7mm US model dabblings

[mickoo]

Mick,

The chimney doesn't look too bad - although I'm working through the same problem at the moment in TurboCAD. I suspect the true enormity of the problem you've got is only really apparent from square on front and sides...

Keep at it - I'm sure there's a solution there somewhere.

Steph

Steph,

I'm sure there is, but it's a case of having the patience to find it. It looks ok in the renders but the base isn't round, it's more oval due to the round barrel intersecting the boiler, then you fillet that joint. In 3DSM you'd just take the base and widen it to make it the right shape Autocad 3d just does not have that tool, no that I've found.

The extrusion only fails when you fillet the joint, before that the base will extrude and chamfer just fine.

The other hard part I've experimented with is smoke box door hinge straps, they are just the devils own work trying to get the strap face to run parallel with the door face underneath, whilst maintaining the same curvature and twist across the door face.

It seems to work really well with basic shapes, something like a buffer housing and mounting flange or cylinder casting etc it'll eat in seconds, but the compound curves are another matter altogether.

MD

 

[Steph Dale]

Hmm, well smokebox door straps I've not had a problem with, it's just the compound curve stuff.

Mind you, I'm still only on assembly number 5, so I've a way to go yet just to find out what all the 'buttons' do:



Steph

 

[mickoo]

I think Autocad would struggle with some of those parts, or more correctly I'd struggle with Autocad to do some of those parts.

On the B1 the door strap is a compound curve of sorts, it has to curve across the door face which is domed, anyway that's one of tomorrows tasks, the dome will suffer the same as the chimney base with the fillet radius, but the real thing isn't a uniform radius so it won't show as badly and there's no flange at the base so it won't have those issues. Safety valves are simple turnings, same for the tapered sand box fillers.

The hardest part is making it all manifold or water tight, if you could just stick the components next to each other it'd be a lot easier.

In 3DSM you worked with vertices and faces and pulled and pushed them to get the shape, with Autocad you have to draw the shape in 2D and then extrude or loft, that's fine until you want to adjust a part of it later, then it gets wearisome!

The pan, zoom gizmos really suck, in 3SDM you just clicked the item you want to roll around and it does it, in AC3D you have to click the little cube and then grab the right part of it and roll around, 9-10 the parts zooms off the screen somewhere and why oh why is top never top all the time! It just looses it's spacial orientation almost constantly, grrrr

MD

 

[simond]

Mick,

Complex 3D curves - use Catia.

Pricy, it was over €14 per hour per seat for my staff in India 10 years back, and that was a large multinational customer, can't imagine the price for a solo artist.

Best
Simon

 

[Steph Dale]

Well I've now tested how well TurboCAD does lofting. This is based on some old and inaccurate artwork (the chimney was drawn for effect, rather than accuracy) but there's some potential here. It is a bit involved though so I'll have to try another way and see if I can find a quicker method:

There are some obvious creases where I didn't do the profile accurately enough and the skirt showing through the flare is just abject laziness. I've got a really good idea how well this works; it's just not particularly quick!

Steph

 

[Brian McKenzie]

. . . . so onto something a bit harder, a simple chimney, the B1 in fact.
View attachment 77138

. . . for some reason Autocad will not extrude the flange at the base nor will it vary the fillet radius at the base, neither can you take the upper fillet line and flatten it which would of course vary the radius.

Mick,
I note you comment later that AutoCAD 3D does 'lofts' apparently. I used that tool in SolidWorks to get this chimney - so perhaps this method could produce the variable radius curvature sought?



The makers plans show the chimney base as a true circle when viewed from the top looking down, yet the hole in the smokebox plate (cut when flat) becomes slightly elliptical when viewed similarly after rolling.
I did wonder if the pattern makers for this base started with a disc of material and then draped it over a suitable mandrel of smokebox diameter - resulting in a marginally different shape to the plan.

Smokebox door hinges can be another challenge as you say. Those for my Sharp Stewart prototype were tapered in thickness - as they wrapped around the compound curvature of the pressed steel door. A note on the makers drawings said, 'grind to fit'.
Furthermore, the hinge arms are not equi-spaced top and bottom from the door's centre - so mirroring one for the other doesn't fit.

-Brian McK.

 

[Big Train James]

Responding on my phone so there's bound to be some auto correct gems....

That being said...AutoCAD has no tapered fillet function.pretty much you're stuck with a work around loft profiles. it can be done.

Not sure which surface exactly you want to extrude but it won't work for warped planes. Again lofting is probably the solution. Also sometimes you can press/pull instead of extrude. Don't know why one works and the other doesn't.

Might try intersection on solids to get some of the saddle forms that you need.

Also consider working with surfaces instead of solids. You can convert them to solids if they form a closed object.

I'll think about it some more and check back tomorrow when I get home.

 

[Big Train James]

Also consider rhino instead. Functions a lot like cad but much more robust and powerful on complex curves. And it supports variable radius fillets.

 

[SimonT]

Mick,
my method for the smokebox door in Rhino would be:

1. Draw a half section of the door. Save a copy of the section
   
2. Revolve the section to produce the door. Switch off this door.
   
3. Take the copy of the door section and offset the curve by the depth of the straps. Connect the two straight parts of the section.
4. Revolve this section; this will become the straps.
5. On a construction plane normal to the door, draw the outline of the straps with three external sides that are bigger than the strap object to form a square with strap openings. Extrude this to form an open cube that is bigger than the strap.
6. Bollean subtract the cube from the strap object to leave two straps.
7. The hinge and pin is a vertical section and a revolve.
8. Filets onto the pin/strap are a construction plane set square and mid level of a strap with a plan section fillet object extruded and then mirrored down onto the other strap. Boolean unite the door, straps, fillets and hinge.

For straps that aren't a constant thickness from the door extrude strap shapes from the back of the door. On a construction plane that is set square above the straps draw the horizontal profile of the straps set into a square, extrude this down to give a cutting object to the front of the straps and then boolean subtract this from the straps.

Brian's method for the chimney is basically how Dave Finney taught me to do chimneys and domes. Dave produced some very tasty ones.

Simon

 

[Steph Dale]

I think we're just exposing differences in packages. I did a smokebox door almost exactly as Simon describes - it's the same way I'd produce a hard pattern in most cases.

But I can't do a quarter chimney base in one shot as it just looks awful - and that's despite TCW having the same engine as Solidworks. 10-12 profiles seems to get a reasonably smooth shape though.

I can do variable radius fillets too, the problem with a chimney or dome is that the base flare tends to be based on an elliptical profile...

Oh well, it may just have to be a job I have to put aside some time for.

Steph

 

[Brian McKenzie]

But I can't do a quarter chimney base in one shot as it just looks awful

Does your package offer 'Start and End Constraints' options for Lofts? I can select 'Normal to Profile' - which puts curvature into the loft - such that the ends are perpendicular off the loft profiles.

-Brian McK.

 

[Dan Randall]

Mick,
my method for the smokebox door in Rhino would be:

1. Draw a half section of the door. Save a copy of the section
2. Revolve the section to produce the door. Switch off this door.
3. Take the copy of the door section and offset the curve by the depth of the straps. Connect the two straight parts of the section.
4. Revolve this section; this will become the straps.
5. On a construction plane normal to the door, draw the outline of the straps with three external sides that are bigger than the strap object to form a square with strap openings. Extrude this to form an open cube that is bigger than the strap.
6. Bollean subtract the cube from the strap object to leave two straps.
7. The hinge and pin is a vertical section and a revolve.
8. Filets onto the pin/strap are a construction plane set square and mid level of a strap with a plan section fillet object extruded and then mirrored down onto the other strap. Boolean unite the door, straps, fillets and hinge.

For straps that aren't a constant thickness from the door extrude strap shapes from the back of the door. On a construction plane that is set square above the straps draw the horizontal profile of the straps set into a square, extrude this down to give a cutting object to the front of the straps and then boolean subtract this from the straps.

Brian's method for the chimney is basically how Dave Finney taught me to do chimneys and domes. Dave produced some very tasty ones.

Simon

That sounds pretty much the same way I'd approach this problem with TurboCAD, though I'm not sure I could have put it into words so well!


Regards

Dan

 

[JimG]

The point being I can / could render, but nine - ten years later it ain't so easy and I'm rapidly coming to the conclusion that AutoCAD 3D is not the best platform for complex manifold shapes

Mick,

Which Autodesk product are you discussing - the "traditional" 3D version of AutoCAD that has been around for years, or a more recent product like Fusion 360?

Jim.

 

[Steph Dale]

Does your package offer 'Start and End Constraints' options for Lofts? I can select 'Normal to Profile' - which puts curvature into the loft - such that the ends are perpendicular off the loft profiles.

-Brian McK.

I'll get back to you on that, the same thought occurred to me first thing this morning - chores done, I'm now back at the m/c.

Steph

 

[Steph Dale]

And this what came up - first impressions are good, and it was fairly quick once I'd generated the sections reliably. However, as the top-down view shows it's not really useable. What a pain - back to lofting by sections I think:



Steph

 

[mickoo]

Jim,

It's Autocad 2016.

Better luck today, started with a new scene and kept things to simple views by clicking the hot spots on the cube, no rolling or flying and only one crash today.

Didn't even bother with the chimney, straight onto the GEVO side frames and they went much better, although a more complex shape is some respects it doesn't any or many compound curves as such. I'm finding AutoCAD works kind of in reverse, you don't necessarily make the shape you want, you make a overall shape and then use other primary shapes to take away bits until you end up with the right shape. In 3DSM you started with a primary shape which you grew into the shape you wanted.

First off this is what I call stand over scale...stand over there and it'll look fine ...as I've no accurate drawings, nor do I reliably trust the HO models I have (photos later), different trucks anyway. Most of the data is from the dozens of photos I took in Florida and thousands collected off the web; there are several variations in detail on the basic truck, this one being fitted to a ES40DC. Even then there are differences due to the build dates, but this is the basic raw frame from which the others can be cloned and detailed.


From below the cups for the axle box springs are easily seen, I've a plan for those later. Basically the axle box is an etched assembly with a hole in it which will sit over the axle end and allows for rotating axle box covers . The springs sit in the axle box cups and are a loose fit up inside the frame which will allow the whole lot to move with the axle springs (CSB)

Two large spigots attach to the 'model' truck frames and the centre rear section with the traction motor suspension bracket is still missing. I don't think it'll cast like that so in the final version I suspect the whole rear turned in end and cross bolster will be one part and the two side frames separate parts, the large spigots may well also need slicing off and being made as individual bits. Luckily there's two lifting holes which will pass right through the spigots and bolt to the model frame inside.


From above more detail is shown, ideally I'd like to split the mold right down the frames along the rear face but there's a couple or big openings in the top which cannot be simply drilled out due to their shape, there are round holds which will just have an indent to centre the drill bit but the two big ones by the brake lever brackets will be a problem.

If I calve off the fixing spigots then the pads for the secondary suspension rubbers will be an issue as they lie directly over the natural joint, they may already be too far in board anyway, I'll know more when I work up the chassis and then insert the rubbers to see how it looks visually.

In reality the box section between the axles is hollow, I could replicate that with a big hollow inside and then simply make a rebate around the opening at the bottom to stick in a plastic card or brass cover sheet, that'd give the lifting holes and brake bracket openings their prototypical flange appearance.

Other details like holes to mount the brake cylinders and anti hunting dampers still need addressing, again just small dimples for drill will be sufficient.


A scene with the basic AutoCAD lighting, pretty bland, shadows are false and too dark and highlights just way too shiny, basically no occlusion lighting.

I need to work with the filleting tool, it works but only in a certain way and you have to follow a certain path for it to work on multi edges where they meet.

The other option to casting is of course just get them done as 3D prints, not sure if 3D printing can do hollows inside the box section so still might need to leave a clear base with ledge for blanking plate; I'm also not quite sure how much they'd cost.

Right, back to looking at calving the rear bolster off and making the motor mount section and perhaps if there's still time, calving off the mounting spigots.

MD

 

[Big Train James]

Couple of things....

For the compound curves of the door hinge straps, I would do as Simon says but change #5 and #6 to extruding the positive geometry of the strap and using the intersection function instead of subtract. But the net result will be the same either way, which is a strap of appropriate thickness bending in more than one dimension.

My other suggestion if you are going to keep using autocad is to stop using the view cube at all. I turn mine off. But then I'm all about short key command aliases and my eye starts twitching and I break out in hives if I have to use any command from a toolbar. Too slow! I never get the right part of the view cube when I have used it.

I typically real-time zoom with the scroll wheel, and pan and orbit by holding the left side button on my mouse either by itself of in combination with the shift key. Normally it would be the middle button (scroll wheel button) but I have a mouse with side buttons so I reassigned the function. This way I don't accidentally zoom (scroll) while pushing down the scroll wheel. I'm surprised you don't use this method already as it is fairly common in most 3d modeling programs I have used, including Sketchup, Maya, Rhino, Revit, and Autocad. The key is remembering what the shift/ctrl/button combos are for each program when you go back and forth.

If you instead mean setting views from top to side or isometric, I either do that from the view tab on the tool palette "home" tab, or the equivalent drop down menu that appears in the top left corner of the work space when the mouse is hovered in that area.



If you need to change the UCS so that the Z axis is properly oriented to draw circles, mirror objects, rotate, or do a host of other commands, then I typically either use the drop down menu from the "coordinates" tab of the "home" tool palette, or typed commands. In certain instances I will use some of the tool buttons to set specify local Z axis orientation at odd angles.

Lastly, changing visual styles from wire frame to shaded and so on can be done from the "view" tab on the "home" tool palette, or from a drop down in the upper left corner of the model space window that appears when the mouse hovers over it. It would be next to the view menu pictured above.

For some of the above, you can customize the .pgp file with your own or additional command aliases so that they can be achieved mostly with keystrokes and minimum mouse movement. But sometimes it doesn't save any time.

I'm still thinking about best methods to achieve the stack to firebox form. I'm confident it will require lofting. Maybe creating some solid forms via intersection or subtraction, with resulting edges or faces extracted from the solids. The problem with most compound curves is autocad will define them as splines, so most of the normal object snaps won't work with them. But intersection (of lines, not solids) will typically work. Some construction line work would need to be created, possibly some ucs manipulation, and so on to get a series of good profiles for lofting.

Hope this helps in some manner, and actually addresses some of your issues. If not, ask away and I will do my best to answer. I've used cad for a long time, and have been doing 3d with it for about three years now. I know I should use something else like Rhino or Solidworks as they are more common in industry for 3d solid modeling, but I am proficient with autocad already. I don't know everything, but I have discovered a lot trying to solve some of my projects. You can see some of what I've achieved if you look back through my sw1500 thread. I think I've posted a couple of my bogie drawings there, and possibly my jacking pad model.

Jim

 

[Brian McKenzie]

And this what came up . . . . . the top-down view shows it's not really useable.
View attachment 77170

Steph

You are nearly there, Steph. If you have the facility to add Guide Curves in association with your Loft profiles, these might eliminate your current cam-like lobes.

You could also take a look at recent Onshape video clip tutorials, where their 'Lofting' has been much improved, and B1 chimneys could be made for free.

-Brian McK.

 

[Steph Dale]

We have lift-off:
There's a slightly radical approach to drafting, I'll admit - forming the curved sections for the fillet uses a geometrical rule to guarantee consistency, but it's done. In the end the method is fairly simple, but I suspect I'll still be allowing an hour or two to do it.


So there we have it. Actually I'm quite impressed with the result; the resulting shape is as consistent as I can make out. I suspect that in anger I'll use rather more profiles to get it absolutely right, but as proof of concept goes I'm happy.

I'm also pleased to note that the learning curve is approachable - this is still less than the tenth item I've drawn in 3D CAD...

Steph

 

[mickoo]

James,

Cheers, my sum total 3D cad in Autocad now amounts to 33 hours in two days, so I don't think I'm doing too bad , 3DSM it was five or so years with three or so in Gmax before and then before that some 3rd party software that evades my memory but it wasn't a bespoke CAD package; the hard part is trying to forget all those old gizmos and learn new ones.

Didn't know about the left and right mouse functions for roll around, I do use the mouse wheel for active zoom though. I do use some text commands but mostly in 2D. My biggest bug bear at the moment is the roll around or fly around, it always centres on the scene and not where the mouse is, so if you've zoomed in at one end of the truck and then roll it around it flips right off the screen, in 3DSM it rolled around where ever you put the cursor so the item you were interested in stayed mid screen.

Autocad has so many other written menus that there are no icons or gizmos for, they're there in the main menu bar at the top if you go looking for them but no gizmos and don't hit F12 by mistake and then spend the next 20 mins screaming wheres my mouse pointer roll out gone! New scenes and re booting AutoCAD don't help either!

The door and straps, well thanks for y'all explanations but I was lost after 1.

Really and seriously you may as well speak Mandarin, just gobbledegook I'm afraid, which is what makes it hard to google stuff, "why doesn't the thingamajig do the flippy whizzy doo dah it's supposed to do" never gets many hits!

The visual styles work well, I tend to use Conceptual or Xray the most as they are the brightest.

So last thing for this evening, closed off the rear end and started on the rear motor mount, wasn't happy with the bulged frame up front so re did all that and as proof of concept opened up the inside of the rear box section; it'll certainly not cast like that as there's an internal dog leg section, but it should 3D print ok.
Also dropped off the fixing spigots and after further peering at photos worked out that the secondary rubber suspension pads could be moved back a bit. I've added a lug on the rear of the frames to locate in a cut out in the spigots which'll help locate it and give better mechanical grip for fixing.





There's a little more left to do on the motor mount, some dimples for drilling for the fitting mounts and calving off the rear bolster and it's pretty much done I think. A decision on whether to leave the box section open or just seal it will be the last step; I've only done the rear part at the moment and will do the front tomorrow. It might pay to try both and see what it looks like.

MD