Sure about the volume?

Attila Strba:
Smith, it is so: we are planning to use this chip for different applications for some very simple one where we will have a simple software (that can be written without any advanced debug possibilities) and also for quite complicated things where we need some debug power.
And for example if you sell 1000 chip for applications where you don't realy need an UART and a JTAG which raises the costs of the ASIC chip by 1$ than you wasted already 1000$.


Roger Steiner
1000pcs isn't that much as 1000$ isn't really a lot of money.
For example, how long can a German Engineer work for just 1000$
You might save some money on the simple projects, but how

There is something VERY strange here: The contradicting concerns are "use this chip for different applications" and "where you don't realy need an UART".

Maybe Attilat need to describe, say, 5 of the intended applications.

If you can not get way beyond 1000 volume why on earth use an ASIC?
you may save $1 (I doubt it) on the chip but the ASIC code is going to cost you min $2 per unit sold (min $2000 total).

Including a controller in an ASIC makes sense in 2 cases
a) the volume of one configuration is very high
b) you have "home" expertise than can make the ASIC unique to different apps very fast, (very fast = low cost)

Let's take Attilas concern about the UART:
if you use reqular '51 chips, you use a LPC chip w/o UART ($0.50 @1000) with an UART ($1.50 @ 1000) both with internal memory so for 2000 units you pay $2000 in uC cost.
Now you mnake an ASIC without an UART and let's say you get it for $1 (including development cost) which I doubt, and now you have $1 for the non-UART app and $1 PLUS $2 for an external UART for the other for a total cost of $3000 in uC cost.

This sounds like someone that "want the latest technology' but do not realize the cost thereof.

Erik

PLEASE do note, the above is based on the information in this particular thread and DO NOT represent a statement against uC cores in general.