone banjo web site includes a photo of all the parts laid out for their cheapest model, I thought it would be interesting to count them and compare the numbers. I counted 69 different parts, but some of them have already been assembled from smaller parts. E.g., the complete neck would have at least the neck itself, the fingerboard, and the inlaid dots. In a kit at that price I wouldn't expect a tensioning bar, but I could well be overlooking some things, so for the sake of argument, let's assume the total actual pieces in an assembled banjo is in the range of 100-150 (the latter maybe if it has an inlaid resonator).

 

Compare that to the above figures for concertinas. I haven't yet tried incorporating various potential design simplifications (including any that might already be used by current "hybrids" or the Jackie). But I would be very surprised if the number of parts could reduced by more than half and still produce a playable "concertina". Even such a 50% reduction would leave still a humble 20-button anglo with nearly 500 separate parts needing assembly by someone (or some machine?) at some stage in the proccess.

 

Edited to correct a few of the numbers.

Edited June 27, 2006 by JimLucas

[JimLucas]

Some time ago I did this calculation, and achieved the figure of 1063 parts for a Jeffries style concertina with six fold bellows but no ebony veneer. Add on (from memory) 33 more per extra bellows fold and 18 per extra button. I included layers in a pad as separate parts, bushed single part buttons, rivetted action. Leaving aside the discrepancy between your figures and mine, which I am happy to accept without need for resolution, the number of parts, as you note, is not necessarily the most accurate way to assess the difficulty in manufacture. The 96 cards in a bellows, for example, are considerably easier to make than 1 reed pan, which in a basic form can require around 75 machining operations.

I did my exercise as a cascading program, i.e.,

Concertina is composed of: bellows, reed assemblies, action, ends, and handles

Action is composed of: boards, arms, force centers, and buttons

Arms are composed of: rods, leather buttons (at the end), and pads

Pads are composed of: leather dots, and pads (with an extra 3 "parts" for the layers of all pads, but not each pad)

Force centers are composed of: posts, springs, and (conditionally) rivets

Etc.

 

One of the advantages of doing it that way is that I can provide the same program with different inputs to get a number of answers far more quickly than I could enumerating and adding things by hand.

 

An even greater advantage is that it becomes far easier to notice -- or search for -- errors in my enumeration process, correct the program, and run it again to get the right (or a "righter" ) answer. And by the same process I can add sophistication. E.g., I realize now that I haven't provided necessary special treatment for metal ends: They require an additional board to guide the buttons and hold the bushings, and if the metal ends are inset into the wooden sides instead of extending all the way to the edge, then there will be additional screws to hold them there.

 

Edited to add: And in fact, I did make several small mistakes in my initial enumeration -- mainly parts that I was overlooking, -- which I found and corrected one by one. As I already said, there probably are still some discrepancies or variations that I haven't yet accounted for, but they should be small, and I'm still able to add them and run the program again.

Edited June 27, 2006 by JimLucas

[JimLucas]

An even greater advantage is that it becomes far easier to notice -- or search for -- errors in my enumeration process, correct the program, and run it again to get the right (or a "righter" ) answer.

I have just done this again. I found a couple of details I had overlooked. I have now corrected my program, and also the relevant numbers in my original post. Several parts counts have increased, though none by what I would consider a significant percentage.

[Theodore Kloba]

Excellent, Jim... It certainly proves the point. Is this program a variation something you're developing for some commercial purpose? It seems like it would have a rather broad application.

 

If you have no commercial aim with it, are you willing to share the source?

[JimLucas]

Is this program a variation something you're developing for some commercial purpose? It seems like it would have a rather broad application.

As written, it has pretty narrow application. It's crudely done, and far too many details are hard-coded.

 

But having done that much, I can see utility in something more flexible and sophisticated, and I'd like to do that. If I can make it past the first week of July, I'll make it an official project.

 

Email me in a couple of weeks to remind me.

[Peter Brook]

What would also be useful is to add some "standard times" for different manufacturing operations. The number of parts is important but the real expensive factor (which of course you know) is the time of the person hand building the instrument.

 

By "standard times" I mean the average time at which most people could complete the operation repetitively without tiring.

[Richard Morse]

I found a couple of details I had overlooked. I have now corrected my program, and also the relevant numbers in my original post. Several parts counts have increased, though none by what I would consider a significant percentage.

Have you included the bellows hinges, bellows frame splines, reedpan blocks, the linen center of the English thumbstraps, wire bale for an anglos handstraps? There's so much....

 

I don't have our list at the moment, but we do have a similar listing (a spreadsheet) of what goes into our concertinas, how many of each part, where the raw material comes from or if it's a process we job out - who to, material/process cost, lead time, shipping costs to get the material/part to us, etc. We include a figure for glue (and in our case, wax), and for the amount of hours it take us to assemble everything, tune, test and give that a dollar figure.... After each batch we readjust all the numbers to see how we did, add in an amount for waste and tool depreciation, administration, overhead, etc.

 

It can be very insightful how a snafu concerning a small but mulititudy part or one multi-operation or one outside jobber problem or training up a new employee can make a sizeable dent in our costs. Also good for streamlining operations, finding alternate materials or methods for producing a part.

 

It would be interesting to add staff-hours into your program, not as a lump like we do, but as applicable to producing the parts, and then as assembling, and then as tuning and testing. Then one could plug in $10/hr and see what happens. And $15 if we'd like a boost in pay (we're talking $US Frank!), or maybe the amount we could make elsewhere using our skills - like $25 or 30/hr. We can also go the other way and see what the same box costs to make in China where the labor rate is 0.64/hour. It'd be great to see a materials/process quality modifier (a percantage thing?) so that your program can be set to reflect the quality costs - like a 75% reduction for swapping cow or plastic sheet goods in the bellows for goat (in the case of inexpensive Chinese concertinas) or a 120% modifier for gold tooling on the bellows...?

 

If you're interested (and have the time!) to put some of my suggestions into your program, give me an e-mail and we can go over them.

[Stephen Chambers]

I'm reminded of this post of Brian's from a couple of years ago, comparing the number of parts/cost of a Wheatstone Hayden duet concertina, with the number of parts/cost of a hand made recorder:

 

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 !!!

It makes concertinas look positively cheap, in comparison.

[JimLucas]

Have you included the bellows hinges, bellows frame splines, reedpan blocks, the linen center of the English thumbstraps, wire bale for an anglos handstraps?

I'm not sure what you mean by "splines", so I'd better open up one of my concertinas and make sure there's not a piece I've missed. But I'm pretty sure I got everything there.

 

All the others you mentioned, yes.

 

There's so much....

There certainly is.

 

...we do have a similar listing (a spreadsheet) of what goes into our concertinas,...

Rich, currently your spreadsheet does far more than my program was ever intended to. Mine was only intended to count the parts. The additional things yours does would make mine much more useful, as would some other enhancements that I thought of while working with it. I would like to eventually construct an application that does all that and more.

 

But I'm not going to promise anything at this stage. If/when I do produce something that I think is worth sharing, it will probably come in slow stages (like Henrik's concertina?), but I'll let you know as it happens. And you in particular (and others with special interest) I would communicate with off-Forum.

 

I have to admit I'm a little surprised at the interest that's been expressed here. Surprised, but pleased. It may motivate me.

[JimLucas]

I'm reminded of this post of Brian's from a couple of years ago, comparing the number of parts/cost of a Wheatstone Hayden duet concertina, with the number of parts/cost of a hand made recorder:

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 !!!

It makes concertinas look positively cheap, in comparison.

Ah, but the shaping of the wood and those holes!

 

Still, I think it is true that a decent concertina costs far more in both materials and labor than even the best recorder.

[malcolm clapp]

I am very disappointed to read this thread!

 

I am now deprived of a simple pleasure in life, that being the ability to easily get off to sleep.

 

Let me explain...

 

When I was young and unable to get to sleep, my parents, like those of many of you no doubt, urged me to count sheep. As an adult, this excercise ceased to work for me; perhaps living in New Zealand had some bearing , but let's not go there....

 

Then I discovered problem solving as an aid to slumber. And my favorite problem was to calculate the number of parts in a 48 button 6 sided 5 fold Wheatstone English treble concertina.

 

To date, I have never succeeded in calculating to my satisfaction the number of parts before being overtaken by sleep.

 

Unfortunately, the figures quoted in this thread now render this excercise futile.

 

I now have no alternative but to turn to the dark side; I shall turn my attention to the number of parts in an accordion in order to get to sleep.

 

If it doesn't work, I shall hold c.net totally responsible!

 

I just hope that nobody on melodeon.net has the same idea!

 

MC

(only slightly tongue in cheek )

Edited June 28, 2006 by malcolm clapp

[JimLucas]

I am very disappointed to read this thread!

 

...I discovered problem solving as an aid to slumber. And my favorite problem was to calculate the number of parts in a 48 button 6 sided 5 fold Wheatstone English treble concertina.

 

To date, I have never succeeded in calculating to my satisfaction the number of parts before being overtaken by sleep.

 

Unfortunately, the figures quoted in this thread now render this excercise futile.

On the contrary. It is now your duty to check my result(s) and determine whether I've made any errors!

 

It is a fundamental principle of science that no result should be accepted as proven unless and until it can be independently verified. That is true of my results as much as anyone else's.

 

The fact that I've already had to amend some of my results since "publishing" them should make you and others cautious of accepting them without question, as should Rich's question and my response.

 

So... get counting!

[Chris Timson]

But having done that much, I can see utility in something more flexible and sophisticated, and I'd like to do that. If I can make it past the first week of July, I'll make it an official project.

The way you have approached this is well-known in concentric database circles, where it goes by the thoroughly apposite name of the parts explosion.

 

Chris

[malcolm clapp]

I am very disappointed to read this thread!

 

...I discovered problem solving as an aid to slumber. And my favorite problem was to calculate the number of parts in a 48 button 6 sided 5 fold Wheatstone English treble concertina.

 

To date, I have never succeeded in calculating to my satisfaction the number of parts before being overtaken by sleep.

 

Unfortunately, the figures quoted in this thread now render this excercise futile.

On the contrary. It is now your duty to check my result(s) and determine whether I've made any errors!

 

It is a fundamental principle of science that no result should be accepted as proven unless and until it can be independently verified. That is true of my results as much as anyone else's.

 

The fact that I've already had to amend some of my results since "publishing" them should make you and others cautious of accepting them without question, as should Rich's question and my response.

 

So... get counting!

 

 

zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz

[Theodore Kloba]

For my double-reed chemnitzer under construction, I need 928 parts, but:

• I considered bellows as 3 separate sections purchased preassembled; if I were to make bellows myself from scratch, add 360 more parts.
  
• I considered reedplates with tongues and leathers attached as preassembled; if I were to make reeds myself, add 642 more parts.
  
• I considered hand straps as preassembled; make them myself, add 12 more parts.
  

Total: 1,728.

[JimLucas]

For my double-reed chemnitzer under construction, I need 928 parts, but:

• I considered bellows as 3 separate sections purchased preassembled; if I were to make bellows myself from scratch, add 360 more parts.
  
• I considered reedplates with tongues and leathers attached as preassembled; if I were to make reeds myself, add 642 more parts.
  
• I considered hand straps as preassembled; make them myself, add 12 more parts.
  

Total: 1,728.

So... our concertinas differ in many details, but the numbers of parts aren't so different from each other. Certainly different from a banjo or recorder, though.

[JimLucas]

The way you have approached this is well-known in ... database circles, where it goes by the thoroughly apposite name of the parts explosion.

Thanks, Chris. I would have been really surprised if it were not a well-known approach. Are there common commercial products to implement this particular sort of system?

[Mark Evans]

I've just finished reading this thread. Whew! My little darlin' concertina is an host. I'm duely impressed and more than a little overwhelmed.