Wednesday 28 October 2015

design - How can I go from prototype to commercial product?


I have made many designs that I would consider "professional grade", but in reality they are still just prototypes. I have used professional board fabs and professional population, but what I haven't had the ability to go through is actually turning that product into a commercial product.


What is needed in order to take the product across this large last step? I can think of obvious things like FCC/CE testing, case design, marketing, distributors, etc. but I am sure there are other things that aren't thought of until going through the process a time or two, as an engineer, what are these things that would come up?



Answer



In my industrial experience, the way things generally go in terms of a design cycle follow this sort of flow (condensed version):



  • Market research/contact with a customer (identification of the what the basics of the product should be)

  • Establishment of a specification (or "spec" as it's often abbreviated)


  • A formal quotation and business agreement between the parties (or between engineering and marketing for a 'standard' product)

  • Initial design / calculation / simulation / debugging

  • Design validation testing, or DVT (you making sure the product meets the spec)

  • Bugfix implementation / regression test

  • Independent DVT (someone else making sure the product meets the spec)

  • Safety and regulatory approvals

  • Design for manufacturability review (DFM)

  • Production release / marketing release


If you have a project that you've developed, and you feel you're ready to market it, you may want to consider some or all of the following points:




  • Write a spec if one doesn't exist. Characterize the typical performance of the product and put this in the spec, as well as any absolute maximums or minimums in terms of 'externals' (volts / amps / degrees) that should be avoided.

  • Make sure that your product meets your spec. Do a formal DVT and make a report of the results. Also, test a quantity of boards against the spec and use statistical analysis (Cp/Cpk) to prove that over tolerances and normal variation, you product can do what you say it can do.

  • Come up with a reduced number of tests to perform per-unit as a manufacturing reliability test. Bad builds, bad parts, new operators, bad equipment can all lead to problems.

  • Make sure that your PCB has in-circuit test pads on as many nets as possible to facilitate in-circuit test (ICT) or manufacturing defect analysis (MDA) - finding problems earlier rather than later is always beneficial.

  • If this product is intended to be high reliability, you may want to perform a life test. Run a quantity of units perpetually until something fails. There are industry standards on life testing (acceleration factors, lot size, etc.) which fall beyond the scope of casual advice.

  • Consider HALT (highly-accelerated reliability testing) as well - testing the product under thermal and mechanical vibration stress can show weaknesses in the design as well as its construction.


Many of these sorts of tests are big bucks. (Tens of thousands of dollars per test). The cheapest things you can do are those that you can do yourself - the spec, the DVT test and statistical analysis, test pads on the PCB, for instance.


You may wish to consult with an experienced manufacturing engineer to figure out what you really need to do, given your budget and the expected revenue of the product.



Good luck!


No comments:

Post a Comment

arduino - Can I use TI's cc2541 BLE as micro controller to perform operations/ processing instead of ATmega328P AU to save cost?

I am using arduino pro mini (which contains Atmega328p AU ) along with cc2541(HM-10) to process and transfer data over BLE to smartphone. I...