*If you're doing Hardware startup, this post is relevant to you. Otherwise, keep an eye out for my Software Prototyping post.
Hardware is a Marathon, not a sprint. You keep running. You don't know when you'll finish. You'll get tired in the process. You may slow down. Slow down so much, you will almost walk.
The patience you need to do a Hardware startup is not something your college could teach you. Neither could your parents or your girlfriend. When you'll see your fellow software entrepreneurs shipping products in the first month of inception, jealousy will hit you. You'll regret your decision. As no one knows what you're building yet, people will question you.
When your software friend builds a prototype overnight for free, you'll be churning out large sums of money just to procure components. No! Arduino will not help.
This is a big problem. My friends started Hardware startups without realising how much money they'll need in the process. No, $500 is not enough. How about $2000? You'll spend this overnight in getting your PCB iterations.
I'll try to narrow my suggestions towards money you'll need for building your first prototype:
1. Find that development kit.
This is your first prototype. It doesn't need to be manufacturing ready. Doesn't need to meet the dimensions of your final design. This is more of a proof of concept prototype. It is very amusing that 60% startups skip this step totally. Even I did. But after 6 months down the line, I had my realization: I made a mistake.
2. Components will burn your pocket.
Development kits will only get you the sensor on board with Bluetooth and chipset, supported with an SDK. If you have any other components that you need the system to support, like IR sensors, or LCD screen, etc, you'll need to buy them at really high prices because single units are pretty expensive.
In my own product, if a detector costs us $10 at 10,000 units, it costs us $90 at single unit.
3. It's time to code - Firmware.
This is the only place you could save some money. Find your EE or EC friend who could code for you. The way I used to do it, I'd go and visit multiple hostels at IIT delhi, especially Jwalamukhi and ask someone to do the job for me for free, and in return he would get the experience. Students at good institutes are hungry for work. Give them something meaningful and they won't disappoint.
CAUTION: Don't ask this guy you meet randomly to be your cofounder. If he asks for money in return, give him Rs. 5-10k. (Approx. $100) You think that's expensive? Try to get that done in US. you'll pay $140 per hour minimum. In China, where it is hell cheap, it is $35 an hour. So shell out some money.
So in a budget of roughly $2000, you should have a working prototype. Now, this whole suggestion is depended on the assumption that you're doing something that is already out there and the development kits for the same are available. If you're trying to do something new, where even the components are hard to source, (As it was in my case) pray every morning and, instead of praying to god for mercy, ask your parents for some money.
In my next post, I shall talk about budgeting for your idea that requires development from ground up. Till Then.....