When I first started trying to design things I had no idea what I was doing, some would say I still have no idea...
With electronic circuits there are two ways of designing that I know of:
Designing from scratch is much harder than basing your design on known circuits, you need to know not only how electricity and electronics work but also what parts are out there to use. When I design from scratch I usually end up reinventing the wheel (or at least a lot of the spokes of the wheel) because there are tons of ICs and circuit blocks I am not aware of and most of the time there is a much better version of those spokes out there. For instance when I first started playing with logic I built all of my own gates using transistors and resistors (all logic gates can be made using discrete, analog components) and my circuits were probably less robust and definately FAR larger than they are now that I use logic ICs. In a way I think doing it the way I did helped me because I know what is going on inside those little black chips we use and now if I need one gate I don't have to waste a whole chip on it, but the fact remains that I was using a very inefficient method.
Designing based on known circuits is far more efficient and usually results in more reliable and robust designs, however there are some different problems to consider. I find circuit blocks in datasheets, published amateur circuits and published professional circuits. In all three cases I have to consider whether the circuit actually works, works the way they say it does (if they even say how it works), and if it can be adapted to fit as a block in my circuit.
For instance the 555 timer IC and the associated circuits from the datasheet are incredibly useful in a huge number of projects and are very general, but you have to be careful about using them in very time sensitive situations because they rely on a capacitor as part of the timing mechanism. The voltage leak across capacitors of the same value and make can vary pretty widely and, worse yet, it changes with age! So if you are using two 555 timers you have to be aware that they might not have the same timing even if you use caps from the same manufacturer and of the same value. This issue gets worse the longer the time delay becomes.
Using circuits built by both professionals and amateur often gets you more robust circuitry because the person has worked out many of the problems and often comments are included discussing them, but these circuits have been specifically built to perform the function the creator had in mind. Often I need a circuit to work a lot *LIKE* the circuit I find but not *EXACTLY* like the circuit. With general datasheet example circuits the engineer who draws them up is trying to make the part the circuit is based on seem promising for as many different applications as possible because the whole point of a datasheet example circuit is to sell the part, whereas a circuit designed for a specific purpose really locks down everything so it works as well as possible for the function it was created for. Testing and tweaking become very important so the circuit will work in your design.
I am not a lawyer and these are only my opinions, but I pretty much go with the idea that if I only used datasheet examples as the *inspiration* for my circuit and did not even use a bit of a circuit published by other designers, I can probably call it my own circuit. If I want to adapt and use blocks from other people's circuits in a *PERSONAL* project that will never be released in any way then I can do so but I can't call it my own. If I want to release a circuit into the public domain (I don't sell circuits, it's a hassle) and there is any part of it based on someone's circuit I must contact them, show them exactly what I changed and kept the same and obtain their *EXPRESS* *WRITTEN* permission.
I don't know if these are the legal obligations I have but they are the moral obligations I feel. Frankly, I feel that most of the problems we have in society are due to people lacking a feeling of moral obligations and lacking a self expectation of being a respectful and respectable person.
So now you have your circuit schematic figured out and you tested it on a breadboard finding that it worked perfectly the ver first time right (*grin*)? Now you need to physically construct it. The first thing I try to do is figure out what it is going to be housed in. A logic probe built into a pen requires very different design considerations than a function generator built into a desktop case. Many people either make their own PCBs, have a company make PCBs for them or use pre-made generic PCB boards. I don't do this much. I started out using perfboard and gradually my projects got bigger but I still haven't reached the point where I feel point-to-point wiring on a perfboard is completely unmanageable. I've come pretty close, though, a recent project I made has over 130 pins and is on a piece of perfboard 2.5" X 2.5". While professional designers may find it simple to do that amount of point-to-point wiring in that amount of space, I would hazard a guess that most people would find it at least a bit daunting.
If you're going to use either a homemade or professionally made PCB then you'll do all your layout in a PCB design program, but if you go with perfboard or generic PCBs then you can do your layout on paper or right on the board. I prefer to do it on the board because once I get it set up how I want I can just start soldering and also because it avoids problems with whether everything will fit physically. I prefer to layout a design so that all the signals come in from one side and leave on the other side, and I like to seperate the circuit into functional blocks.
For instance if I am building something which takes in audio information, processes it, and the sends out the processed audio, I would start on the left side of the board by setting up the input circuitry (filters, jacks, etc.), move to the middle of the board for the processing circuitry, and use the right side of the board for the output circuitry. Now I break it up further and look for obvious problems, for instance do I have signal lines crossing all over each other which may cause interference? Do I have digital seperated well from analog? Do I have access to everything I am going to need to get to later? Do I have components physically close to things which may impact them? After these sanity checks I go on to figure out where my power and ground rails ar going to be, usually by now it is pretty obvious where the best placement is.
For the final phase of layout I ask 'does it look nice?' Usually if it looks symmetrical it is more visually pleasing and it works better for some reason, the blocks on each signal path tend to be near the same blocks and the capacitive and inductive impact on the circuit tends to be more symmetrical I think.
Now that it is built it is time to stuff it into the box... Not so fast! Tons of things go wrong between the breadboard and the built circuit, anything from making a bad connection to putting something too close to something else which makes one of the components act funny. So before you go any further test it on the bench where you can get to stuff and see everything. It really stinks to get a project fully built and find out that you have to take it all apart to figure out why something isn't working!
For me this comes into play mostly when I am working on robotics, and I consider the skeletal structure of a robot to be a form of an enclosure. I tend towards materials like brass and plastics because they are reasonably easy to physically work with. Brass is easy to solder and it can form a very solid structure if you put in the time to figure out where the braces should go. Plastic is easier because it can be glued, but I don't like how brittle it can be. I personally favor using functional parts (motors, circuit boards, etc.) as integrated parts of the skeletal structure, in fact StickyBot's entire skeleton is made up of the servos (gear motors) glued together. Be sure that if you do this you do it in such a way that you can take out components for replacement, nothing is worse than having to completely start over because you glued a circuit board to the motors and now need all new motors and circuit board since the one part that died is impossible to get to because it is under the glued on motor!
Use a CAD or paint program to draw the shapes you need to cut and then paste them onto the brass to use as guides. it takes a little practice but once you get the hang of it your parts will come out much more evenly and usable than just cutting by eye or even drawing on the brass itself. First, though, cut out the pieces from cardboard and fit them together, you'll hate to have to go back to the metal store to get all new stuff because one piece didn't fit right and you couldn't modify the design around changing that piece. I usually buy my metal from either a local hobby store or a really good hardware store.
When working with brass keep in mind that even if the sheet you are using seems flexible once you start adding bracing it will solidify quickly, often thinner rods and sheets are fine. You'll want to have a small torch, lots of solder, flux, visegrips and other clamping tools, hammers, punches to make drill guides, a carpenter's square, levels, lots of measuring equipment, a Dremel type tool with lots of attachments (at least cutting wheels, drill bits, sanding wheels and grinding wheels), an extremely well lighted and ventilated work area with surfaces you can cut on and weld on, cutting oil, and most importantly a fire extinguisher (or at least a bucket of water). I recommend an ABC fire extinguisher because you'll be around all sorts of chemicals and the last thing you want to do is throw water on a grease fire!
About eye protection: Don't rely on your perscription glasses if you wear them, they are not made to deflect flying pieces of metal and often don't protect far enough around the edges of the eye. You'll be creating LOTS of metal shavings and sooner or later you'll slip and a little piece of metal will fly off straight at your eyes. It is worth the money and aggravation of using good eye protection and retaining your sight for the next project! They're like a seatbelt in a couple ways, if you always wear them then it becomes a habit and you just put them on without thinking, the other way they're like a seatbelt is that the one time you need them you REALLY need them!
Venting the fumes out of your work area is also very important, much of what we do involves pretty hazardous stuff. Imagine sitting there and soldering for a few hours, you're melting a lot of lead right under your mouth and nose and you need to get pretty close to see what you're doing. Do you really want all those fumes in your lungs?
When you're using power tools think about what awful thing could happen if you
slip. Don't drill over your legs or cut really close to your hands, use vise grips
to hold your work so if the cutting wheel slips it just nicks the vise grips (they're
a lot more immune to being sliced up than your fingers). Mostly safety is just
about taking a moment to think "Am I about to do something incredibly stupid?"