[Coco] CoCoNet evolving

[gene heskett]

On Saturday, April 30, 2011 05:43:33 PM Roger Taylor did opine:

> At 05:13 PM 4/29/2011, you wrote:
> >On Friday, April 29, 2011 06:03:27 PM Roger Taylor did opine:
> > > Does anyone here use the 3-in-1 or similar type board for their
> > > Super Controller pak, and if so do you use the serial or parallel
> > > port often?  From what I read, the 3-in-1 has a 6551-based serial
> > > circuit with some of the handshaking lines tied active.  In this
> > > case, a simple 3-wire connection can be made to a PC to allow a
> > > CoCoNet link. I'd have to add detection in the CoCoNet ROM, but I
> > > wanted to do a quick survey first to see just how many people (on
> > > this list) use an SC+3-in-1 setup.
> >
> >I have a 4n1, but took it out yonks ago, mainly because the scsi
> >interface was a 1 drive only interface, the serial was an xon/xoff
> >only, and the parport had no handshaking so the data error rate with
> >modern well (FCC specs) filtered input ports rendered that useless
> >about 15 years ago.
> 
> Thanks, Gene.  CoCoNet doesn't use interrupts with the 6551 for the
> virtual drives.  In fact, no hardware handshaking is required.  I
> remember being dragged down a road by my very short hair over
> something like a CTS/RTS issue or what-not even as I explained that
> none of that mattered.  I feel confident that the limited serial
> circuit of the 4-in-1 board will work just fine for another
> high-speed virtual drive connection.  I was just wondering how many
> people actually own and use those things so I can make some decisions.

Well, its basically a 3 wire circuit, so the usefulness as a serial port 
depends on the speeds that can be obtained from the raw hardware, by 
bypassing the internal /16 clocking circuit.  The fastest I have ever run 
one of those chips is midi speed 31,250 baud, which I believe it worked at 
error free.

However, the DW's bitbanger only uses an interrupt to detect the start bit 
of a packet I believe, so once the byte has been received, it can be put 
away in one sta x++, and jump back to looking for the start bit of the next 
byte, until the packet has been received, which is faster than going 
through the whole interrupt (at least in nitros9) with its minimum of about 
15 u-sec latency with a ghost of about 160 u-secs if the IRQ hits when 
nitros9 is in the clock, 60 times a second.  So your max baud without a lot 
of error correcting code too, seems to me has to be slower than DW unless 
you simply hang in a tight loop checking the rrf bit for the next byte, but 
that then suffers in the multitasking dept just as if the bitbanger was 
being used.

It will be interesting to see what you can do with this, Roger, you have 
managed to walk on coco water before without sinking too deep when we were 
watching.  ;-)

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
<http://tinyurl.com/ddg5bz>
<http://www.cantrip.org/gatto.html>
To err is human, but I can REALLY foul things up.