Design development costs for FPGA on PCI board (sorry if sli

[AndyAtHome]

Dear All,

After investigating the price of FPGA PCI boards for a project, it
seems the cost of a board is typically 3 to 4 times the component cost
for mid-range FPGA's like Xilinx 2V3000.

Of course board manufacturers have overheads and need to make a profit
- my point isn't to complain about their Gross Profit margin.

Rather, at what point is it better to meet the initial capital
expenditure of having your own board designed, thereafter only having
to meet board manufacturing costs?

For example, just how much design time and cost does it take to create
a board with -

Quicklogic or PLXTech PCI Bridge
Single FPGA
A PROM
Single bank of DDR memory.

Surely this type of requirement is common and would not require months
of development time or am I being naive?

Any comments appreciated...

Thanks,

Andy.

 

[Andy Peters]

fpgadev@yahoo.com (AndyAtHome) wrote in message news:<29ab33c4.0404290413.2b5a2be5@posting.google.com>...

Dear All,

After investigating the price of FPGA PCI boards for a project, it
seems the cost of a board is typically 3 to 4 times the component cost
for mid-range FPGA's like Xilinx 2V3000.

Of course board manufacturers have overheads and need to make a profit
- my point isn't to complain about their Gross Profit margin.

Rather, at what point is it better to meet the initial capital
expenditure of having your own board designed, thereafter only having
to meet board manufacturing costs?

For example, just how much design time and cost does it take to create
a board with -

Quicklogic or PLXTech PCI Bridge
Single FPGA
A PROM
Single bank of DDR memory.

Surely this type of requirement is common and would not require months
of development time or am I being naive?

Any comments appreciated...

Well, it really depends on how many boards you need, and whether the
hardware design on the off-the-shelf board meets your requirements.

Regarding the former, it's simply the classic make/buy decision. If
you need a handful boards, buying them makes sense because you can
concentrate your development efforts into the FPGA and you know (or at
least hope!) that the board design itself is stable. If you need a
lot of boards, then it may very well make sense to spin your own.

Rolling your own board requires a different skill set from designing
an FPGA, and if you don't have those skills (and the software tools),
then you have to hire them and that cost may bust your budget. You
also have to learn about (or have in place) the entire manufacturing
process, including purchasing, board fab, stuffing, everything, which
will distract you from your FPGA design

Of course, if no off-the-shelf board meets your needs, then your
choice is clear.

Besides, a product that costs only three times the bill-of-materials
is kinda cheap

-a