"mxs46@po.CWRU.Edu" "Michael Sokolov" wrote:

>   Antonio Carlini <carlini@marvin.enet.dec.com> wrote:
>> The MV3100-10/20 and VS3100-30/40 do not correspond in any real way: you
>> cannot convert from one to the other except by swapping the board.
>   Oh sure you can, you don't even have to open up the box, just flip a
>little switch on the back :-). This won't work if you are changing from
>MV3100 M10/10e/20/20e to VS3100 M30/38/40/48 without putting in a GPX or
>SPX video board, but will work in all other cases. I would imagine that all
>NetBSD/vax users are "converting" their VS3100s to equivalent MV3100s this
>way, since NetBSD/vax doesn't work with the graphical console (I think).

Flicking the switch on a VAXstation 3100 Model 38 informs the console that you 
wish to run without the graphics console - it does not turn the VAXstation into 
a MicroVAX. If you do this on a system running VMS I *believe* it will refer to 
the config as a VAXserver of some sort, but it was never supported (AFAIK).

>> (It is possible that one board was a re-design based on the other, but
>> otherwise they are distinct).
>   The system boards used in the MV3100 and VS3100 models in question are
>different versions of the SAME BOARD. All of them use the same PCB, and
>different versions are made by either putting in or omitting some
>components. The VS3100 version has all components in place, while the
>MV3100 version has the monochrome video and the cache memory omitted. And
>the board design allowed for different versions from the very beginning, so
>there was no re-design involved.

This is not true. I have in front of me part number 54-18856-05 [a MicroVAX 3100 
Model 10e CPU (KA41-E, actually a VAXserver 3100-10e but it differs from the 
MicroVAX version only in the ROMS)] and part number 54-19356-01 [a VAXstation 
3100 Model 38 CPU (KA42-B)]. There are many similarites e.g. the corner near the 
ethernet connection and the layout of CPU and associated support chips and also 
the major connectors for add-ons (e.g. memory). However, there are sufficient 
differences such that you could not take one and turn it into the other just by 
adding components; for example, there are some differences near the PSU 
connection (which, BTW, is oriented differently), the chips down the LHS are 
different and the KA41-E has two NCR5380 SCSI controllers whereas the KA42-B has 
no-onboard SCSI and nowhere to mount those chips.

The designs are clearly related, but they are not the same board, they were even 
quoted as having slightly different performance (3.5VUPs for the KA41-E and 
3.8VUPs for the KA42-B).

>   Incidentally, the situation with the 2000s is somewhat similar. The
>system boards of MV2000 and VS2000 are the same except that the VS2000
>version has all components in place and the MV2000 version has the
>monochrome video omitted. But cache memory is not a factor in this case,
>since 2000s never have it. Also the naming is much more technically
>appropriate in this case: the model number is 2000 and the system board
>number is KA410 in both versions.

The MicroVAX 2000 and VAXstation 2000 are the same board: a single jumper is 
moved to tell the console whether it is a VAXstation or not.

>   OK, I didn't know that the MicroVAX version calls itself KA41, I have
>thought that both are called KA42[0]. I'm pretty sure though that the
>software-accessible system board ID is "420" in both versions. (That's why
>I keep writing "KA42[0]".) Note that KA41 looks confusingly close to KA410,
>and this confusion surely looks intentional. For some strange (and probably

The OpenVMS errorlog utility referred to the MicroVAX 3100-10 and 3100-10e CPUs 
as KA420, and also referred to the VAXstation 3100-30 and 3100-38 CPUs as KA420 
too. But that is just text inside the errorlogger - but the XSID field for the 
two class of machines is different (I think - I can't check just now).

>stupid) political reason MV2000 and MV3100 were looked upon as being of the
>same rank (they were nicknamed "TeamMate" and "TeamMate II", respectively),
>while VS2000 and VS3100 were looked upon as being completely different
>("VAXstar" and "PVAX" don't sound even close). And apparently at no point
>had those asses in marketing or whatever even thought of doing The Right
>Thing: putting the MV3100 and VS3100 versions on the same level.

I don't know what you mean here but the codenames don't mean much. I don't 
recall what licencing tiers the various machines were on but licencing has very 
little to do with the CPU implementation.


>> The MicroVAX 3100-30 and 3100-40 (same machine, different box) use a KA45
>> CPU board. (I've never seen any reference to a KA44).
>   MV3100 M30/40 and VS4000 VLC are twins just like the MV3100 and VS3100
>models discussed above. I'm pretty sure that the software-accessible system
>board ID is "440" in both versions. That's why I have thought that the
>system board in MV3100 M30/40 is called KA44. You are saying that it's
>actually KA45. Oh well. But no matter what you call it, it's still useless,

I don't know what the KA44 was, if indeed there ever was one. I have never seen 
a reference to one. I don't believe that the VLC motherboard (KA48) was the same 
as the MicroVAX 3100-30 (KA45) but I don't have a VLC around to check.

>since neither Ultrix nor Berkeley UNIX(R) nor NetBSD will run on it. Your
>(the owner's) best bet is, IMHO, to sell it to some VMSer and buy a VS3100
>M38, M48, or M76, or trade directly.
  
Surely the best option is to sit down and port NetBSD to the MicroVAX 3100-30? 
That is the whole point isn't it?

Antonio

Antonio Carlini                            Mail: carlini@marvin.enet.dec.com
DECnet-Plus for OpenVMS Engineering
Digital Equipment Corporation              Worton Grange, Reading, England