Re: BASIC and System ROM replacements

From: Gerrit Heitsch <gerrit_at_laosinh.s.bawue.de>
Date: Wed, 05 Jun 2013 19:14:40 +0200
Message-ID: <51AF7200.9060901@laosinh.s.bawue.de>
On 06/04/2013 11:01 PM, Jim Brain wrote:
> On 6/4/2013 12:57 PM, Gerrit Heitsch wrote:
>> On 06/04/2013 07:23 PM, Jim Brain wrote:
>>> On 6/4/2013 11:59 AM, Gerrit Heitsch wrote:
>>>> On 06/04/2013 04:14 AM, Jim Brain wrote:
>>>>> As I look at it, I can;t help but wonder if some juggling of the
>>>>> address
>>>>> lines and such would simplify things and remove the requirement for an
>>>>> inverter altogether, but I am not seeing an easy way to do it.
>>>>>
>>>>> http://jbrain.com/vicug/gallery/vicmidi/VIC_MIDI_Schematic
>>>>
>>>> You have a few unused inputs on the LS688. Can't you use them to
>>>> include PHI2 into the _SEL signal already? That would mean you could
>>>> free up some inputs on the other logic simplify the circuit there.
>>> My concern there is that it delays PHI2 all over, and it's typically the
>>> last signal to switch.  I felt putting it through logic would create
>>> timing hazards.
>>
>> That shouldn't be much of an issue with the current logic chips and
>> become less so if you can reduce the logic further.
> I'm struggling with that.  ahct138 is 12nS worst case, and hct688 is
> 34nS worst case (@25C) That's over 46nS of delay, not including the
> ahct14 delay of 8ns for ahct14.  54ns is quite a bit.  Doesn't the VIC-I
> use the other half of the cycle?

Yes, but a halfcycle is 500ns long. You could also ditch the LS688 and 
just use the LS138 even if that means incomplete (but faster) decoding. 
Or you try a faster logic family. Like 74Fxx or 74ALSxx.


>>>> Then you should read the datasheet for the Flash since for most RAMs,
>>>> _WE overrides _OE, meaning you can tie _CE and _OE together. If that's
>>>> the same for the flash, you can free up an inverter for use elsewhere.
>>> That. I did not know.  I will check that tonight.
>>
>> Checked the datasheet for a µPD43256 from NEC, as soon as _WE goes
>> low, _OE becomes 'x' (don't care)
> 39sf010 datasheet specified !OE must be Vih (High) during write cycle.

Annoying... Someone wanted to save a few transistors.

  Gerrit



       Message was sent through the cbm-hackers mailing list
Received on 2013-06-05 18:00:03

Archive generated by hypermail 2.2.0.