[Coco] Blast from the past!

[jdaggett at gate.net]

On 31 Jan 2008 at 16:53, Ciaran Anscomb wrote:

> More in the way the SAM and VDG interact - the VDG itself doesn't
> directly support those modes.
> 
> It does seem odd that they didn't implement the more obvious
> additional modes (that were already documented in the SAM data sheet)
> in the GIME, though I can understand not bothering with the more
> obscure behaviour.
> 
> Hello, I'm new here.
> 
> ..ciaran
> -- 
> Ciaran Anscomb, Perl/C Hacker

ciaran

having been apart of a team that designed a new IC, I can tell you there are trade 
offs in any design. 

My guess is die size and packaging costs dictated what could be done internal to the 
GIME. Increasing the function of the video section probably would have pushed the 
die size beyond that which could be reliably packaged into a 68 pin PLCC. The next 
available size in 84 pins. That would add about another $1.60 minimum to the 
packaging costs and result in a larger die costs. 

My guess initially the GIME chip cost Tandy about $25 - $35 /unit for the first 
production runs. That probably was the pain threshold that dictated how much 
integration was done and what functions were implemented. There is only so much 
area on a 4 inch wafer. the bigger the individual IC is, the less you get per wafer. 
Thus higher die costs. Higher total IC costs.

Back in the 70's and 80's it was more cost effective to do two or more ICs for a 
function or set of functions. Integration was not at the level it is today. 

james