[Info-vax] retro VMS hardware

[Mark Daniel]

Michael Moroney wrote:
> John Wallace <johnwallace4 at yahoo.co.uk> writes:
> 
>> On Sep 26, 9:56=A0am, H Vlems <hvl... at freenet.de> wrote:
>>> ISTR there are two H4000 models, the first model required a jig and
>>> and a drill to get it installed. The other had the same clamp as the
>>> H4005.
>>> The first model was not fit for use with H4005's and thinwire kit,
>>> there's a heartbeat issue that preventsproper operation in mixed
>>> configurations.
>>> Mixed meaning combining 10base5 and 10base2 gear, like fitting a DESPR
>>> to a 10base5 backbone. There is a hardware fix for the newer H4000
>>> (which makes it an H4000-AB IIRC). It involves removing one small
>>> capacitor. The H4005 has this fixed too.
>>> Hans
> 
>> Rings a bell here too. My vague recollection is that the original
>> stuff was Ethernet whereas the later stuff was 802.3, and the relevant
>> difference in this context is the "carrier detect check
>> (failure)" (aka heartbeat) circuit. Very vague though.
> 
> The issue was of "heartbeat", which uses the collision signal.
> I believe the problem was with repeaters in general, not specific
> to mixed configurations.  Repeaters had problems with heartbeat,
> propagating them as collisions IIRC.  There were also similar
> issues with DELNIs, which are old Ethernet devices.

IIRC we had similar issues when implementing an early switched network 
(non-DEC but vendor escapes me) in '91.  Some 'Ethernet' vs 802.3 
subtlety defeated the move requiring a redesign and another vendor. 
Made a local DECUS Symposium session and Proceedings article.