5V tolerant I/O driver that doesn't clamp to 3.3V?

[markp]

Hi All,

I need a logic device (244 buffer in SOIC footprint) that is powered by
3.3V, and which when driving high will actively drive up to 3.3V, then go
high impedance to let a pull-up resistor continue to pull it up to 5V. In
other words, a 5V tolerant I/O device that does not actually clamp to 3.3V
when driving high. I've seen quite a few 5V I/O tolerant logic families, but
all I've seen so far only state up to 5V when tri-stated and 3.3V when
driving high. It's not clear whether LPT from Pericom will do this. Anyone
know of a logic family that can do this?

TIA,

Mark.

 

[Joerg]

Hi Mark,

There are some 5V tolerant open drain versions similar to the 244, such as this
one:

http://www.fairchildsemi.com/ds/74/74LCX760.pdf

Regards, Joerg

 

[Mike Randelzhofer]

Hi All,

I need a logic device (244 buffer in SOIC footprint) that is powered by
3.3V, and which when driving high will actively drive up to 3.3V, then go
high impedance to let a pull-up resistor continue to pull it up to 5V. In
other words, a 5V tolerant I/O device that does not actually clamp to 3.3V
when driving high. I've seen quite a few 5V I/O tolerant logic families, but
all I've seen so far only state up to 5V when tri-stated and 3.3V when
driving high. It's not clear whether LPT from Pericom will do this. Anyone
know of a logic family that can do this?

TIA,

Mark.

See TI's bus switches series SN74CB3T...
http://focus.ti.com/docs/prod/folders/print/sn74cb3t3245.html
http://focus.ti.com/lit/an/scda008/scda008.pdf

Good devices, but expensive, and no '244 but '245 footprints available.

MIKE

 

[markp]

See TI's bus switches series SN74CB3T...
http://focus.ti.com/docs/prod/folders/print/sn74cb3t3245.html
http://focus.ti.com/lit/an/scda008/scda008.pdf

Good devices, but expensive, and no '244 but '245 footprints available.

MIKE

Thanks, but these are Quickswitch like products that don't drive actively to
3.3V (rather they are connected to the input via a FET). I need a buffered
logic level if no pull-up is fitted, or if a pull-up to 5V is fitted I need
the output to raise up to 5V.

Thanks,

Mark.

 

[markp]

Hi Mark,

There are some 5V tolerant open drain versions similar to the 244, such as this
one:

http://www.fairchildsemi.com/ds/74/74LCX760.pdf

Regards, Joerg

Thanks, unfortunately the outputs are clamped to 3.3V when the top FET is
driving (up to 5V is allowed but only in tri-state mode).

Mark.

 

[markp]

as

Thanks, unfortunately the outputs are clamped to 3.3V when the top FET is
driving (up to 5V is allowed but only in tri-state mode).

Mark.

Oops, didn't read your post properly sorry! The problem is that it is an
existing system with no pull-ups, therefore it must drive to TTL high, and
if a 5V pull-up is fitted must raise itself to 5V.

Mark.

 

[Fred Bloggs]

Hi All,

I need a logic device (244 buffer in SOIC footprint) that is powered by
3.3V, and which when driving high will actively drive up to 3.3V, then go
high impedance to let a pull-up resistor continue to pull it up to 5V. In
other words, a 5V tolerant I/O device that does not actually clamp to 3.3V
when driving high. I've seen quite a few 5V I/O tolerant logic families, but
all I've seen so far only state up to 5V when tri-stated and 3.3V when
driving high. It's not clear whether LPT from Pericom will do this. Anyone
know of a logic family that can do this?

TIA,

Mark.

3.3V is TTL high in a 5V circuit-

 

[markp]

3.3V is TTL high in a 5V circuit-

Yes, but I need 5V output.

Mark.

 

[Fred Bloggs]

Yes, but I need 5V output.

Mark.

Really? -And how is that?

 

[Brad Albing]

Hi All,

I need a logic device (244 buffer in SOIC footprint) that is powered by
3.3V, and which when driving high will actively drive up to 3.3V, then go
high impedance to let a pull-up resistor continue to pull it up to 5V. In
other words, a 5V tolerant I/O device that does not actually clamp to 3.3V
when driving high. I've seen quite a few 5V I/O tolerant logic families, but
all I've seen so far only state up to 5V when tri-stated and 3.3V when
driving high. It's not clear whether LPT from Pericom will do this. Anyone
know of a logic family that can do this?

If you can still do this (i.e., haven't already got the PCB all laid out and
in-hand), I'd say use a T.I. SN74AHCT244 (or comparable equiv.) and power it
from +5 VDC. Note the "T" just before the "244." This part expects to see
logic levels coming from a TTL part, so a HI would be => 2.4 VDC. Obviously,
its outputs are LO = 0 VDC (or darn close to it) and HI = 5 VDC (ODCTI).

HTH.

 

[markp]

Really? -And how is that?

Need a CMOS level output, and preferably 5V.

Mark.

 

[markp]

If you can still do this (i.e., haven't already got the PCB all laid out and
in-hand), I'd say use a T.I. SN74AHCT244 (or comparable equiv.) and power it
from +5 VDC. Note the "T" just before the "244." This part expects to see
logic levels coming from a TTL part, so a HI would be => 2.4 VDC. Obviously,
its outputs are LO = 0 VDC (or darn close to it) and HI = 5 VDC (ODCTI).

PCB is fixed unfortunately. The output can drive either of two devices
depending on what's fitted in the box (all of which are also fixed, as is
the box): Either a non-5V tolerant 3.3V logic input (rules out 5V
buffering), or a 5V CMOS input device when a pull-up resitor is added. See
the problem?!

Mark.

 

[Fred Bloggs]

PCB is fixed unfortunately. The output can drive either of two devices
depending on what's fitted in the box (all of which are also fixed, as is
the box): Either a non-5V tolerant 3.3V logic input (rules out 5V
buffering), or a 5V CMOS input device when a pull-up resitor is added. See
the problem?!

Mark.

The problem is your lack of planning- a simple circuit will do both:
Please view in a fixed-width font such as Courier.

+-----<Vdd from
+---------+ +--------+ | destination
| | | | / logic
| 5V | | | Rpu 5V or 3.3V
| | | | | /
| | | | | \
| |\ | | | |

---|---| >---|--+---|---|<|--|-----+-----> To destination

| |/ | | | | logic
| | | | | | 5V or 3.3V
| | | | | |
| gnd | | | |
| | | | |
| | | | |
+---------+ | +--------+
|
TTL INPUT | sd array
BUFFER |
ACCEPTS 3.3 |
OR 5V INPUTS |
|
| +-----<Vdd from
| +--------+ | destination
| | | / logic
| | | Rpu 3.3V or 5V
| | | /
| | | \
| | | |
+---|---|<|--|-----+-----> To destination
| | | logic
| | | 3.3V or 5V
| | |
| | |
| | |
| | |
| +--------+
|
| sd array
|
|
|
to etc

If speed or Pd is a problem then this:
Please view in a fixed-width font such as
Courier.

FOR 3.3V LOGIC

+---------+ +---------+
| | | |
| 5V | | 3.3V |
| | | | | |
| | | | | |
| |\ | | |\ |

---|---| >---|-------|---| >---|-->

| |/ | | |/ |
| | | | | |
| | | | | |
| gnd | | gnd |
| | | |
| | | |
+---------+ +---------+

TTL INPUT 5V INPUT
BUFFER TOLERANT
ACCEPTS 3.3
OR 5V INPUTS


FOR 5V LOGIC

+---------+ +---------+
| | | |
| 5V | | |
| | | | |
| | | | |
| |\ | | JMPR |

---|---| >---|-------|--o---o--|-->

| |/ | | |
| | | | |
| | | | |
| gnd | | |
| | | |
| | | |
+---------+ +---------+

TTL INPUT
BUFFER
ACCEPTS 3.3
OR 5V INPUTS

 

[Tony Williams]

The problem is your lack of planning- a simple circuit will do
both: Please view in a fixed-width font such as Courier.

+-----<Vdd from
+---------+ +--------+ | destination
| | | | / logic
| 5V | | | Rpu 5V or 3.3V
| | | | | /
| | | | | \
| |\ | | | |
| |/ | | | | logic
| | | | | | 5V or 3.3V
| | | | | |
| gnd | | | |
| | | | |
| | | | |
+---------+ | +--------+

[snip]

Memories of old RTL logic there, sending signals
down twisted-pair with 120 ohm terminators/pullups
at the far end. As with RTL though a low impedance
0v-0v connection is required between source and
destination.

 

[markp]

----- Original Message -----
From: "Fred Bloggs" <[email protected]>
Newsgroups: sci.electronics.design
Sent: Saturday, April 24, 2004 10:05 AM
Subject: Re: 5V tolerant I/O driver that doesn't clamp to 3.3V?

The problem is your lack of planning- a simple circuit will do both:
Please view in a fixed-width font such as Courier.

+-----<Vdd from
+---------+ +--------+ | destination
| | | | / logic
| 5V | | | Rpu 5V or 3.3V
| | | | | /
| | | | | \
| |\ | | | |
| |/ | | | | logic
| | | | | | 5V or 3.3V
| | | | | |
| gnd | | | |
| | | | |
| | | | |
+---------+ | +--------+
|
TTL INPUT | sd array
BUFFER |
ACCEPTS 3.3 |
OR 5V INPUTS |

<snip>

Thanks for this! The boards unfortunately are fixed now (as is the box) and
I need another solution (i.e. there's no room for diode arrays or pull-ups,
and re-spinning the PCBs is out of the question). Actually the circuit above
to be strictly accurate reduces low noise margin by 0.3V with shottky
diodes, but in reality Vol for a CMOS output stage is going to be 0.1V or
better so it'll be OK. I have only one type of input to drive, but it could
be either - the CMOS board has a pull-up on its input to 5v but the other
card does not have a pullup and is non-5V tolerant. The only thing I can do
is change the logic family of the 244 driver, I have no control over which
of the other boards is fitted. So, coming back to the original question,
what I need is a driver that actively drives to 3.3V, but will go high
impedance after that if pulled up to 5V.

Mark.

 

[Fred Bloggs]

----- Original Message -----
From: "Fred Bloggs" <[email protected]>
Newsgroups: sci.electronics.design
Sent: Saturday, April 24, 2004 10:05 AM
Subject: Re: 5V tolerant I/O driver that doesn't clamp to 3.3V?




<snip>

Thanks for this! The boards unfortunately are fixed now (as is the box) and
I need another solution (i.e. there's no room for diode arrays or pull-ups,
and re-spinning the PCBs is out of the question). Actually the circuit above
to be strictly accurate reduces low noise margin by 0.3V with shottky
diodes, but in reality Vol for a CMOS output stage is going to be 0.1V or
better so it'll be OK. I have only one type of input to drive, but it could
be either - the CMOS board has a pull-up on its input to 5v but the other
card does not have a pullup and is non-5V tolerant. The only thing I can do
is change the logic family of the 244 driver, I have no control over which
of the other boards is fitted. So, coming back to the original question,
what I need is a driver that actively drives to 3.3V, but will go high
impedance after that if pulled up to 5V.

Mark.

I don't see where you have much room to maneuver here because your board
is done. You have two options:

1) The 5V CMOS Vih is actually 3.5V so that the 3.3V will be recognized
as such 100% of the time with only slightly reduced noise margin;

-or-

2) there are such things as analog switches with input overvoltage
protection- check the Maxim product listing- these will go high
impedance as you describe- and your voltages are no problem for them.

 

[markp]

I don't see where you have much room to maneuver here because your board
is done. You have two options:

1) The 5V CMOS Vih is actually 3.5V so that the 3.3V will be recognized
as such 100% of the time with only slightly reduced noise margin;

No it won't. The minimum Vih is 3.5V, i.e. it needs 3.5V to guarantee seeing
a '1', so the high noise margin in this case is actually negative!

-or-

2) there are such things as analog switches with input overvoltage
protection- check the Maxim product listing- these will go high
impedance as you describe- and your voltages are no problem for them.

Analogue switches such as Quickswitches don't buffer the signal and rely on
the drive strength of the driver predeeding it. In this application it
requires bufferering (i.e. actively driving to 3.3V) due to driving
off-board into a 60 ohm backplane. The internal logic signal is not beefy
enough to drive this load.

Mark.