On the welding table lies a drawing of a platform frame. The welder turns the sheet a quarter turn, hunts through nine views for the position of one gusset plate, finds two dimensions he has to add up himself and writes the sum on the steel with a marker. The drawing is complete, no argument there. It just carries three times more than he needs, and the one number he is looking for is not on it.
A readable technical drawing is not the most complete drawing, but the clearest one: the right view, dimensions where the maker looks for them, and nothing that distracts from the work. In this article we look at drawings from the workbench instead of from the screen. What does a welder, saw operator or machinist actually read, what does he skip, and how do you produce a sheet that works at a glance with the same effort? A toolbox like Thundercad takes away some of the clicking, but readability is above all a matter of choosing.
Read your drawing the way the maker reads it
An engineer reads a drawing as a summary of the model. The maker reads it as a work instruction, and every trade looks for something different. The saw operator wants lengths, quantities and mitre angles, preferably without having to understand an assembly first. The welder looks for positions, setout dimensions and weld information, and wants to recognise the parts lying on the cart next to him. The machinist thinks in setups: what is the reference face, where is the hole pattern, which tolerances actually matter.
Look at your own work through those eyes and you see it immediately: many sheets are written for the checker, not for the maker. Everything is on there, so formally the job is done. But the information is grouped by how the model was built, not by how the work is carried out. That difference decides whether someone searches for ten seconds or four minutes, dozens of times a day.
The right view is half the job
The reflex is: more views, more clarity. In practice it works the other way around. Every extra view is another area the maker has to scan, and every section that adds nothing costs reading time. A good starting point: pick the view in which the maker sees the workpiece the way it lies on his bench or sits in his machine, and only add something when the shape is otherwise ambiguous.
An isometric view deserves a special mention here. It rarely carries dimensions, but it answers the first question of every maker: what is this thing and how does it fit together? One small isometric in the corner prevents more phone calls than a fourth side view ever will. Sections, on the other hand, are suspect by default: any section you cannot tie to a concrete question from the shop floor can usually go.
Dimensions where the maker looks for them
Readability lives to a large extent in where the dimensions sit. Three choices make the difference. First: group dimensions by operation, not by view. The machinist who finds everything for one setup in one place does not have to page around. Second: dimension from the reference the maker uses, such as the stop face or the centreline, so he never has to add numbers up himself. Every sum written on the steel with a marker is a risk you could have removed on the sheet. Third: put repeated features, such as a hole pattern, in one detail with a quantity note instead of dimensioning them again in every view.
Mind you: this is about readability, not about mistakes. A dimension in the wrong place costs search time; a dimension that is missing or ambiguous stops production. That second problem has its own article: Dimensioning mistakes that hold up your shop floor.
You buy time for these choices by shrinking the routine work around your drawing package. The Thundercad toolbox takes that clicking off your hands and can be tested for a month, no strings attached.
Try 30 days freeCut what nobody reads
Every drawing collects ballast over time: the complete parts list on a sheet only the saw operator uses, a revision table going back five changes, notes that applied to a previous project. The maker learns to ignore that noise, and that is precisely the danger: someone who gets used to skipping will one day skip the wrong thing.
Lack of space is a creeping cause of unreadability too. When dimensions collide and views are shoved against each other, the scale is too small or the sheet is too full. Splitting across two sheets often beats cramming, and sometimes a larger sheet is the simplest fix: with Sheet Size Up from Thundercad you switch the sheet size in one action, so an overcrowded A3 becomes an A2 without fuss.
Before and after: three sheets rewritten in words
How this plays out is easiest to show in examples. Three recognisable cases from machine building and sheet metal, before and after the readability pass:
| Situation | Before | After |
|---|---|---|
| Welded platform frame | Nine views, three sections, every tube dimensioned in every view | Front and top view with centre distances, small isometric, weld info at the seam |
| Sheet metal cabinet | Flat pattern and three views mixed together, bend lines without explanation | Flat pattern with bend direction and order, one view of the end result as a check |
| Milled part | All dimensions from one corner, scattered over four views | Dimensions grouped per setup, hole pattern in one detail with a quantity |
In all three cases information came off the sheet, and the drawing got better for it. That is the core: cutting is not a loss of quality as long as you cut what nobody was reading.
Make readability a team agreement
One readable sheet is nice; a readable drawing package takes agreements. Otherwise the style depends on who happened to do the detailing that day. A short set of questions every drawing has to survive before release works better than a thick manual:
- Can the maker tell within ten seconds what this part is and how it is oriented?
- Are the dimensions grouped by operation, measured from the reference the maker uses?
- Does every view and every section have an identifiable reader? If not: out it goes.
- Is there room, or are dimensions and lines colliding?
Agreements like these cost a little detailing time at first and pay for themselves daily afterwards. How to speed up detailing in the meantime without giving up quality is covered in Faster detailing without losing quality: readability and pace turn out to combine just fine.
Frequently asked questions
How many views does a good drawing have?
As few as needed to define the shape unambiguously, and that is often fewer than you are used to. For a lot of welded and sheet metal work, two views plus a small isometric will do; details are only added for a concrete question from a concrete maker.
Is an isometric view just decoration?
No, as long as you use it as a reading aid and not as a carrier of dimensions. The isometric answers the first question at the bench: what is this and how is it oriented? That prevents interpretation errors no amount of extra dimensions can catch.
How do I find out whether my drawings are readable?
Ask the people who work with them, and above all watch while they do: paging, measuring and mental arithmetic are the most honest signals. And if you want to remove lack of space as an excuse, you can try the toolbox with Sheet Size Up free for 30 days.