ASIC Help Needed

Does anyone here have experience in converting a circuit to ASIC?

It's a fairly simple design. Nothing esoteric. Here are my questions:

1 - Who does this? We have zero experience here, and will just farm it out. (We're located in the US.) A search on ThomasRegister yielded 96 "hits", but have not done much with that list yet....

2 - Is there a ballpark guesstimate of what a per-piece ASIC equivalent might run us? For example, if our prototype (using Digikey onesies-twosies) ran $4.50 US including the board but ignoring labor, what's a reasonable ratio?

3 - Finally what about ASIC development costs? I'm presently guessing $250K US to get it converted and then we'll buy units from there. (In other words, the non-recurring engineering costs) Does that seem a reasonable estimate to anyone??

What has been your experience with similar projects? Thanks!!

-mpm

Before I answer _your_ questions, here are some questions that others will ask:

Is it all logic, all analog, or a bit of both?

What are your volumes? You may find that a PLD or FPGA is cheaper.

On to some answers:

1 - Lots of folks. Jim Thompson, on this list, does it. I have an acquaintance who does a lot of mixed-signal IC design & could get you in touch. There are more beyond what's in the Thomas register.

2 - Your ASIC cost will depend a lot on what you want your circuit to do, and your volumes. If that $4.50 is $4.00 for a precision op-amp and $.50 of passives, then your ASIC will cost as much, in quantity, as that precision op-amp. If that's $4.50 of 74HC logic, then you can expect it to be much cheaper, indeed.

3 - Dunno, you'll have to ask your IC design guy, or your foundry. I would suggest you make sure you know _all_ the costs going in. As a newbie you may want to get one point of contact who will guarantee getting _everything_ working for you, even if its at a premium. This will probably mean going through the foundry, unless there's an IC guy out there willing to run that trap line for you.

My only direct experience with an ASIC process was intensely dilbertesque -- upper management fell in love with the idea of an ASIC that would implement the core functionality of their products, without realizing that said core functionality was changing so rapidly that the design couldn't last more than a year or so. Then we all sat around conference tables having pissing contests about requirements. There was absolutely no one in the whole company who had enough will, balls, and authority to get it done.

-- Tim Wescott Wescott Design Services http://www.wescottdesign.com Posting from Google? See http://cfaj.freeshell.org/google/ "Applied Control Theory for Embedded Systems" came out in April. See details at http://www.wescottdesign.com/actfes/actfes.html