You have an idea - now what?
By Mike Tripoli
More times than I care to think about over 30 crazy years designing products, I've run into the same old mistakes being made over and over again. Let's start at the beginning with your "idea".
You have an "idea" for a product, and let's assume (ahem) that you've done your due diligence and know for a fact that your idea is better than, or adds to, a product already on the market. That in itself many times get's one into trouble. Everyone wants to believe that they've come up with a truly new, original idea, believe in it so much that they don't research the market. The first thing I, as a designer will ask you is, who is your competition? What product area are you going after? What is it that you want designed into your product that is different? If the answer is "nothing", that's called reverse engineering and a completely different task.
I tell people all the time, you can so this yourself, quickly, without paying a dime; Amazon and Alibaba.
Inevitably, if I'm given the task, I can find a product that already does what yours does, or something very close, on Amazon or Alibaba. Does that make it a bad idea? Of course not! What it does tell you is that there is a market for your idea (GREAT!) and that you already have competition (not so great). What you have to decide is, does your product add anything to the mix, or greatly improve on a product? This is what the industry calls "value added"; it's the thing that makes the consumer (your target customer) think to themselves "This product offers something extra that the competing products do not" and make them want to purchase it, right now, not later. This, by the way, will also be what a potential distributor or licensor of the product will look for.
So now you've done some research, and ready to move forward. First thing to do is get some 3D models done in CAD and 3D print them, throwing in any of the myriad little electronics modules available, right?
WRONG!
The first thing you do is write a specification for the product. This, I know, sounds daunting. But in reality, it means nothing more than jotting down a list of "features", sketching out what you think it might look like (you don't have to be an artist - that's the job of an Industrial Designer), writing out short paragraphs of ideas. It doesn't have to be perfect; you can add or subtract at will. What it does do is give everyone involved something to work towards during development. Far too many projects died an early death because everyone wasn't on the same page from the start.
Once you've created your "spec", you're now onto a prototype. Read on regarding this often misunderstood and mistreated crucial step in your product development.
Prototypes - What you need to know!
By Mike Tripoli
You have an idea for a product, and the first thing that everyone starts saying is "Do you have a prototype?" Before you start spending money and wasting your time, let's talk about what you should expect from a prototype...
When we started SCARY Design, dinosaurs roamed the Earth and we foraged for food...
No... but it sure seems like it! The fact is, Microchip was a "new" company, and had one microcontroller, the PIC16C84, their first "microcontroller" with on-board EEPROM. In order to "program" it, one needed an in-circuit-emulator (ICE) with a "bond-out" IC. This had to be soldered into the circuit (if using SMT), firmware developed, then, cross your fingers, a "chip" programmed and soldered into the circuit. Things weren't as stable as they are these days; it was not uncommon that the internal oscillator wouldn't start, forcing all kinds of capacitor and resistor changes until it started running. Then, you better hope you didn't create a bug or two for yourself; you'd have to un-solder the part and start again. Thankfully, things have progressed very far in short time. Now, we can build the microcontroller into the circuit and do "in-circuit serial programming" - ICSP. This gives us great flexibility; we can change firmware, or how the microcontroller works, with a code update. BTW, this is also how an "Arduino" works, sortof'...
Things weren't much better on the "hardware" side. If you wanted an actual "looks like" prototype, it meant doing design in 2D (!), drawing multiple views, then, if you had the capabilities, cutting the model from any number of various "modeling woods" like Renshape, or having a "model shop" do it for you. These were often times referred to as "looks like - works like" models, with no real application to production engineering. Can you spell e-x-p-e-n-s-i-v-e?
Fast forward to today; any number of little microcontroller modules are available, with cute names like "Arduino", "Raspberry Pi", "BeagleBone", etc. Cut material? No way, we simply "print" what we like, warts and all. It's the "warts and all" we need to be concerned with...
Given how radically things have changed, "looks like - works like" prototypes should be a thing of the past. Yet I know of very few design houses that will actually state that the work being done on a "prototype" is applicable to real world engineering! Not so at SCARY Design!
Here, everything undertaken on a prototype has a direct correlation and application to the production design. We will guide you through the miasma of available electronics (after all, that is our jobs!) as well as all 3D design files, whether it be plastic injection molded parts, sheet-metal or wood; everything done is targeting a production design. Does this mean we can't make changes? NO! What it does mean is instead of simply "throwing a 3D model together" with none of the features needed to say, make an injection molding tool, like "draft" (draft is the term for slight angles in a plastic part that allows it to easily come out of the injection molding or IM tool easily); it is not uncommon that a "prototype designer" will skip over this critical step. Why? Because if they don't understand IM'ing from the start, they think this can be "added later". For injection molded parts, this means including all the correct draft angles initially; not thrown in later as an "after-thought". All 3D printed parts should reflect an actual production part. Many "design houses" want to do this type of design later, referring to it as "prototype design", then "production design". Why are you paying twice for what should be the same thing?
When it comes to electronics, we will discuss various approaches. Being "firmware" programmers, we'll suggest an appropriate device to be used for both prototype and production. If you want to use one of the available "modules" we'll work with you on it.
In many cases, we've seen an "Arduino" used where a very simple 8-pin device is better.
We try and and make sure that what we're providing will follow to production. Why try, and not do? Glad you asked, Yoda. Sometimes, there may be a device available that can save substantial amounts of money, but not available through standard distribution channels, mostly by programmers and production manufacturers overseas. In this case, we provide a specification and if needed, algorithms that will be used by the programmers of these devices, ensuring that the same functionality is in the production design as in the prototype. Leave you hanging, working with a company overseas? Not ever! We'll be there every step of the way!
So, the next time you're interviewing prospective design houses, ask them if they're doing "prototype to production" designs, or if it's handled separately. If they start waving hands about, making excuses that "things are always changing" or any other such nonsense, tell them that SCARY Design designs for production; no one get's rich "selling prototypes".
Product renderings - What and why.
By Mike Tripoli
Have a look at the above image.
Is it a photo of some "guitar pedals"? No, it's a hastily thrown together "render" of a product I'm working on at my sister company, analog•efx. I say "hastily" as this render is far from "photo-realistic". For instance, the LED's are really light-pipes in the "real-world". Here, I've simply assigned a "light" to the light-pipe model.
The software used to create this render is called "Maxwell Render" and being an "un-biased renderer" (it literally calculates the light rays in the scene as if the scene was being photographed. Other renderers' "fake" things, sometimes pretty well, others, not so much) is capable, with the proper setup and render time, capable of creating images one can't distinguish from an actual photograph.
Renderings are a tool we have available allowing us to take (usually) the 3D models created in engineering software, in this case Solidworks, assign "textures" to the models, then letting the computer do calculations that, not too long ago would have been outright impossible.
Typically, these renderings are created once the design has stabilized, allowing the designer to try on different color schemes, textures, etc.
Not every project requires renders, but when needed, are just one of the "value added" services we provide.
Printed Circuit Board Design - Just a matter of having the software?
By Mike Tripoli
I worked with a guy at one point, a "manager" that had never really done design work. He was forever coming into my office and "looking over my shoulder" to see what I was working on.
One day, while jumping back and forth between OrCAD (electronics CAD) and Solidworks (mechanical CAD) he said out of nowhere, "So, if I had this software, I could do what you're doing". "Sure, I suppose, but the first thing you should work on is making a time machine", I replied. "Yeah, why?" he asked, with a smarmy tone in his voice. "Because, you'll have to go back to when you were 14, which is when I started laying out printed circuit boards on grid paper with a pencil, then transfer that design using "rub-off" traces and donuts directly onto a single-sided piece of copper-clad". Then, I continued, "Spend a bunch of your time doing drafting at home and in school, along with metal and wood shop, learning how to make the stuff you drew". He wanted to say something else cute, thought better of it and left. I thought this was an isolated incident with a "wanna-be" designer. But the more I'm approached by other "designers" the more I think this kind of thinking has become the norm, instead of the exception.
This is the danger of the "We have the software, so we know what we're doing". Um, no.
More often than not, I see PCB's where it literally looks like they dropped the component footprints onto the screen and just "started drawing lines" between pins. Now, will this work? I always ask
"Define 'work'"...
Sure, you may make some boards this way, and it may in fact do what you want of it. If it's a µC board, with all "digital" circuitry, chances are it'll "work". However, once you begin FCC testing (you did know that anything that has a clock frequency of 8kHz or higher must be FCC tested, right?), you'll find out very quickly the difference between "having the intern layout PCB's" and hiring a seasoned pro. If it's an analog or mixed signal layout, well, there are lots of books written specifically on those topics; I can't scratch the surface here. There's a reason, to this day, that "auto placement, auto-routing" of PCB's still only works on the most expensive layout packages. And even then, you better have an experienced person on staff that understands the paradigm of the "auto" whatever, and knows how to correct things when they go wonky (I once watched an autorouter get itself into a corner, where it then created a cornered spiral into itself, then popped a via through the board, shot off and came back up again. Neat...). At the end of the day, it's a lot science, with the "art" thrown in there to keep things interesting. If you don't know where the line is separating science and art, or understand the difference between high and low impedance, and how it effects a board layout, best to call someone. Don't get me started on "little switching power supplies" on PCB's...
Are PCB's created here at SCARY Design better than others? Yes. For the most part, the answer is a resounding yes. There are people out there that do know their stuff; honestly, the ones I've met have been at the same jobs for years. Why? Because they're worth their weight in gold and a company employing a good layout person would be foolish to let them go. Other than a few, I'll state it as fact; I do better work than 95% of the "PCB designers" out there. So, pay $10USD/hour from a "freelancer" site, or get it done right the first time.