X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
MIME-Version: 1.0
In-Reply-To: <20111129003937.7d098958@rainbird>
References: <20111117204524 DOT 10e586f5 AT rainbird>
	<4EC5CBEE DOT 5060904 AT optonline DOT net>
	<CAN0Jx-84K3n+g+ZH29wjzkuusBa=Bk_F=SaY+Z=fdm-JTf=HKQ AT mail DOT gmail DOT com>
	<20111120162218 DOT 555bc069 AT rainbird>
	<4ECA9BF9 DOT 9000601 AT ecosensory DOT com>
	<20111128150702 DOT 2c75a231 AT rainbird>
	<4ED432D0 DOT 8000303 AT ecosensory DOT com>
	<20111129003937 DOT 7d098958 AT rainbird>
Date: Mon, 28 Nov 2011 23:19:38 -0700
Message-ID: <CAN0Jx-886q4KW25yZ0=AzvSTThFdi52J8sCGJadf3Zm_17F59A@mail.gmail.com>
Subject: Re: [geda-user] Modern GAL/PAL design with gEDA?
From: Russell Dill <Russ DOT Dill AT asu DOT edu>
To: geda-user AT delorie DOT com
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id pAT6n6ES024575
Reply-To: geda-user AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: geda-user AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On Mon, Nov 28, 2011 at 10:39 PM, Vanessa Ezekowitz
<vanessaezekowitz AT gmail DOT com> wrote:
> On Mon, 28 Nov 2011 19:18:08 -0600
> John Griessen <john AT ecosensory DOT com> wrote:
>
>> On 11/28/2011 02:07 PM, Vanessa Ezekowitz wrote:
>> > after a week of wrangling with this, I finally turned out a working
>> > design - or at least, it simulates (iverilog + gtkwave)*and*  synthesizes
>> > properly
>>
>> Hey, you did it!
>> Looks like dotclock is a sequence you create after being triggered so as to
>> put out some stored values, right? and that sequence is done with #delay
>> statements. (it's been 20 years since I was doing this daily to make iic
>> ports, serial ports, etc...)
>>
>> John
>
> Actually, you're pretty close to what I was originally trying to do.  The RAM controller chip used in C64 (of course) RAM Expansion Modules only produces enough signals during refresh to handle the 256k x 1 RAMs the use, and in RAS-Only refresh mode at that.
>
> So I had to detect those RAS-Only refresh signals and replace them with CAS-before-RAS pulses.  I originally tried to use a state machine driven by the dot clock, until I realized the 22v10's flip-flops are all clocked by the same line...  so a state machine was out.
>
> I worked out that, assuming my testbench model is accurate, I could use the dot clock to directly drive the RAS output line as long as I pulled all of the CAS lines low first, and as long as I provided it in inverted form.  Since the 22v10 can't invert the clock line, I had to loop it through the chip and invert it that way.
>
> The truly sad thing is the test bench is more complicated than the code it drives. :-)
>

Don't feel badly about that, it is pretty typical for the testbench to
be larger than the code it tests.