Official Luthiers Forum!

Sheesh, now I wish I never opened my mouth about my measurements.
indeed, it was because it was 2.3 micrometers off from the required shape, which is a little over 90 millionths of an inch. So yeah you got me on that one Hesh!
Side not my grandpa worked on that project, I don't know exactly what his contribution was. It was unfortunate he lost his eyesight before the first pictures came back from the hubble.

Mirror grinding and part milling 2 different thing. But no! NASA was not grinding the mirrors closer than .00001" tolerance. NASA contracted the work out as they do 99% of all fabrication. I will see if I still have the Modern Machinist magazine form several years back. If I remember correctly the tolerance used on the mirrors was in the .00025”-.0003” range, I can be a bit off here but not near .00001”. Also it was measured with extremely sophisticated laser instruments because the human eye can not read incrimination that tight more less a machine scribe it onto a dial or scale.

Yes in cad it is possible to draw geometry that is t 9 decimal places accurate. The only reason it is cut off there is they had to stop some where. Back when I was at MTSU my senior year I was involved in the development of MaCado Unigraphis in the early 70's (the first commercially available 3d cad/cam software). This software was developed first for use by MacDonald Douglas, it was later retailed to commercial manufacturing by Mac Donald Douglas. Therefore the name MaCado Unigraphis. It was a true 3d modeling software 10 years before any other 3d modeling software was on the market. That was it’s niche.

I am a 35+ year mechanical engineer and cad designer, and can tell you for a fact that cad systems can be very accurate because every thing in a cad package is just numeric equations. If the developer deemed it reasonable to design the systems to operate in 10,000 decimal places they could have. But software numbers do not relate t real world machining. Everything in a cad package is just mathematical formulas. When you are machining real steel and or plastic, wood, or whatever you have to deal with physics as well as mathematics.

Fair enough, but I had seceded to your point several posts ago, but somehow jokes just done come across all that well on the intertubes. I know it wasn't Nasa, that did so, but Perkins-Elmer the optics company commissioned to do make those mirrors. But not everyone knows who they so the joke wouldn't have made any sense. I think this thread has run its course, and I'm afraid I might inadvertently step on peoples toes due to my naivety of the subject matter. So I figure its time put an end, before I dig a whole too deep.
I've received valuable information from everyone who offered to contribute and it is all very appreciated.

Oh no Greg, I was just discussing the topic and no offence was meant or taken by me at any time. I am sorry if I came across as if I had. That would be my boo-boo.

I made my point strong because a lot of people don’t understand the difference in the machines software capabilities and its manufacturing performance. And unless you have worked with machining it is hard to understand exactly what a manufacturing tolerance is and why variances happen between consecutive cuts.

My last post I was just trying to say that software works with numbers that do not vary in value. But tools and materials add their own physical dynamic to the actual work. The input software is based on just numbers. But the out put is based on interface between tool and material driven at the values that the software provides.

Gees I wish that the software numbers could do the job with adjustment for physical values of actual doing the cut. If so we would near never scrap a part.

Hehe true that, I wish that was true with visual effects too, you'd think that using the math behind light and how it behaves with a surface could easily be calculated if you know all the variables. But more often than not I find myself having to fake things to make them look more real. If that makes any sense. Too many variables to work if your aiming for accuracy. Movies would never get done if that were the case. You rely on numbers to get you 80% there than break it to make it work.
But I understand that numbers in a computer don't equal stuff in reality. I learned that when I was working at Stan Winston studio and was modeling the suits for Iron man. We had the pieces either grown from a stereolithic resin process, or milled out of styrofoam. It was my first crack doing anything outside of the virtual world and I ended up giving no tolerance to some of the joints that met up. So the poor guys in the model shop had to sand down parts that were meant to connect together. So I learned real quick that what you see in the computer is not necessarily what you'll get in real life. But definitely cool to see something you made in a computer actually exist!

Greg, I’d like to apologize also.
Hope I didn’t offend you.
I do like the fact that you’re trying to be as accurate as possible.
Kudos to ya.

_________________
Say what you do, Do what you say.

Oh no don't worry about it, I wasn't offended one bit. I just didn't want to come off as some thinks he knows more than he actually does. Since I wasn't sure how accurate of a model I needed to make, I just decided to be as accurate as possible to be on the safe side. I figured I'd get some flak for modeling the geometry to the nearest atom, just didn't anticipate that much of a response.

Its just that the thread served its purpose, and was beginning to be a topic that should reside in the CNC Lutherie section of the forum. Since the original question asked was regarding kerf width, and ended up being about milling precision.

So don't worry about it, all is good.

The Hubble .... Yes ... its been COMPENSATED !!!!

At a millionth of an inch accuracy, you can be guaranteed that the guitar will play perfectly out of tune in all keys !!! Moreso than us mere mortals can do .

_________________
Tony Karol
www.karol-guitars.com
"let my passion .. fulfill yours"

Hehe, they say a stopped clock is more accurate twice a day then a clock running a minute slow or fast!