Had Another Go

[lee websterParticipant @leewebster72680]

Hi Nigel.

I haven’t posted a comment for a while, but this thread has interested me. Learning to use the sweep tool is a good thing. But, for someting like the simple frame you are trying to draw, why not just creat a sketch on the XY plane and then extrude it into a solid? Sweep takes two sketches on two planes, whereas drawing the frame on the XY plane is one sketch on one plane. Plus of course you would need to use the shell tool if Alibre has one to create the U shape.

If you draw the frame as above can you mirror the sketch to create the othe frame rail? I don’t use Alibre Atom because it requires internet connection.

[David JuppParticipant @davidjupp51506]

Lee,

Atom3D only requires internet connection to validate the licence every 30 days.  Internet is not required during operation.

Yes Atom3D does have a Shell tool.  Nigel has a chassis rail example done using shell.

[Nigel Graham 2Participant @nigelgraham2]

Until someone suggested using “Offset” to reduce the number of steps a bit, I just used rectangles, lines and trimming to form the basic ‘C’. Not the Shell tool – but doesn’t that leave the far end of the profile closed, like the bottom of a box?

I’m a bit puzzled by what Lee means though. The chassis is made up of channel sections not solid bars, and is not a simple rectangle in plan, so I can’t see how it can be extruded in one go either lengthways or top-down as an extrusion.

[Nigel Graham 2]

On Nigel Graham 2 Said:

the Shell tool – but doesn’t that leave the far end of the profile closed, like the bottom of a box?

Depends which face(s) you add to the Remove Faces list.  By default Shell doesn’t remove any faces from the model, it just hollows out the inside.

With some thought about faces to remove, and perhaps thickness overrides, the Shell tool can be a shortcut to modelling some shapes.

It’s very rare for there to be only one single way to model a particular part.  Different users may find very different approaches to make most sense.

[Nigel Graham 2Participant @nigelgraham2]

I was trying to think what I’d used Shell for, but eventually remembered it was for the wagon’s coal bunkers, open-topped boxes fabricated from steel sheet. Not its chassis.

[David JuppParticipant @davidjupp51506]

Error correction time:

I stated previously that in a Sweep, the position of the Path sketch doesn’t matter.  Jason did some tests which showed I was wrong.

I had always believed that the path is followed even if it is offset from the profile.  Turns out the finished sweep actually sits parallel to the path sketch, so the relative lengths of portions of the result can be affected by the position of the path in some cases.

So not only for understandability, but also for accuracy the path should coincide with the profile – you’ll have to decide if touching the profile is OK, or if the path should run through the centreline of the profile.

Apologies for any confusion caused.  I’m checking with Alibre QA if this is working as intended, and if the Help file needs a little more adding to cover this.

[Nigel Graham 2Participant @nigelgraham2]

I had thought the sweep was parallel to the path, so the same layout and length, but displaced.

I’d read the manual as saying the sketch does not need be on the path, which seemed strange but I took it to mean the calculations would place it on the path. Only how would it be centred? On the sketches’ axis of symmetry, on one corner if a polygon, on some arbitrary point on the circumference of a circle?

Tht may not matter if the object is floating about as a Part, I suppose.

[David JuppParticipant @davidjupp51506]

Nigel,

The sweep feature always goes through the profile sketch.

The path that the sweep feature follows is parallel to the path sketch, and anchored to the profile sketch.

If the path sketch is offset from the profile sketch, the parallel application of the path sketch can result in some lengths changing from those in the path sketch, when the feature is generated.  It’s generally not a big effect, but can make a difference.

For your chassis rail, it can result in the angled portion getting moved along the rail slightly.

[JasonBModerator @jasonb]

These are the couple of sweeps that I sent David with my Query.

In both cases the lower square has the path against it’s edge. However the upper square has the path sketch offset. Looking from above shows that the upper square remains a coonstant distance from the path so in teh case of teh angled one the main leg of teh chassis would be longer and the front section shorter. In the case of teh bent one then the vbend radii change as they are equally spaced from a springing point further away so the radius alters.

 

 

There are times when you want it parallel to the path such as this bead running around the opening but set 0.5mm from the edge which was the path.

[Nigel Graham 2Participant @nigelgraham2]

Oh Lor’! Gets worse.

I thought I was beginning to understand how to use that Sweep tool but you have just warned me it holds a confusing trap hard to avoid, making it more difficult still.

I’ll have to plan any models either to use a work-round in the CAD models (simplifying or omitting such awkward shapes), or by designing the project itself without them physically.

.

Having investigated the possibility, I am tempted to overcome all this palaver by scrapping the chassis I have made and replacing it with one with fully-parallel sides. The narrowing for the steering clearance is then by making each side rail as two straight lengths bolted together, overlapped, with a plate spacer. The displacement each side is 1-1/4″, so the spacer is only 1/4″ thick . No awkward angles, easier machinery mounting, more room for it. The problem apart from material cost and work time, is losing space needed for the steering-gear.

Ironically this CAD model is supposed to help me work out how to install the engine and road-gears in a rather cramped space.

 

I’ve also another design task. I have lost the long gib-head key between the rear wheel and axle, a rather awkward one because the axle and hub keyway widths differ slightly, probably <0.01″ total error. So back to the books to determine the correct dimensions then make a new one, 1-in-100 taper and all, but of dual width.

[JasonBModerator @jasonb]

There is no trap. If done how I first suggested then it will work fine and a bit of forethought will avoid any problems.

The only thing I would say is that I drew the path for the inside edge of the chassis to follow, if your dimensions were taken from the outside edge then that is where the path and C section sketches should meet. To me it made sense to use the inside edge as that is what all the other main cross rails fit to.

Don’t give up on sweep as it is a feature that can be used a lot and giving up will restrict you.

[Nick WheelerParticipant @nickwheeler]

It’s only a trap if you blunder into it unprepared. Being prepared enables you to swerve around and carry on unaffected.

 

To be prepared requires more practice by repeating exercises different to the couple you’ve recently accomplished. Doing this is a long established learning and teaching method. Moving on before you’ve mastered, or at least understood the fundamentals, is deliberately setting yourself up for disappointment.

Jason’s new examples demonstrate how more sophisticated techniques can make complex parts easier.

I designed my hotrod chassis rails using the external dimensions as they need to fit inside the body that I already have. When I actually build the thing, I’ll pull the internal dimensions as they’ll be easier to measure on the build table. And when I have the flat parts laser cut, I’ll have some templates cut for the angles as the chassis taper turns out to be a hard to measure 7.9°

[Nigel Graham 2Participant @nigelgraham2]

Well, that “blunder into it” is the point. I’d not even known this can happen until Jason described it, and it seems very easy to forget these subtle points, be confused and make a mess of things as a result.

I have tried setting up that frame in 3D many times by now by different ways and it either just goes wrong or sort of “works” but in a very clumsy and inaccurate way.

I was using the inside dimensions all along, but because the frame tapers, I found I could not constrain those two cross-members accurately to both sides in their right places.

…….

Several years ago I fabricated two hexagonal frames from black (well, very rust-brown) steel bar, about three feet diameter. They didn’t need very close accuracy, but I still wanted reasonably good hexagons that would screw to other components correctly.

The very “pre-loved” steel was too rough and the bars too long for reliable setting by combination-square.

So I made each as three pairs, aligned to pencil lines set out on a sheet of 12mm plywood by standard geometrical construction methods and long beam-compass.

 

[Nigel Graham 2]

On Nigel Graham 2 Said:

 

I have tried setting up that frame in 3D many times by now by different ways and it either just goes wrong or sort of “works” but in a very clumsy and inaccurate way.

I was using the inside dimensions all along, but because the frame tapers, I found I could not constrain those two cross-members accurately to both sides in their right places.

The taper is a distraction. You know where it is, how long it is, and the widths either side of it; the angle is available on inspection but probably isn’t helpful. You construct and constrain the crossmembers using that known geometry. Project it into new construction sketches if necessary. That is another fundamental skill that would be a good face-to-face lesson.

 

The references for my entire chassis are the centre of the lower front edge of the rear crossmember and the axis perpendicular to it. That point is placed definitively – with a constraint – on the origin so it cannot move. I chose it because it will be easy to measure from when constructing the real part.

[Nigel Graham 2Participant @nigelgraham2]

Well, I decided trying to make that chassis into an exercise in using Atom as Alibre’s programmers intend, was going nowhere.

I’ve not yet tried to model keyways part-way along shafts though David did try to describe to me. Every time I tried the extruded cavity just evaporated.

However I had a go at partially erecting the engine, concentrating on just one half, and adding three “temporary” parts modelled simply to hold things together and assess how to squeeze the rest in. The temporary parts are the base-plate (with a big chunk cut out of the end so I could see where the big-ends go), back wall (upside-down!) and the cross-bar holding the guide-bar to the wall.

The base-plate is already made but needs modifying, as I made it before really knowing how the crankshaft was going to look and where it will sit in the box. The conencting-rod shown is a simplified version. My first try had it with its bearing-brasses, but when I tried turning the shaft their effect was to throw everything of alignment. So I made a special copy of the rod with no liners.

This image is hours old and a complete second-attempt, with forever running into “over-constrained” errors as the main and most frustrating problem, and having to cope with the crankshaft axis not on the primary axes. I’ve turned off (by the eye sign) many of the planes I concocted as they were covering the image with foliage.

 

[JasonBModerator @jasonb]

There is very little you need in the way of measurements for the cross members, just the cross section and their position along the chassis.

You can use the CAD to give you the length by using the geometry of the main rails, these will also give you the angle of the ends. I still don’t know the angle but the two fit together perfectly in the right position. I could pull the angle off the model or do a 2D drawing with it on.

I did it by constructing the new cross rail within the assembly but could also have put a couple of guide lines in to get the length and angle so that I could go and model it in it’s own screen. One sketch for the “C” and I also did an endplate.

It is saved as a separate part

[JasonBModerator @jasonb]

Is that lower half of the engine done as a part or an assembly? Would need to see the file to get the crankshaft where it should be as you probably have some constraints that are stopping it from being positioned.

Also way too many planes if an assembly.

[David JuppParticipant @davidjupp51506]

There’s no need to cut chunks out to see where things go – use section views to do that, or temporarily hide parts.

As for cuts that ‘evaporate’ – save the file (with the problem) and let someone take a look.  The cause is probably straightforward once spotted.

[JasonBModerator @jasonb]

If Atom has it you can also got to the colour of an individual part and reduce its opacity so it becomes semi transparent which lets you see inside. Though section is often better as you can get the mouse in to click on internal parts which you can’t do with a transparent part.

For me cuts that disappear are usually due to me clicking a solid extrude when I meant to click cut extrude so it does nothing to the part.

[Nigel Graham 2Participant @nigelgraham2]

The engine model is all-Assembly. The largest single Part is the crankshaft, but I’d not noticed it is away from the main axis until far too late – when I had put everything else together and was adding the base and wall.

On that I placed one pair of eccentrics, one sheave (simplified to a plain disc), the two outside main bearings (ball-races simplified to plain cylinders) and oil-seals, and crosshead.

Those Parts and the base-plate are all ready-made in reality, but some may need modifying.

 

I’d also added the connecting-rod, as a new design not yet made physically. At 6 inch centres it threatens to make the entire engine too high. The whole thing is not supposed to be more than 16 inches tall, 17 at most; and that back-plate and guide-bar top are already 12″ above the underside of the base.

I’ve made a pair of connecting-rods at 4-inch centres but the resulting centres / stroke ratio of only 2 is too small. It should be at least 3. They would also need a longer piston-rod to stretch the engine enough to fit the valve-gear, which I’ve not yet tried to design (adapted from a published design).

.

The guide-bar is shown in temporary or unfinished form. I created that, the back wall and guide-bar support purely as representative to help build up the CAD model.

.

I’ve just looked to see if Atom has opacity as well as colour controls. It does but I’d not spotted that at the time.

All those planes were the only way I could see to try to align everything hopefully in the right places, and centrally on them; but there are still location errors somewhere in there. The bearings and possibly the whole crankshaft are displaced slightly towards one end, and probably because I could not constrain the bearings to their tiny location steps on the shaft. (The eccentrics are bored to 0.80″, the bearings are of 0.750″ bore.) The model’s probably full of needless, wrong or missing planes and constraints.

E.S. Hindley & Sons’ design department never had these problems….

[Charles LamontParticipant @charleslamont71117]

Nigel, it is good to see you are making progress. I just tried out the offset sweep thing in Onshape and this may make it clearer how it works.

[JasonBModerator @jasonb]

Never too late to alter things but best done as soon as possible.

David and I have said about right clicking a part and selecting “show geometry” this will give the main planes and axis that were used to model the part.

So in the case of your crankshaft it’s own axis could have been aligned with one of the main axis on the assembly screen. This will then allow it to be rotated about that axis so you can watch the conrods move and the cross heads go up and down.

You do not generally need to add planes to position the various parts as they are inserted. You should be using an offset constraint or the above mentioned geometry. So for that back plate you should select the Horizontal plane and the top of the plate and entered a distance. Again as your crankshaft is not central to the engine you could have selected the vertical plane and the end of the crankshaft and entered a offset distance.

If something is “small” and you are having difficulty selecting it then just zoom in to make it larger. It is something you have to do a lot.

If you want to send me the individual part files for those items and the cylinder I will assemble them so you have an example to look at, right clicking any constraint down the left will show what I did.

 

This is an example of a transparent crankcase so I can see what is going on inside.

 

[Nigel Graham 2Participant @nigelgraham2]

Thankyou –

Charles,

but a) I don”t use Onshape though I expect the Sweep tool is broadly similar if it does the same thing; and b) I can’t view that attached image anyway.

 

Jason –

Well, I did say I’d not aligned the crankshaft with the X axis but had not spotted that until I’d nearly finished. Those planes at the end were my attempt to stop the shaft slithering lengthways. It seemed to work.

I’d also tried to set the shaft symmetrically to the vertical plane, as that seemed logical at the time.

I set the base first, with a lot of trial and error, then constrained the back plate to that. The extra horizontal plane on its top locates the top of the guide-bar.  The base, back and cross-bar were the last three pieces I added, and had to model them as Parts first. Directly as trying to use the New Part way in Assembly only confused me, as I found when sometimes accidentally selecting that instead of the New Design tool next to it.

 

I did later examine the Colour menu to find the transparency setting. Now I need go back to the Parts folder to see if used a cut-away copy of the base or had cut the original!

Very kind invitation but do you really want to spend all that time re-creating that engine Assembly, something like twenty Parts?  I know it’d take you well under an hour but it’s still a fairly involved model. It was something like five hours for me, perhaps more. Incidentally I tried to align the eccentrics with their keyways by the groove edges, but they were already too constrained.  (The keyways should not go to the end of the shaft either, but I could not show them properly).

[JasonBModerator @jasonb]

It should not take long, though may depend on what hidden things you have in the files to complicate things!

I thought your bores were offset in the cylinder that I why I said the crank shaft is not likely to be symmetrical to the mid plane.

[Nigel Graham 2Participant @nigelgraham2]

I’ve just re-visited the model to correct the mis-alignment, then plonked the block on top.

The bores are slightly offset to be able to disguise the engine as a compound, and I’ll compensate for that in the block and valve-chests. The crankshaft including eccentric centres is symmetrical and can be placed centrally in the engine.

[Author]

Posts