Re: Modern emulation solutions

On Jan 4, 2019, at 8:54 AM, "alan@xxxxxxxxxxx" <alan@xxxxxxxxxxx> wrote:

This list seems pretty dead, but also seems the most apropos to my
question.

I recently acquired a 5362 System/36 that actually IPLs and runs with an
internal 30 and 60 MB hard disk. It came with a 3179 terminal. But I
started looking at some more practical (and lighter) communication
options - particularly for data transfer - and only found the suite of
products by twindata.com. Which seems ok for basic emulation but
overpriced for a data transfer solution.

After browsing countless eBay listings for ISA & PCI twinax emulation
cards, I visually noticed a common hardware pattern in the component
selection so I started digging a little deeper. It turns out the wire
level protocol on midrange twinax ports isn't that complex.
Electrically it's basically differential TTL. The protocol is
essentially 1 Msps Manchester encoding with a default idle wire state,
simple pre and post ambles, and simple framing. A lot of the cards I've
found use a common framer ASIC but that could easily be replicated with
a small FPGA and some line drivers / opamps. And iCE40up5k could do the
entire job including USB and frame buffers. A BoM cost (not including
twin-ax dongle) would be < $25.

It would be fairly easy to build a completely opensource USB to twinax
adapter (or Ethernet or ...) but the protocol payload seems like an
ocean I don't want to fill with a kitchen tap of time.

How feasible would it be to leverage something like the tn5250 code base
on top of a custom hardware bridge to create a modern twinax emulation
solution? Are the payload formats over twinax and the IP encapsulation
similar?

Other than a few retro enthusiasts, how much interest would there be in
such a design? The last thing I want to do it tank segment of
TwinData's market to make what is probably an already shaky business
unstable then I run out of time and can't support things. I would hate
to leave the community with less options by creating more.

Just curious on thoughts from what seems like a 5250 core user base.

In the meantime, I may build a simple sniffer to eves-drop on the line
and capture a communication dump of the S/36 and terminal to see if the
payload looks sane.

Thanks,

-Alan
--
This is the Linux 5250 Development Project (LINUX5250) mailing list
To post a message email: LINUX5250@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/linux5250
or email: LINUX5250-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/linux5250.