High(er) voltage Schmidt Trigger buffers with lower Threshold Voltages?

[Smokey]

RS232 receivers!  I love it.  Of course an answer to a stupid high voltage tolerant application like I'm posing would come from an era when everything wasn't reliable low voltage. 

https://www.analog.com/media/en/technical-documentation/data-sheets/MAX3180-MAX3183.pdf

3V-5.5V supply
0.5-5uA supply current
-25V to 25V input voltage range
Switch Thresholds = VTL:1.2V, VTH:1.5V
1.5MBit data rate

https://www.digikey.com/en/products/detail/analog-devices-inc-maxim-integrated/MAX3180EUK-T/1516332
$3.24 in 1s ($1.86 in 500s) is not ideal

[spostma]

GD75232 from Texas Instruments with 5 RS232 receivers and 3 drivers
is on stock at LCSC and TME for about $0,50:
https://www.lcsc.com/product-detail/C543039.html
https://www.tme.eu/en/details/gd75232dbr/rs232-rs422-rs485-integr-circ/texas-instruments/
and has +/- 30V input voltage range

[langwadt]

RS232 receivers!  I love it.  Of course an answer to a stupid high voltage tolerant application like I'm posing would come from an era when everything wasn't reliable low voltage. 

https://www.analog.com/media/en/technical-documentation/data-sheets/MAX3180-MAX3183.pdf

3V-5.5V supply
0.5-5uA supply current
-25V to 25V input voltage range
Switch Thresholds = VTL:1.2V, VTH:1.5V
1.5MBit data rate

https://www.digikey.com/en/products/detail/analog-devices-inc-maxim-integrated/MAX3180EUK-T/1516332
$3.24 in 1s ($1.86 in 500s) is not ideal

with the 5k pull down does it really gain much over a HCT14 with series resistors?

[Zero999]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876

I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 

The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

[Smokey]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876

I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 
The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

Problem with push/pull is I need the comparator Vcc at the highest input voltage so I'm not diode clamping anything, and then need it to feed a 3.3V microcontroller.  The open collector pull up resistor is part of the hysteresis. (Unless I'm missing something, which I probably am)

[PCB.Wiz]

The more recent ESD dual mosfets have lower thresholds even at 60v and the 20v gate gives wide range.
The gain is very high, so they switch well even with ms slew drive.
If HW hysteresis is critical you can use two fets per channel, but it would be rare cases where that was critical like quadrature hw encoding.

Attached are spice examples of Low C, ESD MOSFETS with low thresholds, and a means to add HYST if that is really needed. (needs 1 more fet and 2 more R's per channel)

[Zero999]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876

I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 
The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

Problem with push/pull is I need the comparator Vcc at the highest input voltage so I'm not diode clamping anything, and then need it to feed a 3.3V microcontroller.  The open collector pull up resistor is part of the hysteresis. (Unless I'm missing something, which I probably am)

I'm confused.

Please state the following:

Input high voltage.

Input low voltage.

Output voltage.

Power supply voltage(s).

Only then, can we think about designing something.

[Smokey]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876
I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 
The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

Problem with push/pull is I need the comparator Vcc at the highest input voltage so I'm not diode clamping anything, and then need it to feed a 3.3V microcontroller.  The open collector pull up resistor is part of the hysteresis. (Unless I'm missing something, which I probably am)

I'm confused.
Please state the following:
Input high voltage.
Input low voltage.
Output voltage.
Power supply voltage(s).
Only then, can we think about designing something.

Sorry.  I was never really asking anyone to design anything here.  I have a viable solution, which I noted before I even asked the question about why a single chip doesn't exist.  Like I said up front, It's just bigger footprint than I would like it to be, and it made me make some design compromises, but it should work fine.  Just big and the history of electronics is that things that take up a lot of board space in discrete implementations often get made into dedicated self contained ICs.  And I was complaining that this particular problem apparently has evaded that path.  If someone here happened to know of a chip that actually did what I was looking for then I wanted to hear about it, but I was fairly confident it didn't exist.
I really like the suggestion of RS232 receivers in this context since it's close (but not exactly the same, or exactly viable here) considering they solve a similar problem from obviously a time that didn't care as much about bandwidth or quiescent current.  I wasn't trying to say I'm converting everything to RS232 receivers.
Both the resistor network and the self biased transistor followed by something with a Schmidt trigger is also pretty interesting but I would need to see if a two chip design would be bigger or smaller than a comparator and 4 passives per channel, since that's not obvious enough for me to redo the layout I already have.

[langwadt]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876
I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 
The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

Problem with push/pull is I need the comparator Vcc at the highest input voltage so I'm not diode clamping anything, and then need it to feed a 3.3V microcontroller.  The open collector pull up resistor is part of the hysteresis. (Unless I'm missing something, which I probably am)

I'm confused.
Please state the following:
Input high voltage.
Input low voltage.
Output voltage.
Power supply voltage(s).
Only then, can we think about designing something.

Sorry.  I was never really asking anyone to design anything here.  I have a viable solution, which I noted before I even asked the question about why a single chip doesn't exist.  Like I said up front, It's just bigger footprint than I would like it to be, and it made me make some design compromises, but it should work fine.  Just big and the history of electronics is that things that take up a lot of board space in discrete implementations often get made into dedicated self contained ICs.  And I was complaining that this particular problem apparently has evaded that path.  If someone here happened to know of a chip that actually did what I was looking for then I wanted to hear about it, but I was fairly confident it didn't exist.
I really like the suggestion of RS232 receivers in this context since it's close (but not exactly the same, or exactly viable here) considering they solve a similar problem from obviously a time that didn't care as much about bandwidth or quiescent current.  I wasn't trying to say I'm converting everything to RS232 receivers.
Both the resistor network and the self biased transistor followed by something with a Schmidt trigger is also pretty interesting but I would need to see if a two chip design would be bigger or smaller than a comparator and 4 passives per channel, since that's not obvious enough for me to redo the layout I already have.

since you didn't answer the questions no one can tell you if a chip exist

[Smokey]

Closest I could find is a dual-voltage bus transceiver, but the voltage are too low (3 to 5V) and voltage tolerances too low (input V + 0.5).  https://www.ti.com/lit/ds/scas585r/scas585r.pdf?ts=1781837273876
I assume you're trying to harden some generic inputs that you expect customers to hookup a variety of things to, all at different voltage levels?

Ya, I have a system that needs to operate with mixed 12V, 5V, and 3.3V stuff. 
The comparator with configurable hysteresis thresholds works fine, but it's 4 resistors per channel and you can't share those configuration components between channels like you can with just a voltage divider to set the single trip of a basic comparator on the whole package.  4 x 15 is a lot of parts, even as 0402.

For non-inverting configuration, it's only two resistors per channel if a comparator with an push-pull output is used and all channels can use the same reference.

Problem with push/pull is I need the comparator Vcc at the highest input voltage so I'm not diode clamping anything, and then need it to feed a 3.3V microcontroller.  The open collector pull up resistor is part of the hysteresis. (Unless I'm missing something, which I probably am)

I'm confused.
Please state the following:
Input high voltage.
Input low voltage.
Output voltage.
Power supply voltage(s).
Only then, can we think about designing something.

Sorry.  I was never really asking anyone to design anything here.  I have a viable solution, which I noted before I even asked the question about why a single chip doesn't exist.  Like I said up front, It's just bigger footprint than I would like it to be, and it made me make some design compromises, but it should work fine.  Just big and the history of electronics is that things that take up a lot of board space in discrete implementations often get made into dedicated self contained ICs.  And I was complaining that this particular problem apparently has evaded that path.  If someone here happened to know of a chip that actually did what I was looking for then I wanted to hear about it, but I was fairly confident it didn't exist.
I really like the suggestion of RS232 receivers in this context since it's close (but not exactly the same, or exactly viable here) considering they solve a similar problem from obviously a time that didn't care as much about bandwidth or quiescent current.  I wasn't trying to say I'm converting everything to RS232 receivers.
Both the resistor network and the self biased transistor followed by something with a Schmidt trigger is also pretty interesting but I would need to see if a two chip design would be bigger or smaller than a comparator and 4 passives per channel, since that's not obvious enough for me to redo the layout I already have.

since you didn't answer the questions no one can tell you if a chip exist

I mostly did in the OP.  But it's all good.  If you managed to surface something we haven't already talked about you would be both my new personal hero and likely a miracle worker   Like I said, I'm good either way for now.

[PCB.Wiz]

....  Like I said up front, It's just bigger footprint than I would like it to be, and it made me make some design compromises, but it should work fine.  Just big and the history of electronics is that things that take up a lot of board space in discrete implementations often get made into dedicated self contained ICs.  And I was complaining that this particular problem apparently has evaded that path.  If someone here happened to know of a chip that actually did what I was looking for then I wanted to hear about it, but I was fairly confident it didn't exist.

The volumes of what you seek, are not large, so do not expect too many offerings.

I already listed 2 parts above with 7 channels of voltage tolerant buffering.
TI TLP7407 
3Peak TPM2003

Other parts that swallow many passives are dual channel ones like OnSemi's SZNUD3160DMT1G and Toshiba's SSM6N357R
These have pull down resistors, and output clamps and are > 12V rated in and > 60V rated out.

Of course, all those are not widely second sourced, so anyone needing such universal input buffers in high volume, might craft a more widely sourced solution.

The lowest Cin part with clamping appears to be 74HC2G17/14, which have ±20mA clamping and 2pF Cin.
That low Cin means they can switch in under 50ns even with 10k series R, which is then good to ±200V

A 74HC2G17 plus a resistor network can manage 2 channels in 2 placed small parts. Fast, flexible and robust and outside the scope for full integration into low cost IC processes.

[Smokey]

....  Like I said up front, It's just bigger footprint than I would like it to be, and it made me make some design compromises, but it should work fine.  Just big and the history of electronics is that things that take up a lot of board space in discrete implementations often get made into dedicated self contained ICs.  And I was complaining that this particular problem apparently has evaded that path.  If someone here happened to know of a chip that actually did what I was looking for then I wanted to hear about it, but I was fairly confident it didn't exist.

The volumes of what you seek, are not large, so do not expect too many offerings.

I already listed 2 parts above with 7 channels of voltage tolerant buffering.
TI TLP7407 
3Peak TPM2003

Other parts that swallow many passives are dual channel ones like OnSemi's SZNUD3160DMT1G and Toshiba's SSM6N357R
These have pull down resistors, and output clamps and are > 12V rated in and > 60V rated out.

Of course, all those are not widely second sourced, so anyone needing such universal input buffers in high volume, might craft a more widely sourced solution.

The lowest Cin part with clamping appears to be 74HC2G17/14, which have ±20mA clamping and 2pF Cin.
That low Cin means they can switch in under 50ns even with 10k series R, which is then good to ±200V

A 74HC2G17 plus a resistor network can manage 2 channels in 2 placed small parts. Fast, flexible and robust and outside the scope for full integration into low cost IC processes.

Ya, you got me there.  These are all really interesting solutions and I should have addressed them better.

The TPL7407, I wouldn't have thought to look at a low side driver as an input buffer, but ya, it has 40V tolerant inputs.  I'm not sure what the expected application is there.  There must be a story behind this.  What made you think of this part?  Previous project?  The 50k series input resistor is limiting current into the clamp diodes.  I would also need to add hysteresis somewhere else.  Funny using a 500mA rated driver backwards as a level shifter.  So this is another two chip thing (TPL7407 + Schmidt trigger).  The 7Channel 30V tssop16 TPL7407LAPWR is only $0.45 in 1s, which is pretty nice.  One potential issue is the data sheet doesn't appear to have a supply current spec.  Not sure how that got through.

3Peak TPM2003: I admit I have no evidence to justify this, but my design intuition isn't comfortable with 3Peak right now.  Is everyone comfortably designing in 3Peak stuff into commercial products now? 

I already said the integrated transistor as an input buffer was worth looking into.  That covers SZNUD3160DMT1G and Toshiba's SSM6N357R.  It would be nice to have larger gate counts per chip than 2 though.

The 74HC2G17 with only 2pF of input capacitance is pretty fast.  I agree this would probably be fine with just a series resistor.  The 74HCT2G17 (TTL-Level version has 20 times more supply current than the nonT CMOS one?).
SN74AC17 is also 2pF.  I shouldn't have assumed they were all much higher.

[PCB.Wiz]

I would also need to add hysteresis somewhere else.  Funny using a 500mA rated driver backwards as a level shifter.  So this is another two chip thing (TPL7407 + Schmidt trigger).  The 7Channel 30V tssop16 TPL7407LAPWR is only $0.45 in 1s, which is pretty nice.  One potential issue is the data sheet doesn't appear to have a supply current spec.  Not sure how that got through.

3Peak TPM2003: I admit I have no evidence to justify this, but my design intuition isn't comfortable with 3Peak right now.  Is everyone comfortably designing in 3Peak stuff into commercial products now? 

They are pin compatible, so you can try both.
The 3Peak one does show HYST on the drawing, but only gives a MAX Icom, no typical.
TI's TPL7407LA specs 17uA typ and 30uA max, on Icom, & does not show HYST, but it does have quite tight <0.9 (5uA) and >1.5V(100mA) spec points, so that plus the non-zero Icc suggests more active silicon than a simple 7 MOSFETS seen in N channel MOS x62003A parts.
Could be worth measuring TI's part for hysteresis.
   

[David Hess]

I forgot to mention that one advantage of the RS-232 receivers is that most if not all of them have built in hysteresis.