No. SketchUp is capable of greater than .001" precision and is being used across the globe to design and build everything from skyscrapers to machine parts. But SketchUp is a tool, and a very complex one at that, so until you learn how to use it you are likely to be a little confused and frustrated and yield unexpected results. A table saw can make perfect dadoes, or it can cut your fingers off, the only difference is in the knowledge of the operator, not the tool. SketchUp can totally do what you want, and if you put in the time to learn it a bit better you will laugh at these earlier struggles and get down to work. Here is a good place to start. https://learn.sketchup.com/track/sketchup-fundamentals-part-1
Post your model here and I am confident you will get some helpful, constructive, tips on how you can improve you process and achieve your goals.
Riley, Anssi, John McC, Steve of the long essay (that I truly appreciate), RTCool, Box, and all the other sages, learned ones and basecamp Attendees I may have overlooked.
Back in my dotcom days, after work, we’d all get on the network and play DOOM. I wasn’t as good as my fellow workers—not as fast on the keyboard or whatever. I kept getting killed. I knew the time had come to stop playing DOOM (and learned something about myself) when I finally managed to kill someone and spent the next 2 minutes pumping bullets into him! This is how I’m beginning to feel about SketchUp.
Here is my file.
It is unbelievably simple. I started by creating a proof of concept document showing the way the kitchen is now (remnants of which exist in the Current Configuration layer), moving the fridge, and showing how it might work with the additional cabinetry and shelves.
I want to end up with a drawing that I can dimension (to 1/16” accuracy), print, and use to construct the cabinetry. I understand that no drawings are completely flawless (the ones I drew for the two story house I constructed [VectorWorks] came close) but I want to be able to depend on the drawing’s printed dimensions and not to have to remeasure everything in situ.
I also find using the program extremely engaging! (not so the banging my head on the wall)
I don’t want you to solve my problems. I want to learn to solve them myself.
(To answer the question of this email) I continue to switch from one view to another. Usually the reason I do this is to locate, select, and manipulate something that’s proving elusive to handle.
General questions:
I use the tape measure to check measurements. Sometimes it’s completely accurate (e.g., ¾” member thickness); sometimes it gives different readings (albeit “~” approximations) for dimensions of items that were copy/pasted. I don’t understand why they aren’t as exact as any other measurement.
Currently, despite shelves not fitting vertical members, etc, I thought I’d try putting shelf standards (two different kinds) on the cabinet without the shelves. I drew a rectangle (which I planned to route ¼” deep), copied it, and pasted it—unfortunately on the other side of the plank—and I can’t get to it to put it back on the correct side, or keep it from aligning to an edge of another nearby piece of wood.
Oh by the way, by way of maybe being able to help you guys out, Might it not be a good idea to insert an ellipses after Window/Model Info… and Window/Preferences… ?
Please excuse all the mistakes in my file. I look forward to learning.
Immediate quick reaction - you need to learn how to use components more consistently, and also not leave ‘loose geometry’ in the model.
Give each component a meaningful name as you make it. Looks as though you just accepted the default “Component #xx”.
Start with the overall size of a box for a whole cabinet. You can later add subcomponents for shelf or shelves, sides, top, bottom, kick plate etc. Keep the whole assembly as a single parent component, with subcomponents.
Your cabinets are mostly not exact heights. One countertop is 3’ 0 1 2/" - is that intentional? Or should it be an exact 3’, which is a common US countertop height.
And a tall cabinet measures approximately ~ 6’ 8 15/32". That’s a very odd height, and probably shows you have a difficulty in drawing exactly.
When you are drawing a line or a rectangle, or using the PushPull tool, you can input exact measurements by typing. Just type. What you type will show in the Measurement box bottom right, but do NOT click in there before typing.
And what’s the purpose of the shrunken copy of your main model, down near the origin?
You have the view style set to Hidden Line. If I change it to Monochrome, I can see you have the wall face reversed (showing grey/blue to the front, instead of white).
In the spirit of learning… Some of this may duplicate what @john_mcclenahan already wrote. Sorry to say, but at this point the model is enough of a mess that you would likely be better off to trash it and start over. Fixing small errors can be quite fiddly, tedious, and error prone.
What’s with the itty-bitty version near the model origin?
I see lots of evidence in the model that you need to work on improving the following skills:
inputting exact measurements as you draw “free” objects (i.e. ones not constrained by needing to fit against or into others). In SketchUp you start a basic shape such as a rectangle, let go of the mouse, and type in the required dimensions + return. DO NOT click in the measurements box. Similarly, if you want to pushpull something to 3/4", start the pushpull in the appropriate direction, let go of the mouse, and type 3/4" + return. There are multiple instances of items in your model with weird sizes that you probably didn’t intend. For example:
using inference snaps to make things align with, match size with, or touch other pre-existing things. Watch for the little inference snap markers before you click to set a position. For example, if you are drawing a dado to receive a 3/4" shelf you’ve already drawn, use snaps to the corners of the shelf. Or conversely, if you drew the dado first, use snaps to match the shelf to it. If you are drawing a cabinet you want to fit up against another, use inference snaps from that other. There are multiple places in your model where something like a shelf does not meet up with the sides of a cabinet for no evident reason. Some examples:
you might find it better on something like this model to work in either fractional inches or metric. There is a known problem with architectural units that when the inch portion is not exact, there is no ~ shown to indicate the displayed dimension is approximate. This exacerbates the previous points because you can be off a small fraction of an inch without knowing it!
I’d work in monochrome mode, not wireframe. Wireframe suppresses display of faces, and that can mislead you as to what is really there. You can always orbit to a better viewpoint or hide objects (or turn off their tags) when you can’t see a needed snap point through them.
I have absolutely no idea what the itty-bitty version is all about. I assumed it was a (useful in some way I was yet to learn) SketchUp artifact.
If I’m about to start the project again, I would like to do so by establishing another layer and encasing the cabinets in large boxes. I don’t want to have to go back to the kitchen and start measurements from scratch again.
For some reason your images aren’t showing up. I’ll wait a while and see if it is an issue with the forum software trying to reformat them (it sometimes does that). Until then I have nothing to go on!
I imported the dwg into an empty file, and I can already see issues both with placement of things and inexact dimensions. Alas, you may be starting from an imprecise dwg, and those problems will just propagate as you draw in SketchUp!
Here’s a dose of reality: see attached PLEASE EXCUSE THE MESS!!
Those images (partial screen captures) are the way I’ve communicated all along. I assumed you’d all seen those (I wouldn’t have understood my emails without them.)
Blinding all the other layers except “Fixed Cabinets” (the minimal anchor I need to start from)
Copying all on layer untagged (Cntl A/Cntl C)
Blinding all layers except “Starting over” and pasting…And I end up with nothing on that layer
In the process of doing this I copied an object from layer Untagged (I want that layer empty) pasted it into “Fixed Cabinets” and discovered it was now a component tied into another object ON ANOTHER LAYER!
Q1) What is your concept of Layers? How do you want people to think of them?
Q2) Having closed out the file (I can understand temp files might be created while the program file is open) I now find I now have two identical files (filenames) which I didn’t think was even possible in Windows! Which one is my operational file and can I expect more instances in the future?
Q3) How do you transfer objects between layers?
Q4) How do you transfer all objects on one layer to another?
This has been a major source of confusion for people coming to SketchUp from other CAD or 2D graphic apps. It is the main reason the name was changed to “tags” in SU 2020. Unfortunately, the SketchUp literature is crammed with wording that misleads folks about this. Even many experts, out of habit, keep saying “this object is on LayerX”, which is not really correct. Many of us have been trying for some years to guide a change in this language to things like “uses LayerX” or “references LayerX”.
In SketchUp, layers/tags are properties that an object can reference to determine its visibility (and a few other visual effects). Multiple objects can reference the same Layer/tag at the same time. The relationship is backward from what people with 2D background expect: when SketchUp goes to render an object, it looks to the layer/tag used by that object to decide whether to show that object. The layer/tag does not “own” anything. In fact, there is no way to go from a layer/tag to the objects using it other than probing each object to see what layer/tag it uses. Through the GUI that would mean turning off all the other layers/tags to see what is left (and assuming things are not invisible due to being hidden).
Further, in SketchUp layers/tags do not provide any kind of isolation between edges and faces. Edges and faces will stick to and intersect each other regardless of what layers/tags they use and regardless of whether those layers/tags are currently visible. Groups and components are what SketchUp uses to prevent this interaction: edges and faces will interact only within the same “context” - model, group, or component. That is why all the experts say that with only a few very special exceptions, all edges and faces should use Layer0/untagged. Layers/tags should be applied only to groups or component instances and non-geometric things such as dimensions, screen text, and images. Doing otherwise will almost always eventually lead to confusion when you edit one edge and an invisible one gets changed along with it. Unfortunately, because that’s not how layers are used in CAD, we very often see imported edges and faces with associated layers in SketchUp.
Each time you save a file, SketchUp relabels the original version as .skb as a backup (on a Mac it follows the UNIX convention of inserting a ~ before .skp instead). Unless you change the extension to .skp or change the file associations, SketchUp won’t normally open a .skb.
See answer to Q1. This question is essentially meaningless because layers don’t own objects. “Transfer” isn’t applicable because it implies possession or placement. Instead, you select a desired object and via the Entity Info window tell it to use a different layer.
Again, not the right concept. However, in SketchUp you can open the Layers window, select a Layer, and click the - button to delete it. SketchUp will ask you what you want to do with the objects currently using that layer. One of the replies is to give them a different layer.
Not the case. This .DWG is a bit of a mess full of stray lines, misaligned objects and some odd dimensions. Starting with this as your base is certainly contributing to your problems, anyone would be hard pressed to develop a working model from this. Garbage in, garbage out. I’m guessing the tiny version comes from importing the .DWG once using model units which imports about 3" tall, then you re-imported using inches.
I think you would really benefit from following the step by step learning course at the SketchUp Campus. Get the basics down so you understand what the tools do and how to use them. Start with making a rectangle, make it exactly the size you want with %100 precision by typing in the measurements, then make that a box using push/pull and again type the measurement. Select all of that and make it a component so it is isolated from other geometry, give it a relevant name. You have made an exact sized working object, now make another… This is the beginning workflow, master this and then move on.
Another reason (thanks for calling me an ”expert”, I haven’t heard that in a while) there might be a confusion is because when a .dwg file is imported from another CAD environment, its Layers automatically become SketchUp Tags.
From what you say, I now understand why the whole Tag concept can be confusing for someone previously involved with layers. But it still begs the questions:
Q1) On a very simple level: Why apply a term more indicative of a 1 item (1 face, 1 edge, 1 object, 1 component)—a name of sorts—to a stage where multiple faces, edges, objects and components can reside?
Q2) Why use (multiple) Tags at all? Why not translate everything in a .dwg file into one “Layer”? I know it would be confusing, terribly confusing to someone who went through all the trouble of organizing things into different layers, but if that’s the way-of-SketchUp one might have to abide by the rules of that bible.
Q3) Have you ever considered that Labels are terribly terribly useful things. They allow one to instantly select/deselect large categories of a drawing (buildings from landscape; floor1 from floor2; things one feels comfortable about from new experimental or troublesome things…) Layers reflect a mode of organization! Is there anything (e.g., actual word tags which one can multiply assign to objects, etc. and choose to color, select/deselect, lock/unlock, or otherwise restrict/allow ) in SketchUp that might correspond to what the rest of us call Layers?
I now want to ‘encase’ each of the objects (on whatever level I believe an object to be useful) into their own ‘box’.
This is the way I originally began the concept-drawing in nanoCAD and the way a sensible person might want to explore a concept (construct the landscape before concentrating on the bushes and trees),
Q4) Is it even going to be possible to LOCK the cabinets that are going to stay in place and call them visible/invisible when one needs to reference them?
Q5) Is it possible to enclose the new items in boxes that I might be able to extrude down to separate sides, tops, and bottoms?
I’m going to try Riley’s suggestion, but it’s going to be very hard. I can’t, and don’t know of anyone that would, construct each item in a house without having any idea of the limitations, relationships, and existing items are in the house already. I have to start out with big boxes, exactly placed the way they are. Doing a forest by drawing the trees isn’t a CAD design project, it’s evolution.
I’m going to try your suggestion, but it’s going to be very hard. I can’t, and don’t know of anyone that would, construct each item in a house without having any idea of the limitations, relationships, and existing items are in the house already. I MUST start out with big boxes, exactly placed the way they are now and work from there. Doing a forest by drawing the trees isn’t a CAD design project, it’s evolution and that’s god’s problem.
I’m NOT telling you which objects to start with. Start with the foundation, then make the walls, then work on the kitchen. I would. Make geometry in whatever order you want. I’m telling you the basic workflow for successful modeling. Model things in place, to the correct dimensions, as separate entities. Start with an object, say the floorspace of the kitchen, make it correct, make it a component or group, then move on.
Yes, of course Labels and Tags are use-full things, they exist in SketchUp and they work great. They are a key part of good model organization. You do not understand how to use them yet and so simply jumping in the drivers seat and pulling random levers and turning random knobs is not going to get you where you want to go. Honestly I would not go down the Tags rabbit hole just yet until you master correct dimensioning and groups and components. Tags are not much use until you have that down.
This is what groups and components are for. Making raw geometry into a group or component is the only way to isolate that geometry and prevent it from interacting with other geometry. Every object in your model should be a group or component. This is why I have suggested you master the basic workflow above.
This why the campus tutorials are a good place to start, because they introduce the core concepts in a step by step manner that builds on previous knowledge, instead of guessing at how the program works or using inapplicable knowledge from other programs.
I can’t justify any of these things. I am just an unpaid volunteer who tries to help out here on the forum, not a SketchUp employee or developer. I didn’t create or have any input regarding the design of the dwg importer. I can only try to explain it as best I understand, in the hope it will help others to use SketchUp effectively. I will say, though, that many of your questions and the issues in the models you have shown indicate that you don’t yet understand how to follow a good workflow in SketchUp. You are fighting with it trying to make it be CAD.
Yes. That is no doubt an effort to at least preserve part of the organization that was in the original dwg file. But it comes with potentially confusing consequences. Unless the geometry was in blocks in autoCAD (in which case they import as groups), edges and faces from different layers will already have intersected during the import and already be entangled despite the layers!
The term “Layer” originated in 2D graphics, image editing and CAD programs, in which it is natural to think of a drawing as composed of a stack of sheets or layers. You see the image as the cumulative action of the layers starting at the “bottom” and working to the “top” (even though those terms already imply a third dimension in which to stack!). Turning off a layer is like removing one of the sheets from the stack. So, people would say things like “let me see the plumbing layer”, which meant “remove all the other sheets from the stack”. With multiple layers turned on, the image is the “sum” of whatever they contribute to the image, often in a transparent overlay sort of additive presentation.
In a 3D interactive modeling app, the concept must be understood somewhat differently. Although you can control what parts of a model are visible based on which layers are turned on, there is no stack-like accumulation of effects in some particular direction. Things block visibility of other things based on what is closer to the camera from its current position, which can be changed interactively. So, the term “layer” is already misleading and used only based on pre-existing convention.
And in CAD, interaction between edges and faces is governed by whether you run commands to make them interact, not by what layers they are “on”. In SketchUp, they always interact unless you put them into separate groups or components, regardless of what layers they use.
And again, in SketchUp a layer is not a place where anything can “reside”.
I have long thought that each layer from the dwg should get a separate group during import. In fact, I have an unpublished Ruby script that does exactly that after the fact, though edges and faces may already have been fragmented. That way the layer-based organization of the dwg would be preserved without running afoul of SketchUp’s different handling of edges and faces. I have no idea whether that is feasible in the importer before the damage has been done.
Of course! I use them in every model I build. The issue isn’t to avoid using layers, it is to use them in accordance with their behavior in SketchUp else you face consequences. Build groups or components first to isolate edges and faces that should belong to distinct objects, then use layers to determine visibility of groups and components that are conceptually related. SketchUp is geared toward meaningful representation of whole “thngs”, not to pencil-and-paper-like representation of lines and polygons.
It is possible to lock a group or component, which makes it impossible to edit it in any way. The only “gotcha” is that assigning a layer/tag must be done before locking, as the assignment is a variety of edit. Turning a layer on or off is not an edit and can be done with a group or component that is locked.
I’m not sure I know what you mean by “boxes”. In a good SketchUp workflow, each actual object - a cabinet, a wall, a stove,… - is modeled as a group or component. That isolates that collection of geometry as an individual thing that others can snap to for reference, but can’t modify accidentally. So, if you have a space where a new cabinet will go, and it is bounded by other pre-existing cabinets, you can draw a rectangle on the floor for how it fits between the others, pull that up to whatever height, and make a group or component from those newly drawn edges and faces. Nothing except them will be captured in the new object. Then if needed you can open the object for edit and add more details.
You might not be part of the crew but you deserve to be.
Your previous tome (TLDR, which I had to look up) beautifully captures essences about the program that people deserve to know. Also your honesty (“SketchUp is lazy (and in some cases buggy”) shows that you understand the program under the hood—a valuable skill the SketchUp crew should respect. It brings back memories of some of the best programmers I knew while I was designing software and later taking on the responsibility of quality assuring it. Fortunately, I understand the implications of things like “clipping and z-fighting”.
If the program weren’t so ■■■■ addictive and if I hadn’t been in contact with some other users who praise its ability I might be tempted to let it go.
Right now I worry that without a coherent concept of layers, and a developers’ intention that people get used to a paradigm suggesting “all edges and faces should use Layer0/untagged. Layers/tags should be applied only to groups or component instances and non-geometric things such as dimensions, screen text, and images” that any file of even moderate complexity will immediately devolve into a rat’s nest of unwanted or misunderstood location snaps, and hours spent trying to search the proverbial haystack for the right edge or surface. Or, just as bad, that the preponderance of tags, if a new tag is going to be applied to each and every object will make tag management as complex as drawing the scene itself.
CAD has been a good thing for the world of design. I’m slightly fearful of the implications of your statement:
My impression with the tutorials I followed was that they were simple yet spectacular. They featured single/few-item designs and said little about the whys or wheretofors of using SketchUp in complex situations. They seemed to portray someone who knew exactly what it was he wanted to do and how he wanted to do it; spending his time on one particular thing (e.g., a table) until he got it exactly as he’d intended—not with someone confronting an existing situation and experimenting with different designs he might use to change that situation, shuffling them in and out of the general picture until he came up with the one that works best.
Bring up blank canvas in the correct view (no easy matter, since getting to the right view involves going iso and then you lose the label that tells you what view you’re using) and placing a 2D wall on it.
a. Labelling the wall component: Wall
Showing that canvas in overhead iso and placing a non-rectangular 2D floor plane onto it (non rectangular so that it can be distinguished from the rectangular wall)
a. How do I draw a rectangle with a chopped off corner? Can’t find a tool capable of drawing a non-regular rectilinear line except the Freehand line tool and that doesn’t seem to allow for straight lines at all! Can’t draw a polyline that’s not a regular object (like a rectangle with a chopped off corner).
b. Apparently I have to settle with another rectangle.
Go to overhead view and place a 30”x 30” rectangle (the refrigerator, closed) on the floor. (later I will resize it when I add and rotate its door)
a. Then I try to identify it:
i. Selecting it, right clicking, and going to Entity Info does nothing.
ii. Going to Edit/Make Component… allows me to label it “refrigerator” but the odd relocation of the UCS (for lack of a better word) at this point in time is totally confusing
b. Trying to rotate it is odd. I select it but can’t lock to its center and when I move the protractor nothing appears in the Angle field (not true, there have been times a number does appear, but I can’t figure out when) (The user guide only says: “Rotate Tool Use the Rotate Tool to rotate, stretch, distort, or copy entities along a rounded path. Activate the Rotate Tool from the Modification Toolbar (Microsoft Windows), the Tool Palette (MacOS X) or the Tools menu.”
i. )
Want to continue to place approximate objects in that view in 2D; and later change their dimensions; expand their heights; and move them into their current position.
a. How do I redimension a rectangle or any 2D object? Selecting the rectangle doesn’t bring up its current dimensions to change.
b. Can I click a rectangle and get its dimensions? Every time I try and do so I get its area which is less than useless.
c. Also, I understand the eye and the lock, but what do those other icons mean (no tool tips)
i. Oh, when I select the eye it—AND THE TOGGLE!–disappears completely. And when I select the lock how do I unlock it?
None of this is acting like a “sketch” process.
a. Having to input exact dimensions and place objects in exact locations reminds me of pre-GUI CAD programs and not being able to rough things in and later refine them to exact dimensions is daunting.
b. Also, having to make everything into a named-component is terribly time consuming.
I’m getting nowhere fast.
a. Currently, I was trying out the eye and lock toggle and succeeded in losing the floor and everything on it, with no clue how to get it back.