24 Aug
2014
24 Aug
'14
4:07 p.m.
On Sun, Aug 24, 2014 at 2:06 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
This is quite a nroaml way to do it. A TTL output can
sink a larger
current tahn it can source,
By which Tony means "can sink a **MUCH** larger current than it can source".
If you look at the spec for the TI SN7404, the difference is a factor of 40:
max -0.4 mA source (high), vs. max 16 mA sink (low). Generally speaking,
0.4mA is nowhere near enough to be usable for an LED, so driving the low
side of the LED was the only practical approach with normal TTL.
Even the max sink spec of 16mA wasn't enough to drive a typical mid-1970s
red LED very brightly. For example, a Monsanto (later GI) MV5020 red LED
at 20 mA would only give you 2.0 mcd typical (and 0.6 mcd minimum!).
That very low source current spec (-0.4 mA) is about the same for all true
TTL families, excepting some buffer parts that are spec'd able to source more
current.
Modern CMOS parts tend to have closer or even identical source and sink
limits, so it has become somewhat more common to see high-side drive
for LEDs, though many designers have stuck with low-side drive out of
tradition.
Many modern LEDs are also much more efficient than those from the 1970s,
which is partially because different chemistries producing lower wavelengths
are commonly used for red, e.g., 620-340 nm vs. 660 nm, and the human
eye is significantly more sensitive to the lower wavelength. However, the
modern chemistries are also more efficient in terms of absolute light output
per current. A representative modern T1 3/4 LED with comparable spatial
distribution to the MV5020 is the Cree C503B-RCS-CW0Z0AA1, which has a
typical wavelength of 624 nm, and typical luminous intensity of 5100 mcd
(minimum 3000 mcd) at 20 mA.
When comparing LED specs, it is important to look at the spatial distribution.
A narrow viewing angle LED will give higher luminous intensity but only
on-axis.
Eric
24 Aug
24 Aug
5:03 p.m.
Thanks for that, Eric! Very illuminating info.. err, better say
informative. I had no idea of the order-of-magnitude difference between
source & sink currents in TTL stuff. Now I see why all those LEDs are
driven off the low side as you say. Makes perfect sense to me.
The difference in LED efficiency / intensity is truly stunning, though..
the new LEDs required 20X the resistance of the old 70s units - two of
which had "just failed". When was the last time +I+ ever saw an LED go bad?
That would be second to never, pretty much..
But yeah, 3K9 (call it 4K) vs. 220R (call it 200).. a nominal factor of
twenty, and just as bright if not more so off-axis! I just kept cranking
that sub-box knob going "wow, really?".
On Sun, Aug 24, 2014 at 6:07 PM, Eric Smith <spacewar at gmail.com> wrote:
On Sun, Aug 24, 2014 at 2:06 PM, Tony Duell <ard at
p850ug1.demon.co.uk>
wrote:
This is quite a nroaml way to do it. A TTL output
can sink a larger
current tahn it can source,
By which Tony means "can sink a **MUCH** larger current than it can
source".
If you look at the spec for the TI SN7404, the difference is a factor of
40:
max -0.4 mA source (high), vs. max 16 mA sink (low). Generally speaking,
0.4mA is nowhere near enough to be usable for an LED, so driving the low
side of the LED was the only practical approach with normal TTL.
Even the max sink spec of 16mA wasn't enough to drive a typical mid-1970s
red LED very brightly. For example, a Monsanto (later GI) MV5020 red LED
at 20 mA would only give you 2.0 mcd typical (and 0.6 mcd minimum!).
That very low source current spec (-0.4 mA) is about the same for all true
TTL families, excepting some buffer parts that are spec'd able to source
more
current.
Modern CMOS parts tend to have closer or even identical source and sink
limits, so it has become somewhat more common to see high-side drive
for LEDs, though many designers have stuck with low-side drive out of
tradition.
Many modern LEDs are also much more efficient than those from the 1970s,
which is partially because different chemistries producing lower
wavelengths
are commonly used for red, e.g., 620-340 nm vs. 660 nm, and the human
eye is significantly more sensitive to the lower wavelength. However, the
modern chemistries are also more efficient in terms of absolute light
output
per current. A representative modern T1 3/4 LED with comparable spatial
distribution to the MV5020 is the Cree C503B-RCS-CW0Z0AA1, which has a
typical wavelength of 624 nm, and typical luminous intensity of 5100 mcd
(minimum 3000 mcd) at 20 mA.
When comparing LED specs, it is important to look at the spatial
distribution.
A narrow viewing angle LED will give higher luminous intensity but only
on-axis.
Eric
25 Aug
25 Aug
12:36 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
The difference in LED efficiency / intensity is truly
stunning, though..
the new LEDs required 20X the resistance of the old 70s units - two of
Somethign that mazd me : Take a modern white LED and a 9V battery. Hold
the cathode lead of the LEd with one hand and touch the anode lead on the
+ve terminal of the battery. Not touch the -ve terminal of the battery
with your other hand. In effect you are using your body as the LED series
resistor. When I tried it I got a noticable glow from the LED. Not
bright, but I could certianyl tell if it was on or off.
which had "just failed". When was the last
time +I+ ever saw an LED go bad?
That would be second to never, pretty much..
Iv'e had LED segnemnts fail in old displays (e.g in HP calculators), but
to be fiar it's often the bonding wire that breaks not the LED die iteslf.
LED toruhes (flashlights) over here tend to just conenct white LEDs
acorss a 4.5V battery (at least the cheaper ones do). This is nto kind to
the LEDs, and I've had a fiar number fail. I must have replaced at least
as many LEDs in cheap LED torches as I ever replaced filament lamps in
traditional torches. Hmmm...
-tony
24 Aug
24 Aug
9:48 p.m.
On Sun, Aug 24, 2014 at 5:07 PM, Eric Smith <spacewar at gmail.com> wrote:
By which Tony means "can sink a **MUCH** larger
current than it can source".
If you look at the spec for the TI SN7404, the difference is a factor of 40:
max -0.4 mA source (high), vs. max 16 mA sink (low). Generally speaking,
0.4mA is nowhere near enough to be usable for an LED, so driving the low
side of the LED was the only practical approach with normal TTL.
This is also why many parts offered "open collector" outputs, where the high
side transistor of the totem pole was omitted, allowing wire-OR (negative logic)
or wire-AND (positive logic) with a pullup resistor, but there weren't TTL chips
that only had the high side drive, since that was too wimpy to be very useful.
Even with CMOS, where the equivalent is "open drain", there usually aren't
high-side only outputs, though with the symmetric drive of modern CMOS
that would work just as well.
That's also why strobes are almost always negative logic (active low).
Eric
25 Aug
25 Aug
8:22 a.m.
Interesting, but I'm not sure what a strobe (signal) is.
I'm surprised that no one on the list has turned up with an S-100 SRAM card
for me! Those things are actually all over eBay, must be 20-30 listed right
now, in the $59-300 range. The top-end of the prices are for a genuine MITS
4K DRAM board - the short-lived, ill-fated one that about lost them the
business.
On Sun, Aug 24, 2014 at 11:48 PM, Eric Smith <spacewar at gmail.com> wrote:
On Sun, Aug 24, 2014 at 5:07 PM, Eric Smith
<spacewar at gmail.com> wrote:
By which Tony means "can sink a **MUCH**
larger current than it can
source".
If you look at the spec for the TI SN7404, the
difference is a factor of
40:
max -0.4 mA source (high), vs. max 16 mA sink
(low). Generally speaking,
0.4mA is nowhere near enough to be usable for an LED, so driving the low
side of the LED was the only practical approach with normal TTL.
This is also why many parts offered "open collector" outputs, where the
high
side transistor of the totem pole was omitted, allowing wire-OR (negative
logic)
or wire-AND (positive logic) with a pullup resistor, but there weren't TTL
chips
that only had the high side drive, since that was too wimpy to be very
useful.
Even with CMOS, where the equivalent is "open drain", there usually aren't
high-side only outputs, though with the symmetric drive of modern CMOS
that would work just as well.
That's also why strobes are almost always negative logic (active low).
Eric
10:25 a.m.
On Sun, Aug 24, 2014 at 11:48 PM, Eric Smith <spacewar at gmail.com> wrote
of asymmetric TTL drive capability:
That's also why strobes are almost always negative
logic (active low).
On Mon, Aug 25, 2014 at 9:22 AM, drlegendre . <drlegendre at gmail.com> wrote:
Interesting, but I'm not sure what a strobe
(signal) is.
A signal that causes an action by being briefly in an active state, and the
rest of the time in an inactive state.
For example, the chip select signal to a memory or I/O chip, or the
read and write strobes.
11:36 a.m.
Huh, ok. So.. strobe is a meta-term for a line that pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
On Mon, Aug 25, 2014 at 12:25 PM, Eric Smith <spacewar at gmail.com> wrote:
On Sun, Aug 24, 2014 at 11:48 PM, Eric Smith
<spacewar at gmail.com> wrote
of asymmetric TTL drive capability:
That's also why strobes are almost always
negative logic (active low).
On Mon, Aug 25, 2014 at 9:22 AM, drlegendre . <drlegendre at gmail.com>
wrote:
Interesting, but I'm not sure what a strobe
(signal) is.
A signal that causes an action by being briefly in an active state, and the
rest of the time in an inactive state.
For example, the chip select signal to a memory or I/O chip, or the
read and write strobes.
11:44 a.m.
On Aug 25, 2014, at 11:36 , drlegendre . <drlegendre at gmail.com> wrote:
Huh, ok. So.. strobe is a meta-term for a line that
pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
Any signal can pulse. The term "strobe" usually carries the connotation that the
pulse latches or otherwise qualifies values present on other signals. I guess that you
might say that "pulse" is what is does, and "strobe" is why it does
it.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
12:20 p.m.
Perhaps:
Clock - line changes state at a constant rate drives other logic
Strobe - line changes state when an event occurs, drives other logic
In chip terms they are both 'clock' lines they tell the rest of the
mechanism to cycle, but from a use standpoint one is constant and one
is intermittent.
--Chuck
On Mon, Aug 25, 2014 at 11:44 AM, Mark J. Blair <nf6x at nf6x.net> wrote:
On Aug 25, 2014, at 11:36 , drlegendre . <drlegendre at gmail.com> wrote:
Huh, ok. So.. strobe is a meta-term for a line
that pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
Any signal can pulse. The term "strobe" usually carries the connotation that
the pulse latches or otherwise qualifies values present on other signals. I guess that you
might say that "pulse" is what is does, and "strobe" is why it does
it.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
11:48 a.m.
Not precisely. Strobe specifically means a signal that is active (active level or active
edge, it depends) to indicate the instant at which other signals are meaningful and should
be latched or acted on.
For example, in the Unibus you can refer to the MSYN signal as a strobe: when the selected
Unibus slave, in an output cycle, sees MSYN assert (rising edge), the data lines are valid
and it will latch them or otherwise act on their value.
?Regular? isn?t the key point. The key point is that it?s a pulse or edge that triggers
and action.
paul
On Aug 25, 2014, at 2:36 PM, drlegendre . <drlegendre at gmail.com> wrote:
Huh, ok. So.. strobe is a meta-term for a line that
pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
On Mon, Aug 25, 2014 at 12:25 PM, Eric Smith <spacewar at gmail.com> wrote:
> On Sun, Aug 24, 2014 at 11:48 PM, Eric Smith <spacewar at gmail.com> wrote
> of asymmetric TTL drive capability:
>> That's also why strobes are almost always negative logic (active low).
>
> On Mon, Aug 25, 2014 at 9:22 AM, drlegendre . <drlegendre at gmail.com>
> wrote:
>> Interesting, but I'm not sure what a strobe (signal) is.
>
> A signal that causes an action by being briefly in an active state, and the
> rest of the time in an inactive state.
>
> For example, the chip select signal to a memory or I/O chip, or the
> read and write strobes.
>
12:19 p.m.
On 2014-Aug-25, at 11:36 AM, drlegendre . wrote:
Huh, ok. So.. strobe is a meta-term for a line that
pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
Strobe: As derived from stroboscope. To briefly make information visible.
In optics or photography a strobe (flash) briefly illuminates an otherwise dark scene.
In digital electronics a strobe briefly gates signals onto otherwise idle lines.
12:36 p.m.
Also remember the dead simple strobe/ack protocol on the "centronics" parallel
printer port:
place 8 bits of data on output port
Wait some period of time for signals to settle (usually a simple counter)
Assert STROBE
Wait for ACK
Deassert STROBE
Wait for !ACK
and repeat until done ;-)
10 bit interface, guaranteed flow control.
On Aug 25, 2014, at 1:19 PM, Brent Hilpert <hilpert
at cs.ubc.ca> wrote:
On 2014-Aug-25, at 11:36 AM, drlegendre . wrote:
Huh, ok. So.. strobe is a meta-term for a line that pulses with some
regularity? Seems like kind of a redundant term, since we have 'pulse' as
well.
Strobe: As derived from stroboscope. To briefly make information visibl
In optics or photography a strobe (flash) briefly illuminates an otherwise dark scene.
In digital electronics a strobe briefly gates signals onto otherwise idle lines.
1:22 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
On Sun, Aug 24, 2014 at 11:48 PM, Eric Smith <spacewar at gmail.com> wrote
of asymmetric TTL drive capability:
I would say it's rather moee than that, it ahs to start something happening.
Consider the Centronics interface. Pin 1 is called stb/, clearly standing
for 'strobe' (the '/' means actuve low). A low pusle on that causes the
priter to read the chracter on the data lines and do soemthing with it
(store it in the printer's buffermemeory, print it directly, depending on
the type of pritner).
Theacknowledge signal, Ack/ on the same interface is also only briefly
active (it pulses low after the printer has processed the current character
and is ready for the next one), but I would not call that a 'strobe;
-tony
That's also why strobes are almost always
negative logic (active low).
On Mon, Aug 25, 2014 at 9:22 AM, drlegendre . <drlegendre at gmail.com> wrote:
Interesting, but I'm not sure what a strobe
(signal) is.
A signal that causes an action by being briefly in an active state, and the
rest of the time in an inactive state.
12:07 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions, if
On Sun, Aug 24, 2014 at 2:06 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
Although to be fari, if yo get it to source more current (sensibly more
current, like the 10mA to light an LED), the IC will not be damaged but
the output voltage may well drop below the specification for a logic 1
output.
Plenty of (commercial) units drove LEDs form a TTL output to ground, even
though it is not good practice. (and here 'TTL' excludes the CMOS
families)
-tony
This is quite a nroaml way to do it. A TTL output
can sink a larger
current tahn it can source,
By which Tony means "can sink a **MUCH** larger current than it can source".
If you look at the spec for the TI SN7404, the difference is a factor of 40:
max -0.4 mA source (high), vs. max 16 mA sink (low). Generally speaking,
2:43 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions, if
I wrote:
By which Tony means "can sink a **MUCH** larger
current than it can source".
If you look at the spec for the TI SN7404, the difference is a factor of 40:
max -0.4 mA source (high), vs. max 16 mA sink (low). Generally speaking,
On Mon, Aug 25, 2014 at 1:07 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
Although to be fari, if yo get it to source more
current (sensibly more
current, like the 10mA to light an LED), the IC will not be damaged but
the output voltage may well drop below the specification for a logic 1
output.
Plenty of (commercial) units drove LEDs form a TTL output to ground, even
though it is not good practice. (and here 'TTL' excludes the CMOS
families)
It's beyond "not good practice" and into the realm of "can't be
expected to
work consistently". While it's true that sourcing more than the rated
current from a 74xx output will pull its voltage below the Voh
spec, there's no telling how much lower it will pull it. You could
characterize the voltage of a typical 7404 output sourcing 10 mA
under various conditions (e.g., supply voltage, temperature), and find
that driving the LED works will all the parts you buy today, but parts you
buy tomorrow or from a different vendor might fully meet specifications yet
not be able to drive the LED.
I would consider any product that high-side drives an LED from a TTL
output beyond the rated source current to be defective in design.
Some moron at TI has revised the SN5404/SN54LS04/SN7404/SN74LS04
datasheet to claim that all unused inputs must be held at Vcc or GND, which is
good practice, but to further cite their app note "Implications of
Slow or Floating
CMOS Inputs", which is completely irrelevant to a bipolar TTL device. Sigh.
26 Aug
26 Aug
1:46 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
Plenty of
(commercial) units drove LEDs form a TTL output to ground, even
though it is not good practice. (and here 'TTL' excludes the CMOS
families)
It's beyond "not good practice" and into the realm of "can't be
expected to
work consistently". While it's true that sourcing more than the rated
Some moron at TI has revised the SN5404/SN54LS04/SN7404/SN74LS04
datasheet to claim that all unused inputs must be held at Vcc or GND, which is
good practice, but to further cite their app note "Implications of
Is it? I thought that for plain TTL and (I think) LS it was good practice
to pull unused inputs up via a resistor and not link them straight to
Vcc. Something about supply line spikes causing breakdown of the input
transistor.
Slow or Floating
CMOS Inputs", which is completely irrelevant to a bipolar TTL device. Sigh.
-tony
3:27 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
I wrote:
Some moron at TI has revised the
SN5404/SN54LS04/SN7404/SN74LS04
datasheet to claim that all unused inputs must be held at Vcc or GND
On Tue, Aug 26, 2014 at 2:46 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
Is it? I thought that for plain TTL and (I think) LS
it was good practice
to pull unused inputs up via a resistor and not link them straight to
Vcc.
For "plain" TTL yes. One of the purported advantages of LS was not
needing to do that. I forgot about that detail when commenting on the
TI datasheet lossage.
Something about supply line spikes causing breakdown
of the input
transistor.
I've heard that explanation and I suppose it sounds plausible but I would like
to find an authoritative statement from TI (i.e., in a TI data sheet
or application
note).
Fairchild AN-363 "Designing with TTL" dated June 1984 mentions
standard TTL, LS, S, ALS, AS, L, and F (but omits H), and states
without reference to a specific family:
A good design rule is to tie unused inputs to a solid logic level.
Inputs are usually tied to V CC through a 1 k? to 5 k? resistor,
since tying them to ground means supplying the I IL current
instead of the I IH current. I IL is several orders of magnitude
greater than I IH . The resistor is recommended to protect the
input against V CC voltage surges and to protect the system
against the possibility of the input shorting directly to ground.
A single 1k resistor can handle up to 10 inputs.
The other thing commonly done was to tie unused "don't care" inputs
to some other logic output, provided that the fanout wasn't
exceeded. Even when the input wasn't a "don't care", sometimes
a spare gate was used, e.g., one inverter from a 7404 with the input
tied to ground, to drive multiple other TTL inputs that needed to
be high.
5:09 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
On 2014-Aug-26, at 3:27 PM, Eric Smith wrote:
I wrote:
(Sent this an hour earlier in response to the earlier response, but it hasn't shown up
yet, so replying here instead)
Did you explicitly want a reference from TI, or something explanatory from an
'authority' ?
The characteristics and design treatises at the beginning of Fairchild TTL databooks
address it in a little more detail.
(from the 1978 version):
"For a permanent high signal, unused inputs can be tied to Vcc.
A current limiting resistor, in the range of 1K to 5K, is recommended for emitter-type
inputs since these break down
at some unspecified voltage over 5.5V and power supply misadjustment or malfunction can
cause damage unless
current is limited.
.. diode-type LS-TTL inputs have breakdown voltages above 15V and thus protective
resistors are not normally required."
Sidetracking, it's interesting to note that most LS-TTL isn't 'really'
TTL, it's DTL with Schottky input diodes.
Some moron at TI has revised the
SN5404/SN54LS04/SN7404/SN74LS04
datasheet to claim that all unused inputs must be held at Vcc or GND
On Tue, Aug 26, 2014 at 2:46 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
Is it? I thought that for plain TTL and (I think)
LS it was good practice
to pull unused inputs up via a resistor and not link them straight to
Vcc.
For "plain" TTL yes. One of the purported advantages of LS was not
needing to do that. I forgot about that detail when commenting on the
TI datasheet lossage.
Something about supply line spikes causing
breakdown of the input
transistor.
I've heard that explanation and I suppose it sounds plausible but I would like
to find an authoritative statement from TI (i.e., in a TI data sheet
or application
note).
Fairchild AN-363 "Designing with TTL" dated June 1984 mentions
standard TTL, LS, S, ALS, AS, L, and F (but omits H), and states
without reference to a specific family:
A good design rule is to tie unused inputs to a solid logic level.
Inputs are usually tied to V CC through a 1 k? to 5 k? resistor,
since tying them to ground means supplying the I IL current
instead of the I IH current. I IL is several orders of magnitude
greater than I IH . The resistor is recommended to protect the
input against V CC voltage surges and to protect the system
against the possibility of the input shorting directly to ground.
A single 1k resistor can handle up to 10 inputs. 
5:28 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
On 26/08/2014 23:27, Eric Smith wrote:
The TTL Data Book for Design Engineers, 1982 ed, page 5.2 says
Some possible ways of connecting unused inputs are:
a. Connect unused inputs to an independent voltage supply. Preferably
this voltage should be between VOHmin and 4.5V. Series 54LS/74LS
devices with diode inputs may be connected directly to Vcc.
b. Connect unused inputs to a used input if maximum drive capability of
the driving output will not be exceeded. Each additional input
presents a full load to the driving output at a high-level
voltage but adds no loading at a low-level voltage.
c. Connect unused inputs to Vcc through a 1k resistor so that if a
transient that exceeds the input maximum rating should occur, the
impedance will be high enough to protect the input. One to 25 unused
inputs may be connected to each 1k resistor. Series 54LS/74LS
devices with diode inputs may be connected directly to Vcc.
d. Connect unused inputs to any fixed-high-level compatible output such
as the output of an inverter or NAND gate that has its input(s)
grounded. Maximum high-level drive capability of the output should
not be exceeded.
There it is in part (c).
There's some more esoteric stuff about circuit parameters including
worst-case testing in Chapter 3 of TI's "Designing with TTL Integrated
Circuits", 1971, but it doesn't actually say to use 1k pullups.
Calculation of values of various things for various circumstances is
left as an exercise or two for the reader :-)
--
Pete Peter Turnbull
Network Manager
University of York
On Tue, Aug 26, 2014 at 2:46 PM, Tony Duell <ard at
p850ug1.demon.co.uk> wrote:
Is it? I thought that for plain TTL and (I think)
LS it was good practice
to pull unused inputs up via a resistor and not link them straight to
Vcc.
For "plain" TTL yes. Something
about supply line spikes causing breakdown of the input
transistor.
I've heard that explanation and I suppose it sounds plausible but I would like
to find an authoritative statement from TI (i.e., in a TI data sheet
or application note).
5:36 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
On 2014-Aug-26, at 3:27 PM, Eric Smith wrote:
On Tue, Aug 26, 2014 at 2:46 PM, Tony Duell <ard at
p850ug1.demon.co.uk> wrote:
Here's a quote from TI, mentioning the same 5.5V extreme as Fairchild, but not
providing the explanation:
Something about supply line spikes causing
breakdown of the input
transistor.
I've heard that explanation and I suppose it sounds plausible but I would like
to find an authoritative statement from TI (i.e., in a TI data sheet
or application
note). From "TTL Integrated Circuits Catalog from Texas
Instruments" / 1 Aug 1969:
"Unused inputs of AND and NAND gates and unused preset and clear inputs of
flip-flops:
Tie directly to +Vcc where Vcc is guaranteed to *always* be <= 5.5V; (emphasis
TI's)
or
Tie to Vcc through resistor >= 1Kohms. Several unused inputs may be tied to one
resistor;
or
Tie to unused input of same gate if maximum fan-out of driving device will not be
exceeded;
or
Tie to unused gate output where unused gate input is grounded.
"
4:08 p.m.
New subject: TTL source/sink currents and LEDs (was Re: Some Altair questions,
On 2014-Aug-26, at 1:46 PM, Tony Duell wrote:
This is addressed in, for example, the characteristics and design treatises at the
beginning of Fairchild TTL databooks.
(from the 1978 version):
"For a permanent high signal, unused inputs can be tied to Vcc.
A current limiting resistor, in the range of 1K to 5K, is recommended for emitter-type
inputs since these break down
at some unspecified voltage over 5.5V and power supply misadjustment or malfunction can
cause damage unless
current is limited.
.. diode-type LS-TTL inputs have breakdown voltages above 15V and thus protective
resistors are not normally required."
Sidetracking, it's interesting to note that most LS-TTL isn't 'really'
TTL, it's DTL with Schottky input diodes.
Some moron at TI has revised the SN5404/SN54LS04/SN7404/SN74LS04
datasheet to claim that all unused inputs must be held at Vcc or GND, which is
good practice, but to further cite their app note "Implications of
Is it? I thought that for plain TTL and (I think) LS it was good practice
to pull unused inputs up via a resistor and not link them straight to
Vcc. Something about supply line spikes causing breakdown of the input
transistor.
3769
days inactive
3772
days old
20 comments
9 participants
participants (9)
-
ard＠p850ug1.demon.co.uk -
chuck.mcmanis＠gmail.com -
drlegendre＠gmail.com -
hilpert＠cs.ubc.ca -
nf6x＠nf6x.net -
paulkoning＠comcast.net -
pete＠dunnington.plus.com -
roeapeterson＠gmail.com -
spacewar＠gmail.com