[Coco] Socat drivewire relay to CoCo1

[Gene Heskett]

On Sunday 01 March 2015 05:09:36 jon bird wrote:
> In article
> <CAA6uQZRrhJ=wREEHy9+9S6kD90j7w-xsOaN6dHbN2r-qhay5RQ at mail.gmail.com>,
> Aaron Wolfe <aawolfe at gmail.com> writes
>
> >On Sat, Feb 28, 2015 at 6:21 AM, jon bird <news at onastick.clara.co.uk>
> >
> >wrote:
> >> Hi,
> >>
> >> Now I've got my CoCo 1 up and running again I thought I'd give
> >
> >Drivewire
> >
> >> a go. A couple of initial queries then - I made the cable up iaw
> >> with the diagram here:
> >
> >http://www.cocopedia.com/wiki/index.php/Getting_Started_with_DriveWir
> >e
> >
> > I'm not entirely sure of the significance of the CD/"turbo mode"
>
> option
>
> > and there seems to be some differences in the baud rate you need to
>
> set
>
> > the server end to - either 38400 or 57600. Currently the only
> > configuration I get anything sensible with is 38400 with or without
> > CD connected.
>
> [...]
>
> >The coco 1 is a bit of a special case because it cannot reliably
> >operate at 57,600 bps (slow level shifters or some such hardware
> >deficiency compared to coco 2).  Because of this, it must remain at
> >38,400bps.  The double speed "turbo" mode using the CD pin is not
> >possible as far as I know.  It is also sensitive to cable length, so
> >make sure the serial cable is as short as possible.  Cabling and
> > speed problems often manifest as checksum errors like you are
> > seeing.

This is primarily because the op-amp used to drive the bit banger is too 
slow.  IIRC its a 741, which in op-amp history is somewhat below the KT 
boundary.  It is not hard to replace if you can run a soldering iron, 
and there are, in the chip books, an almost unlimited choice of op-amps 
that will drop right in, probably use less power, and are literally 
millions of times faster while being 100% footprint compatible.

The 741 is so slow from slew rate limits (.5 volts per microsecond) that 
its open loop gain of 100,000 is usable only up to 10 hertz!  Modern 
op-amps have open loop gains that in some cases aren't that high, some 
as low as 20,000, but the -3 db gain rolloff is above 2 gigahertz.

I expect a google search for "fast 741" might suggest some good ones in 
the one or two dollar range.

One of the links spit out for a "faster 741 op amp replacement"

<http://www.allaboutcircuits.com/vol_3/chpt_8/14.html>

Has a good description without an excess of $100 words, looks like an 
lm356 in the right packaging might be a good substitute.  You only need 
5 active pins, plus and minus inputs, plus and minus supply rails, and 
an output pin.  Anything else the 741 needed can be left unused as its 
100% internal and more precisely done in the newer ones.  And its 30x 
faster.

> >To ensure the coco is operating at 38,400bps you need to use the
> >correct ROM or DW binary on the coco, the special version for coco 1
> >should only operate at 38,400.  Of course the serial port on the pc
> >side should also be set to 38,400.  The DW server does this when it's
> >told a coco 1 is connected, but since you're working over IP that
> > will need to be done on the machine running socat.
> >
> >FWIW, I wrote DriveWire using a prolific adapter :)  Many people do
> >have trouble with them, so that may be a factor in your
> > configuration. However the checksum issue makes me suspect its
> > either trying to work at 57,600 or the cable itself is causing
> > trouble.
>
> Thanks for the info. I am indeed using the CoCo 1 binary, the serial
> cable is ~30cm long. Given that it seems to work reliably in a "local"
> configuration it sounds unlikely that it is the cable itself causing
> the problem. Given the network traffic looks ok the only difference I
> can think of is the timing differences due to it sending over
> Ethernet.
>
> As an aside though, I also have a CoCo 2 in the cupboard, it needs a
> bit of work before I can use it but (once I fix it) presumably that
> will then run at 56K, does that also need this "turbo mode"
> configuration on the serial cable?
>
> A couple of other observations which I forgot to mention earlier, I
> couldn't actually get DW to work when connected directly to the
> Prolific adaptor. The first time around it generates a "Unsupported
> com operation whilst opening /dev/ttyUSB0", then repeated "Handler #0:
> Serial device in use" messages. Given that I don't intend to run DW in
> this configuration, it doesn't affect me but just thought I'd mention
> it in case there is a bug here.
>
> Rgs,
>
>
> Jon.
>
> --
> == jon bird - software engineer
> == <reply to address _may_ be invalid, real mail below>
> == <reduce rsi, stop using the shift key>
> == posted as: news 'at' onastick 'dot' clara.co.uk

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>