For a new design is it best to start off with DIP parts and get everything working on a bread board? Some parts for my design are only available as surface mount parts, so prototyping becomes more complicated.

Do most of you guys just create a PCB board and cross your fingers that everything works properly?

Thanks for the info!

-Henk

Hi Henk

I hope its a bit more hard work than just crossing fingers

For high speed memory and a processor like the AT9200 DIP prototyping can lead to more problems than it solves.
Any area we are unsure about we talk with technical support of the chip manufacturer and get sample schematics, until we are sure we understand. A week behind schedule to prototype saves time compared to having to redo a design.

After a design is complete we spend about a week checking and cross checking our design, net lists and layout issues (3 engineers) before doing a limited run (usually 4 to 10 boards).
Then we do the EMC checking before going to a production run.
At the end of the first 1000 boards we do a review and discuss with production and purchasing to highlight any modifications in Bill of Material or placing issues

Last year we did 4 designs using the AT9200 with video decoders, network PHYs TWI and external IO ports. Each one worked first time. For two we did slight changes to pads for production after 1000 units to reduce build cost.

Once we have an area of board working, e.g. processor, flash, DRAM and power supply we lift it complete to a new board. This minimises risk.

We have not had to scrap a design for about 5 years. But I still keep my fingers crossed when a new board arrives.

Simon

Thanks for the great response!

-Henk