On Feb 9, 21:56, Robert Schaefer wrote:
> Ooops-- I assumed you had one in front of you too. ;) The raylan is a
> 15-port managed concentrator with (currently) 10bFL hot-swap cards
> installed. Maybe concentrator is the wrong word-- that's what was used
to
> describe it to me. I was going from the tranceiver directly into the
> raylan. I'm pretty sure I had Rx & Tx right, as I got a link light on
the
> raylan, and the error on the transceiver.
Bad assumption on my part. When someone mentions "concentrator", I tend to
think of something with several places to connect AUI cables, and I assumed
that's how you'd connected them. It's basically a repeater, with it's own
transceivers (instead of just AUI) and probably management/control of
individual ports.
It ought to work, with Rx on one connected by fibre to Tx on the other, and
v.v.
> Ok. They have a three-positon DIP switch, labeled `SQE TEST', `ALTERNATE
> COLL MODE' and `FULL STEP'. Does that narrow it down?
SQE TEST is the "heartbeat" function, or Signal Quality Enable setting,
which is normally off if connecting the AUI to any kind of repeater,
otherwise on. Try it both ways.
ALTERNATE COLL MODE is to do with collision detect timing. On 3Com units
(and they bought ORnet/Chipcom) turning it on is the default, and arranges
that a single collision-presence (CI) signal is sent to the AUI interface.
If off, CI is active all the time the collision condition persists, which
is the normal setting for IEEE 802.3 compliance (the 3Com setting is an
optimisation for certain interface chips).
FULL STEP is to do with the signal sent at the beginning of a packet. HALF
STEP is normally used for 802.3 Ethernet, FULL STEP is used for DIX
Ethernet. It won't hurt to try both ways.
Another possibility is that you have the wrong type of AUI cable (old ones
use different shielding arrangements, and most "office" type cables are way
below spec) or even the wrong kind of AUI interface for this device.
Still another is that your fibre is too lossy, and the transceiver is
seeing something, but not recognising it (there's a continuous series of
pulses at 1MHz rate when it's active). Try cleaning the ends with
iso-propyl alcohol, and swap the sides over. Blow out the connectors with
dry compressed air and/or IPA.
What is the ORnet transceiver connected to?
--
Pete Peter Turnbull
Network Manager
University of York
At 07:46 PM 9/02/2002 -0500, Julius Sridhar wrote:
>I would prefer DECnet, but a lot machines don't grok DECnet. Most of mine
>do, though.
No reason why we couldn't do DECnet as well. As I alluded to in a previous
e-mail, there are other network protocols that we should try to preserve,
just like we are trying to preserve the hardware.
I understand that there's a reasonable DECnet implementation for Linux as
well as real implementations for lots of DEC machines :-) There's also
third party DECnet stuff from Xi? we used to run DECnet on a couple of our
AIX boxes to maintain connectivity to the _real_ systems running VMS (this
well before I "went with the flow" and put TCP/IP on the VAXen).
Huw Davies | e-mail: Huw.Davies(a)kerberos.davies.net.au
| "If God had wanted soccer played in the
| air, the sky would be painted green"
I'm sure there are other folks on the list that have systems that use
twist-lock connectors, so I am posting here before I contact a surplus
company to work out a trade.
I have the following Hubbell twist-lock connectors that are surplus to my
needs:
Qty Description Hubbell Part # NEMA
4 125V 30A Insulgrip Plug HBL2611 L5-30P
3 125V/250V 20A Insulgrip Plug HBL9965C N/A
1 125V/250V 20A Insulgrip Flanged Receptacle HBL3326C N/A
1 125V/250V 30A Insulgrip Plug HBL3331C N/A
1 125V/250V 30A Insulgrip Flanged Inlet HBL3334C N/A
1 125V/250V 30A Insulgrip Flanged Receptacle HBL3335C N/A
2 125V 15A/250V 10A Insulgrip Plug [Not UL Listed] HBL7567C N/A
I am looking for these connectors to get power to my old SGI systems:
Qty Description Hubbell Part # NEMA
2 125V 20A Insulgrip Plug HBL2311 L5-20P
3 125V 20A Insulgrip Connector Body HBL2313 L5-20R
4 125V 20A Single Flush Receptacle HBL2310 L5-20R
2 250V 30A Insulgrip Plug HBL2621 L6-30P
2 250V 30A Insulgrip Connector Body HBL2623 L6-30R
3 250V 30A Single Flush Receptacle HBL2620 L6-30R
Hubbell twist-lock catalog:
http://www.hubbell-wiring.com/new/sectionb.pdf
If anyone wants to make a trade, email me directly.
-Toth
Brian Chase <vaxzilla(a)jarai.org> wrote:
> For the scenerio of running a private netnews network, it'd probably
> make the most sense to just use NNTP over TCP/IP. Again, the "in
> network" servers would have to only allow traffic between themselves.
> Maybe, we could gateway in some of the better groups from the regular
> Usenet hierarchy, and then create a vunet.* (or whatever) hierarchy
> which is only carried within the network.
If you want to run a private netnews network, you just set up another
hierarchy and control who you feed that hierarchy to. This is pretty
much transport-independent and you can use either NNTP or UUCP to do
the feeding. You can also run it in parallel with less-private
netnews feeds (e.g. Big 8, alt, geographic hierarchies, Usenet II) and
even exchange feeds for the less-private traffic with hosts that don't
have anything to do with the private netnews hierarchy.
The difficult bit is keeping it private. It's way too easy for a
site's admin to set up a new feed of everything to a site that might
not be welcome at the party, and about the only recourse that the rest
of the private network has is to stop feeding the loosely administered
site if its admin won't fix the leak. It is also hard to detect
outgoing leaks if they do not permit posts to leak in.
One rule that you must have is: no cross-posts to groups inside and
outside the private hierarchy (e.g. cross-posting to
alt.folklore.computers and vunet.classiccmp is not to be done and
preferably not to be allowed). It is generally suggested that you set
up the feed for the private hierarchy as a separate feed (separate
lines in the newsfeeds file, if you're running INN) so that this rule
can be enforced.
It might be worth looking at http://www.usenet2.org/ and the links
there as something of a how-to.
> But now we're not talking about vintage computers or software. We're
> talking about vintage values--where there's a global network with the
> intent to further education, research, collaboration, and general
> discussion (pointless or otherwise). Personally, I'm sick of the over
> commercialized monstrosity that the internet has become.
Yes, you really should go look at http://www.usenet2.org/ . Consider
that the plan was to have a viable alternative when Usenet became
unusable, which was expected to happen but didn't (hasn't yet).
Usenet II exists but is not really used today.
-Frank McConnell
On Feb 9, 16:29, Robert Schaefer wrote:
> Interesting. I don't know too much about the kit I have (I would
> desperately love info on the raylan fiber concentrator &| it's SNMP card)
> but from what I did see it's pretty standard 10bFL. The tranceivers are
> `ORnet FIBER OPTIC TRANCEIVER' by chipcom. Model 9301T-ST, about 4" x 4"
x
> 2", and every bit of a pound and a half. when I connected the tranceiver
to
> the concentrator, the diag led on the tranceiver indicated `invalid
data'.
> That was just the carrier-- nothing else was plugged into the
concentrator.
> Perhaps I didn't test it enough-- there are three DIP switches on the
end.
A concentrator is the wrong thing to connect it to. Is your concentrator
something you would normally plug some computers into? If so, it's like
several transceivers. In other words, you're plugging something meant to
connect to a computer (the transceiver), into something meant to connect to
a computer (the concentrator). That's like connecting two computers on a
serial line with neither modems nor a null modem cable between.
Does the transceiver have a male 15-pin D-connector? Does the concentrator
have a female 15-pin D-connector with a slide lock, or a male one with
pins? If both are male, they don't match, and it won't work.
BTW, if it's ORnet, it's old, and almost certainly FOIRL not 10baseFL. The
size you mention bears that out, too. 10baseFL transceivers are mostly
small, about half the size of a cigarette pack.
--
Pete Peter Turnbull
Network Manager
University of York
A while back, I posted a message about a mystery board that came with
my VAX-4000/200 manufactured by a company named Talon. This board has
one 15-pin connector, a couple of 32x8 SRAM chips, a 68K CPU, assorted
logic chips, and a couple of chips that I can't identify, such as:
- Rockwell R65NC22J4 11484-39 9244 94314-3 - is this just a Rockwell
version of a 6522 interface chip?
- (unknown mfr.) 9206 SC11290CN
There's one area where a second, much smaller, circuit board is
attached to this one with chips between the two boards; however, there
are many wires soldered between the boards holding the two boards
together, and I'd rather save unsoldering them all as a last resort.
Any thoughts on what this board might be?
--
Copyright (C) 2001 R. D. Davis The difference between humans & other animals:
All Rights Reserved an unnatural belief that we're above Nature &
rdd(a)rddavis.org 410-744-4900 her other creatures, using dogma to justify such
http://www.rddavis.org beliefs and to justify much human cruelty.
On 09-Feb-2002 Tony Duell wrote:
>> OK, lets say the PSU is beyond repair for me and everyone else I asked.
> I've lost count of the number of 'beyond repair' things that have left my
> workbench working again :-)
[...]
> I can't believe it's any worse than the DEC PSU in the PDP11/44. That
> darn thing has _3_ independant choppers, a full H driver for the (AC)
> fans, and a fair amount of control electronics. And I've managed to work
> on those.
OK, you win ;-)
>> I'm sure it's repairable with a lot of time and many years of experience.
>> But neither I nor my fellow students have that experience.
>
> I don't want this to sound like a flame, but surely a 'student' is
> supposed to be learning. In which case you might as well learn about PSU
> repair :-). When I was a Ph.D. student I learnt how to fix SMPSUs, laser
> printers, and so on. Because somebody had to do it :-)
I knew this would backfire... :-)
Well, this discussion has motivated me enough to sit down in front of
that darn PSU again :-)
At least I will learn something...
[...]
> If the PSU will run at all, there's probably not a lot wrong with it. My
> first suggestion would be (as ever) dried up electrolytics.
OK, I'll take that into account.
bye
--
To invent, you need a good imagination and a pile of junk.
-- Thomas Edison
Ron Hudson <rhudson(a)cnonline.net> wrote:
> I don't know how UUCP works, is there a good primer out
> on the net somewhere?
Try these:
http://www.uucp.org/http://www.airs.com/ian/uucp-doc/
uucp is really a file transfer mechanism (Unix-to-Unix CoPy) with some
remote-job-execution stuff layered on top of it. One host sends a
file to another host along with instructions to run some commands that
use the file, e.g. "rmail" to process a received mail message or
"rnews" to process a received news batch. Hosts generally have some
idea about what commands they will allow other hosts to ask them to
execute.
uucp can use any of several different protocols; the administrators at
each end of a link choose a protocol based on the sort of link (e.g.
whether it's 7- or 8-bit clean, perhaps other packetization
characteristics of the link).
Then there's all the infrastructure that was built to make this
livable when it was all you could get. If you know I'm user "fmc" at
node "west", and you're "rhudson" at "cnonline", how do you get here
>from there? Well, maybe you know which hosts talk to which hosts
between here and there and can construct a bang-path for the route,
but you probably don't. This is where the UUCP Map Project and
pathalias fit in: hosts would publish uucp map entries, these map
entries would get collected into regional maps by the UUCP Map
Coordinator for that region, and the regional maps would get posted to
comp.mail.maps every so often. Interested nodes would take a feed of
this group and feed the articles to pathalias, which would work out a
database of how to get there from here. This would allow you to
address your mail to west!fmc, and the mail software at your end would
consult the database to figure out the best way to route the mail to
west, and construct a !-path from that.
When that Internet thing and the domain name system caught on, it
became possible to set up a domain with a mail exchanger that would
know to queue received mail for that domain as uucp jobs. That is how
mail to reanimators.org works: it goes to one of my uucp upstreams and
I poll them every hour or so to get e-mail. This removed a lot of the
need for uucp maps and pathalias, and the maps generally fell into
disuse (most folks didn't bother to delete their map entries when they
moved on from uucp connections to IP connections) so the UUCP Map
Project shut down about two years ago.
-Frank McConnell ...!daver!west!fmc
It didn't sell because it was listed in the wrong categories. Fatal error.
If you are interested, send him an email with an offer. This is how you find
those bargains.
Paxton
Astoria, OR