21 Oct
1999
21 Oct
'99
2:52 a.m.
IIRC, the 8080 was about 4000 MOSFET transistors. If
you implemented it
with individual FETs, and packed it densely, I think you could fit it in
a 10.5" high rack space easily, and a 5.25" high rack space with difficulty.
Of course, you'll need plenty of forced air cooling. From a serviceability
point of view, building it less densely is clearly better.
True.
If you implemented it with bipolar transistors
configured as saturating
logic, it would require perhaps twice as many transistors and a lot more
resistors for TTL logic, or 50% more transistors, a lot of diodes, and a
lot of resistors for DTL logic.
Also true. A lot of the power estimates have been based on the assumption that
we have bipolar transistors. What about discrete mosfets? We could do NMOS or
even CMOS designs directly that way...
I liked Hans's suggestion (which I have now deleted, alas) of a museum exhibit
with three identical computers, but with processor as single chip, board of
gates, and rack of trannies for comparison. It does provide a service that you
couldn't get another way...
... or could you? Didn't PDP8 come in all three versions?
I've often wondered if you could build a transistorised computer without pcbs at
all. You know, trannies on tag board, little plugs bolted onto the ends (or on
flying leads) and so on... I was assuming ECL for two reasons - if you use
early transistors (Ge), ECL would probably be the only way to go at all fast;
and ECL gives the advantage of easy differential line drivers and receivers for
long interconnects.
Philip.
This E-mail message is private and confidential and should only be read
by those to whom it is addressed.
If you are not the intended recipient, you are hereby notified that any
dissemination, distribution, copying, reproduction, modification or
publication of this communication is strictly prohibited. Please delete
the message from your computer and destroy any copies.
This message is not intended to be relied upon by any person without
subsequent written confirmation of its contents. This company therefore
disclaims all responsibility and accepts no liability of any kind which
may arise from any person acting, or refraining from acting, upon the
contents of the message without having had subsequent written
confirmation.
If you have received this communication in error, or if any problems
occur in transmission please notify us immediately by telephone on
+44 (0)2476 425474
22 Oct
22 Oct
12:16 a.m.
> IIRC, the 8080 was about 4000 MOSFET transistors.
If you implemented it
And how many depletion load resistors.?
Also true. A lot of the power estimates have been
based on the assumption that
we have bipolar transistors. What about discrete mosfets? We could do NMOS or
even CMOS designs directly that way...
MOSFETS solve half the problem. There are still load resistors on the
chip never minding all of the registers are constructed like DRAM IE: they
are capacitors. That dynamic logic and memeory is very hard to simulate
at tyhe discrete level. If you simulate them using static cells the
transistor count goes up a factor of 6 for each BIT of register plus a
lot of resistors that dynamic logic does not use.
Allison
5:18 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1548 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991022/c7e66cc9/at…
7:25 a.m.
... or could
you? Didn't PDP8 come in all three versions?
Yes it did, but there were _slight_ differences in the instruction set
IIRC (rather like the differences in the PDP11 instruction set between
similar-ish models). But having a straight-8, a PDP8/e and a DECmate all
running side by side
would be an interesting exhibit. Pity I can't do it...
a PDP-5, straight-8, 8I, 8E, 8A and DECMATE (I/II/III?). Each represents
a level of difference on the base design. For example the 8A had
mechanism in hardware for stacks (actually IOTS so it's retrofitable!).
The DECmates represent two different chips the 6100 (base 8Eish) and the
6120 (more 8Aish) both having a control pannel mode unique to them.
The commonality is great from the 5 though the DMIII that you can wite
program that run on both. This made the PDP-8 series likable (software
investment).
Allison
3:31 p.m.
New subject: Calculator Comparisons
Tony wrote:
If we consider calculators for a moment, you can get 2
out of the 3
machines by :
Casio AL1000 (discrete transistor, programmable, not scientific)
Casio AL2000 (ICs, much the same functionality as the AL1000)
???? (All-in-one-chip equivalent)
To check these out, see:
http://www.geocities.com/oldcalculators/commal-1000.html
This is a Commodore AL-1000. It's a Commodore-badged OEM version identical
to the Casio AL-1000.
and
http://www.geocities.com/oldcalculators/casal2k.html
This is the Casio AL-2000.
You can see quite a difference in size and number of boards as a result
of the IC technology used in the AL-2000. Most of the IC's are
small scale, but there are a few medium scale (adder and complimenter)
IC's in the AL-2000, whereas the AL-1000 all transistor.
The third one would be a little more difficult to find, as Casio pretty
much abandoned the architecture of their early programmable machines
in their later machines, so there's much less of a one-to-one
correspondence. The programming capability on the AL-1000 and
AL-2000 was extremely simple-minded.
Or :
HP9100 (discrete transistors, programmable, scientific)
??? (Equivalemt with simple ICs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
How about the Cintra/Tektronix 909? It's got much
of the functionality that the HP9100 has (but not quite as much
memory, and lacks branching functions), and is implemented
with mostly SSI and MSI TTL & DTL, along with MOS RAM and ROM.
Information on it is at:
http://www.geocities.com/oldcalculators/tek909.html
HP65 (Almost the same functionality in a handheld)
Only worrying thing would be putting the AL2000 and the
HP9100 alongside
each other. They're almost the same size, but the HP uses 'older'
technology and does a lot more...
No doubt that HP pulled a coup on the 9100 machines...the other
calculator makers couldn't come close in features, speed, and power
until a couple of years later...and by that time, HP had the 9800-series
machines, which were again a quantum leap in technology.
The AL-2000 is actually quite a bit smaller than the 9100, though.
Rick Bensene
The Old Calculator Web Museum
http://www.geocities.com/oldcalculators
23 Oct
23 Oct
6 a.m.
New subject: Calculator Comparisons
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 4018 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991023/57494d03/at…
22 Oct
22 Oct
7:07 a.m.
Philip.Belben(a)pgen.com wrote:
I was assuming ECL for two reasons - if you use early
transistors
(Ge), ECL would probably be the only way to go at all fast; and ECL
gives the advantage of easy differential line drivers and receivers
for long interconnects.
You're absolutely right. These are the exact reasons that IBM's early
transistor computers (7090, Stretch/7030) used non-saturating logic.
But they did you PCBs with wire-wrapped backplanes.
Their tube computers did not use PCBs, but instead eight-tube pluggable
units that were hand-wired.
And someone will no doubt feel obliged to correct me if I fail to
mention that the earlier IBM machines such as the 603 multiplier and 604
calculator didn't use the eight-tube pluggable units. That was a later
innovation. I'm not sure whether the 650 used them.
9193
days inactive
9195
days old
6 comments
5 participants
participants (5)
-
allisonp＠world.std.com -
ard＠p850ug1.demon.co.uk -
eric＠brouhaha.com -
Philip.Belben＠pgen.com -
rickb＠bensene.com