crew
Kent Trammell (theluthier)

1,343 replies · posted

Citizen Collaboration: Modeling Heavy Equipment (Backhoe)

Accomplishing a BIG project TOGETHER

Starting in March 2020, I began a large stream project to model a comprehensive, highly-detailed Caterpillar 434E backhoe. Please watch the first stream (first 3 minutes of abridged version at least) for an overview.

It's a straight-forward project of building a complex model over an extended period of time. The spin on the project is that I want your help to accomplish it! If I build some parts and you build some parts, we will finish this thing much quicker. Plus the contribution format will include reviews, the potential of having your piece(s) assimilated into the final model. Not to mention large quantities of XP are at stake 🤑

NOTE: This is an involved project reserved for Citizen members.

Collaboration Details

The general idea is that I kickoff stages of the project via live stream, which is typically once per month. For the time in between streams, you choose a piece of the backhoe and apply what you learned from the stream to that piece. For example, the first stream covered initial block out. So between stream 1 and stream 2, your job is to pick a piece and block it out.

This is the repeating protocol for each Assignment Period (between-streams):

  1. Signing up for pieces of the model. I will be maintaining a Google Sheet to organize piece selection and grading. Following this URL will give you the ability to comment. The sheet is sectioned vertically and color-coordinated with a backhoe image. To sign up for a piece, right click on an empty cell below your desired piece section and click "Comment" (CTRL + ALT + M). Comment your name or username and I will fill it in officially as confirmation.

    Remember that it is OK for multiple people to model the same piece. But lets try to have at least one person assigned to each piece before doubling up.

    2. Project file structure. We're going to use Google Drive as a means to sync a single working directory among all contributors. See this thread about details and how to get set up.

    3. Submitting your piece for review. Exactly one week before the next stream is scheduled to be broadcast, your piece is due. In the "Attempt" cell of the spreadsheet, paste a download link to your .blend file via Dropbox, Google Drive, or equivalent hosting service. This will earn you at least 10 XP each week.

    4. Review. On stream I will review each submitted piece and decide which will be assimilated into the master model (in the case of multiple submissions of the same piece). The purpose of the review is to generate feedback for you work and advice for improvement.

    5. Assimilation. If your piece makes the cut and is assimilated into the master model, your name / username will be added to the object in Blender. You will go down in history as a definitive contributor to this epic backhoe model!

    Once the model is completely done, I plan to render a short demo reel showing off the model and crediting all contributors. It should be a rewarding conclusion to all the work we will do!

Assignment Prompts

I will reply to this thread after each stream with a [big] assignment post denoted by a  📣 emoji in the title. There I will clarify instructions about each Assignment Period.

Online Resources

Reference Images

Livestreams


This is THE thread

We will centralize our collaborative communication between the streams and this thread. Ask any and all questions pertaining to the Backhoe project here.

  • Ok I did some digging today on my CAT website. Here is the maintenance manual for combined models that includes the 434E. The illustrations may be helpful to gleen some details. There are some decent axel shots as well and it talks about operation a little regarding the individual parts.

    https://dealer.cat.com/content/dam/dealer/ps/repair-options/self-service-options/Service%20Instructions/backhoeloaders/PM3-JBA_NBA_EME_SEF_DPH_SJL_MAW.pdf


    Let me know if the link doesn't work. It works for me but I am an employee haha.

  • theluthier You might benefit from this close up video it showcases some nice details not shown on prints.


    https://www.youtube.com/watch?time_continue=35&v=_2vPdctN8pw&feature=emb_logo

  • I struck it rich!

    I found the page that I think Jake may have stumbled upon but from the Legit Cat SIS 2.0 website. 


    I did a search for the 434E and there are all kinds of diagrams: There's a lot here......

    https://sis2.cat.com/#/refine?keyword=434E 


    Refining the Search to the "Parts Identification" page gets really into detail. There is a whole left section of stuff to click on and it is sorted by Parts/Repair/Service sections that changes the left side options.
    https://sis2.cat.com/#/detail?keyword=434E&serialNumber=SJL&mediaNumber=SEBP5078&tab=1 

    There are even options to make PDF's

  • crew

    📣 Assignment #2: Finalize 25% of your chosen Part

    April 15, 2020  - May 11, 2020

    For the assignment period between stream 2 and stream 3 the focus is finalizing aspects of our chosen backhoe parts. By "finalizing" I mean that individual pieces have all details accounted for and are build for subsurf smoothing:

    In stream 2 it took me 2 hrs to model one large piece of the digger arm. Since this kind of complex modeling is slow and tedius, my expectation for assignment #2 submissions is to complete at least 25% of your chosen part. This should get us in the groove of the workflow, hopefully making for smooth sailing for assignment #3 and beyond.

    Everyone start from the same file

    I still think this approach is best. After each stream I will upload a .blend for everyone to the corresponding stream page. For example, on the stream 2 page you will find the MASTER_caterpillar-434e_model_02.blend file available to download.

    That file features the part I modeled on stream as well as the stabilizer / chassis part from  adrian2301. Try to model organized; keeping all your parts within your part's collection.

    Going forward I will continue to add each iterative MASTER .blend file to the respective stream pages in sync with assignment posts:

    • The stream 3 page will be home to the MASTER_caterpillar-434e_model_03.blend model file (when the time comes)
    • The stream 4 page will be home to the MASTER_caterpillar-434e_model_04.blend model file (when the time comes)
    • And so on...

    Assignment Submission

    I've added a new tab to the Google Sheet labeled: #2: Finalize 25%

    The names of contributors remain from the first assignment. To submit for assignment #2, contributors just need to make a comment containing a link to their .blend file in the Attempt cell beside their name

    **The DUE DATE for assignment #2 submissions is MAY 10, 2020**

    Resource Links

    I've added links to the bottom of description of this thread (appears at the top of every page wihin): Online Resources and Reference Images. Thanks to everyone researching online, we're developing a dense collection of helpful web pages including technical diagrams of individual pieces! Shoutout to jjakeblended for uncovering that gem 👏

    I've outlined several relevant links from the directory but there's plenty more where those come from.

    The reference images are simply the modeling sheet and photo collage that I've been using on the stream for your download convenience. Note that these are now packed into the .blend file I mentioned above.

    I will continue to update add links to this list as we uncover more online resources.

    Organization

    I'm trying to tie all the lose ends so we all can be on the same page through this collaboration. But I'm figuring it out as we go so don't hesitate to ask questions or make suggestions!

    • My only regret is that we won't be able to take this thing for a ride once we're done, maybe smash some ATM's. But a render with alpha on the driver seat, we can easily put ourselves taking it for a ride.

  • Thanks to jjakeblended 's treasure trove discovery, I have made some more adjustments to the proportions of the chassis and we can now see how the different parts will eventually come together. 👍


  • The front axle and the back axle of the Caterpillar 434E are - as I've already supposed - significantly different:

    By the way, if you wonder how to rotate your viewport in Blender 2.8, it's called "view roll":

    counterclockwise: Shift Numpad 4
    clockwise: Shift Numpad 6

    (see: Preferences => Keymap => 3D View => 3D View (Global) => View Roll)

  • Kent the array modifier is not working for me properly 

    When I try to put the object offset to a bezier curve it goes in a random direction and gets smaller and smaller 

    how to fix this

    • sshanr: you don't need a curve. For your treads, take an Empty and rotate it 36 degrees around the global x-axis (which is our wheel's rotation axis). The object origin of your tread must be in the same place as your Empty (that is where the axes of your Empty are crossing). Add an Array Modifier to your tread and only select "Object Offset" as offset in your Array Modifier). Set the "Fixed Count" to 10 (since 10 * 36 degrees is a full circle of 360 degrees). Rotations and scaling should be applied (with CTRL + R and CTRL + S) because otherwise these transformations would be applied for every duplicate in relation to its predecessor in the duplication order. You can analyze it here in this file:

      https://www.dropbox.com/s/5z9nvdcyaegl3kh/ArrayModifier_Demonstration.blend?dl=0

      Select the original monkey head (top middle; this also selects the duplicates) and scale it, rotate it or move it in object mode and see how it influences the duplication. Doing the same transformations in edit mode also applies to every duplicate but the duplicates now change identically in an absolute manner and not relatively to the preceding object instance.

    • No it does the same

      spikeyxxx do you have a idea

    • duerer is absolutely right.

      The only thing that can change the size of your object with Object Offset is when the Scale of your object or of the offset object is not 1 in x, y and z.

      If you make sure that the origin of your object and of the offset object are in the same place you should get something like this:

      Personally I would not use Object Offset in the Array Modifier for a wheel or tyre, because you'll have to make a tapered (and curved) section in order to not get holes (like in above picture) or overlaps and that is harder to model than something straight. Also making adjustments to that section later on is a lot easier with a straight model. That is why I always use an Array with Relative Offset followed by a Simple Deform Modifier set to Bend (360° Angle):

      You'll have to put your Empty at the 'bottom' of the wheel, but no need to calculate the rotation of an Empty like in the Object Offset;)

      Modifiers can be tricky (you have to get everything correct), but once you've figured them out, they are like a found treasure and you'll be using them all the time!

      If you still can't get it to work, feel free to post your .blend, so that we can see where the problem is exactly.

    • Another possibility would be a "Curve Modifier" in conjunction with a Bezier circle, but for that you would need to manually adapt the distorted object so that it fills a whole circle unless you know the exact length of your curve (I know that there was a Python script to calculate that value in Blender 2.7x). That's why the "Simple Deform Modifier" is clearly the better choice for circular deformations since you only need to specify the angle as 360 degrees. The "Curve Modifier" is the choice for curves that are no circle.  I like the "Simple Deform Modifier" for things like this:


    • I just had to smile when reading in the Blender manual about the "Bend" mode in the "Simple Deform Modifier":

      "The Bend mode is more complex and less intuitive." 😄

      see: https://docs.blender.org/manual/en/dev/modeling/modifiers/deform/simple_deform.html#bpy-types-simpledeformmodifier

  • Trying out the Shrink Wrap Mod on the fuel (gas) filler cap as suggested by theluthier during the live stream. Looks OK.

  • theluthier Kent I added my part and commented on the thing with the link on the thing. Let me know if it worked please. 

    I think I kept it at 25% completion, maybe I got to 30%, an argument can be made for 32%.

    EDIT: I updated the file, I forgot about something but now it's ok. I hope so. So if you downloaded it, downloaded it again please.

    • crew

      I got it 👍 Thank you, sir!

    • theluthier I do have a question, about consolidating the edge loops when they're side by side, when 3 loops are side by side, you did that during the stream on my battery box and I was like oooohhh, that is some sweet technique, you can remove lots of excess geometry, I don't know if you remember, you merged them 3 into 1 and left a quad and removed lots of vertices. I was going to do that to optimize the mesh, but then I though, what if I have to change something later on? Am I committing too soon when stuff ain't done yet?

      So I guess my question is, should that be left to when things are more polished and on a finished state?

    • crew

      dostovel Good question! Indeed you only want to consolidate 3 edges side by side if there's no use for them beyond a certain detail. I think this is a great talking point to cover in the next stream, but I'll try to answer generally here:

      In the case of your batter boxes (which look awesome btw!) there's several planar sides that don't appear to have any additional geometric detail and in some cases are visually covered by other pieces:

      Therefore the orange and blue sides are ideal candidates for "triple-edge optimization" -- Let's call it that so we both can refer to this technique quickly.

      Those are the principles that govern my decision making process about optimization. But you're the expert on this part of the backhoe, so please apply the principles accordingly!


      The last thing I'll say is that this optimization process is not completely necessary. Best case scenario is it simplifies the object as much as possible for workability and should reduce render computation / RAM usage by a fraction at least. It's a "best practices" kind of thing. For a quick and dirty project I would skip this process.

    • theluthier Oh my god there they are, you have them now, technology is like super awesome, one day they are in Santo Domingo, the next day in the USA. I feel like other battery boxes are like mama! come quick, your son is on the tv. 

      Thanks for the great answer, I like best practices, it's like an OCD to leave things super tidy up. Leave no dirty dishes on the kitchen kinda modeling for me.

  • spikeyxxx Has just fixed my problem Hip Hip Hooray for Spikey

    Here is my model

  • spikeyxxx , one more question concerning the cutting-in of sharp edges into curved surfaces has come to my mind:

    How do I manage to connect the cut-in egde loop smoothly with the surrounding mesh and avoid pinches?

    • I am not sure what you mean exactly.

      By 'beveling' the cut-in edge loop it has supporting edge loops inside and outside of the cut. It does create some 3- and 5-sided poles at the corners, but that's where the Shrinkwrap comes in and that is also a reason to have a higher subdivision level on the target mesh. Due to the fact that the original and target mesh are already at the same place the Shrinkwrap Modifier does a really good job (and the projection method is not so important I think).

      As long as the original mesh has a clean topology and you create the cut from existing edges, there should not be any pinching.

      I've used this method a lot and never had any pinching issues, but if you do, please let me have a look at it.

    • spikeyxxx I'm thinking at the edge of the fuel tank cap in the first of your three instruction images where diagonal edges are visible in the upper left and lower right corner of the cap outline. These edges are obviously not part of the original rectangular grid cell mesh. In this case, the cut-in edge loop has a comparatively simple structure, but what if there is a more complex pattern with a lot more of direction changes and details?

      This has been cut in only with "Knife Project". Now I use "Triangulate Faces" and then "Tris to Quads", but this dosen't guarantee you a good result, as the edge beveling shows:


      The original file with the cut-in edge can be found here.



    • diagonal edges are visible in the upper left and lower right corner of the cap outline. These edges are obviously not part of the original rectangular grid cell mesh

      Actually they are! I just edge-slid those verts a bit to show that you can change the shape of the cap a bit, even after you created the target mesh.

      As for the complex model: that is not how this was intended (did I forget to mention that you should start out with a clean topology?).

      As this is a Subsurf method, I wouldn't use such an already rounded shape with 153(!) verts. make it as simple as you can to still get that shape when subdivided and then maybe subdivide the Sphere one more  time and the move  ( Merge) the verts of the sphere to make that shape. Or just create that (simplified!) shape directly on the Sphere by sliding some verts. That ensures that you keep the all-quad topology (at least in that area).

    • spikeyxxx You are absolutely right, I should follow more the "keep it super simple" (k.i.s.s.) principle . Only one question arises for me when you propose vertex sliding for creating the shape of the cut-in edge: How do you ensure that you don't loose the round shape in the area of your cut-in piece? With light inciding on your sphere at a sharp angle you can even see a very subtle surface unevenness.

    • duerer that will be taken care of by the Shrinkwrap (again, make sure that the target has a higher subdivision level!). For an even better result you can move your vertices with surface snapping on (still using Shrinkwrap as well, of course).

    • spikeyxxx you are a fountain of knowledge. I had forgotten about the magic V hotkey as rarely used. I think theluthier was going to cover this in the next stream which I am looking forward to, mainly to see if my attempt is as you suggested.

    • I originally selected the faces and separated by selection, then used a shrinkwrap using vertex group to get the desired result but with subsurf could not get the smooth corners. But CTRL+B and then V and a shrink wrap, so simple, genius.

    • adrian2301 I had to make a lot of those panels/seams/... in the Vehicle Contest here on CGCookie:

      So I had to find a way to be able to comfortably do those kind of things;)

      I'm sure I'm not the first or only one to come up with this technique, but I found it by myself, which makes me kinda proud...

    • OMG.... I'm seeing all this and going....panic.  I look at what I've done and I'm going... yikes.  I ended up starting over and it still looks bad...better but not like this.   I'm so out of my element up in here 

    • ketre

      'DON'T PANIC'

      (Imagine that, written in large, friendly and reassuring letters..)

    • ketre Why don't you post what you have, no matter how it looks like, and tell us where you are having trouble. 

    • Here you go, this took me 3 day to get this... The sides bulge out in a way from the back view that they don't from the front, (not frong) lol ugh... if I take in to perspective view it doesn't look like that. 

    • You should be proud spikeyxxx , your vehicle looks awesome, did you win the contest? Your technique defines the panels so well with clean lines. I will probably use this technique again, every time I do, I will give a little nod in your direction.


      ketre like dostovel says post it here, I'm sure somebody can give you some tips, as was the case with me.

      'You are not alone'

      (Imagine me singing that like Micheal Jackson), actually DON'T

    • hahahaha Imagine the uhhh dirty words flitting through my mind in my frustration. I actually turned the jams on while doing this, thinking yea imma rock this doggy....hahahah NOT

    • ketre as you can see from this image, the left most window does curve quite significantly from top to bottom and from front (above the high point of the wheel arch (fender)) towards the bottom left corner, which would look like a bulge from the rear view. The important thing is what it looks like in 3D / perspective view, when you enter front, back or side view you switch to orthographic view which is flat 2D. Try and asses your model by panning around in 3D to see where you can smooth out any areas, maybe add a subsurf with 1 division. Keep it up. It's not a contest we are all here to learn and support each other. 👍

    • ketre It must have a slight slant, I think is the word. Can you post a top view? Or better yet, can you share the .blend? 

    • ketre try to start with less geometry; we are going to use a  Subdivision Surface Modifier on this, so do not use too many edge loops where they are not necessary...

      You will need some holding edges, but the basic shape doesn't need that many vertices. 

      It's easier to add geometry than to remove it, if that makes sense;)

    • Here's the file. 

      https://drive.google.com/file/d/1pTAvujvlRl20-Ff7fhwuedbx446uQkab/view?usp=sharing

      If I have to start over again...ugh  I just tried adding the subsurf to my existing file and omg...

    • That bulge you see is just because of the angle, it slants so things begin to bunch up in orthographic mode and you can see the geomerty. It's like someone peeking out the edge of a door. Everything is fine there.


      If the side was perfectly flat, without the slant, you wouldn't see the geometry, because it would be occluded. No one peeking out the side of the door there.



      Let me now see the other thing. Give me a sec.

    • Yes, here we have a problem.


      You have a case of very deadly F'ed up faces.



      That happens by accident a lot of the time. You make a selection and without realizing it you select vertices on the back and move stuff around and without you know you're moving stuff you don't want to move. Then you see something like that and you're like what the hell happened. It happens to us all. 

      Or maybe you made an inset faces and it went pass the desired amount.


    • Really just delete those problematic faces, and connect them over to the good ones over there. 

    • Another good thing to do as Kent always says, is to inspect your mesh from lots of angles, you'll see the contour and dips of vertices misaligned.

    • Now it follows a nice arc and when subD will smooth nicely. 

    • Told you guys I am seriously out of my element here, lol  and ya I see the f*ked up faces and trying to fix it has had me frustrated. Yall make it look so easy.   I've not really watched anything about hard surface modeling. I guess I'm more into Organic...yep I know its just  excuses. Lately my eye sight has been blurring, after the issues I had with health I'm doing better.

      I'm not giving up, but I'm pretty sure yall gonna have to fix my mess before you show the world lol

    • Or it could be like a Modern Art piece. We can say " Yes we made the driver cabin like that unpurpose, as an expression representing our heads of state" It would sell for millions of dollars, trust me .

    • lol dostovel  I just love you guys

    • I haven't been here for the last 16 hours and already 22 new replies, a new record!🥇🏆📀😁

    • Hi Guys I am am going to send a photo of my model.  What do you think?

    • theluthier  and spikeyxxx  as well as dostovel 

    • The treads are looking very 'thin' compared to the references. There should also only be 22 per wheel (or 22 left and 22 right if you want), it looks like you have 50 or more.

      But you're getting there, it's improving.

    • sshanr Looking good. Though it looks like you forgot to download the updated model from livrestream #2. Download that one and copy your parts to the updated file.

      You can also look on YouTube for wheel modeling tutorials, there are good techniques that will help you make the wheel one solid mesh. 

      Also here on CG Cookie there's this one from when Jonathan Williamson modeled the wheels of the post apocalyptic truck.

      https://cgcookie.com/lesson/modeling-the-tires

  • Hey guys I'm trying to figure this out on my own but I'm having trouble making this look right. I have a  solidifier on (and its off in this screen shot) but when I turn it back on I have an even bigger mess. Am I doing the topology wrong? And why are those faces blank ? HELP!!!

  • theluthier Kent can you tell The Big W to see if it's possible to implement a Down arrow here in the forum? There is an Up arrow at the bottom left of the screen which takes you up once you're down, but when you enter the forum and you have to go down... man it's a long way down and there is no Down button. Unless I am missing one?

  • DOG Manufacturing Group,

    I need your advice on  this part of the bucket. I think I have this side correct, but there isn't good reference for the opposite side of this should be shaped like. The part "inside" the bucket.

    Should I just solidify this section? It would interrupt the bucket's otherwise smooth shape inside.

    Here's what I mean, it doesn't seem right, but for the back to function it kind of has to be this way?

    Maybe these "innards" are hidden by another sheet of metal, so the bucket can stay smooth?
    I'm hoping there's an obvious solution I'm just not seeing.

    Everyone's pieces are looking so nice!

    - Parker

  • Ok. I need some hard surface modeling guidance from the gurus of the group. I am having a hard time making smooth and crisp transitions from square/hard edges to circular topology. Three instances on the loader arms are causing me grief and I am unsure how to proceed. It acts like the Bevel is "bleeding" over areas where it shouldn't. I'm sure it's an easy fix but not sure what it is and I've tried a few things but with no success. I think I need maybe some more supporting geometry or "holding" edge loops in more places but maybe it's something else. It's hard to add to the circular cross section parts to give more edges to them to work with and keep them circular haha. I'm dreading that is perhaps the solution.

    Link to my progress file: Loader Arms - Blanchsb


    I turned on the normal matcap because I feel like it highlights this best: Areas circled in yellow are where I am reaching the limits of my "crisp yet looks soooo smooth" skills. I have my modifiers ordered like so: Bevel (with weighted edges I pick, all set to a value of 1 and percent of bevel of 5%) then Subdiv with divisions set to 2. That mod stack seems to work to keep things looking smooth yet sharp, except in these places.


    Wireframe for reference

    • Hi Blanch, that is looking great!

      About those problem areas, this is how I would do it (maybe not the best way, but I am terrible with Creases and the Bevel Modifier with Bevel Weight only works  in some cases for me...):

      Having edge loops on both sides of the selected loops gives full control on how sharp/rounded they look, without artefacts.

      The round cylinder has 8 vertices, while the rectangular piece has 12, so you need a 4 to 2 junction on both sides.

      I just did this small part and only on one side, but you get the idea.

      Personally, I would not use a Bevel Modifier at all, probably and only work with holding edges, but that's just preference.

      Didn't show it here in Object Mode, with Subsurf, but it looks just as good as the rest of your model!

    • As far as I know, the biggest problem you are having, is the 5-sided pole directly on the edge:

      I basically moved that pole on a flat area, where it cannot do a lot of harm;)

    • Wait! Where did you move that 5 sided pole and can you show me? I m having a hard time thinking how to fix that part. It totally makes sense verbally but my mind is not working on the same level as you.


      My brain keeps telling me "that's pole has to be there because of all of the surrounding edge complexity" but I know that is flawed thinking haha.

       

    • Does this clear it up? (probably not,,)

    • No you're explanation is great but I think that removes some geometry with that fix on that particular section.

    • Oh, sorry I removed that detail;) I removed alot more and then re-made the connection and forgot that 'step', but that can easily be re-added with changing the topology of that area.

      Just select the white face loop and do something like E, ALT+S, SHIFT Z...(or any other magical incantation that works for you):

    • I don't know what the technical term is called for that? A shoulder perhaps?

    • Bevel (with weighted edges I pick, all set to a value of 1 and percent of bevel of 5%)

      blanchsb Remember Kent said to try and aim for software agnostic  modeling, meaning don't use stuff that are only a Blender thing. I think weighting the edges you pick is a Blender thing. The only modifier we should use is the SubD. 

      Let me try and have a look at your problem areas, that pinching in rounded from flat ares are always a headache.  

    • blanchsb I think I misread the situation a bit...

      Do you want this to be flat:

      And this to be round:

      ?

    • I could give you something like this:

      Which has this horrible topology( so far):

      The thing is: you have a blue edge with only one holding edge to control the sharpness turning into a green edge that is supposed to be completely round, so the orange holding edge need to get away from the green edge to make that round if that makes any sense at all;)

      The point is that an edge (corner) should have two holding edges, one on each side for full control...

      I think you want the green part of the 'block' to be round on the top and flat on the side, but if you move the orange edges towrds the green ones, that will make the top flatter..

      This is all very hard to explain in words, but I hope you get the drift.

    • Yes! One side flat, the other side round; And I want that Lip/Shoulder on the "round" piece in the bottom image to stick out on the side (sticking out past the edge)

    • Yes, sorry your answer updated on my first reply haha. I do want the cylinder section to remain a cylinder, which is why I am afraid to add any holding edges laterally along the cylinder (it doesn't subdivide as a circle anymore when I do that or I am doing it wrong). I basically want a cylinder with nice crisp cut on the side that is completely flat and everything from that point forward will have flat/square features going towards the main pivot point Pin Hole on the right side.

      Your updated reply looks much more like what I am after on the top "metallic"image.

    • Crap dostovel I must have missed that. I just wanted the bevel modifier there "for now" because it makes it so easy to go in and change simple shapes. I can pick what I want to bevel and what I don't. I can always apply the bevel mod at any time and then it is just a subsurf.

      But you and @spikeyxxx are right. Maybe I just need more"holding edges.

      Haha I am feelig like ketre now. Like a noob amongst pros. Can I join your Team Karen?

    • blanchsb 

      Maybe I just need more"holding edges.

      But holding edges at the right place!

      I will have a go at it tomorrow, if dostovel doesn't come up with the perfect solution before that;)

      I think trying to tweak this to get the result you want, is like trying to fix a broken vase with tape; it might look okay, but the structure will be awful.

      I'll just start from scratch (on this specific part) knowing what I(/you/we all) want as result!

      That will give cleaner topology and therefor cleaner  result.

    • Ha! Million Dollar Question. Where is the right place for the holding edges haha. Yeah. Complex yet simple shapes give me such a hard time. I spend way too long on them. Thanks guys for the assistance. If anything I am learning what not to do and what to do to get results I want.

    • blanchsb If I understand correctly I think this is what yo are looking for.

      Added some edge loops and changed the settings on the Bevel Mod.

      Changed from percentage to offset at 0.1 and from weight to angle and adjust the angle until it looks good, ended up at 60d. 

      With this just need to keep an eye on the rest of the model, but usually its just a matter of adding edge loops or changing the angle of the face slightly by moving  edges by literally fractions.


    • blanchsb or maybe this is what you meant after rereading:

      As you can see I added more edge loops to help define the shape of the cylindrical shape and scaled on the local x and y. same settings for the Bevel Mod as I suggested earlier, minus the other loops I added previously. It's all about the angles. Hope this helps.

    • adrian2301  Yeah your second image is closer. Everything except for the part on the top where the cylinder meets the square. That top needs to remain circular and only the side remains the square.


      Yeah, how did you "scale" the extra lateral loops on the cylinder to make it circular (did you just "eyeball it" while scaling in the local X and Y)? 

      I'm hoping for a precise solution as opposed to "eye balling it" haha. But yours and Spikey's are really more like what I am trying to get at.

      When I make those extra loops using Ctrl R they don't magically turn into a circular cross section and the sub-div now makes them appear even more-so like an octagon. Hope that makes sense? I guess I could bust out some of my engineering math to move them precisely but I was hoping for a modelers approach haha.


      Maybe starting over with a circle cross section with more edge loops would help give more geometry. Maybe the limits of the basic shape 8 sided shape are being pushed beyond their ability?

    • Or,    

      just a late night test....

      far from perfect, but...

    • blanchsb I eyeballed it but if you want to be more precise , with the loop tools add-on loop select a ring of verts and click on circle ( loop tools add-on), for the top squarish part  I would vertex snap to line up with the others. I see now on spikeyxxx 's post , I didn't include that edge when I scaled out . But I would say more geometry is required or  at least one edge loop as spikeyxxx has done on that face to define the shape.

    • spikeyxxx is this a 6 sided pole I am seeing? That actually looks decent on the sub-div though!


    • blanchsb It's a 5-sided pole, but it's totally in the wrong place here!

      Finally managed to move it out of the way (see post on page 4)...

  • I've done some things that seem janky to wrangle edge loops.

    Are these okay? Is there a better way to handle these situations?

    • I do not know, what exactly you want to achieve, but this doesn't look right to me (but who am I to judge...):

      In red: a vertex that doesn't make any sense to me; no matter where you put it, it will always make a concave quad:

      in blue are the 5-sided poles. We try to avoid those and although that is not always possible, it looks like you've got too many of them and at non-planar points...I don't know what is happening at the purple question mark (yes, it's supposed to be a question mark!)

      Not sure about the green 3 sided pole, might be ok, or not?

      In short: there is a better way to do this, but it is not clear to me what exactly you want to do...



    • spikeyxxx My goal is to keep those holding edges to keep the shape, and just fill in the gaps. 

      There's a weird mix of concave and convex shapes all really close together.
      I used GG to make sure that the blue vertex was directly in the middle, and neither face would be concave.

      Here's my second attempt at this section. I'm not sure it's much better.

    • ppfbourassa I'm still confused as to what I am looking at;)

      Anyway a quad that comes in the shape of a triangle is not considered good practice...

  • Waves at everyone...yall so good :)