10 Dec
2010
10 Dec
'10
5:36 a.m.
Hi everyone,
we've had a look inside our IME 122 calculator and discovered that it is
full of SN14xx logic ICs. They are mainly from TI, but there are also some
from Motorola and others. It seems that they have the
same function and
pinout as the SN74xx parts but there must be a difference since
the
machine has quite a lot of SN1401 (the SN7401 is a quad open-collector
NAND), but there are no pullup resistors anywhere!
Some of the types are SN1400, SN1401, SN1474, SN1490; the ALU is made up
of SN1482 and SN1483.
Anyone knows this series? BTW the supply voltage is 5V.
Christian
11 Dec
11 Dec
2:11 p.m.
On 2010 Dec 10, at 2:36 AM, Christian Corti wrote:
we've had a look inside our IME 122 calculator and discovered that it
is full of SN14xx logic ICs. They are mainly from TI, but there are
also some from Motorola and others. It seems that they have the same
function and pinout as the SN74xx parts but there must be a difference
since the machine has quite a lot of SN1401 (the SN7401 is a quad
open-collector NAND), but there are no pullup resistors anywhere!
Some of the types are SN1400, SN1401, SN1474, SN1490; the ALU is made
up
of SN1482 and SN1483.
Anyone knows this series? BTW the supply voltage is 5V.
I can't find a reference for those numbers, and I haven't seen them
before, however I have seen TI inexplicably producing series identical
or similar to more-common series, but numbered differently. For
example, the SN3900 and SN4500 series are very similar to more-common
DTL series such as the 700/800/900 series, but I have never seen a
reference for the 3900 or 4500 series in TI databooks. One suggestion
might be they were a 'consumer-grade' series, a step below the standard
commercial-grade stuff.
Found your museum's page here:
http://computermuseum.informatik.uni-stuttgart.de/dev_en/ime_122/
index.html
Is PHxxxx a Philips or Siemens number?
Just some suggestions regarding the 1401 outputs and load resistors:
- make in-circuit measurements between the gate output and Vcc, but
with both polarities on the meter leads. This will sometimes indicate
whether there is a resistive element vs. measuring the semiconductor
junctions.
- look for an unused gate on which to measure R of the output.
- perhaps a bit of poor design and/or they were being used at a slow
enough speed they didn't need pull-ups.
17 Dec
17 Dec
12:05 p.m.
It's possible these were bulk untested devices that the buyer could then
test and choose those the met their requirements or they had been tested to
a different standard than the usual 74 series.
I have my 1974 TI TTL data books from my design engineer days and there's no
reference to them there.
Regards
?
Rod Smallwood
?
?????
?
-----Original Message-----
From: cctech-bounces at classiccmp.org [mailto:cctech-bounces at classiccmp.org]
On Behalf Of Brent Hilpert
Sent: 11 December 2010 19:12
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: Unknown TI logic series
On 2010 Dec 10, at 2:36 AM, Christian Corti wrote:
we've had a look inside our IME 122 calculator and discovered that it
is full of SN14xx logic ICs. They are mainly from TI, but there are
also some from Motorola and others. It seems that they have the same
function and pinout as the SN74xx parts but there must be a difference
since the machine has quite a lot of SN1401 (the SN7401 is a quad
open-collector NAND), but there are no pullup resistors anywhere!
Some of the types are SN1400, SN1401, SN1474, SN1490; the ALU is made
up
of SN1482 and SN1483.
Anyone knows this series? BTW the supply voltage is 5V.
I can't find a reference for those numbers, and I haven't seen them
before, however I have seen TI inexplicably producing series identical
or similar to more-common series, but numbered differently. For
example, the SN3900 and SN4500 series are very similar to more-common
DTL series such as the 700/800/900 series, but I have never seen a
reference for the 3900 or 4500 series in TI databooks. One suggestion
might be they were a 'consumer-grade' series, a step below the standard
commercial-grade stuff.
Found your museum's page here:
http://computermuseum.informatik.uni-stuttgart.de/dev_en/ime_122/
index.html
Is PHxxxx a Philips or Siemens number?
Just some suggestions regarding the 1401 outputs and load resistors:
- make in-circuit measurements between the gate output and Vcc, but
with both polarities on the meter leads. This will sometimes indicate
whether there is a resistive element vs. measuring the semiconductor
junctions.
- look for an unused gate on which to measure R of the output.
- perhaps a bit of poor design and/or they were being used at a slow
enough speed they didn't need pull-ups.
12:36 p.m.
It's possible these were bulk untested devices
that the buyer could then
test and choose those the met their requirements or they had been tested to
a different standard than the usual 74 series.
I do not think TI would allow anything with a TI logo untested out of
their factory.
They are interesting, but no more than weird numbers on semicustom
devices. Motorola does the same thing, with their SC series of parts
(not MC or XC)..
--
Will
1:30 p.m.
On 17 Dec 2010 at 12:36, William Donzelli wrote:
They are interesting, but no more than weird numbers
on semicustom
devices. Motorola does the same thing, with their SC series of parts
(not MC or XC)..
It was very common back then. I have some 54175 cerDIPs that I
bought way back when that bear cryptic numbers, but they're National
parts, for certain--they were much cheaper than the correctly-labeled
versions.
Of course, this comes from a time when "glue" logic cost real money.
--Chuck
4:17 p.m.
It was very common back then. ?I have some 54175
cerDIPs that I
bought way back when that bear cryptic numbers, but they're National
parts, for certain--they were much cheaper than the correctly-labeled
versions.
This goes back to transistor, and even tube manufacture.
Given enough money or a big enough order...
--
Will
5:08 p.m.
On 17 Dec 2010 at 12:36, William Donzelli wrote:
Don't forget that many large manufacturers had their own 'house numbers'
for ICs. Teh HP 1820-series is the best know, but Xerox, IBM, ICL, etc
had them too. It's certainly quire common to see a DIL package with a TI
or National logo and an HP 1280- number.
They are interesting, but no more than weird
numbers on semicustom
devices. Motorola does the same thing, with their SC series of parts
(not MC or XC)..
It was very common back then. I have some 54175 cerDIPs that I
bought way back when that bear cryptic numbers, but they're National
parts, for certain--they were much cheaper than the correctly-labeled
versions.
Of course, this comes from a time when "glue" logic cost real money.
It still does... Or at least i find it does...
-tony
6:20 p.m.
Wang Laboratories had house numbers that were put on their IC packages.
(http://oldcalculatormuseum.com/t-wangic.html). Some chips from vendors
other than TI had the original part number covered up (usually by a
white strip of paint), and an internal part number printed elsewhere on
the package. Not clear if this was done by Wang, by some reseller, or
perhaps by the original manufacturer. The parts with TI logos (mostly
7400-series TTL) actually had Wang part numbers put on them by TI, or at
least so it appears.
-Rick
1:26 p.m.
Christian Corti wrote:
Hi everyone,
we've had a look inside our IME 122 calculator and discovered that it
is full of SN14xx logic ICs. They are mainly from TI, but there are
also some from Motorola and others. It seems that they have the same
function and pinout as the SN74xx parts but there must be a difference
since the machine has quite a lot of SN1401 (the SN7401 is a quad
open-collector NAND), but there are no pullup resistors anywhere!
Some of the types are SN1400, SN1401, SN1474, SN1490; the ALU is made up
of SN1482 and SN1483.
Anyone knows this series? BTW the supply voltage is 5V.
Christian
TI and others often produced "House Numbers" for companies that wanted
to hide the 'real' part numbers of components. I would assume this is
what you have here. Thus the SN1474 may not be a flip-flop (7474) as a
result...they may well have jumbled the numbers around to make it
difficult to copy the circuit.
John :-#)#
18 Dec
18 Dec
2:03 p.m.
we've had
a look inside our IME 122 calculator and discovered that it
is full of SN14xx logic ICs. They are mainly from TI, but there are
also some from Motorola and others. It seems that they have the same
function and pinout as the SN74xx parts but there must be a difference
since the machine has quite a lot of SN1401 (the SN7401 is a quad
open-collector NAND), but there are no pullup resistors anywhere!
Some of the types are SN1400, SN1401, SN1474, SN1490; the ALU is made up
of SN1482 and SN1483.
Anyone knows this series? BTW the supply voltage is 5V.
Christian
TI and others often produced "House Numbers" for companies that wanted
to hide the 'real' part numbers of components. I would assume this is
what you have here. Thus the SN1474 may not be a flip-flop (7474) as a
result...they may well have jumbled the numbers around to make it
difficult to copy the circuit.
19 Dec
19 Dec
12:53 p.m.
While this posting doesn't address the SN14xx-chips the original poster
queried about, there is another machine that I have that has some
TI-made SN12xx chips that I was able to find the story of that may be of
interest to others.
The Friden 115x (Printing) and 116x (CRT Display) calculators use three
interesting TI chips that have SN12xx part numbers that I know to be
custom chips made for Singer by TI.
The Friden 130/132 calculators, the predecessors to the 115x/116x lines,
didn't have adders in the classic sense. These machines are all based
on counters. The original architecture of the EC-130 was based on a
series of four five-bit counters that were interconnected in different
ways (by mostly gating logic) to allow them, by counting up and down, to
perform the math operations. After the 130 was introduced, in the
process of developing the EC-132 (which added square root), it was
realized that all of the functionality could be done with three counters
rather than four, and a rework was done of the EC-130 boardsets to use
this realization to reduce the part count, and thus the manufacturing
cost. A few EC-132's were also made with the four-counter architecture,
but soon into production the EC-132 was also changed to the
three-counter design. Each of the three counters had a slightly
different logic design.
The goal of the 115x/116x calculators was basically to shrink the
machines down by using IC logic rather than discrete transistors.
These machines were implemented using a mix of small-scale DTL and TTL
devices. Friden didn't seem picky about vendor...there are chips made
by Signetics (7400-series TTL), Motorola (MC8xx DTL), TI (7400-Series
TTL, SN158xx DTL) all in 14 & 16-pin plastic DIPs.
Along with all the SSI devices, there are three 24-pin MSI devices, part
number SN1286, SN1287, and SN1288. These chips, which were custom made
for Friden by TI (this is mentioned in discussions I've had with one of
the engineers that worked on the development of the machines), implement
the three counters.
Is it possible that the SX12xx-series TI part numbers were for custom
devices?
Rick Bensene
The Old Calculator Museum
http://oldcalculatormuseum.com
2:50 p.m.
On 2010 Dec 19, at 9:53 AM, Rick Bensene wrote:
While this posting doesn't address the
SN14xx-chips the original poster
queried about, there is another machine that I have that has some
TI-made SN12xx chips that I was able to find the story of that may be
of
interest to others.
The Friden 115x (Printing) and 116x (CRT Display) calculators use three
interesting TI chips that have SN12xx part numbers that I know to be
custom chips made for Singer by TI.
The Friden 130/132 calculators, the predecessors to the 115x/116x
lines,
didn't have adders in the classic sense. These machines are all based
on counters. The original architecture of the EC-130 was based on a
series of four five-bit counters that were interconnected in different
ways (by mostly gating logic) to allow them, by counting up and down,
to
perform the math operations. After the 130 was introduced, in the
process of developing the EC-132 (which added square root), it was
realized that all of the functionality could be done with three
counters
rather than four, and a rework was done of the EC-130 boardsets to use
this realization to reduce the part count, and thus the manufacturing
cost. A few EC-132's were also made with the four-counter
architecture,
but soon into production the EC-132 was also changed to the
three-counter design. Each of the three counters had a slightly
different logic design.
The goal of the 115x/116x calculators was basically to shrink the
machines down by using IC logic rather than discrete transistors.
These machines were implemented using a mix of small-scale DTL and TTL
devices. Friden didn't seem picky about vendor...there are chips made
by Signetics (7400-series TTL), Motorola (MC8xx DTL), TI (7400-Series
TTL, SN158xx DTL) all in 14 & 16-pin plastic DIPs.
Along with all the SSI devices, there are three 24-pin MSI devices,
part
number SN1286, SN1287, and SN1288. These chips, which were custom made
for Friden by TI (this is mentioned in discussions I've had with one of
the engineers that worked on the development of the machines),
implement
the three counters.
Is it possible that the SX12xx-series TI part numbers were for custom
devices?
I can't find mention of the SN12xx numbers, but together with the
SN12xx a vague correlation I see in the 1969 TI databook is that
SN1<anything> are designs or numbers which do not originate with TI,
that is, they are either custom or design/numbering originated with
another manufacturer:
SN179xx 900/800 series RTL from Fairchild
SN178xx " " " " "
SN159xx 900/800 series DTL from Fairchild
SN158xx " " " " "
SN154xx custom digital logic circuits
20 Dec
20 Dec
5:14 a.m.
On Sun, 19 Dec 2010, Rick Bensene wrote:
The Friden 115x (Printing) and 116x (CRT Display)
calculators use three
interesting TI chips that have SN12xx part numbers that I know to be
custom chips made for Singer by TI.
Interesting, will have to have another look to our 1154 and 1162...
perform the math operations. After the 130 was
introduced, in the
process of developing the EC-132 (which added square root), it was
realized that all of the functionality could be done with three counters
rather than four, and a rework was done of the EC-130 boardsets to use
this realization to reduce the part count, and thus the manufacturing
cost. A few EC-132's were also made with the four-counter architecture,
but soon into production the EC-132 was also changed to the
three-counter design. Each of the three counters had a slightly
different logic design.
I think that this claim is wrong as we have repaired three EC-132s some
weeks ago, and all have four counters (A, B, C and D), they are clearly
visible and marked on the PCBs. The machines also had very different
serial numbers.
Even the boards on your site (http://oldcalculatormuseum.com/friden132.html)
show the four-register version although you've labelled them "three
register architecture". You'll see that the test connectors at the top of
the sandwich-PCB have marking like A->B, A->D, B->C etc., and also ADV A
and ADV D (advance A resp. D).
The four counters are built all the same way, i.e. the flip-flop with
2N2635. They differ slightly in the logic driving the flip-flop drivers,
the B and C can't shift, and D shifts in the inverse direction of A (it's
more like a rotate with "top" bit inversed and fed back to the bottom FF).
devices. Friden didn't seem picky about
vendor...there are chips made
by Signetics (7400-series TTL), Motorola (MC8xx DTL), TI (7400-Series
TTL, SN158xx DTL) all in 14 & 16-pin plastic DIPs.
Yes, I remember the 15xxx parts in those machines.
Is it possible that the SX12xx-series TI part numbers
were for custom
devices?
Possible, but later the "real" custom devices were mostly MOS and had the
TMC prefix (=TI MOS Custom).
Christian
9:50 a.m.
New subject: Friden EC-130/132 boardsets (Was: Unknown TI logic series)
I wrote:
manufacturing
To which Christian replied:
perform the
math operations. After the 130 was introduced, in the
process of developing the EC-132 (which added square root), it was
realized that all of the functionality could be done with three
counters
rather than four, and a rework was done of the
EC-130 boardsets to
use
> this realization to reduce the part count, and thus the cost. A few
EC-132's were also made with the four-counter
architecture,
but soon into production the EC-132 was also
changed to the
three-counter design. Each of the three counters had a slightly
different logic design.
I think that this claim is wrong as we have repaired
three EC-132s
some
weeks ago, and all have four counters (A, B, C and D),
they are
clearly
visible and marked on the PCBs. The machines also had
very different
serial numbers.
Even the boards on your site
(http://oldcalculatormuseum.com/friden132.html)
show the four-register version although you've labelled them "three
register architecture". You'll see that the test connectors at the top
of
the sandwich-PCB have marking like A->B, A->D, B->C etc., and also ADV
A and ADV D (advance A resp. D).
The four counters are built all the same way, i.e. the flip-flop
with
2N2635. They differ slightly in the logic driving the
flip-flop
drivers,
the B and C can't shift, and D shifts in the inverse direction of A
(it's
more like a rotate with "top" bit inversed and fed back to the bottom
FF).
I will double check the photos of the boardsets in the EC-132 exhibit.
The photos were from a spare boarset that I have, and I may have
misidentified it. Thanks for pointing this out.
The information relating to the different architectures I provided was
taken from Friden's internal service bulletins, which I have copies of.
Friden EC-130s with serial number 8500 and above were three-counter
machines. Machines prior to #8500 were four-counter machines.
Friden EC-132s with serial numbers 3902 and above were three-counter,
with earlier machines four-counter
I have seen both EC-130's and EC-132's with both architectures. I have
three EC-132's, one of which is SN 1984, and it's definitely a
four-counter machine, while the other two (SN 10663 and 11945) are
definitely three-counter machines.
There were some wiring differences between the machines, which did limit
the compatibility of boardsets (e.g., not a good idea to put a
four-counter boardset in a calculator made with a three-counter
boardset, but there are specifics in the service bulletins that give a
matrix of compatibility between the machines based on the individual
board pairs.
I need to scan these service bulletins and put them online on the museum
website, but haven't gotten to it yet.
-Rick
21 Dec
21 Dec
5:29 a.m.
New subject: Friden EC-130/132 boardsets
On Mon, 20 Dec 2010, Rick Bensene wrote:
Friden EC-132s with serial numbers 3902 and above were
three-counter,
with earlier machines four-counter
I've just checked our two machines (the third was from a friend), these
are the serial numbers on the bottom front of the base (and also somewhere
inside the machine):
* 132-33797
* 132-4638-1910
There were some wiring differences between the
machines, which did limit
the compatibility of boardsets (e.g., not a good idea to put a
four-counter boardset in a calculator made with a three-counter
boardset, but there are specifics in the service bulletins that give a
matrix of compatibility between the machines based on the individual
board pairs.
I need to scan these service bulletins and put them online on the museum
website, but haven't gotten to it yet.
The Fridens are very interesting machines, I'm always impressed by the
ingenuity of the designers of early electronic calculators. The Diehl
Combitron series is another example, clearly bearing the hallmarks of
Stanley Frankel.
Christian
19 Dec
19 Dec
8:22 a.m.
On Fri, 17 Dec 2010, John Robertson wrote:
TI and others often produced "House Numbers"
for companies that wanted to
hide the 'real' part numbers of components. I would assume this is what you
have here. Thus the SN1474 may not be a flip-flop (7474) as a result...they
may well have jumbled the numbers around to make it difficult to copy the
circuit.
But as I said they are wired just like the 74xx parts, i.e. the 1474 or
1476 have all the signals at the "right" pins like clock, inputs and
outputs. A gate from a 7400 is often used as inverter by connecting the
two inputs together, the same applies to some 1400. All pins from each
14xx type matches the corresponding 74xx counterpart.
Christian
4967
days inactive
4978
days old
15 comments
8 participants
participants (8)
-
ard＠p850ug1.demon.co.uk -
cc＠informatik.uni-stuttgart.de -
cclist＠sydex.com -
hilpert＠cs.ubc.ca -
pinball＠telus.net -
rickb＠bensene.com -
rodsmallwood＠btconnect.com -
wdonzelli＠gmail.com