"sokolov@alpha.CES.CWRU.Edu" "Michael Sokolov" wrote:

>   So far I have been assuming that VAXstations are being viewed as
>drastically different from both MicroVAXen and VAXservers, which are also
>different from each other, but not as much. Also I was assuming that
>MicroVAXen are more "mainstream" than VAXservers. However, points 2 and 3
>contradict these assumptions, especially the latter one. Is it more correct
>that MicroVAXen kind of stand apart while VAXstations and VAXservers are
>closer together?

The first thing to remember is that when going from one "family" to another, 
things may not remain consistent: for example the VAXstation 2000 and MicroVAX 
2000 are the same CPU but with one jumper moved, but the VAXstation 3100 and 
MicroVAX 3100 are different CPUs.

I don't know what you mean by "mainstream" but I expect that Digital sold more 
MicroVAX 3100s than VAXserver 3100s.

VAXserver 3100s could have been produced in one of two ways. Take a KA41-A CPU 
board and swap the appropriate EPROM(s) (one or two - I cannot remember) and you 
have a KA41-B: you have now converted your MicroVAX 3100 into a VAXserver 3100. 
Now take a VAXstation 3100 KA42-A and flick up the alternate console switch and 
you have something that OpenVMS calls a VAXserver 3100. In both cases the 
original hardware is unchanged except for either a bit in EPROM that the OS can 
get to or a hardware switch that the OS can read the setting of. It is up to the 
OS to decide what, if anything, this difference means. With a VAXserver made 
from a KA41 the OS cannot decide to use video because there isn't any. With a 
VAXserver made from a KA42 the OS could activate and use the video (i.e. you can 
*probably* run a VAXstation as a VAXserver but still use it as a workstation but 
with a serial console ... I've never done this but I see no reason why it could 
not be done in principle).

>   Now we are getting to the heart of the matter. What IS the difference
>between "a workstation" and "a microvax"? People were suggesting to me that
>it's licensing.

Correct (or you could call it marketing instead.

>                You are saying that it isn't. It isn't the console, is it?
>The console is determined by pins 8 and 9 on the DB9 and not by the jumper
>(there are a lot of people who change the console without even knowing
>about the jumper). WHAT IS IT then?

Don't confuse anything that happens on a MicroVAX 2000 with anything that 
happens on a MicroVAX 3100 or VAXstation 3100 or anywhere else. These things are 
not consistent. 

On the MicroVAX 2000/VAXstation 2000 you cannot use both serial console and a 
workstation because of physical constraints: you have to either use a video 
connector or plug in the MMJ-converter-pack to get serial lines.

On a MicroVAX II you add a few extra boards (VCB01 or VCB02) and the console 
will notice and use them and you have a workstation. Playing with the DB9 
console connection has no effect - in fact in a BA213 chassis has no DB9 it has 
MMJ instead and yet the behaviour is the same as in BA23 or BA123.

On MicroVAX 3100 there is no video circuitry. KA41-A/KA41-B or KA41-D/KA41-E 
difference is just a single bit in the EPROM which the software reads via a 
system call. That's all there is to the VAXserver 3100.

On the VAXstation 3100 there is a switch you can flick which the software can 
read. The video always exists. I don't know whether the console code still turns 
on the video when in VAXserver mode but there is nothing (AFAIK) to stop the OS 
doing the initialisation itself (whether it actually does it is another 
matter!). The console code itself will write to the video when in VAXstation 
mode and to a serial line when in VAXserver mode. (Actually, it may be the 
presence or absence of a keyboard directly connected to the VAXstation box that 
determines whether the OS calls it a VAXserver or not ...).

>> The MicroVAX 3100s (KA41 series) do not have video circuitry so there is
>> no question of converting a KA41 to a KA42 or vice-versa.
>   What I'm really wondering about is whether such circuitry exists in the
>first place or is it just a software construct. Take KA410 for instance. By
>now I have learned that it has never had any components omitted. However,
>imagine hypothetically that you wanted to make a MicroVAX-only version of
>it. Is there some actual circuitry that's JUST for the video and could be
>hypothetically removed? Or is the frame buffer taken out of main memory and
>whatever chip produces the video signal also does some vital functions like
>DRAM refresh?

On the KA42 there obviously is video circuitry. On the KA41 there is no video 
circuitry that I can see. I can't imagine why there would be any: it is a 
different board layout. The KA41 and KA42 may well have been put together by 
related design teams, possibly even the same team, but given that the layout has 
been changed, why waste money putting extra circuitry into the MicroVAX knowing 
that it will never be used?

As far as a hypothetical KA410 goes, it could have been designed without video 
circuitry. I don't know how the hardware is arranged but if I remember (and if 
noone else posts it first) I'll look in the tech manual tonight.

>   There is one more twist to this hypothetical issue. Imagine there were a
>version of, say, KA410 without base mono video. Then maybe it would be
>possible to make that beast use a graphical console by adding a GPX or SPX
>video board. Assuming the DB15 connector and the circuits leading to it
>were left in place, the added board would be able to send its video output
>to the monitor, right? Keyboard and mouse would work, since they are purely
>software constructs made out of fully generic serial ports, right? If these

I suspect that the GPX and SPX addon boards require the base mono video 
circuitry which the KA42/KA43 have. I don't remember whether the VAXstation 2000 
uses the same GPX board as the VAXstation 3100s. However, if you added a GPX to 
your hypothetical KA410-without-mono-video then it would probably not work.

>arguments are correct, they lead to an important observation that the real
>meaning of KA41 not supporting video is not that it doesn't have base mono
>video but that it can't have any video at all. Suppose that I were to stick

The real meaning of KA41 not supporting video is that it has no video.

>a GPX or SPX board in a KA41 system. It should be mechanically possible,
>since on KA410 video boards and communication options are inserted into the
>same connector, and KA41 should take the same communication options as
>KA410. So KA41 should be mechanically compatible with GPX and SPX boards,
>but it's a completely different story that they wouldn't work because of
>the lack of the DB15 connector. If the arguments above are correct, this is
>the real meaning of KA41 not supporting video. Are they?

You can mechanically insert the GPX/SPX boards into a KA41 but although it has 
the same connectors as a KA42 they are connected to different circuitry: in this 
case they are for the asynch/synch comms board options. Looking at the boards, I 
think that where the KA42 has video circuitry, the KA41 has a SCSI controller 
chipset.

If you mean "it could have been designed like this" then the answer is that 
maybe it could (assuming sufficient pins on the connectors for all the required 
signals and enough space on a relaid out board to add back the base video) but 
in fact it wasn't done this way.

>It's easy to tell if a system uses this daughterboard design. This
>daughterboard sits in the upper-left corner of the system unit if viewed
>from the back, it has this funny external connector, and the latter is
>covered by a very characteristic cover. These characteristics of the
>external connector almost certainly guarantee that the daughterboard design
>is used. It is possible in theory to make the external connector look right
>and be in the right place while keeping the SCSI chips on the system board,
>but it's far from easy and I'm sure no one would do that. In this case the
>external connector would have to be crimped on a ribbon cable. The
>connector is a high-density one, so the ribbon cable would have to be high-
>density too, which causes other problems. The connector would have to be

The VAXstation 3100-30 and 3100-38 use that motherboard but the 
VAXstation 3100-76 does not ... it has a strange internal cable that carries the 
internal SCSI bus around the system box and carries the external SCSI bus to the 
external 68-pin connector. So you cannot tell just from looking at the box. I 
have not seen the insides of a DECstation 3100 recently enough to remember what 
happens in that box.

>   In the quote above you are saying that this daughterboard design is used
>in M30 and M38. I have to include M40 and M48 too. In the manual for VS3100
>M40 the back of the system unit is drawn very well, and I clearly see that
>the external connectors looks just like the one I'm used to seeing on my
>M38s, and is located in the right place under the right type of cover. Plus
>you were assuring me yourself that M40 uses exactly the same system board
>as M30 and M48 uses exactly the same system board as M38. Well, the same
>system board implies the same daughterboards accepted, right?

Yes. The distinction between VAXstation 3100-30 and VAXstation 3100-40 is just 
the box size: the software cannot tell the difference. Same goes for VAXstation 
3100-38 and VAXstation 3100-48. The VAXstation 3100-76 came in 
one-size-fits-all. The MicroVAX 3100-10 and MicroVAX 3100-20 are the same CPU in 
a different box and the same is true for the 3100-10e and 3100-20e. And also the 
MicroVAX 3100-30 and MicroVAX 3100-40 are just the same electronics in a bigger 
box. In all cases what works in one box should work in the other, subject to 
mechanical constraints.

>> The VAXstation 3100-76 has its SCSI interfaces on the main board.
>   Immediately upon reading this I put my coat on and took a walk to the
>next building where there is an M76. It isn't mine, so I can't open it up
>and see it directly, but the back panel suggests the SCSI daughterboard is
>there, since it looks EXACTLY like that of my M38s (you can't tell which
>machine it is without looking at the front).

You cannot tell which machine it is by looking at the fron either, it is quite 
possible to take apart enough of the casing to get that little medallion off and 
swap it for another one :-) When you do get to look inside you'll see that cable 
I referred to above ... it is quite weird knowing that there are two independent 
SCSI busses and seeing the two NCR chips on the motherboard (right by the 
connector) but finding only one cable!

>connector is up there in an extremely inconvenient location. By the way,
>what's the difference between DS2100 and DS3100?

IIRC: one runs at 12MHz and one at 16.7MHz but I haven't played with these much. 
There was a DECsystem 3100, which was just a variation on the DECstation 3100: 
I'm guessing that it was a DECstation 3100 with no monitor or keyboard but I 
don't actually know.

>   I'm pretty sure that there was a significant change in this area between
>KA41-based ones and M30+. As you say yourself, the former use NCR 5380.
>This means SCSI-1. I'm pretty sure that the latter use NCR 53C94, which is
>SCSI-2.

I have a MicroVAX 3100-80 motherboard in the cupboard behind me and it uses an 
NCR53C94. Given that it is of the same vintage (and from the same design team) 
as the 3100-30/40 it seems likely that those use that same chip.

>   I doubt that KA41-based boxes have SCSI-2 external connectors, since
>they are SCSI-1 after all. Plus Thordur says that on his box it's
>Centronics-50 (the standard connector for SCSI-1). You are referring to
>Amphenol, but I don't quite know what it is. Is it the SCSI-1 one

I thin kwe are talking about the same connector ... the Digital supplied 
terminators for that connector are made by Amphenol (or at least that is stamped 
on the connector) and I find I plug fewer printers into the wrong socket if I 
call it that!

>Plus Thordur indicates that it's above this line. What I'm assuming is that
>the system board has only internal SCSI connectors, and a ribbon cable runs
>up from one of them to an upper opening where it ends with a Centronics-50
>connector crimped on it. Is this the case?

The SCSI connector is as Thordur indicates and is supplied by a ribbon cable 
internally (at least that's what I remember, it certainly is not on the 
motherboard).

>MicroVAX 3100. They match in serial ports, communication options, the
>location of the SCSI chips, nicknames, and, as it turns out now, the
>external SCSI connectors. By the way, do you remember that in the SHOW DEV

Finding matches between designs is not unusual ... when designing a new widget 
it makes sense to see what has been done before and to reuse parts/interfaces 
etc. where appropriate. Internal codenames mean nothing: the idea (I presume) is 
to allow teams to talk about "Comet" without having to describe what "Comet" is 
each time. The final marketing name change more times than anyone can remember 
and will only stabilise the day before the product is announced. (In the case of 
the AlphServer 2100 series the name changes even after the box has shipped :0).

Teammate vs Teammate II means nothing other than perhaps TeammateII was a 
follow-on to Teammate i.e. something to fill the same marketing space.

>   Now I also see that there was a re-design involved. Originally I have
>thought that even if KA41 and KA42 are different, they were still designed
>at the same time by the same people. Now it seems to me that KA42 is the
>original mainstream version and KA41 is a side branch. Their ideologies are
>completely different, so it's clear that they were designed by people with
>completely different ideological views, and the ideology of KA42 is much
>more in line with the rest of DEC stuff, which suggests that KA41 is a side
>branch and not vice-versa.

At the time (late 1980s) worstations and low-end systems were seen as different 
things. Taking one basic design and adapting it to two different markets by 
slightly redesigning would make sense. These days you can take a PC164LX and 
throw in a decent PCI video card and voila - system is now workstation.

>   Note that KA42-B's performance of 3.8 VUPs equals that of KA655, and
>KA42-A's performance of 2.8 VUPs equals that of KA650. KA650 has a 64 KB
>cache, so I'm assuming that KA655, KA42-A, and KA42-B do too. KA41-A/B's
>performance of 2.4 VUPs, on the other hand, matches that of KA640, which is
>like KA650 but has 4 MB of RAM and the DSSI interface on-board. I guess
>that KA640 is a bit slower than KA650 because when squeezing in the RAM and
>the DSSI interface the designers had to take something out, and they
>decided to take out the 64 KB cache. It is plausible that the same was
>necessary when squeezing in the NCR 5380 chips and the supporting logic to
>go from KA42 to KA41.

Here's what I have:

VAXstation 3100-30: CVAX chip @ 90ns, 64KB external cache
VAXstation 3100-38: CVAX chip @ 60ns, 64KB external cache
VAXstation 3100-76: Rigel chip @ 28ns, 128KB external cache
MicroVAX 3100-10:   CVAX chip @ 90ns, unknown external cache
MicroVAX 3100-10e:  CVAX chip @ 60ns, unknown external cache
MicroVAX 3500/3600: CVAX chip @ 90ns, 64KB external cache
MicroVAX 3800/3900: CVAX chip @ 60ns, 64KB external cache
MicroVAX 3300/3400: CVAX chip @ 100ns, unknown external cache

>   As for MV3100 M30+ and VS4000, it seems to me that the system boards of
>different models, however different, should be plug-compatible. First, I
>believe that DEC allowed all M30+ models to be in-cabinet upgraded to

The MicroVAX 3100-30 onwards can probably all be swapped for one another (i.e. 
the boxes are the same) but since all of them are essentially a single board it 
just means the designers were given a goal of making their new baby fit in the 
existing box. The NVAX based ones are all variations on the same design anyway 
(maybe a faster NVAX chip here, more cache there etc.). Only the very latest 
ones (MicroVAX 3100-88/98 and VAX 4000-108) differ in box design and that was 
presumably done to share boxes with one of the Alpha line.

VAXstation 4000 boxes are different (at least the VLC and VAXstation 4000-60 
are, I haven't seen a VAXstation 4000-90).

>   As for SCSI, a DEC paper says that KA50 (MV3100 M90) and KA52 (VAX 4000
>model 100) use NCR 53C94 (the SCSI-2 chip), as I have expected. I think

The SCSI interface is on board, but the SCSI connector is on the box and is 
connected to via a ribbon cable.


Antonio


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