Is there a part?

Not necessarily at the same time.

"Supply Voltage (Min): 4.5 V"

Nope. It'll work in the spot where I've got a 5V supply (sometimes), but not where I have a 3.3V supply.

I mean -- it'll _probably_ work at 3.3V, but I'd rather risk the customer stuffing the wrong part than risk a part not functioning correctly when presented with out-of-spec voltages.

run

eds

t,

ion.

d text -

Looks like all the CMOS Schmitt's set their worst-case VT(+) at

The Fairchild FDV301N is cheaper than cheap and available from Mouser- to be used as open drain low voltage inverter easily compatible with

Seems like the issue is that BOTH cases have 3.3V input swings, but one output needs to be full 5V ???

Pretty much a piece-a-cake in my custom world ;-)

In the discrete world the problem is that a 5V (VDD) inverter may be unhappy at a 3.3V VIH, drawing overlap current.

Thus my still unanswered question: SPEED requirements? ...Jim Thompson

solution.

Is your requirement to have an inverter that accepts 3.3V input but works on either 3.3v or 5V VDD?

SPEED? ...Jim Thompson

A buffered three-stage inverter should be ok, only the first small geometry stage would draw a little. That's why I suggested to use to resistors to scoot the range up so it lands somewhat symmetrical around the VCC/2 threshold and safely covers the process extremes. Of course no more input current than the substrate diodes can safely stomach, and one has to think about the case when the 5V supply is off or has failed. Very likely ok though.

Since Tim has looked into 74HCT it can't be supersonic :-)

In CMOS, you need to be less than a VTH from rail to avoid overlap current. HCT is a trap for people not understanding CMOS. All HCT means is that the threshold is optimized to keep the output symmetric.

Above threshold only means that the output is a low. Note that nand gate mentioned above states (in the datasheet) VIH > 0.7*VDD. For 5V VDD that's 3.5V :-( ...Jim Thompson

Probably somewhere in the middle :-)

HCT was introduced so people could interface to LSTTL which cannot pull up to the rail. Do you know why the HCT datasheets say that VCC must be

That's why I suggested intoducing an offset. Then the input won't go to zero anymore but since min threshold is 0.3*VCC that can be ok.

Use the LVC part supplied from 3V3 for both and make your "high voltage output" with a small cascode fet with its gate tied to 3V3. It'll be open drain, but you sure can manage that.

Maybe simply use a 2N7000 with a resistor pull-up... or get fancy and use that plus a pair of inverters to give it snap action. ...Jim Thompson

r to

ve

If 5v is just to drive the LED, a 3v3 part does that fine. I posted an ASCII schematic.

-- Cheers, James Arthur

operate a single 3.3v chip with 2 or more inverters, just clamp one input with a zener or R diver for the 5 volt input?

Jamie

We're driving blue LEDs from as little as 3.3V. I don't like only having a 10 or 25ohm ballast resistor, but it works.

JFET? NPN?

Wait until they change to the next reel or the ambient temperature changes big time :-)

I've warned them about that. They said they'd buy another reel. Management doesn't like to listen (but they do like blue LEDs). Where I had headroom (Li-Ion battery) I stuck in a current source so at least it's covered at the high end of the battery range.

stor to

leave

ng a 10

I don't like the small ballast resistors either. At 5mA that's just

-- Cheers, James Arthur

Then they need a consultant. "Do you know what time it is?" ... "Give me your watch and I'll tell you" :-)

Then you must work on an audio design or on products that are marketed to people under 30. When I bought a new electric razor it had only a red and a green LED, no blue, so I figured it must be for geezers.

Hey, a free battery indictor of the lottery kind. LED is very dim ->

Battery is low ... maybe ... or maybe not. Probably wouldn't have cost more than a buck to make a higher voltage for all the blue LEDs.