PartDesign Bearingholder Tutorial I/de



Wie der Warnhinweis oben auf der Seite bereits andeutet, funktioniert dieses Tutorium NICHT, es sei denn, du kompilierst einen speziellen, sehr experimentellen Zweig aus dem FreeCAD Quellcode und ist ein einführendes Tutorium zur Modellierung mit dem PartDesign Arbeitsbereich in FreeCAD unter Verwendung von Datumsebenen, die in den meisten FreeCAD Versionen noch nicht vorhanden sind. Der Zweck des Tutoriums ist es, dich in zwei verschiedene Arbeitsabläufe zur Erstellung eines Gussteils mit Entwürfen und Verrundungen einzuführen. Je nachdem, welche anderen CAD Programme du bisher verwendet hast, könnte dir das eine oder das andere vertraut sein. Als Arbeitsbeispiel werden wir ein einfaches Lagergehäuse modellieren.

Dies ist der erste Teil des Tutoriums. Es wird der so genannte "Ein Körper" Arbeitsablauf verwendet, wobei der (einfachere) obere Teil des Gehäuses als Beispiel dient.

Um dieses Tutorium durcharbeiten zu können musst du offensichtlich den PartDesign Arbeitsbereich aktivieren.

Du findest meine Version des von diesem Tutorium erstellten Teils [hier]. Die Datei ist nicht mehr verfügbar, eine neue wird zu einem späteren Zeitpunkt bereitgestellt werden.

Konstruktionsdaten
Das Gehäuse sollte in der Lage sein, ein Lager mit einem Durchmesser von 90 mm und einer Breite von bis zu 33 mm aufzunehmen (z.B. DIN 630 Typ 2308). Das Lager erfordert eine Schulterhöhe von mindestens 4,5 mm im Halter (und auf der Welle). Der obere Teil des Halters wird mit zwei 12mm Schrauben an der Unterseite verschraubt. Auf beiden Seiten des Lagers sollte eine Nut vorhanden sein, die einen Standard Wellendichtring DIN 3760 aufnehmen kann: 38x55x7 oder 40x55x7 auf der einen Seite, 50x68x8 auf der anderen Seite.

Das Gehäuse wird ein Sandguss mit einer Mindestwandstärke von 5 mm, einem Entformungswinkel von 2 Grad und einem minimalen Verrundungsradius von 3 mm sein.

Einrichten der Skelettgeometrie


.

.

.

.

.

.

.

.

.

.

.

.

After you have created the sketch, pocket it up to the datum plane marking the bearing side plus 5mm wall thickness. If you ever want to modify the holder to be able to hold wider bearings, all you have to do is to change the dimension of these datum planes, and the cut-out depth will follow along.

To reduce the amount of machining required, we also want to cut away some material inside the holder. Again, duplicating the sketch of the first pad is convenient. It doesn't even have to be remapped. Again, the only important dimensions are the machining allowance (3mm) and the outer diameters: 84mm for the place where the bearing will be (90mm - 2x machining allowance), 49mm for the smaller sealing ring (55mm - 2x 3mm) and 62mm for the larger sealing ring.

After creating the sketches, pocket them: Symetrically 28mm for the bearing cut-out (34mm - 2x machining allowance) and one-sided 23mm for the cut-outs for the sealing rings: 34mm / 2 for half the bearing width + 9mm for the sealing rings - 3mm machining allowance.

Your part should now look like the picture on the right. Note how the different cut-aways combine to create an almost uniform wall thickness, which will make the casting easier and less liable to have pores.

Now all that remains is to create some material for the bolts to go through. You might be tempted to sketch these as a circle and then pad them, but this will head you for trouble when you try to put the draft onto them later (I assume that is a weakness of OpenCascade). So to circumvent the problems, it is better to create a sketch with the draft angle included and then rotate it through 360 degrees.

Here again the skeleton planes come in useful. You will need the bolt axis plane and the bolt head plane as external geometry. Then, create a straight line for the rotation axis and make sure it is constrained to the bolt axis plane reference. Toggle it to be construction geometry. Then, sketch the rest of the contour. The important dimensions are the machining allowance at the top and bottom and the radius of 12mm: 7mm for the hole radius + 5mm wall thickness.

Create a revolution feature from the sketch and then mirror it on the YZ-plane. This is all the solid geometry we need to model. The rest is draft and fillets.

Applying draft to the side faces
The next step is to apply drafts on all faces. Its important to consider the location of the neutral plane, that is, the plane which the face is "rotated" around. If we choose as neutral plane the bottom of the holder, then we will have a problem with the wall thickness in the top part of the holder. Therefore, we create a datum plane at an offset of 40mm from the XZ plane as a compromise between the top of the holder becoming to thin and the bottom becoming to wide.

To put draft on a face, select this face and create the draft feature. You can then select more faces to apply the draft on. If you have a large part, it is advisable to draft only one face at a time. This means that if you change the geometry and a draft fails, only this one feature will fail, whereas if you put all faces in one draft feature, then the whole feature might fail because of one face failing. For a small part like the bearing holder, its sufficient to create two draft features: One for the four outside faces, and one for the inside faces.

The dialog will force you to select a neutral plane before completing. You can leave the pull direction empty, in this case it will be normal to the neutral plane. Don't forget to set the draft angle to 2 degrees.

Filleting the holder
We can now fillet the part. The picture shows the first set of fillets. Start with the small circular fillets and make them 4mm radius. Even though 3mm would be enough as per specification of the part, a radius of 4mm means that after machining 1mm of the fillet is left, reducing the sharp edge produced by the machining. The large fillets are 6mm radius to help spread the force from the bolts into the rest of the part. It would be nice to make this radius even larger, but unfortunately OpenCascade can't handle overlapping fillets yet.

As with drafts, in a complex part you should fillet only one edge at a time to avoid unnecessary failures if the base geometry changes.

The rest of the fillets are simply 3mm radius. Looking at the picture on the right, the two highlighted fillets could actually be filleted with 5mm to achieve a more uniform wall thickness for the casting. After machining, the minimum wall thickness of 5mm would still be maintained. But again the fact that OpenCascade can't handle overlapping fillets prevents us from doing this for the inner of the two highlighted fillets.

Filleting the inside of the part presents us with a difficulty that cannot be solved with the current tools in the PartDesign workbench. The highlighted edge cannot be filleted at all, again because the rounds would overlap. This could be worked around by creating a sweep instead of a fillet, except that sweeps are not implemented in PartDesign yet. For the time being, we are forced to leave the edge as it is.

The picture on the right shows the finished part in the state it will be before machining (except for the one edge that is impossible to fillet). You will notice that one edge that runs around the whole part has been left unfilleted on purpose. This is the edge where the bottom and the top of the mould meet. Here, no fillet is possible (and none is required anyway).

Machining
Now we can cut away the material that will be machined off the raw cast part. This is very easy with the skeleton geometry defined. The idea is to create all machining features (Pockets and Grooves) using datum features only. This means they will be totally independent of the solid geometry of the bearing holder, which gives us some big advantages:
 * No matter how you change the solid geometry, the features for the machining can never fail.
 * You can create the machining geometry before finalizing the solid, which gives you useful visual feedback.
 * If you move the skeleton datum planes, then both the solid geometry and the machining will adapt automatically.
 * If you make a mistake in your solid geometry, the machining will still be in the correct position, and very likely the mistake will become glaringly obvious (e.g. a wall thickness becoming 2mm instead of 5mm). Whereas if you reference the machining to the solid geometry, it will adapt to the error in the solid and e.g. maintain the 5mm wall thickness, just in the same wrong location as the solid is.

Before starting on the machining geometry, I like to place a datum point in the tree and name it something like "Machining_starts_here". This is useful if you want to switch between the raw and the machined state of the part because you can see at a glance where to move the insert point to get the raw state.

To machine the bottom of the holder, just sketch a large rectangle on the XZ plane and pocket it. For the top, sketch a circle on the datum plane defining the bolt head location, and then mirror the pocket on the YZ plane. In the same way, create a pocket for the hole which the bolt will go through and mirror it. To machine the inside of the holder, create a sketch on the YZ plane and groove it.

Once you have done the machining, you can have a nice visual effect by colouring all the machined faces so that you can see at one glance which parts are raw casting and which are machined after casting.

Final notes
We have modelled the bearing holder top with the dimensions it will have after casting. To create the casting mould, you need to apply shrinkage to your part because after casting, when the hot metal cools down, it will shrink by a few percent (depending on the material). Usually it is best to leave the application of shrinkage to the foundry making the part because they have the required special knowledge. They should also tell you if your part has problematic areas, e.g. very thick walls suddenly joining to very thin sections without a properly tapered section between them.

Part Two
PartDesign Bearingholder Tutorial II