Universal logic modules vs NAND-like modules

[Candida Ferreira]

Hi everyone,

Is there a proper name to distinguish NAND-like universal modules from
ordinary ones, that is, the ones formed by a function F, plus NOT, plus ONE,
and ZERO?

Thanks,
Candida
---
Candida Ferreira, Ph.D.
Chief Scientist, Gepsoft
http://www.gene-expression-programming.com/author.asp

GEP: Mathematical Modeling by an Artificial Intelligence
http://www.gene-expression-programming.com/gep/Books/index.asp
Modeling Software
http://www.gepsoft.com/gepsoft/
Get APS 3.0 Std free with the book!

 

[martin griffith]

Hi everyone,

Is there a proper name to distinguish NAND-like universal modules from
ordinary ones, that is, the ones formed by a function F, plus NOT, plus ONE,
and ZERO?

Thanks,
Candida
---
Candida Ferreira, Ph.D.
Chief Scientist, Gepsoft
http://www.gene-expression-programming.com/author.asp

GEP: Mathematical Modeling by an Artificial Intelligence
http://www.gene-expression-programming.com/gep/Books/index.asp
Modeling Software
http://www.gepsoft.com/gepsoft/
Get APS 3.0 Std free with the book!

Wow, formal description.....
I'd call it an IF gate, I always have problems with formal names, I
still imagine XOR as a Diff gate, since it shows if there is a
difference.

sorry, but that wont be any help



martin

After the first death, there is no other.
(Dylan Thomas)

 

[Candida Ferreira]

Wow, formal description.....
I'd call it an IF gate, I always have problems with formal names, I
still imagine XOR as a Diff gate, since it shows if there is a
difference.

I'm not sure which IF gate you are referring to, but the one defined by IF
A=1, THEN B, ELSE C, which is also by definition a ULM, can not by itself
describe a NAND gate as it is unable to create a NOT gate. But there are
other functions that behave exactly like NAND or NOR gates in the sense
that, by themselves, they can also describe any other function. I think
there is something special about them and I would like to distinguish them
from the ordinary ULMs.

Candida
---
Candida Ferreira, Ph.D.
Chief Scientist, Gepsoft
http://www.gene-expression-programming.com/author.asp

GEP: Mathematical Modeling by an Artificial Intelligence
http://www.gene-expression-programming.com/gep/Books/index.asp
Modeling Software
http://www.gepsoft.com/gepsoft/
Get APS 3.0 Std free with the book!

 

[PeteS]

The type of device you describe is common in FPGA architectures
(indeed, it looks almost exactly like the a Logic Block, as found in
Altera and some Xilinx devices).

Such a logic block (with a function F, usually from a lookup table)
plus the ONE, ZERO and NOT, is a subset of what is used on those
devices. (They also have some straight AND/NAND gating as well).

For want of a better term, I would call it a universal logic *function
block*, as opposed to a universal logic *gate* (which a NAND is, by
definition, as well as NOR).

Cheers

PeteS

 

[Candida Ferreira]

The type of device you describe is common in FPGA architectures
(indeed, it looks almost exactly like the a Logic Block, as found in
Altera and some Xilinx devices).

Such a logic block (with a function F, usually from a lookup table)
plus the ONE, ZERO and NOT, is a subset of what is used on those
devices. (They also have some straight AND/NAND gating as well).

For want of a better term, I would call it a universal logic *function
block*, as opposed to a universal logic *gate* (which a NAND is, by
definition, as well as NOR).

Cheers

PeteS

This is a good suggestion. Thank you for the input.

Candida
---
Candida Ferreira, Ph.D.
Chief Scientist, Gepsoft
http://www.gene-expression-programming.com/author.asp

GEP: Mathematical Modeling by an Artificial Intelligence
http://www.gene-expression-programming.com/gep/Books/index.asp
Modeling Software
http://www.gepsoft.com/gepsoft/
Get APS 3.0 Std free with the book!

 

[keith]

I'm not sure which IF gate you are referring to, but the one defined by IF
A=1, THEN B, ELSE C, which is also by definition a ULM, can not by itself
describe a NAND gate as it is unable to create a NOT gate. But there are
other functions that behave exactly like NAND or NOR gates in the sense
that, by themselves, they can also describe any other function. I think
there is something special about them and I would like to distinguish them
from the ordinary ULMs.

I really can't figure out what you're looking for, but fixed functions are
often referred to as "unit logic". Programmable logic (AND/OR/ZERO/NOT)
is a ROM, or perhaps even a multiplexor.

 

[Ratch]

The type of device you describe is common in FPGA architectures
(indeed, it looks almost exactly like the a Logic Block, as found in
Altera and some Xilinx devices).

Such a logic block (with a function F, usually from a lookup table)
plus the ONE, ZERO and NOT, is a subset of what is used on those
devices. (They also have some straight AND/NAND gating as well).

For want of a better term, I would call it a universal logic *function
block*, as opposed to a universal logic *gate* (which a NAND is, by
definition, as well as NOR).

The NAND and NOR gates are universal logic building blocks, but not by
definition. One can easily construct AND, OR, and NOT gates from NANDs and
NORs. From AND, ORs,and NOTs, one can build anything logical. That is what
makes them universal. Their inbuilt NOT property cannot be duplicated from
only ANDs and ORs. Ratch

 

[PeteS]

Semantics perhaps

My point is that given sufficient NAND gates, one may construct any
logic function (ditto for NOR), leading to my comment that a NAND is,
by definition (mathematical, that is) a universal gate. I can see the
issue with calling it a gate (it's really two), but in common parlance,
at least, we call it a gate. Mathematically, I would agree it's a
function.

Cheers

PeteS

 

[Greg Neff]

Semantics perhaps

My point is that given sufficient NAND gates, one may construct any
logic function (ditto for NOR), leading to my comment that a NAND is,
by definition (mathematical, that is) a universal gate. I can see the
issue with calling it a gate (it's really two), but in common parlance,
at least, we call it a gate. Mathematically, I would agree it's a
function.

Cheers

PeteS

This whole discussion is semantics. As you said, NAND/NOR/XNOR
functions are really AND/OR/XOR followed by NOT. Also, there are
AND/OR/XOR constructs with NOT(s) at the inputs, but there is no
common label for these. For example, Xilinx calls a 2-input AND gate
with one NOT input an AND2B1 (B is for bubble, I presume).

Fairchild (and others) make "universal configurable 2-input logic
gates" that allow any 2-input function to be implemented. In this
datasheet the same example I used above is called "2-input AND with
inverted input". See:

http://www.fairchildsemi.com/ds/NC/NC7SP57.pdf

We don't have convenient labels to describe all possible 2-input
functions, let alone n-input functions. It is probably best to use
schematic or equation representation to be clear. Failing that, maybe
use ARB to designate an arbitrary function, but the function still
needs to be described somewhere. When I use one of the Fairchild
universal gates in a design it is drawn as a box, and I put the
function (equation) on the schematic so that the logic can be
followed.

================================

Greg Neff
VP Engineering
*Microsym* Computers Inc.
[email protected]

 

[Keith Williams]

This whole discussion is semantics. As you said, NAND/NOR/XNOR
functions are really AND/OR/XOR followed by NOT.

AND/OR are really NAND/NOR followed by a NOT. ;-)

Also, there are AND/OR/XOR constructs with NOT(s) at the inputs,
but there is no common label for these.

Sure there is. They're called NOR/NAND/XOR. ;-)

For example, Xilinx calls a 2-input AND gate
with one NOT input an AND2B1 (B is for bubble, I presume).

I call it O <= NOT a AND b and let the software figure out what to put
in the lookup table.

Fairchild (and others) make "universal configurable 2-input logic
gates" that allow any 2-input function to be implemented. In this
datasheet the same example I used above is called "2-input AND with
inverted input". See:

http://www.fairchildsemi.com/ds/NC/NC7SP57.pdf

Wow. That seems like a waste. I guess if one only needs one gate and
has no power to spare it is useful, but I've never been in that
situation.

 

[Jim Thompson]

[snip]

This whole discussion is semantics. As you said, NAND/NOR/XNOR
functions are really AND/OR/XOR followed by NOT.

I think you have that backwards. In the real (physical) world, NAND's
and NOR's are native (minimum device count). It takes an extra
inverter to get to AND's and OR's.

Also, there are
AND/OR/XOR constructs with NOT(s) at the inputs, but there is no
common label for these. For example, Xilinx calls a 2-input AND gate
with one NOT input an AND2B1 (B is for bubble, I presume).

[snip]

Again, notational, not physical.

...Jim Thompson

 

[Greg Neff]

(snip)

Wow. That seems like a waste. I guess if one only needs one gate and
has no power to spare it is useful, but I've never been in that
situation.

These gates have the very useful property of handling overvoltage
inputs. These (and similar) devices are handy when you need both
voltage translation from a higher voltage domain to a lower voltage
domain, and you need some simple logic.

================================

Greg Neff
VP Engineering
*Microsym* Computers Inc.
[email protected]

 

[Keith Williams]

(snip)

These gates have the very useful property of handling overvoltage
inputs. These (and similar) devices are handy when you need both
voltage translation from a higher voltage domain to a lower voltage
domain, and you need some simple logic.

They don't look like they're very robust to me. The Recommended input
voltage is 0 to 3.6V with an absolute maximum of -.5V to 4.6V. I've
found quick-switches to be useful for translating from 5V to 3.3V
logic. I guess some find these things useful though (they seem to sell
them). <shrug>

 

[Winfield Hill]

Keith wrote...

Greg says...

They don't look like they're very robust to me. The Recommended
input voltage is 0 to 3.6V with an absolute maximum of -.5V to 4.6V.
I've found quick-switches to be useful for translating from 5V to
3.3V logic. I guess some find these things useful though (they
seem to sell them). <shrug>

The nc7sp57 and nc7sp58 gates are spec'd down to 0.9V, so imagine
you're in 2005 and working in a 1.2V logic environment (14ns prop
delay). A 3.6V input-overvoltage tolerance looks rather high...
more than high enough to deal with your 3.3V high-voltage supply!

If you want a similar capability (but only spec'd down to 1.65V),
with a 5.5V max supply rating, use the nc7sz57 and nc7sz58 parts.