[Coco] So, how do I make Drivewire go?

Aaron Wolfe aawolfe at gmail.com
Mon Sep 23 23:26:07 EDT 2013 

On Mon, Sep 23, 2013 at 11:14 PM, Christopher Smith <csmith at wolfram.com> wrote:
> Well, I did have a chance to try this, but increasing the read wait didn't really seem to change things at all.  I'm beginning to wonder whether I have some kind of a connectivity problem.  I can start a terminal program on the host side and see things coming across, though I don't recognize them as being anything near ASCII.  Are they supposed to be readable in any sense?

No, data on the line will not generally be human readable.  The DW
opcodes are not.  Sector data may be.

>
> Here's a log of a pretty typical session from the DW server:
>
> Mon Sep 23 2013 22:02:20.444  WARN   DWProtocolHandler   dwproto-0-10        UNKNOWN OPCODE: 192 À
> Mon Sep 23 2013 22:02:26.573  WARN   DWProtocolHandler   dwproto-0-10        UNKNOWN OPCODE: 192 À
> Mon Sep 23 2013 22:02:26.575  WARN   DWProtocolHandler   dwproto-0-10        UNKNOWN OPCODE: 176 °
> Mon Sep 23 2013 22:02:33.454  WARN   DWProtocolHandler   dwproto-0-10        UNKNOWN OPCODE: 192 À
> Mon Sep 23 2013 22:02:41.294  WARN   DWProtocolHandler   dwproto-0-10        UNKNOWN OPCODE: 192 À
>

This means the bytes your DW server sees are not valid instructions.
That can be due to speed mismatch, and when the unknown opcode is
consistent a speed mismatch is pretty likely.  It could mean a broken
PIA or bad cable, but I would expect more random results or just
nothing in that case.  There is a small possibility that your serial
port on the PC is to blame, either just borked or has timeouts and
other settings that don't work well at low speeds.  Some USB adapters
come with settings that are suited to Mbps block operation and don't
work (or work very well) at low speeds with exchanges containing only
a few bytes.

You can have DW log every byte sent or received on the wire.  This
might be helpful.  To do so, set the logging level to "Debug" and
enable "LogDeviceBytes".  You will have to restart the server for this
option to take effect as it is shortcircuited at boot up for speed
purposes.
You might also want to add the option -logviewer to the java command.
This will open a separate log viewing tool that is much more powerful
than the simple list found in the main UI.


> ----- Original Message -----
>> From: "Luis Antoniosi (CoCoDemus)" <retrocanada76 at gmail.com>
>> To: "CoCoList for Color Computer Enthusiasts" <coco at maltedmedia.com>
>> Sent: Monday, September 23, 2013 3:17:35 PM
>> Subject: Re: [Coco] So, how do I make Drivewire go?
>>
>> When I tested with my umodded coco1, I remember i tried several times
>> and
>> it didn't work until I found that changing the advanced configs and
>> adding
>> a Read Wait of 1 or 2 solved the problem.
>>
>> But then I decided to replace the 741 and I have no regrets since
>> them
>> using it at the same coco2 speed. This makes things even easier like
>> only
>> one ROM cartridge.
>>
>>
>> On Mon, Sep 23, 2013 at 4:11 PM, Christopher Smith
>> <csmith at wolfram.com>wrote:
>>
>> > That's the file I'm using, though I can't _guarantee_ that my CoCo
>> > is
>> > unmodded.  I'm not the first owner, and somebody did throw a CoCo 3
>> > keyboard onto it, so there's a chance they changed some other stuff
>> > around
>> > as well.  Will using the HDBDW3CC1 ROM file if they put a newer
>> > op-amp in?
>> >  I could always grab the CoCo 2 ROM, set the servers to CoCo 2
>> >  mode, and
>> > see what happens.  Worst case I suspect is precisely the same
>> > problem I
>> > have now.
>> >
>> > Chris
>> >
>> > ----- Original Message -----
>> > > From: "Luis Antoniosi (CoCoDemus)" <retrocanada76 at gmail.com>
>> > > To: "CoCoList for Color Computer Enthusiasts"
>> > > <coco at maltedmedia.com>
>> > > Sent: Monday, September 23, 2013 3:03:27 PM
>> > > Subject: Re: [Coco] So, how do I make Drivewire go?
>> > >
>> > > on an umodded coco1 you should use the HDBDW3CC1.BIN/ROM/WAV
>> > >
>> > > This is because of the 741 op-amp it has. If you change it to a
>> > > JEFT
>> > > TL081
>> > > or any other newer op-amp, then you can use the
>> > > HDBDW3CC2.BIN/ROM/WAV.
>> > >
>> > >
>> > > On Mon, Sep 23, 2013 at 3:48 PM, Christopher Smith
>> > > <csmith at wolfram.com>wrote:
>> > >
>> > > > The machine is a CoCo 1, 64k.  It had one floppy drive hooked
>> > > > up
>> > > > locally
>> > > > through a Tandy controller; I don't think that should matter.
>> > > >  HDBDOS did
>> > > > manage to work the local drive when I typed DRIVE OFF.  The ROM
>> > > > image I was
>> > > > using came from cloud-9's Drivewire 3 page, so I assume it's
>> > > > not
>> > > > the DW4
>> > > > ROM.
>> > > >
>> > > > I thought about running a serial sniffer on the host side and
>> > > > seeing
>> > > > what's going on.  Trouble is that I don't really have a
>> > > > terminal
>> > > > program
>> > > > for the CoCo.  I could download one and send it over with
>> > > > Drivewire
>> > > > of
>> > > > course, but if that worked I wouldn't need it... ;)
>> > > >
>> > > > Chris
>> > > >
>> > > > ----- Original Message -----
>> > > > > From: "Tormod Volden" <lists.tormod at gmail.com>
>> > > > > To: "CoCoList for Color Computer Enthusiasts"
>> > > > > <coco at maltedmedia.com>
>> > > > > Sent: Monday, September 23, 2013 2:42:33 PM
>> > > > > Subject: Re: [Coco] So, how do I make Drivewire go?
>> > > > >
>> > > > > On Mon, Sep 23, 2013 at 7:27 PM, Christopher Smith wrote:
>> > > > > > Since everyone else is talking about it, I thought I'd
>> > > > > > write in
>> > > > > > about my experience with Drivewire last week.  I thought
>> > > > > > I'd
>> > > > > > set
>> > > > > > up Drivewire and get the CoCo actually doing something.
>> > > > > >  Pulled
>> > > > > > the HDBDOS WAV file for the CoCo 1 from the cloud-9 web
>> > > > > > page.
>> > > > > >  Also installed the Drivewire 4 server on my Linux box and
>> > > > > >  the
>> > > > > > MacOS X Drivewire 3 server on a different machine, just in
>> > > > > > case.
>> > > > > >  The cable, as far as I can tell, is correct.  I made it
>> > > > > >  after
>> > > > > > this diagram:
>> > > > > >
>> > > >
>> > http://www.mediafire.com/view/c91gj67kw7q6yzu/CoCo_DriveWire_3_Serial_Cable.pdf
>> > > > >
>> > > > > It sounds like you did everything that is needed. Note that
>> > > > > above
>> > > > > cable is not prepared for DATurbo mode, so you must not use a
>> > > > > DW4
>> > > > > WAV
>> > > > > file. What kind of CoCo are we talking about?
>> > > > >
>> > > > > > Anyway, I load up either Drivewire server, tell it to serve
>> > > > > > some
>> > > > > > floppy images, load HDBDOS off cassette, and everything
>> > > > > > looks
>> > > > > > ok
>> > > > > > but the only thing I can get HDBDOS to tell me about any
>> > > > > > disk
>> > > > > > device is "I/O Error."  It flashes lights on the serial
>> > > > > > interface
>> > > > > > when it tries to access the disk, but the server doesn't
>> > > > > > seem
>> > > > > > to
>> > > > > > do anything useful in response.  I'm afraid it's been a
>> > > > > > couple
>> > > > > > of
>> > > > > > days, but I remember seeing a lot of OP_NOP in the log.  Am
>> > > > > > I
>> > > > > > missing something obvious here?  How does one debug this
>> > > > > > stuff?
>> > > > >
>> > > > > How to debug depends a bit on your equipment and skills. I
>> > > > > would
>> > > > > have
>> > > > > run some terminal emulation program on both ends to verify
>> > > > > that
>> > > > > you
>> > > > > have the physical setup correct. I don't know what's
>> > > > > available on
>> > > > > CoCo, but on the other computer you can use Hyperterminal on
>> > > > > Windows
>> > > > > or "screen /dev/ttyUSB0 57600" on Linux.
>> > > > >
>> > > > > The flashing light and NOP in the logs indicate that
>> > > > > communication is
>> > > > > at least working in one direction.
>> > > > >
>> > > > > Tormod
>> > > > >
>> > > > > --
>> > > > > Coco mailing list
>> > > > > Coco at maltedmedia.com
>> > > > > http://five.pairlist.net/mailman/listinfo/coco
>> > > > >
>> > > >
>> > > > --
>> > > > Christopher Smith
>> > > > Systems Engineer, Wolfram Research
>> > > >
>> > > > --
>> > > > Coco mailing list
>> > > > Coco at maltedmedia.com
>> > > > http://five.pairlist.net/mailman/listinfo/coco
>> > > >
>> > >
>> > >
>> > >
>> > > --
>> > > Long live the CoCo
>> > >
>> > > --
>> > > Coco mailing list
>> > > Coco at maltedmedia.com
>> > > http://five.pairlist.net/mailman/listinfo/coco
>> > >
>> >
>> > --
>> > Christopher Smith
>> > Systems Engineer, Wolfram Research
>> >
>> > --
>> > Coco mailing list
>> > Coco at maltedmedia.com
>> > http://five.pairlist.net/mailman/listinfo/coco
>> >
>>
>>
>>
>> --
>> Long live the CoCo
>>
>> --
>> Coco mailing list
>> Coco at maltedmedia.com
>> http://five.pairlist.net/mailman/listinfo/coco
>>
>
> --
> Christopher Smith
> Systems Engineer, Wolfram Research
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco

More information about the Coco mailing list