My first computer was a Litton Automated Business machine. It used drum memory to store data and had an instruction register and maybe two other registers. I purchased it for $100 my first summer home from college.
It was a remarkable machine. It was fun to work with, but you really couldn’t do much with it. It had dual paper tape readers, a printer, and a paper tape punch. I wrote an inventory control program for my father with just that, in something that looked very much like machine code.
It wasn’t a usable machine for my father.
The computer I told my parents to get was a Macintosh. They just worked out of the box. Plug them in and you had a word processor, a paint program, and I think a spreadsheet. It all just worked.
They got a PC and fought with it for years.
Today I can buy a piece of hardware, load an operating system on it, and have it fully functional as a general purpose computer or acting as an embedded machine in just about an hour.
The biggest time sink is removing and inserting screws to hold everything in place.
3D Printer
10 years ago or so I purchased a 3D printer kit. I’m sure I never got a fully successful printout of that damn thing. I had to do so much to just get it to do something. I spent more time trying to make it work than I did printing. And it was fragile.
Today, I believe that the kit instructions had the count of the number of teeth on one of the drivers wrong. Which meant that cubes were squished.
Today we have Macintosh printers. After 6 months of research, I pulled the trigger and purchased a Bambu Lab’s P2S printer with AMS.
Setup took around 2 hours. Every step was clearly documented. All the tools to do setup were included. Mostly setup consisted of removing packing, tape, and shipping screws.
Thereafter, it was plugging in one cable, 2 tubes, and the power. Turning on the power brought up the screen that forced me through an initial setup process that calibrated everything.
Finally, I pressed a few buttons on the control panel, and it printed a tool.
From there I used the phone app to scan a QR code, which took me to a cloud version of a storage box to print. That’s printing as I write this.
Again, there is no effort on my part to do any of this. It is pick, click, and print.
Calibration
My kit had no bed leveler. That was done by putting a piece of paper on the bed, lowering the nozzle until it just touched, and then clicking the next button to repeat. I think there were 20 or more sample points. And I still don’t know if that was enough.
The automated version required the printer to be working to print a new piece for the printer, which would hold a switch. The switch used a paperclip as a probe. This went much faster, but it still didn’t work.
The motors were noisy, but it was a joy to watch them move the hotend around.
Today’s calibration took around 45 minutes. This included using the built-in lidar to measure the distance to the bed, and then I think it used a pressure sensor to determine when it actually touched the build plate. It took samples every centimeter or so in a grid. That went rapidly.
It then went into a noise tuning calibration. For 20 minutes it ran the hothead around in diagonals, working to find the correct stepper speeds at different head speeds and then tuning them to be quiet.
It worked. These are stepper motors you can’t really hear. It blows my mind.
From there it did vibration calibration. This thing can accelerate so fast that it will cause the printer to move. For every action, there is an equal and opposite reaction.
This thing figured out, for this printer, on this surface, just how much the printer reacted to head movement to be able to offset that motion during the deceleration stage.
The first print after calibration took only a few minutes to check calibration before it started printing.
Slicers
To create a 3D print, you start with an idea. You build a 3D model. I use FreeCAD; people use many CAD systems; Fusion 360 is a popular one.
Once you have a “solid”, a completely closed volume, you export that solid as an STL or a STEP file.
An STL is a triangulated file; a STEP file still retains geometry. For example, an STL file will represent a cylinder as a mesh of triangles, while STEP represents the same geometry as a cylinder or as a curved surface; regardless, STEP is the cleaner format.
Now that you have a surface representation of your solid, you import that into a slicer. I’m using OrcaSlicer which is a fork of the Bambu Studios.
This allows some manipulation of stl/STEP objects. The important part is to position the object on the build plate with no overlaps. Once that is done, you can slice the volume.
This is where things have come so far.
The solid is sliced into layers, generally 0.2mm high. The slicer then calculates the path of the print head over the object at the same height. It knows where edges are and uses loops to make solid walls, it adds internal fill to keep the print light yet strong.
3D prints can’t print in thin air, sort of. They can span short distances before the plastic droops too much. To print with an overhang, or to put a top on something, the slicer has the hot end create a raft across infill or across supports. Once that layer is completed, it will put a more finished layer, then an actual finished layer.
The slicers are pure magic. It really is easy.
All the hard work remains back in the CAD package, which is the same package I’m using for all my other engineering builds.
If you are interested in 3D printing, decide why you want it. Then pick any of the plug and play printers out there. I strongly suggest getting one with an enclosure. An enclosure will be needed for certain types of filament.
A good set of starting projects are GridFinity, an organizational system for flat surfaces, including shelves and drawers, and MultiBoard, which is a hyped up pegboard system.