Name a negating chip...

[Ignoramus24489]

I would like to find a chip that takes 1/0 input (like 16V being 1 and
ground being 0), and generates two synchronous outputs f(input) and
g(input), where f(input) = input and g(input) = NOT input.

What is important is that there is the same time delay between these
outputs, that the time when they are equal should be as short as possible.

Sorry for a long winded explanation, I bet that the answer is very
simple. I would prefer something that can live in a 16V supply rail.

thanks

i

 

[Iwo Mergler]

I would like to find a chip that takes 1/0 input (like 16V being 1 and
ground being 0), and generates two synchronous outputs f(input) and
g(input), where f(input) = input and g(input) = NOT input.

What is important is that there is the same time delay between these
outputs, that the time when they are equal should be as short as possible.

Sorry for a long winded explanation, I bet that the answer is very
simple. I would prefer something that can live in a 16V supply rail.

thanks

i

Look for a differential driver.For instance
http://www.analog.com/en/prod/0,,759_786_AD815,00.html

Kind regards,

Iwo

 

[Tim Shoppa]

What is important is that there is the same

time delay between these outputs, that the
time when they are equal should be as short as possible.

Differential transmitters and receivers are designed to have very equal
propogation times. Typical specs for a transmitter from a few decades
ago is that the skew between inverted and not-inverted is 2
nanoseconds. Typical specs for a LVDS transmitter have a skew in the
low hundreds of picoseconds.

I would prefer something that can live in a 16V supply rail.

Most data transmission differential transmitters work on 5V or 3.3V
rails, not 16V...

If what you want is a power driver, then maybe you want an H-bridge or
H-bridge driver (typically hundreds of ns to a few microseconds
switching time). Usually these go from "f on, g off" to "f off, g off"
to "f off, g on" with a known (controlled) dead time.

If you can quantify your "short as possible" then you can probably get
a much more concrete answer. Otherwise we'll start proposing
thyratrons etc!

Tim.

 

[[email protected]]

There is very little logic made that can run from a 16V supply rail.
Unless your tolerances on the difference in time delay are around the
microsecond level, or slower, none of them would be much help.

ECL logic line receivers, on the ther hand, are designed to do exactly
that job.

Take a look at the ON Semiconductor (used to be Motorola) MC100EL16D
which you can buy off the shelf from Motorola/Newark

http://www.onsemi.com/pub/Collateral/MC100EL16-D.PDF

The maximum propagation delay from input to output s 425psec - which is
thus the absolute upper limit on the output skew, which isn't specified
for this device.

IIRR they have similar parts, explicitly intended for clock
distribution which do have more detailed specification.

The MC100EL16D is intended to run between the 0V rail and a -4.5V.
Motorola makes a fuss about using it as "PECL" which is to say, between
a +5V rail and 0V, where it runs just as well (and slightly faster) but
you should bear in mind that the ECL output is simply an
emitter-follower. If you short the output to 0V when the chip is runing
between 0V and -4.5V, you don't do any damage. If you short it to
ground when the chip is running between 5V and 0V, you blow up the
output transistor.

Hewlett-Packard used PECL-connected ECL in their laser interferomenter
system, and around 1985 I was told that 90% of the in-service repairs
to that bit of electronics involved replacing exploded ECL chips ...

 

[Ignoramus24489]

Differential transmitters and receivers are designed to have very equal
propogation times. Typical specs for a transmitter from a few decades
ago is that the skew between inverted and not-inverted is 2
nanoseconds. Typical specs for a LVDS transmitter have a skew in the
low hundreds of picoseconds.


Most data transmission differential transmitters work on 5V or 3.3V
rails, not 16V...

If what you want is a power driver, then maybe you want an H-bridge or
H-bridge driver (typically hundreds of ns to a few microseconds
switching time). Usually these go from "f on, g off" to "f off, g off"
to "f off, g on" with a known (controlled) dead time.

If you can quantify your "short as possible" then you can probably get
a much more concrete answer. Otherwise we'll start proposing
thyratrons etc!

Thanks Tim. I think that what I need is called a "comparator". I will
do some reading now.

i

 

[[email protected]]

Look for a differential driver.For instance
http://www.analog.com/en/prod/0,,759_786_AD815,00.html

This happens to be two single-ended drivers. Not really what was asked
for.

One might think about using it for boosting the differential outputs of
a line receiver to 16V levels. Something that got closer to
rail-to-rail might be more useful.

 

[[email protected]]

Thanks Tim. I think that what I need is called a "comparator". I will
do some reading now.

There are comparators with differential outputs, and they would do what
you ask for (apart from coping with 16V). Unfortunately, all the
comparators with differential outputs that I can think of are
ECL-compatible parts.

You'd get a better class of advice if you told us more about what you
are trying to do - or at least what sort of time delays you have in
mind when you say "as short as possible".

 

[Frank Bemelman]

There is very little logic made that can run from a 16V supply rail.
Unless your tolerances on the difference in time delay are around the
microsecond level, or slower, none of them would be much help.

[snip]

I was thinking of a 4070 exor gate. Take an input from
each exor gate and ties these together. The other input
of the two gates go to ground and +V respectively.

 

[[email protected]]

There is very little logic made that can run from a 16V supply rail.
Unless your tolerances on the difference in time delay are around the
microsecond level, or slower, none of them would be much help.

[snip]

I was thinking of a 4070 exor gate. Take an input from
each exor gate and ties these together. The other input
of the two gates go to ground and +V respectively.

You are right Frank, one microsecond was a gross exaggeration.

The propagation delays for the HEF4070B at 15V range from 25 to 55nsec,
which is a lot less than a microsecond

http://www.semiconductors.philips.com/acrobat_download/datasheets/HEF4070B_CNV_3.pdf

and the part can be used a with a 16V supply - the absolute maximum
supply voltage for the HEF series is 18V

http://www.semiconductors.philips.com/acrobat_download/various/HEF_FAMILY_SPECIFICATIONS.pdf

The original RCA CD4070 was almost as fast, though it couldn't take
18V.

Fairchild, and ON Semiconductor sell equivalent parts (CD4070B,
MC14070B.

Still not something you could could recommend to somebody who wanted
the smallest possible skew between the differential outputs.

 

[Pooh Bear]

I would like to find a chip that takes 1/0 input (like 16V being 1 and
ground being 0), and generates two synchronous outputs f(input) and
g(input), where f(input) = input and g(input) = NOT input.

What is important is that there is the same time delay between these
outputs, that the time when they are equal should be as short as possible.

Sorry for a long winded explanation, I bet that the answer is very
simple. I would prefer something that can live in a 16V supply rail.

Wrong approach.

Stick with a dedicated half / full bridge driver.

Graham

 

[Ignoramus24489]

Wrong approach.

Stick with a dedicated half / full bridge driver.

Graham, you made an assumption that I want to replace a gate driver
with something not made for that purpose. That is an incorrect
assumption. I will use a gate driver. However, the gate driver that I
chose, a bad ass IR22141SS chip, needs two inputs, HI and LO, to turn
high and low sides respectively. I want to generate these inputs
(which should be the logical opposite of one another) based on a
single input. That was the reason for my question.

So, I would like to repeat my question. I have a 1/0 signal, call it
x. I need a chip that would make two signals out of it, x and NOT x.

i

 

[John Woodgate]

I read in sci.electronics.design that Ignoramus24489

So, I would like to repeat my question. I have a 1/0 signal, call it
x. I need a chip that would make two signals out of it, x and NOT x.

Isn't that a D flip-flop with CLOCK tied to DATA? HEF4013B or something?

 

[Jim Thompson]

I read in sci.electronics.design that Ignoramus24489


Isn't that a D flip-flop with CLOCK tied to DATA? HEF4013B or something?

That changes on the clock EDGE. 74HC04 INVERTS.

...Jim Thompson

 

[Iwo Mergler]

This happens to be two single-ended drivers. Not really what was asked
for.

You're right, I have only seen the marketing blurb (differential or dual
single ended), never bothered to check the datasheet.:^(

Iwo

 

[Ken Smith]

you ask for (apart from coping with 16V). Unfortunately, all the
comparators with differential outputs that I can think of are
ECL-compatible parts.

LT1016 has TTL outputs

 

[Ken Smith]

assumption. I will use a gate driver. However, the gate driver that I
chose, a bad ass IR22141SS chip, needs two inputs, HI and LO, to turn
high and low sides respectively. I want to generate these inputs
(which should be the logical opposite of one another) based on a
single input. That was the reason for my question.

So, can we assume that you care a lot that the outputs never be true at
the same time and less about the false state?

If so solving this gets a lot easier. You can do it like a
"non-overlapping clock" circuit.

 

[Ignoramus24693]

So, can we assume that you care a lot that the outputs never be true at
the same time and less about the false state?

That's correct. If both inputs to the gate driver chip are true, then
the gate driver would shutdown with an error condition, and would need
to be reset. That's a safety feature. So, while both signals equal to
true will not destroy any equipment, they would result in an annoying
interruption.

I have selected two chips, XR2206 as a timer with pulse width control
(negative vs. positive cycle), and LM339 as a dual comparator. Then
can all happily live on the 15V supply bus.

In fact, I have enough stuff to get them to work and to connect them.
I may be able to do something late at night and test the pulse
generator circuit. I will use fixed frequency for now, I think.

I demolished a piece of military surplus equipment this morning and
salvaged a nice 6A regulated power supply for 15V DC. I am very happy.

i

If so solving this gets a lot easier. You can do it like a
"non-overlapping clock" circuit.

--

 

[Ken Smith]

That's correct. If both inputs to the gate driver chip are true, then
the gate driver would shutdown with an error condition, and would need
to be reset. That's a safety feature. So, while both signals equal to
true will not destroy any equipment, they would result in an annoying
interruption.

So the non-overlapping circuit would be the right way to go then

I have selected two chips, XR2206 as a timer with pulse width control
(negative vs. positive cycle), and LM339 as a dual comparator. Then
can all happily live on the 15V supply bus.

Why not use an LM339 and make the timing signals with the other two
sections?

 

[Ignoramus24693]

So the non-overlapping circuit would be the right way to go then

Yes, and I suppose that LM339 would take care of it.

Why not use an LM339 and make the timing signals with the other two
sections?

No good reason, I did no think about it, for a $3 XR2206 anyway, plus
it is easy to make it do what I want (pulse width control, frequency
etc).

i