Switching from 10base2 to AUI (4000/705)

[al.johnston]

I have a vax 4000/705. The 10base2 seems to have gone bad, so I am trying to switch to using the AUI that is also built in to the cpu box. I threw the 10base2/AUI selector switch but AUI is not working (carrier check failed). Is there any change I need to do w/in VMS to do this change? NIC shows up as a EZA (>>> show Ethernet) irrespective of the switch setting. Not sure if that should be the case.

[Hoff]

There is probably a mismatch with the transceiver.

Toggle the heartbeat setting on the transceiver, as a start.

This VAX should also accept a DELQA-SA or DESQA adapter, if you can locate one of these.

Nemonix Engineering was (IIRC) was (is?) offering network adapters for certain of the old VAX boxes.

And not to state the obvious, this system and this hardware is long past its replace-by date for replacement, either with VAX emulation, or with an Alpha or Integrity port and upgrade, or with a more general port.

Some related material I have posted on VAX, VAX emulation, the Q-bus, and on upgrades and ports, is here:

http://labs.hoffmanlabs.com/node/70
http://labs.hoffmanlabs.com/node/10
http://labs.hoffmanlabs.com/node/225
http://labs.hoffmanlabs.com/node/226

The age of this hardware implies there might be some odd-ball Q-bus widget around. If so, then see these:

http://www.arduino.cc/
http://www.pjrc.com/teensy/
http://en.wikipedia.org/wiki/SheevaPlug
http://en.wikipedia.org/wiki/GuruPlug

[al.johnston]

Thanks... I have placed the media converter, cable, and 10baseT swich on another 4000/705 we have and it works fine. Guessing it is a problem with the NIC inside of the VAX's cpu box. Will be swapping out the card that is on the cover of this box and see if that helps.

[Hoff]

That's the S-box I/O bulkhead, and not the NIC.

[al.johnston]

Thanks Hoff. I have been looking for something that would give me a block dia. or schem., but so far all I have found is the IPB... Guess I need L4006 (cpu card) then.

[al.johnston]

finally found one (block dia). Enet ctrl is on cpu card. Thanks again.

[Steven Schweda]

> [...] Guess I need L4006 (cpu card) then.

If you have a free Q-bus slot, then you might
find a DELQA or DESQA cheaper than a CPU.
There'd be a little more effort involved in
rejiggering the network software to use a
different interface (EZA0 -> XQA0?), but
probably not a dreadful ordeal.

[Bob Blunt]

Al, to answer your question, there isn't anything in OpenVMS that you need do to change from the coax to the AUI. However, I have, in the past had much better success if I was able to power cycle a VAX of this type when changing from one to the other.

If you're using a "media adapter" between the AUI connector remember that they work best if setup to provide heartbeat. There are also issues with switches synchronizing to lower speed (10MB), half duplex. Moving the adapter to another system is a pretty good test but when you make that move it should cause the switch to negotiate the connection. Is the adapter connected directly to the AUI interface or are you using an AUI cable so the cabinet will close? I've had those cables go bad and replacing or removing them fixed the problem (my VAX4000-600s were rack-mounted so the cabinet wasn't a problem). You can also test the coax with a doubly terminated T-connector or the AUI connector with a AUI turnaround test connector (Ethernet Loopback connector, DEC part 12-22196-02). In either case, if the ethernet interface on the VAX CPU card is good, a DECnet Phase IV circuit setup to use the on-board ethernet should "come up" and stay up.

Hoff mentioned Nemonix. I think they offer a 10/100 Q-Bus NIC. And you definitely do NOT want a DEQNA. DELQA or DESQA are your best bet if you need an all DEC NIC replacement but can't get the CPU module.

bob

[al.johnston]

Guys, I would like to double check something. The bulkhead (3604) has a circuit card built in to it that has some active (and a heck of a lot of passive) componints. What I wiould like to double check is if any of this (particularly the active componets) are involve in interfaciont the RG-58/AUI connector to the system. This excludes the switch of course. I have already checked the fuses.

As alway, thank you.

[Hoff]

Carrier check is usually the transceiver setting, a failed transceiver, or a cable short, or a screaming NIC out on the coax.

Possibly a bad AUI cable or loose connection.

It can be the bulkhead or the host connection or the host controller, too.

I'd rule out the easy stuff before going after the bulkhead and the NIC, but if the easy stuff has been ruled out...

There was some wonky stuff back in this era with carrier check; mixing the older Ethernet gear and the newer 802.3 was problematic in this particular area. When Ethernet became 802.3, one of the changes was in the heartbeat detection. (Fossil-grade memory here. That switch-over was twenty years ago...)

Add a parallel Q-bus controller as a quick hack to get the box back online, and then get to work on replacing this box, and with porting whatever it is here that has kept this VAX box in use this long. VXworks or odd Q-bus iron or whatever.

[al.johnston]

Switch the two halfs of the wifi hardware from what is obvious.

Connect each of the UFH-R's to its own wifi/10baseT media converter or to its own wifi/10baseT router/switch. The latter is what you connect to the 10baseT to Cable internet (from cable) media convert. (It is just over kill in this situation, but will work fine and will be easier to find, and possibly cheaper because of the volume that are sold)

We would config that wifi box via a normal pc and then hook it to one of the UFH-R.
Thank you for your anser on the two parts. Greatful.

pretty sure I have ruled out everything at this point. I have done this stuff before, but it was a while back.

Problem with using the 10base2 connector was that everything on the net scream about to many collisions. This was true (for the problem system) even when all that was connected was a tee and two terminator.

Problem with using AUI was the carrier absent or carrier framing errors.

[al.johnston]

Da--! Please ignore the above message. Here is what should have been posted:

Thank you for your anser on the two parts. Greatful.

pretty sure I have ruled out everything at this point. I have done this stuff before, but it was a while back.

Problem with using the 10base2 connector was that everything on the net scream about to many collisions. This was true (for the problem system) even when all that was connected was a tee and two terminator.

Problem with using AUI was the carrier absent or carrier framing errors.