[Coco] What do you make of this non-approved HSCREEN mode?

jdaggett at gate.net jdaggett at gate.net
Thu Jun 11 21:49:06 EDT 2009

> I've just stumbled onto the list for the first time in a while... best
> thread in quite a while. Just wondering if this behavior is consistent
> with both '86 and '87 GIMEs.

My thoughts also.

> My instinctive theory would be that it is just incomplete
> implementation. Maybe management put the rush on the design team to
> "finish up," so they had to ship it, no time to even close the doors
> or squash a bug. Time is money!

Yes and no. Incomplete maynot be the correct term. The HRES and CRES bits are encoded 
bits to yield different modes. What may have happened is that there is incomplete decoding 
or decoded ports that are not implement and yield really garbage when encountered. 

Hard to say without some idea of what the designers thought. It is possible the the prototype 
board may yield some insight. I may have to review what I have. It would be nice to have 
detailed pictures withour the dots and labels on the covered ICs also. Better to ID them and 
thus function. 

> This may be totally irrelevant, but is the design loosely based on the
> 6847 architecture? If you issue PMODE4,x and SCREEN1,x, then issue
> PMODE0,n without a SCREEN command, then draw, for example, concentric
> circles, the circles show up in two columns simultaneously. You issue
> one circle command and get two circles on screen, side by side. I'm
> just saying, it seems like sort of similar behavior.

Yes some of the  modes of the MC6847are duplicated. You could say that it was at least an 
inspiration for design. Also all modes of the 6847/6883 are not implemented. Most likely 
would not fit and/or to difficult to implement on the GIME chip. 

This is indeed some of the best ideas in a while. It is exactly what needs to be done to truly 
start to understand the behavior of the GIME chip. That along with the prototype a more 
complete understanding of the internal workings can be better ascertained. 


More information about the Coco mailing list