Hole dialog discussion

[ToniTen]

So if you could deliver a norm in which hole sizes for dowels and pins are normed, I can have a look if I can add this feature.

I don't have access to the standards themselves either. I hope someone in the community has them. I would hate for someone to buy the standard and then have it not include this information, and instead refer to a different standard.

I can repeat the guidance pin manufacturers give, which might or might not be what the norm states, but I assume you already got that. They give different values based on hole material (hardened steel, mild steel, aluminium, brass...) and fit (interference, transition and clearance).

As a first step, I would think that it would be best to include the different norms with the hole diameters, that are the same as the diameters of the pins (before tolerance). I could compile for you a spreadsheet with the different norms and the different pin/hole diameters for each norm.

As for tolerances, standard DIN/ISO pins that I have dealt with have m6 tolerance (non-standard ones might differ), but since hole tolerances depend on fit and material, I would leave them freely available to the designer

[RatonLaveur]

Link for metric:
https://www.spirol.com/library/sub_cata ... ins_us.pdf

Link for freedom units:
https://www.spirol.com/library/sub_cata ... ins_us.pdf

Link for ISO-8752:
https://www.spirol.com/library/sub_cata ... ins_us.pdf

And last but not least: all these specs give a recommended hole size

[uwestoehr]

Link for ISO-8752:
https://www.spirol.com/library/sub_cata ... ins_us.pdf
And last but not least: all these specs give a recommended hole size

Very good. I'll put it on my Todo list.

[RatonLaveur]

I'm no coder, but if i can parse that data for you (into a csv structure of your choice for example) let me know I'm happy to help anyway i can.

[uwestoehr]

Very good. I'll put it on my Todo list.

I was too eager. I talked to one of our technicians and he convinced me that the hole dialog should not propose hole diameters for press fits and spring pins:

Depending on the material the hole diameter needs to be adjusted. This applies for almost all cases. So a DIN 7 pin in hardened steel is something different than in aluminum. People are lazy, they create a hole and see, "cool, I already have the diameter". But in fact they have it in most cases not. So we should not offer something that make the users feel safe, despite we know that this is not the case. So by not providing diameters for press fits, we force the user to have a look at his material and the diameters recommended by the pin supplier, handbooks etc. And there they write that it strongly depends on the material. So the user will then also think about his material. As result he designs holes that are probably working well.

Also from my experience I know that a working fit press hole diameter is often a matter of only a dozen microns more or less. There are therefore drilling tools available with diameter steps of 50 microns.

[ToniTen]

On the TechDraw forum, there's a thread on "threads in TechDraw" with what I think is a wonderful idea to get proper threaded holes support in FreeCAD, as outlined in this post. Crossposting it here because the idea actually just requires some minor modifications to the hole function and solved many problems in one go (quoting myself from that thread):

The process is as follows

• You create the sketch
• You select the hole dialog
• You select threaded. At this point, it asks for thread size, depth, and core depth.
• This creates a cylindrical surface without end caps (not a solid) that acts as a placeholder for the thread, with a diameter equal to the thread size (16 for M16), and a height equal to thread depth

This accomplishes the following goals

• Having the cylinder there in 3d view makes threaded holes instantly recognizable. This is an issue, because right now there's no way to see if a hole is threaded without checking the properties of the hole operation that created the hole
• Even unmodified, TechDraw should plot the cylinder outline in both top down and section view. This would not fit the ISO standard exactly (the top down view for threads is a circle arc, not a full circle), but it would be a step in the right direction. Adding a dimension to the circle would give you the thread size, but the designer would need to remember to add the M in front of it
• TechDraw could then be modified to recognize this face cylinder and display it and it's dymensions as the ISO mandates

How feasible is this? how many issues does this cause?

[chrisb]

Cossposting is not a very good idea. At maximum a link should do.

[uwestoehr]

The process is as follows

Yes, this is the state we currently have. There is an issue: the proposed diameters for threaded holes are not always norm-comform. Fixing this is on my Todo list.

I did not quickly act because the question is why FreeCAD is not modeling the thread directly. What are the arguments?:
Pro:
* In other areas FreeCAD creates models as they will later be. This allows you to make everything with it, simulations, ray-traced views etc.
Contra:
* if you have a part with dozen threaded holes, creating objects for real threads makes recomputing slow
* the model tree is "polluted" with objects

So my current opinion is to:
- make the boring diameters norm-conform, but by default don't model the threads
- add an option to make the thread modeled as PartDesign object that appears in the tree

What are your opinions?

[ToniTen]

Cossposting is not a very good idea. At maximum a link should do.

Ok, sorry, won't happen again. I did include a link to the original post, I just could not figure a way to link to my follow-up post to that.

I did not quickly act because the question is why FreeCAD is not modeling the thread directly. What are the arguments?:
Pro:
* In other areas FreeCAD creates models as they will later be. This allows you to make everything with it, simulations, ray-traced views etc.
Contra:
* if you have a part with dozen threaded holes, creating objects for real threads makes recomputing slow
* the model tree is "polluted" with objects

Another con: norm conforming drawings don't need the actual threads, they need the an arc of a circle on the top view, and a rectangle on the section view. TechDraw would need to remove the thread geometry for this.

So my current opinion is to:
- make the boring diameters norm-conform, but by default don't model the threads
- add an option to make the thread modeled as PartDesign object that appears in the tree

What are your opinions?

I think we're confusing things here:
Currently, threaded holes are no different from regular holes in either 3D view or TechDraw. My suggestion (based on the post I linked in my previous postt, it's worth checking the original thread, for the actual proposal from Maker with UI changes to the hole dialog) is to have the hole tool create one very simple symbolic 3D "fake" geometry to serve as the basis for both 3D view and TechDraw drawings: a cylindrical surface without top and bottom faces, that will be around the actual hole, and that will have a height equal to thread depth (not necessarily hole depth, blind holes are often not fully threaded). This fake geometry will have the following advantages:

• Serve as a visual hint that the hole is threaded in 3D view, since it will have a double ring around the hole. Currently there's no indication in 3D view that a hole is threaded, and the only way to see that is to find the hole operation that created the hole and check it's properties
• Serve as the rectangle required by the norm for TechDraw section view
• Serve as a placeholder for the arc that the norm requires around threaded holes once TechDraw is updated to support it. I know that TechDraw has some very fancy functions to add arcs around holes, but right now, I don't think they're used for this

I'm not knowledgeable enough about the inner workings of FreeCAD to know if this is a good idea, but it has the benefit of simplicity (all you're doing is adding a simple cylindrical surface around the hole, which should be very cheap compute-wise), but it would at the very least move FreeCAD closer to being norm conforming.

As for actual modeled threads, I do often export my models to Blender and add actual threads for rendering there, so having the option to have the exported geometry include the thread is awesome and will save me loads of work, but it's IMHO less of a priority, plus it has all the cons outlined above.

Of course, adding "fake" geometry comes with it's own issues, that I can't properly evaluate myself since I can't foresee all the unintended consequences that implementing this would have.

[uwestoehr]

I finally found the time to fix the hopefully last major issues of the dialog:
- the diameters for threaded holes are only norm-conform for metric fine threads
- some UNC/F/EF sizes are missing

With this PR both will be fixed: https://github.com/FreeCAD/FreeCAD/pull/4091

For the diameters I took ISO 2306 since it covers the original definition DIN 336 and also the UN threads defined in ANSI B1.1. Note that due to rounding some tables you find in the web differ by up to 10 microns.

Here are some tables I found and used:
Metric, UNC, UNF: https://motorang.com/schraubereien/PIC/ ... explus.pdf
Collection of thread and bold standard: http://www.mettex.co.uk/wp-content/uplo ... ndards.pdf
thread tables: https://innorat.ch/Gewindetabellen_u1_20.html
another thread table collection: http://www.mikroskopfreunde-nordhessen. ... 011-12.pdf