Maker Pro
Maker Pro

50 Ohm, 5v driver needed

N

Norm Dresner

Jan 1, 1970
0
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
The hard part is that the pulse at the receiving end has to go from 0 to 5V
(more or less). Standard 50 ohm TTL drivers (74S140) are only spec'd to
produce a TTL-high of 2.5 volts and that's what I've observed; paralleling
74128 (75 ohm drivers) doesn't do much better. The basic specs are 0-5v
into 50 ohms with rise/fall times on the order of (but not much greater
than) 20 ns. The board we're using has a 5v supply output so we can use
that for any TTL(or equiv)-logic and a single voltage external supply to
provide the higher voltage needed to get rail-to rail 0-5v output. We have
on hand some 6v/1A regulated wall-warts which should be adequate.

My prototype used a 7407/7417 (HV OC inverter) to drive a 2N2219 transistor
operating from a 5.7v separate supply, and while this works, I think there's
got to be a neater way. I'm hoping to find a single IC that will accept a
TTL-compatible input and produce the necessary output, but so far I've
failed to find anything. Since we're only producing single-digit
quantities, cost is not a primary factor.

TIA
Norm
 
J

John Larkin

Jan 1, 1970
0
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
The hard part is that the pulse at the receiving end has to go from 0 to 5V
(more or less). Standard 50 ohm TTL drivers (74S140) are only spec'd to
produce a TTL-high of 2.5 volts and that's what I've observed; paralleling
74128 (75 ohm drivers) doesn't do much better. The basic specs are 0-5v
into 50 ohms with rise/fall times on the order of (but not much greater
than) 20 ns. The board we're using has a 5v supply output so we can use
that for any TTL(or equiv)-logic and a single voltage external supply to
provide the higher voltage needed to get rail-to rail 0-5v output. We have
on hand some 6v/1A regulated wall-warts which should be adequate.

My prototype used a 7407/7417 (HV OC inverter) to drive a 2N2219 transistor
operating from a 5.7v separate supply, and while this works, I think there's
got to be a neater way. I'm hoping to find a single IC that will accept a
TTL-compatible input and produce the necessary output, but so far I've
failed to find anything. Since we're only producing single-digit
quantities, cost is not a primary factor.

TIA
Norm

Just use an HC or HCT part. Paralleling several/all sections of an
HC(T)04 should do what you want. AC-series parts are even a bit
fiercer, but maybe *too* fast.

John
 
N

Norm Dresner

Jan 1, 1970
0
John Larkin said:
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
[SNIP]


Just use an HC or HCT part. Paralleling several/all sections of an
HC(T)04 should do what you want. AC-series parts are even a bit
fiercer, but maybe *too* fast.

Umm.. NO. A 5v pulse into 50 ohms requires 100 ma and that's beyond
the spec for even the 74HC(T)244 octal buffer.

Norm
 
F

Frank Bemelman

Jan 1, 1970
0
Norm Dresner said:
with
board.
[SNIP]


Just use an HC or HCT part. Paralleling several/all sections of an
HC(T)04 should do what you want. AC-series parts are even a bit
fiercer, but maybe *too* fast.

Umm.. NO. A 5v pulse into 50 ohms requires 100 ma and that's beyond
the spec for even the 74HC(T)244 octal buffer.

Perhaps a nice opamp, line National LMH6639 in SOT23?
 
J

John Larkin

Jan 1, 1970
0
John Larkin said:
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
[SNIP]


Just use an HC or HCT part. Paralleling several/all sections of an
HC(T)04 should do what you want. AC-series parts are even a bit
fiercer, but maybe *too* fast.

Umm.. NO. A 5v pulse into 50 ohms requires 100 ma and that's beyond
the spec for even the 74HC(T)244 octal buffer.

Norm

Norm,

Try it.

John
 
W

Winfield Hill

Jan 1, 1970
0
Frank Bemelman wrote...
Norm Dresner wrote ...
John Larkin wrote ...
Norm Dresner wrote...

We need to drive a terminated by a load that's equivalent to
50 ohms with a pulse created by the TTL-compatible output of
a PC counter/timer board. [SNIP]
Just use an HC or HCT part. Paralleling several/all sections of
an HC(T)04 should do what you want. AC-series parts are even a
bit fiercer, but maybe *too* fast.

Umm.. NO. A 5v pulse into 50 ohms requires 100 ma and that's
beyond the spec for even the 74HC(T)244 octal buffer.

Perhaps a nice opamp, line National LMH6639 in SOT23?

There are many ways to drive 50-ohm lines, including a few rather
attractive ones we mention in AoE. But let's reconsider the '244.
Each of the eight sections has a roughly 50-ohm Ron for the pullup
FET (see the SGS datasheet, also pullup is higher than pulldown).
Eight in parallel should be under 6 ohms, marginally-suited for
driving a single 50-ohm far-end termination to about 4.5V at 90mA.
Various manufacturer's parts have Vdd and Vss supply-pin maximum
ratings of 50 to 75mA (some Fairchild parts). OK, but will they
work? They'll no doubt work just fine. But with a large number
of parts over a long time period, perhaps some can be expected to
have electromigration problems.

As for myself, I like 50-ohm source termination with open-circuit
destination termination. This only requires a few HC gates. But
should I need to provide a zero-ohm source, 50-ohm destination
termination, I'd consider a tc4427 mosfet driver IC, both outputs
paralleled for Ron < 7 ohms (5V supply).

Or I'd use my favorite AoE single-transistor solution (page 612).

Thanks,
- Win

whill_at_picovolt-dot-com
 
R

Rene Tschaggelar

Jan 1, 1970
0
Norm said:
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
The hard part is that the pulse at the receiving end has to go from 0 to 5V
(more or less). Standard 50 ohm TTL drivers (74S140) are only spec'd to
produce a TTL-high of 2.5 volts and that's what I've observed; paralleling
74128 (75 ohm drivers) doesn't do much better. The basic specs are 0-5v
into 50 ohms with rise/fall times on the order of (but not much greater
than) 20 ns. The board we're using has a 5v supply output so we can use
that for any TTL(or equiv)-logic and a single voltage external supply to
provide the higher voltage needed to get rail-to rail 0-5v output. We have
on hand some 6v/1A regulated wall-warts which should be adequate.

Usually, not 5V are required, but a TTL_Hi level.
So a Phillips 74F0037 would do the job.
The 100mA required at 5V are an extremely bad choice,
a better choice would be a conversion to ECL or LVDS
and back at the other side.

Rene
 
W

Winfield Hill

Jan 1, 1970
0
Rene Tschaggelar wrote...
Usually, not 5V are required, but a TTL_Hi level.
So a Phillips 74F0037 would do the job.

74F3037

Thanks,
- Win

whill_at_picovolt-dot-com
 
F

Fred Bloggs

Jan 1, 1970
0
Norm said:
We need to drive a terminated by a load that's equivalent to 50 ohms with a
pulse created by the TTL-compatible output of a PC counter/timer board.
The hard part is that the pulse at the receiving end has to go from 0 to 5V
(more or less). Standard 50 ohm TTL drivers (74S140) are only spec'd to
produce a TTL-high of 2.5 volts and that's what I've observed; paralleling
74128 (75 ohm drivers) doesn't do much better. The basic specs are 0-5v
into 50 ohms with rise/fall times on the order of (but not much greater
than) 20 ns. The board we're using has a 5v supply output so we can use
that for any TTL(or equiv)-logic and a single voltage external supply to
provide the higher voltage needed to get rail-to rail 0-5v output. We have
on hand some 6v/1A regulated wall-warts which should be adequate.

My prototype used a 7407/7417 (HV OC inverter) to drive a 2N2219 transistor
operating from a 5.7v separate supply, and while this works, I think there's
got to be a neater way. I'm hoping to find a single IC that will accept a
TTL-compatible input and produce the necessary output, but so far I've
failed to find anything. Since we're only producing single-digit
quantities, cost is not a primary factor.

TIA
Norm

That is just a simple OC high current switch- use a Shottky B-C clamp to
eliminate storage time delay. If the termination at the receiving end is
to GND, but still a DC connection, then go with an pnp. OnSemi and Zetex
have many candidates for the job-use the selector guide under
"switching", they work:).

Please view in a fixed-width font such as
Courier.


5V
|
/
50R
/
\
SD |
+-|>|----+---------------+
| |
| c +------------+
RB | |/ | |
o---/\/\--+------| | |
|\ | |
e | |
| | |
+--+ |
| |
--- ---
gnd gnd
 
N

Norm Dresner

Jan 1, 1970
0
Rene Tschaggelar said:
Usually, not 5V are required, but a TTL_Hi level.
So a Phillips 74F0037 would do the job.
The 100mA required at 5V are an extremely bad choice,
a better choice would be a conversion to ECL or LVDS
and back at the other side.

There are many other choices, but they weren't mine. After the
equipment was built and they wanted to build the production test set, I was
called to write a major Windows Emulation program, a device driver for the
counter/timer board, and to provide a way to get this signal from the
program to the equipment. They have an optoisolator at their end that has a
rather heavy current requirement as well as the 50 ohm termination. Based
on their design, we need at least 4.6V across the termination to get the
required current through the optoisolator and its series resistor.

Norm
 
Top