This might be a good adapter for T/R to FE, I've just not had luck with
the system it runs in... pFsense or something...

On 8/8/2021 11:15, RickE wrote:

Back in January 2002, I picked up these four Token-Ring 100/16/4 adapters, thinking that I would use them to connect some machines in a Token-Ring network as an extension to my standard 100/10 Ethernet network. I never got further than that bit of

thinking, they've been sitting in the attic for the last 19 years. I got them at the IBM Endicott PC Sale, they are all returns, probably from a customer that realized that 100 Mbps Token-Ring was just "too weird" after they placed their order. I'm
thinking about trying to sell them as a lot on eBay, but I'll offer them here first for just the cost of shipping in case one of you is interested in the "weird" side of Token-Ring networking. Each box contains:

1 booklet of safety instructions (worthless)
1 Token-Ring 100/16/4 PCI adapter (sealed in the anti-static bag, probably re-sealed as part of the return processing by IBM)
1 RJ-45 wrap plug for diagnostics (be still my beating heart)
1 Installation and CD-ROM Guide 31L3803 (includes drivers and a program to generate install/update/diagnostic diskettes)
1 76H7254 Wake-On-LAN cable
1 42H2397 Wake-On-LAN cable

The 42H2397 cable is intended for IBM PC Type-Models that have a 2-pin WOL connector on the system board, 76H7254 is used with the more common 3-pin WOL connector on the system board. If you don't care about the Wake-On-LAN feature, these cables are

just so much junk.


--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 10:15 AM, RickE wrote:

Back in January 2002, I picked up these four Token-Ring 100/16/4
adapters, thinking that I would use them to connect some machines
in a Token-Ring network as an extension to my standard 100/10
Ethernet network. I never got further than that bit of thinking,
they've been sitting in the attic for the last 19 years. I got them
at the IBM Endicott PC Sale, they are all returns, probably from a
customer that realized that 100 Mbps Token-Ring was just "too weird"
after they placed their order.

Please elaborate on what "too weird" is in this context. I'm especially curious about examples of "too weird".

I'm thinking about trying to sell them as a lot on eBay, but I'll
offer them here first for just the cost of shipping in case one of
you is interested in the "weird" side of Token-Ring networking.

I don't /need/ them any more than I need another hole in my head.

But I am interested in them. Particularly for my retro computing
"pasture". You know, where old but lovable things go to have an easy life.

Each box contains:

1 booklet of safety instructions (worthless)
1 Token-Ring 100/16/4 PCI adapter (sealed in the anti-static bag,
probably re-sealed as part of the return processing by IBM)
1 RJ-45 wrap plug for diagnostics (be still my beating heart)
1 Installation and CD-ROM Guide 31L3803 (includes drivers and a
program to generate install/update/diagnostic diskettes)
1 76H7254 Wake-On-LAN cable
1 42H2397 Wake-On-LAN cable

The 42H2397 cable is intended for IBM PC Type-Models that have a
2-pin WOL connector on the system board, 76H7254 is used with the
more common 3-pin WOL connector on the system board. If you don't
care about the Wake-On-LAN feature, these cables are just so much junk.

I'd be curious to have all type / model / part / etc. numbers off of
things. I'll add them to my organically growing database.



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, August 8, 2021 at 4:44:53 PM UTC-4, Grant Taylor wrote:

Please elaborate on what "too weird" is in this context. I'm especially curious about examples of "too weird".

Most Token-Ring infrastructure equipment will not handle the 100 Mbps speed setting, they only work with 16 and 4. A basic IBM 8228 MAU will work, because it's just a collection of relays, it "has no brains". CAUs and LAMs work just fine with 16/4,
they puke on 100 (not designed for it). So you pretty much have to isolate the 100 Mbps Token-Ring segments, then have a bridge machine with both a 100/16/4 card and a 16/4 (or a 100/16/4 set for 16 or 4) so that the bridge can connect to "normal" Token-
Ring segments. Plus another bridge machine to go between Token-Ring and Ethernet, if you're so inclined. Token-Ring lost, Ethernet won.

I'd be curious to have all type / model / part / etc. numbers off of
things. I'll add them to my organically growing database.

I listed the part numbers of the cables and CD-ROM, the adapter is part number 25L5572, FRU number 30L5980. The "raw card" (PCB) has part number 30L6899, the large quad flat-pack module that is the "brains" of the adapter is 25L5484. The back plate of
the adapter has both 9-pin D-shell and RJ-45 connectors. Each adapter has a serial number and a MAC address. The first 3 octets of the adapter I pulled out of the anti-static bag are 00 06 29.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, 8 August 2021 at 22:24:29 UTC+1, ekb...@vnet.ibm.com wrote:

On Sunday, August 8, 2021 at 4:44:53 PM UTC-4, Grant Taylor wrote:

Please elaborate on what "too weird" is in this context. I'm especially curious about examples of "too weird".

Most Token-Ring infrastructure equipment will not handle the 100 Mbps speed setting, they only work with 16 and 4. A basic IBM 8228 MAU will work, because it's just a collection of relays, it "has no brains". CAUs and LAMs work just fine with 16/4,

they puke on 100 (not designed for it). So you pretty much have to isolate the 100 Mbps Token-Ring segments, then have a bridge machine with both a 100/16/4 card and a 16/4 (or a 100/16/4 set for 16 or 4) so that the bridge can connect to "normal" Token-
Ring segments. Plus another bridge machine to go between Token-Ring and Ethernet, if you're so inclined. Token-Ring lost, Ethernet won.

I'd be curious to have all type / model / part / etc. numbers off of things. I'll add them to my organically growing database.

I listed the part numbers of the cables and CD-ROM, the adapter is part number 25L5572, FRU number 30L5980. The "raw card" (PCB) has part number 30L6899, the large quad flat-pack module that is the "brains" of the adapter is 25L5484. The back plate of

the adapter has both 9-pin D-shell and RJ-45 connectors. Each adapter has a serial number and a MAC address. The first 3 octets of the adapter I pulled out of the anti-static bag are 00 06 29.

FYI, there are 100Mbps rated RJ45 to ICS adapters for the ICS 8228 MAUs. They have Green "100" stickers on the side. I have enough 100/16/4 PCI bus Token-Ring adapters. Madge and IBM made versions, the 25Mbps Token-Ring technology never really made it to
the market.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 5:25 PM, WBSTClarke wrote:

On Sunday, 8 August 2021 at 22:24:29 UTC+1, ekb...@vnet.ibm.com wrote:

On Sunday, August 8, 2021 at 4:44:53 PM UTC-4, Grant Taylor wrote:

Please elaborate on what "too weird" is in this context. I'm especially
curious about examples of "too weird".

Most Token-Ring infrastructure equipment will not handle the 100 Mbps speed setting, they only work with 16 and 4. A basic IBM 8228 MAU will work, because it's just a collection of relays, it "has no brains". CAUs and LAMs work just fine with 16/4,

they puke on 100 (not designed for it). So you pretty much have to isolate the 100 Mbps Token-Ring segments, then have a bridge machine with both a 100/16/4 card and a 16/4 (or a 100/16/4 set for 16 or 4) so that the bridge can connect to "normal" Token-
Ring segments. Plus another bridge machine to go between Token-Ring and Ethernet, if you're so inclined. Token-Ring lost, Ethernet won.

I'd be curious to have all type / model / part / etc. numbers off of
things. I'll add them to my organically growing database.

I listed the part numbers of the cables and CD-ROM, the adapter is part number 25L5572, FRU number 30L5980. The "raw card" (PCB) has part number 30L6899, the large quad flat-pack module that is the "brains" of the adapter is 25L5484. The back plate of

the adapter has both 9-pin D-shell and RJ-45 connectors. Each adapter has a serial number and a MAC address. The first 3 octets of the adapter I pulled out of the anti-static bag are 00 06 29.

FYI, there are 100Mbps rated RJ45 to ICS adapters for the ICS 8228 MAUs. They have Green "100" stickers on the side. I have enough 100/16/4 PCI bus Token-Ring adapters. Madge and IBM made versions, the 25Mbps Token-Ring technology never really made it

to the market.


They were still using ICS at IBM Poughkeepsie a couple years ago with a
cable like that ICS on one side, maybe a balun?, and RJ45 on the other
end.. gigabit Ethernet though.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 6:25 PM, WBSTClarke wrote:

FYI, there are 100Mbps rated RJ45 to ICS adapters for the ICS 8228
MAUs. They have Green "100" stickers on the side. I have enough
100/16/4 PCI bus Token-Ring adapters. Madge and IBM made versions,

Interesting. Good to know.

the 25Mbps Token-Ring technology never really made it to the market.

Twenty-Five Megabits per second Token Ring? I've never heard of this
before.



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 6:51 PM, Kevin Bowling wrote:

They were still using ICS at IBM Poughkeepsie a couple years ago with a
cable like that ICS on one side, maybe a balun?, and RJ45 on the other
end.. gigabit Ethernet though.

I'd be shocked to learn that it's Gigabit Ethernet. Mostly because I
thought that Gigabit Ethernet /required/ eight wires. And the last time
I looked, ICS connections were only four wires.

Is there any chance that it was the elusive Gigabit Token Ring that
never made it out of the labs? -- I can see how IBM Poughkeepsie
/could/ have things from IBM labs that the rest of the world wouldn't
have access to. Maybe it could run on four wires.



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 3:24 PM, RickE wrote:

Most Token-Ring infrastructure equipment will not handle the 100 Mbps
speed setting, they only work with 16 and 4. ...

*nod*

That makes perfect sense.

So you pretty much have to isolate the 100 Mbps Token-Ring segments,
then have a bridge machine with both a 100/16/4 card and a 16/4 ...
so that the bridge can connect to "normal" Token-Ring segments.

I would expect this to be the case.

It's my understanding that dumb / passive Token Ring MAUs (e.g. 8228)
couldn't mix speeds period. So, needing to isolate 100 Mbps Token Ring
on it's own ring makes perfect sense to me. Likewise with 16 Mbps and 4
Mbps (on dumb / passive MAUs).

Plus another bridge machine to go between Token-Ring and Ethernet,
if you're so inclined. Token-Ring lost, Ethernet won.

Or router. ;-)

I listed the part numbers of the cables and CD-ROM, the adapter
is part number 25L5572, FRU number 30L5980. The "raw card" (PCB)
has part number 30L6899, the large quad flat-pack module that is the
"brains" of the adapter is 25L5484. The back plate of the adapter
has both 9-pin D-shell and RJ-45 connectors. Each adapter has a
serial number and a MAC address. The first 3 octets of the adapter
I pulled out of the anti-static bag are 00 06 29.

Thank you Rick.



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/8/21 10:19 PM, Grant Taylor wrote:

On 8/8/21 6:51 PM, Kevin Bowling wrote:

They were still using ICS at IBM Poughkeepsie a couple years ago with
a cable like that ICS on one side, maybe a balun?, and RJ45 on the
other end.. gigabit Ethernet though.

I'd be shocked to learn that it's Gigabit Ethernet.  Mostly because I thought that Gigabit Ethernet /required/ eight wires.  And the last time
I looked, ICS connections were only four wires.

Is there any chance that it was the elusive Gigabit Token Ring that
never made it out of the labs?  --  I can see how IBM Poughkeepsie
/could/ have things from IBM labs that the rest of the world wouldn't
have access to.  Maybe it could run on four wires.

Unsure I didn't look closely at the connector or configuration, I just
happened to pass through while visiting and thought it was peculiar.
Maybe they ran 100mbit to desktops and that was "good enough" for
employee systems.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 8/9/21 12:48 AM, Kevin Bowling wrote:

Unsure I didn't look closely at the connector or configuration, I just happened to pass through while visiting and thought it was peculiar.

ACK Using ICS connections after about 2005 would likely be something I
would notice, and definitely make mental note of, if I had seen it.

Maybe they ran 100mbit to desktops and that was "good enough" for
employee systems.

I could see that. I still think that what most people do would work
just fine over 100 Mbps connections. After all, much of what people do
now is over the Internet and most Internet connections are 100 Mbps or
less. So ... ya ... I think that 100 Mbps to the desktop could suffice
for many ~> most corporate users.



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

This might be a good adapter for T/R to FE, I've just not had
luck with the system it runs in... pFsense or something...

There would need to be support for Token Ring under FreeBSD, and that was pulled before pfSense came on the scene - Here's the site that was building drivers, and it hasn't been active since 2003: http://www.jurai.net/~winter/tr/drivers.html

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, August 8, 2021 at 8:51:16 PM UTC-4, Kevin Bowling wrote:

They were still using ICS at IBM Poughkeepsie a couple years ago with a cable like that ICS on one side, maybe a balun?, and RJ45 on the other
end.. gigabit Ethernet though.

Endicott is still using ICS to most of the offices in Building 256 and we have the RJ-45 converters for Ethernet (with large shrink tubing around the sides so that people don't pull them off the wallplate). Yes, that only supports 100/10 Ethernet, while
our backbone is all Gigabit with Power-Over-Ethernet to support the Cisco telephone handsets that were mostly removed from the facility in March 2021 -- IBM Global Voice Services wants all of us to use Cisco Jabber instead of having a physical phone. If
you've ever had the "pleasure" of using Jabber, you know that's a joke in poor taste. Thankfully, I still have an analog phone on my desk from my modem-wrangling days, but GVS will remove the analog phone hardware soon, probably within a year (just as
soon as they can finally drive a stake through the heart of FAX). New cable runs for the last 20+ years have been Ethernet cables, and those support Gigabit. The Token-Ring standards folks even created a 1000 Mbps Token-Ring specification, but no one
implemented it after 100/16/4 fell flat.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)