Maker Pro
Maker Pro

Answers to old-fart electronics quiz

J

James Meyer

This brought to mind a fairly long running discussion, from a few years
ago, that never came to a definitive conclusion. (That you probably
know the answer to).

Why was the 7400 TTL power supply 5+/-5% volts?

Weren't the precursors to TTL, RTL and DTL, also 5 volts? Perhaps it's
a legacy issue.

Jim
 
T

Tweetldee

James Meyer said:
Weren't the precursors to TTL, RTL and DTL, also 5 volts? Perhaps it's
a legacy issue.

Jim

ISTR that there was a flavor of RTL that ran on 3.5V.
--
Tweetldee
dgmason44 at comcast dot net (Just subsitute the appropriate characters in
the address)

Never take a laxative and a sleeping pill at the same time!!
 
S

Spehro Pefhany

Weren't the precursors to TTL, RTL and DTL, also 5 volts? Perhaps it's
a legacy issue.

Jim

RTL was 3.6V, IIRC. DTL was 5.0V, and ECL used -5.2V originally, and
4000 series CMOS works from <5 to around 15V. All were available by
around 1970, I think, with RTL and DTL dying off around that time.

Best regards,
Spehro Pefhany
 
S

Spehro Pefhany

Richard said:
The 1101?

i1103, IIRC. 5V, -5V and 12V. A whopping 1024 x 1 *bits* in a DIP-16
with the familiar RAS/CAS multiplexed address bus. One of the most
financially successful chips of all time.

Best regards,
Spehro Pefhany
 
Y

YD

I read in sci.electronics.design that Jan Panteltje
01.evisp.enertel.nl>) about 'Answers to old-fart electronics quiz', on
Thu, 22 Jan 2004:



No, that a triode-hexode. There wasn't a B9A based heptode, AFAIK, but
the code for that would be EK8x. There was an EK90; only 7 pins! Some
'K's were octodes.

I have an EK2 in my box of old valves. Big red coke-bottle thing with
a grid cap on top. Took it out of my beloved 1938 Ericsson that sadly
burned out something and I was too young to know how to fix it. I've
kept about 40 of the things, ranging from the 30s to the 50s, really
should think about making a display case some day.

- YD.
 
M

Michael A. Terrell

Spehro said:
i1103, IIRC. 5V, -5V and 12V. A whopping 1024 x 1 *bits* in a DIP-16
with the familiar RAS/CAS multiplexed address bus. One of the most
financially successful chips of all time.

Best regards,
Spehro Pefhany

The 1101 was the earlier 256 bit version, and the 1102 was a 512 bit
version. I remember seeing a news story when the 1101 was released.
"King core is dead!" There were a lot of timing issues with the 1101,
and by the time they were ironed out, the improved, and larger chips
were available. Remember computer terminals with shift register memory?
 
J

John Woodgate

I have an EK2 in my box of old valves. Big red coke-bottle thing with
a grid cap on top.

And a side-contact base instead of pins. It's not big by comparison with
its predecessors.
 
J

Jan Panteltje

[repost]
8. Traditional op-amp ICs were designed whenever possible to run on +- 15
volts. Why that voltage?

Allowing a clean output swing of +- 10V, a convenient scale for their
original service as analog-computer computation elements. Vaccuum-tube op
amps set the trend earlier with +- 100V standard output swings. See also
Note 3 on op amps.

Welcome back.

This brought to mind a fairly long running discussion, from a few years
ago, that never came to a definitive conclusion. (That you probably
know the answer to).

Why was the 7400 TTL power supply 5+/-5% volts?


Mark Zenier [email protected] Washington State resident
5 was easier to count on the fingers of one hand.
 
L

Leon Heller

Jan said:
[repost]
Max Hauser said:
8. Traditional op-amp ICs were designed whenever possible to run on +- 15
volts. Why that voltage?

Allowing a clean output swing of +- 10V, a convenient scale for their
original service as analog-computer computation elements. Vaccuum-tube op
amps set the trend earlier with +- 100V standard output swings. See also
Note 3 on op amps.

Welcome back.

This brought to mind a fairly long running discussion, from a few years
ago, that never came to a definitive conclusion. (That you probably
know the answer to).

Why was the 7400 TTL power supply 5+/-5% volts?


Mark Zenier [email protected] Washington State resident

5 was easier to count on the fingers of one hand.

The previous DTL logic family already used 5V and it had to be
compatible with that. I don't know for sure why DTL used 5V, though. It
was probably a design compromise between speed, ease of manufacture and
power consumption. IIRC the Mk II Datapac DTL logic modules (discrete
components) that were designed at English Electric, Kidsgrove, when I
worked there in the 1960s used 6V.

Leon
 
J

John Woodgate

I read in sci.electronics.design that Leon Heller <[email protected]>
The previous DTL logic family already used 5V and it had to be
compatible with that. I don't know for sure why DTL used 5V, though. It
was probably a design compromise between speed, ease of manufacture and
power consumption.

Weren't these early digital transistors gold-doped to increase speed, at
the cost of having a Vceo (Vcbo?) not much greater than 5 V?
IIRC the Mk II Datapac DTL logic modules (discrete
components) that were designed at English Electric, Kidsgrove, when I
worked there in the 1960s used 6V.

Discrete devices wouldn't have the same restrictions.
 
L

Leon Heller

John said:
I read in sci.electronics.design that Leon Heller <[email protected]>
wrote (in <[email protected]>) about 'Why
does TTL run on 5 volts (was Re: Answers to ...)', on Sat, 24 Jan 2004:




Weren't these early digital transistors gold-doped to increase speed, at
the cost of having a Vceo (Vcbo?) not much greater than 5 V?

I don't think so, the systems were quite slow. I think they used the
2N706 a lot - nothing special about that.

Leon
 
J

Jim Thompson

I don't think so, the systems were quite slow. I think they used the
2N706 a lot - nothing special about that.

Leon

Gold-doping kills minority carriers, thus significantly cutting down
on the storage time component of the switching speed... device comes
out of saturation more quickly. The down-side is that gold-doping
also reduces beta. Caused me significant heart burn back in the early
PLL days.

TTL was gold-doped until the Schottky variants came along.

CMOS is majority-carrier-based.

I don't think conventional TTL will fit in less than 5V because of
resistive pull-ups/bias-tracking. Schottky will fit in less voltage
because you can use current mirrors.

...Jim Thompson
 
M

Max Hauser

There is a story or two below, beyond immediate topic.

Mark Zenier said:
Welcome back.

Thanks. Sehr freundlich. (I am still responding to a presumption about my
name in another thread.)
This brought to mind a fairly long running discussion, from a few years
ago, that never came to a definitive conclusion. (That you probably
know the answer to).

Why was the 7400 TTL power supply 5+/-5% volts?

No idea. I do remember that DTL ran on 5, DTL was standard before TTL (and
many firms designed their TTL families to interoperate and intermarket with
their established TTL -- Fairchild "CCSL" 3000 series and so on). And that
RTL like Fairchild's uL900, 914, 923, 926 (I think I still have some of
those in my Junk Box), as already cited, around 3.6V. The commercially
successful DEC discrete-component logic PCB's, une of many things nicknamed
over the years "flip chips" (R series -- red -- "2 MHz;" W series --
white -- "10 MHz") used in some versions of the seminal PDP8 and other
computers -- had *if memory serves* 0 and -3V logic levels and multiple
supplies including +10V, the logic based on PNP discretes.

We received a couple of PDP5's -- predecessor to PDP8, larger -- for the
Berkeley Computer Club when I was an undergraduate there in the middle
1970s. They were not functional and neither were we, by which I mean that
none of us students had time to repair them. A few of us students (not I)
had discovered the Internet then and one of us students achieved the
doubtful distinction of arrest by FBI for some of the earliest hacking into
forbidden sites, a story in its own right with spicy elements that I'll save
for in person some time. (I do like good stories.) Here's another good one
anyway. The two PDP5's stood idle in our basement office in the student
union building, an office we'd obtained along with funds (used for a public
educational program) through the skills of the founder (or restarter
actually) of the club, who had instinctive political ability, and maneuvered
to get annually for us a slice of the ASUC budget that was eagerly contested
among very many student organizations, many of them political and (to my
unsophisticated eyes) similar looking. But as it happens our little office
was located right between those of the (Trotskyist) Spartacus Youth League
and the (Stalinist) Revolutionary Communist Youth Brigade, high-profile
bitter rivals always suspicious of each other. (At the time, newspapers
reported one of the groups, sporting rakish red berets -- very revolutionary
looking -- having meetings that got broken up by the other group showing up,
uninvited, carrying axe handles. That's how it was. I didn't know these
groups well but from appearance and statements I got the impression that
most of the members were middle-class suburban college kids, some of them
probably from well-off families. It is often so. That was not my own
background by the way, my parents had been "beatniks" in the 1950s, trained
in fine arts and hung out with a bohemian crowd that included various
Wobblies and Communists and Lincoln-Brigade veterans but there I go, getting
off track again.) Anyway, one of our Computer Club members, the same one
who got the attentions of the FBI, had a fine idea for the idle computers,
which though inactive were formidable looking. They stood against a wall
facing the door of the little office so you could see them from the partial
glass in the door. The mischievous member posted a sign inside the door
glass, leaving the computers visible, that "Both computers are on loan to
the Campus Surveillance Project" -- there was of course no such nonsense,
but the SYL and RCYB could be reasonably relied on to not know that (it is
often so) and to get paranoid about this operation right between their own
two offices. The note went on to refer curious members to a Sergeant
so-and-so of the Campus Police with a contact number that was, I believe,
one of the handy local TelCo test numbers, a busy-beep generator probably.
Hoped to offer the intrepid comrades some mild distraction from beating each
other up.

Have a follow-up account of early voltage standards but one story at a time.

Max
 
N

N. Thornton

John Woodgate said:
I read in sci.electronics.design that N. Thornton <[email protected]>
'Answers to old-fart electronics quiz', on Fri, 23 Jan 2004:

OK. For extra credit, what is a 'pi mode' output stage?

I'm not very up on microwave stuff. I had a looksee. Just call me
creditless. :( I did find EER amplifiers tho, and posted under I think
it was Amplifier Classes or something.


Regards, NT
 
M

Max Hauser

General point on digital vs analog power supplies.

Mark Zenier said:
Why was the 7400 TTL power supply 5+/-5% volts?

Though I don't know good answer to that, I do observe over many decades of
parallel development of modular circuits for digital logic and analog op
amps (and related analog building blocks) that the digital circuits of a
given generation always tend toward lower power supply voltages. This may
be simply because of the inexorable physics of (C * V squared * F) that
repays lower logic swings quadratically with lower fundamental power
requirement. (At least for those logics whose power is limited by the
fundamental C Vsquared F, which many, traditionally, were not).

Thus today important digital chips run off low (and even
device-to-device-optimized) single-digit supply Vs while a great many analog
functions need at least 5V. When digital logic ran off 5V, op amps ran
standardly from +- 15. When VT digital logic as described in some older
books used 150V supplies, VT analog modules (including the popular K2 and
USA series op amps from Philbrick) routinely took +- 300 to deliver reliable
+- 100V output swings that could be interpreted as numerical values in an
analog-computer simulation. The analog world tends to lag the digital
world, very broadly speaking of course, in moving to lower supply voltages.
 
J

John Larkin

Thus today important digital chips run off low (and even
device-to-device-optimized) single-digit supply Vs while a great many analog
functions need at least 5V. When digital logic ran off 5V, op amps ran
standardly from +- 15. When VT digital logic as described in some older
books used 150V supplies, VT analog modules (including the popular K2 and
USA series op amps from Philbrick) routinely took +- 300 to deliver reliable
+- 100V output swings that could be interpreted as numerical values in an
analog-computer simulation. The analog world tends to lag the digital
world, very broadly speaking of course, in moving to lower supply voltages.


It is annoying that nowadays the really fast stuff has pitiful voltage
swing capability. I'm looking for a wideband linear amp that will put
a 7-8 volt shaped pulse into a 50 ohm load with, say, 3 GHz bw, and
there's nothing around, so I'll have to build the damned thing.
Similarly, I'd love to find an opamp that would swing +-10 volts into
maybe 100 ohms at above a volt per ns, but that's not in the cards
either. Only TI seems to have a reasonably fast bipolar process that
will work from +-15 rails.

Even in digital, the standard 5 volt (or even 3.3 volt) rail is dead.
I'm having to do digital designs with a slew of different regulators
for different chips: 5 volts for bus interface and uP, 3.3 for FPGA
i/o and ram chips, 2.5 for one fpga core, 1.8 for another. Keeping
track of logic level compatibilities and supply sequencing is a
nightmare. I have one mixed-signal VME board with eight regulators.

Snarl.

John
 
S

Stefan Heinzmann

John Larkin schrieb:
[...]
Similarly, I'd love to find an opamp that would swing +-10 volts into
maybe 100 ohms at above a volt per ns, but that's not in the cards
either. Only TI seems to have a reasonably fast bipolar process that
will work from +-15 rails.

Have you seen the LM7171? Don't know whether you're happy with the BW,
though...
 
Top