[Coco] A Website For My Projects + A Neat CoCoNET Idea

[jdaggett at gate.net]

On 9 Jan 2009 at 18:41, John Eric wrote:

>  During research, I found two suggested
> fixes. The first and easiest was to replace capacitor c64 with a 220pf
> cap (it is originally 150pf). I've already performed this fix, but
> honestly, I don't see it preventing the contention when the ROMs are
> accessed.
|==========

C64 serves two purposes. First is to establish a cutoff frequency in which to 
prevent a mode jump in the crystal. A 28MHz crystal is generally a 3rd 
overtone crystal and some cuts can jump and osscilate on the 5th 
overtone. The inductor and C64 prevent that. 

Secondly the series combination of C64 and L1 in paralell with C60 form a 
net capacitance that is in series with TC1. The resulting capacitance is the 
load capacitance for the crystal that sets the frequency. Changing C64 can 
allow an increase in the range of frequency adjustments for TC1. 

The chances of mode jump are minimal but are possible. Yes instability of 
the oscillator can cause "sparkles". Also you have to remember that the 
timing constraints for RAM access for the GIME and the CPU are very 
tight. In the CC3 the CPU rate is 1/16th or 1/8th that of the crystal. So to 
effect the RAM access timing by pulling the oscillator has to take the 
divisor in context. 

james