PDA

View Full Version : Where to start wheel model?



unstable
07-07-2014, 06:14 PM
I wanted to model this wheel, but I quickly realized that I have no idea where to start with the inner disks. Every time I try to drill an off-center hole in a disk and add the wires to keep the inner hole round, I end up destroying the roundness of the outer disk. Would someone mind sharing where I should start?

erikals
07-07-2014, 07:22 PM
maybe this >
http://www.trhlogos.dk/main.html

go to "tutorials"

jeric_synergy
07-07-2014, 07:31 PM
IMO the easy way to start these kinds of objects is to start with a WEDGE, not a circle.

Do ONE wedge, like a pie, with a hole in it, then CLONE ARRAY it around.
+++++

::whew::, Well, it's ugly, but it's "done". All over but the weeping. Started with a 1/5 of a circle. -- I'mma not gonna tell you how long this [email protected]@rd took....
122842

122843

unstable
07-07-2014, 09:20 PM
-- I'mma not gonna tell you how long this [email protected]@rd took....
122842

122843
LOL... Wow, I never would have thought of doing all those ploys within the wedge. I did consider the wedge solution, but wasn't thinking of that design. That's amazing! If I wanted to model the inner sprocket, there are nearly 100 holes or buttons. I don't think I can use this same approach. I'll try to replicate your solution. Thanks

jeric_synergy
07-07-2014, 09:42 PM
Well, it's certainly easier to get ONE wedge correct, than five. Actually, you only need, and it's better I JUST FOUND OUT :cry: , to model one TENTH of the wheel-- although that's not practical because it's too difficult. But at some point you'll want to slice your wedge in half and mirror it, so that when you array it around the edge points match up.

I tried to use AXIS SCALE to mirror it, but that did not fly, wah. AXIS-MIRROR would be nice, don't tell me we already have it?

AND, check THIS out!!! All Quads!!!! WoooHOOO!!! --One thing I've noticed is that as you work on making these subd objects, suddenly you'll see a path to the holy state of All Quaddom. I didn't really think I'd get there, but WALLAGH!
122844

Here's the object so you can see my shame as archived in the layers...
122845

erikals
07-08-2014, 12:14 AM
ah,,! sorry my eyes crossed...

never mind...

jeric_synergy
07-08-2014, 01:22 AM
122846

JoePoe
07-08-2014, 09:54 AM
Well Jeric, I admire the puzzle work! :)

Just wanted to show a simpler solution as another option..... 122854

To the OP:
Don't let the inner sprocket intimidate you. If you can do one circle you can do a thousand. Array/clone is your friend. :thumbsup:
Edit: I personally would try to leave a "clean" section in between each circle just so you don't have a six star after a six star after a six star....

jeric_synergy
07-08-2014, 10:24 AM
Yeah, but I wanted a smooth edge in poly mode on the outside, so I think I have roughly 2x as many perimeter divisions. Probably not smart.

JoePoe
07-08-2014, 11:21 AM
Yeah, but I wanted a smooth edge in poly mode on the outside, so I think I have roughly 2x as many perimeter divisions. Probably not smart.

Interesting, cuz clearly you're in SubD. Hence the "All Quad" eureka! :D

But that makes me realize... nobody even asked if the OP wanted SubD or not. :foreheads :ohmy:

unstable
07-08-2014, 01:04 PM
Thanks JoePoe, your solution is more what I was expecting. I was actually thinking of making an inner ring with holes like you have, but then extruding the points outward and inward. But when I tried this, I couldn't get a very good circle. I was planning to do this in subD. I thought it would make for a better model, but it really doesn't matter. I just wanted to give this a shot. I haven't had time to give it much of an attempt yet. I think I have an idea for tackling the inner socket with all the holes. I think I'll combine both your solutions to start with and see where that gets me. It may be the weekend before I have any results, but I'll try and post if its worth it. Thanks folks.

JoePoe
07-08-2014, 04:21 PM
Hmmm, just took a closer look at the inner socket.
Absolutely do a section and array.
I would do a section that includes both the inner and outer ring of holes. 122860

Arrayed (and a quick merge of tris). 122861

And finally. 122862

jeric_synergy
07-08-2014, 05:37 PM
Interesting, cuz clearly you're in SubD. Hence the "All Quad" eureka! :D
Well, I wasn't quite thinking that thru, was I? :grumpy: --- ;)

Once in SubD, it appears anything over 8 makes a dandy circle. Since I knew I was going to divide my starting disk into 5 parts, I think I just went for 10 divisions per wedge, total of 50. Yours has four, for a total of 20. Now I know better. FWIW, I started w/a disk.

It was certainly good practice, especially the bit about how INITIAL CONDITIONS can really affect the difficulty of the meshing process. I think this point has been neglected in the tutorials: choosing the wrong point# of disk can really bone you later in the modeling process.

Since HOLES are the major sticking points for most of this type of modeling, I'm surprised people don't just do a hole, arrange and copy as needed, and then mesh them together.

JoePoe
07-08-2014, 06:19 PM
It was certainly good practice, especially the bit about how INITIAL CONDITIONS can really affect the difficulty of the meshing process. I think this point has been neglected in the tutorials: choosing the wrong point# of disk can really bone you later in the modeling process.

Completely :agree:

My first disc (for the five hole example) had two sections for each hole and one spacer poly in between.... but that made 15 sides and that wasn't going to be nice in the middle. So I added another spacer to each section to get an even number... that's how I got to 20.


Since HOLES are the major sticking points for most of this type of modeling, I'm surprised people don't just do a hole, arrange and copy as needed, and then mesh them together.

That's exactly how I did the Inner Socket. Circular Arrayed a disc twice (inner and outer loops of holes) connected a small group of them up. Took the appropriate section of that and arrayed again. :hey:

unstable
07-13-2014, 10:44 AM
Okay. I'm close, but the rivits are a little too far apart. My problem is that if I bring them closer together, there are too many going around the wheel. There should be 30 in each row. If I make the rivits bigger so they are closer together, then the rivits are too big. I don't have any specifications for this wheel other than the outer rim diameter and the rim width. So I have estimated the silver disk to be 4.5 inches and the black inner disk to be 7 inches. Maybe I need to leave the rivits the same size and reduce the inner disk to 6 inches? I've spent hours on this so far. But I'm learning. Thanks to both of you. I've attached some shots to show my progress. I look forward to your feedback.

Surrealist.
07-13-2014, 02:15 PM
Perhaps just a matter of terminology but those are not rivets. Those are actually the spokes.

But I only bring it up because visually there is a subtle difference between what you are modeling and what is there in the real world.

123015

Here are some more extreme examples. Notice how there is depth and a sense of wiggle room for the spoke:

123017

123016

123018

Also it is not perfectly flat and the depth is at a slight angle because of the angle of the hub.

If you look closely at your photo you can see this on the right. Looking at the top and bottom the lighting and angle is good to see that it has more depth and is uneven.

And what fits in it looks something like this:

123019

The polyflow looks pretty good though at first glance.

XswampyX
07-13-2014, 02:55 PM
That's a great resource there Surrealist.

If you select the inner surface of the holes, and then use Modify -> Translate -> More -> Point normal move. You can 'resize' the holes without mucking up the mesh.

http://i465.photobucket.com/albums/rr16/xXswampyXx/Point_Normal_Move_zpsea2f1ea4.jpg (http://s465.photobucket.com/user/xXswampyXx/media/Point_Normal_Move_zpsea2f1ea4.jpg.html)

unstable
07-13-2014, 04:20 PM
Thanks Surrealist for the education and the additional reference material. You make a very good point. So after learning how to make it 'look' similar to my reference image. I should probably work on something closer to swampy's solution. That looks very good swampy and thanks for that bit on using transform. I did not know that. My latest solution is just kind of thrown together. I did resize the inner disk. After looking closely at my reference image on the right, I think it is closer to 6 inches. I also thought I saw the inner row of spokes slightly raised, but mine has a harder edge than theirs. The surfacing and such is just experimenting.

Feedback is welcome. I will probably restart this next weekend. :D

jeric_synergy
07-13-2014, 07:01 PM
Here's one thing: there's absolutely no reason to make the spoke-heads/rivets the same LW part as the hub-- after all, they aren't in RL, so....

If making them a part of the hub is a chore or complicates the design/math, you can skip it entirely. Just do the holes. (EDIT: and maybe not even them.)

++++
::waxing philosophic::
There's something I think is interesting in both JoePoe's and I's mesh: one bit of it is 5x symmetrical, and so is the other, but not on the same axis.

The outermost bit has an obvious repeating pattern, and so does the innermost, but there's a place where they are offset from each other-- simpler to see than to type.

What that means is: you wouldn't get all quads from any one bit of the pie wedge-- it's only after they have been Array'd around and you do some simple merging that you achieve All Quaddom (does the NSA perk up everytime I type that?).

Surrealist.
07-14-2014, 07:13 AM
Thanks Surrealist for the education and the additional reference material. You make a very good point. So after learning how to make it 'look' similar to my reference image. I should probably work on something closer to swampy's solution. That looks very good swampy and thanks for that bit on using transform. I did not know that. My latest solution is just kind of thrown together. I did resize the inner disk. After looking closely at my reference image on the right, I think it is closer to 6 inches. I also thought I saw the inner row of spokes slightly raised, but mine has a harder edge than theirs. The surfacing and such is just experimenting.

Feedback is welcome. I will probably restart this next weekend. :D

Yeah I live for reference material... can never have enough... :)

Looking forward to updates. Nice project. :)

unstable
07-14-2014, 06:20 PM
For a little better understanding, if I'm creating a wheel that will be used in a picture or within an animation but the wheel isn't going to fly apart or anything, why should I make it in parts verses simulating the look? Don't get me wrong, I love the input Surrealist and swampy provided, but I was just wondering. I can understand making the spokes fit the wheel as swampy has shown if the wheel were going to be used in an animated crash where the spokes may come loose and flop around. Like jeric mentioned, I could do the spokes without creating the holes. I'm going to try creating it swampy's way just for practice, but I was wondering what is the deciding factor whether to do it one way or the other?

jeric_synergy
07-14-2014, 07:05 PM
Speed in construction. Holes not required.

Surrealist.
07-14-2014, 09:01 PM
For a little better understanding, if I'm creating a wheel that will be used in a picture or within an animation but the wheel isn't going to fly apart or anything, why should I make it in parts verses simulating the look? Don't get me wrong, I love the input Surrealist and swampy provided, but I was just wondering. I can understand making the spokes fit the wheel as swampy has shown if the wheel were going to be used in an animated crash where the spokes may come loose and flop around. Like jeric mentioned, I could do the spokes without creating the holes. I'm going to try creating it swampy's way just for practice, but I was wondering what is the deciding factor whether to do it one way or the other?

The deciding factor is realism. Obviously in CG you do a lot of things that fake a look. You cheat a lot of things.

When it comes to modeling, it is just a best general practice to model something as it was constructed. It is not only actually easier and gives you more options, it is something that the eye subconsciously picks up. We know something is wrong but not quite why perhaps. But when you take the time to research how something was constructed and model it that way it just has a feel of "looks right". Even though you could probably get by with less, it is the subtle difference that makes something just OK, great.

unstable
07-15-2014, 07:34 AM
Jeric, I believe your statement is correct. If time is an issue and it represents the object well, then simulating may be the choice. But I also strongly agree with Surrealist. It probably would make the object more realistic to assemble it as it actually was in real life and thereby making it more believable. Thanks

jeric_synergy
07-15-2014, 10:25 AM
Not to belabor this dead horse but: because of the unique complexity of drilling holes in subpatch meshes, avoiding doing so represents a significant productivity gain.

In general I do agree with both you and Surrealist, but up against the clock, that's the first thing I'd look to eliminate.

After watching 3dPowers Subpatch Booleans though, I may have to rethink that. I'd love to see how it would handle 5 cutters going into a disk.

Surrealist.
07-15-2014, 09:42 PM
Well there are certainly good arguments from both sides that is for sure. And the balance is what you have to achieve in order to do anything or all you wind up ever doing is perfecting and getting nothing accomplished.

My rule of thumb - hard won from experience in my studio lately - is 99% of the time when you try to save time by cutting corners it will come back to bite you as work gets rejected. It is better, more efficient, to simply do it right the first time.

For personal work or work of course there is the simple satisfaction that it looks "right". And this transcends to the viewer in a more pleasant image.

It depends on what effect you want to create. If it is to convey a sense of realism even if only a small part of a larger project, you make the holes.

If you look at some of the great 3D art you see around, you will find a combination of compromise and attention to extreme detail. How well you balance that is what makes the difference.

unstable
07-16-2014, 07:51 PM
I made some time to give this a shot. I think this is coming along well. I need to make the spokes yet and I'm not sure what I'm going to do about all those teeth for that either are part of this inner sprocket or fit around it. I don't think it can be part of the sprocket because I'm pretty sure it will mess up my roundness. So I think the teeth should be a separate part on its own. Thoughts?

unstable
07-27-2014, 08:01 AM
Ok, after listening to your feedback I reworked the wheel so that the spokes are one piece. I rendered it out and it looked pretty good, but I hadn't made any holes in the interior wheel where the spokes fit and I felt something was looking right because of it. So I struggled to get the holes where they should be. Of course this meant I had to increase the number of polys used in the wheel so I had the right number of lines for the holes. At least that's what I thought. If there is a better way to do this, please let me know. Overall, I'm pretty satisfied.:thumbsup: I think this is one of the best models I've done in my short LW experience. I'm even pretty happy with the materials and render, although its not perfect. Always happy to receive feedback that makes me better.

jeric_synergy
07-27-2014, 08:43 AM
The brushed steel on the flange is great: how'd you do that?

djwaterman
07-27-2014, 09:51 AM
You poor thing, that certainly turned out to be a lot of work, however it's now a thing of beauty, and I'm still only talking about the mesh. Nice render also.

unstable
07-27-2014, 10:20 AM
Thanks for the positive feedback. I've very happy that you think the wires are looking good dj. Makes me feel like I'm on the right track. Now I'll press on and put a tire on this thing.:D

Jeric, the creation of the steel material was rather easy. The hardest part was positioning the material and then I had to break the disk up into parts. The large flat part was one section for materials, another was the sloped ridge, and then the interior axel hole. At any rate, I attached the material so you can try it. I think you have all that is needed in the zip.

XswampyX
07-27-2014, 10:50 AM
Pity you have the wrong number of spokes....

Only kidding, it looks great!
You can clearly see that by going that extra bit, you can get a model that renders beautifully and will look good at any angle.

Surrealist.
07-27-2014, 12:10 PM
Yep, agree. Looks great!

jeric_synergy
07-27-2014, 12:30 PM
AFC, thanks for the material unstable! I'll look at it when I get home. Great model, and it sparked really good subd discussions.

unstable
07-27-2014, 06:30 PM
Pity you have the wrong number of spokes....



Thanks guys. Man, swampy, just for a second you made my heart skip a beat. LOL!!