Fun with CAD

[Terry Downes]

Good point. I am not sure if Terry mesh fixed out of Templot or not, I have presumed he did.
I certainly mesh fixed my Templot exports. when trying to get them to work in Freecad. but I could not get them to do what I wanted.
cheers
Phil,

FYI, Yes, I did mesh fix using 3D Builder. SJ1 chair attached (00-SF default settings, Templot stl Mex fixed via 3D Builder, imported into SolidWorks and exported as AP203 .STEP)

 

[Phil G]

Hi Martin,
With a bit of manipulation and some rounding up and down in certain places I believe I have, more or less super imposed a full size drawing of the side elevation of an S1 REA chair on to the 3D model of the same chair scaled up to full size.
There very close given my drawing is full size even the 0.0907 interference fit of the key side translates to a 0.0012 thou inference fit,
when you then factor in both rail extruder variation, and resin chair shrinkage you can see why it all fits nicely together.

The only three differences I can see to the profile, which are not at all major, are the key jaw seems to be a scale 0.0041 thou smaller in height, and the gauging jaw is slightly squarer, and goes deeper down the rail, which makes total sense as more strength is an advantage in the resin.
My question is have a more or less placed the REA profile over your stylised chair correctly this time?

cheers
Phil,

 

[Martin Wynne]

My question is have a more or less placed the REA profile over your stylised chair correctly this time?

@Phil G

You're a hard taskmaster, Phil.

There isn't a "correct" answer because I set out to make the chair look like an REA chair, not be an exact scale model of one. It can't be that, because of changing the rail to vertical and using model rail sections.

The prototype has the top of the outer jaw fractionally below the top edge of the key. If the jaw was higher than the key, rainwater would lodge behind it and rot the key:

Rotating the rail to vertical pushes the key lower than the REA design, and to keep the prototype appearance I pushed the top of the jaw down too. The key intentionally overlaps the jaw face to ensure integrity after slicing (they are inserted as two separate blocks), and it also represents the slight prototype crushing of the key as it is driven home:

They're very close given my drawing is full size even the 0.0907 interference fit of the key side translates to a 0.0012 thou inference fit

I think some of your discrepancies are due to rounding effects in whatever scaling process you are using. For example you have the seating thickness showing as 1.7498", and it is definitely exported by Templot as 1.750" (scale).

p.s. I can't keep explaining every tiny detail of the design, otherwise I never will make any progress. Writing this reply and creating the attachments has taken up most of my computer time for today. I'm hoping to make a release of 244c later because I found another bug. I want to reach a stage of a stable release so that I can forget endless explanations and get on with the next stage.

cheers,

Martin.

 

[Phil G]

@Phil G

You're a hard taskmaster, Phil.

There isn't a "correct" answer because I set out to make the chair look like an REA chair, not be an exact scale model of one. It can't be that, because of changing the rail to vertical and using model rail sections.

The prototype has the top of the outer jaw fractionally below the top edge of the key. If the jaw was higher than the key, rainwater would lodge behind it and rot the key:


View attachment 9047


Rotating the rail to vertical pushes the key lower than the REA design, and to keep the prototype appearance I pushed the top of the jaw down too. The key intentionally overlaps the jaw face to ensure integrity after slicing (they are inserted as two separate blocks), and it also represents the slight prototype crushing of the key as it is driven home:
View attachment 9046



I think some of your discrepancies are due to rounding effects in whatever scaling process you are using. For example you have the seating thickness showing as 1.7498", and it is definitely exported by Templot as 1.750" (scale).

p.s. I can't keep explaining every tiny detail of the design, otherwise I never will make any progress. Writing this reply and creating the attachments has taken up most of my computer time for today. I'm hoping to make a release of 244c later because I found another bug. I want to reach a stage of a stable release so that I can forget endless explanations and get on with the next stage.

cheers,

Martin.

Hi Martin,
Sorry I am bugging you for an explanation, I will leave you in piece now promise I was not suggesting you were wrong either, only getting clear in my head a reason for the discrepancy. (I knew there had to be one)
I figured there must be a reason for the key jaw being a bit lower for example.

I understand there stylised chairs, that said there a very very good representation of the real thing. Which if my memory serves you mentioned you had a real cast iron chair in your garden.

By the way the slight errors are because the cad is measuring in 3D and some of the z values are not on the same plane,( they sometime pick up the back edge of the triangle for example, hence some manor value discrepancies, I should have mentioned that, I was just being a bit lazy and not pulling them all back onto the same Z plane.
cheers
Phil,

 

[Martin Wynne]

.
See also this topic:

https://85a.uk/templot/club/index.php?threads/cad-basics-from-james-walters.911/

Martin.

 

[Rusty]

Hi Phil,

FreeCAD seems to offer what you are looking for -- a free program which will convert an STL mesh into a solid:

For information there is an interesting new kid on the block: Ondsel.

https://ondsel.com/

It's built on top of FreeCAD so it has all the features (as far as I can tell) but with a more intuitive, consistent user interface, than FreeCAD. The design process is a little different in FreeCAD which has been a bit of a barrier for people moving to FreeCAD from other applications. I've just downloaded Ondsel to give it a trial run - fortunately a MacOS arm package for MacBooks M1/2/3 chips is available.

The solo version is free to use - the costed packages are more for team working - cloud storage and computing time etc.

 

[Martin Wynne]