invert digital signal: BJT or MOSFET

[earckens]

Not necessarily for TTL, yet for a digital signal: how would you invert the signal: using a BJT or a MOSFET?

 

[Grossel]

I find it most easy using pull-up resistors and bjt's, mosfet works as well. For TTL's be aware that many of the old circuits have open collector output and input will be regarded as high if not connected (not recommended).

 

[MikeMl]

Need more information:

What is the amplitude of the input signal?
Where do you want the trip point to be? ~Half of the input amplitude?
How much delay can you tolerate?
Does the delay from rising edge of input to falling output have to be the same as going the other way?
What does the output have to drive? Source current same as Sink current?
What is the power supply voltage used on the output side?
Are you simultaneously require level shifting from input to output?

I could come up with about 20 more...

 

[earckens]

Need more information:

What is the amplitude of the input signal? depends on application, but mostly 12V
Where do you want the trip point to be? ~Half of the input amplitude? it's square wave, so halfway surely would do
How much delay can you tolerate? does not matter, frequency <200Hz
Does the delay from rising edge of input to falling output have to be the same as going the other way? does not matter
What does the output have to drive? Source current same as Sink current? source = sink, impedance about 100Ohm
What is the power supply voltage used on the output side? 12V
Are you simultaneously require level shifting from input to output? I do not understand?

Answers written behind your questions.

I could come up with about 20 more...

Go ahead

 

[earckens]

I find it most easy using pull-up resistors and bjt's, mosfet works as well. For TTL's be aware that many of the old circuits have open collector output and input will be regarded as high if not connected (not recommended).

The reason I prefer MOSFET is that it is easier to drive

 

[MikeMl]

Harkens back to day of RTL.

About your current sink, current source requirement? Do you mean you want a single circuit that can both source 12V/100Ω = 120ma and sink 120mA. Or do you mean that sometimes you would like to source 120mA into a grounded load (using circuit A) and other times you would like the circuit to sink 120mA from a load returned to +12V, and you would be willing to change (to circuit B) depending on the nature of the load?

 

[dknguyen]

The reason I prefer MOSFET is that it is easier to drive

I do too because no gate resistor needed like the base resistor in a BJT.

 

[MikeMl]

I do too because no gate resistor needed like the base resistor in a BJT.

Hardly puts the trip point halfway between 0V and 12V so has crummy noise immunity. With an NPN, it starts turning on at about 1V unless you create a two resistor voltage divider. Same with an NFET except that it begins to turn on at Vth and the resistors can have higher resistance...

If you turn it upside down and use PNP or PFET, the thresholds move to ~11V and (12-Vth) respectively.

 

[dknguyen]

Hardly puts the trip point halfway between 0V and 12V so has crummy noise immunity. With an NPN, it starts turning on at about 1V unless you create a two resistor voltage divider. Same with an NFET except that it begins to turn on at Vth and the resistors can have higher resistance...

Sorry, that post seemed a little incomprehensible to me. Sort of like it was typed on a phone.

 

[AnalogKid]

To invert a 12 V signal and produce another 12 V signal, you will get more consistant performance with a BJT and a zener diode than with a MOSFET, meaning a more predictable trip point with less drift due to temperature changes and better repeatability from one component to another.

ak

 

[earckens]

It should not be that complicated? Isn't it possible just to list "pro" and "con" in either general terms for each approach?

 

[earckens]

To invert a 12 V signal and produce another 12 V signal, you will get more consistant performance with a BJT and a zener diode than with a MOSFET, meaning a more predictable trip point with less drift due to temperature changes and better repeatability from one component to another.

ak

thanks, that I do understand.

 

[earckens]

To invert a 12 V signal and produce another 12 V signal, you will get more consistant performance with a BJT and a zener diode than with a MOSFET, meaning a more predictable trip point with less drift due to temperature changes and better repeatability from one component to another.

ak

And what if the trip point does not matter, as in a simple low frequency circuit (less than 200Hz)?

 

[dknguyen]

I wonder how that compares to a MOSFET or BJT totem pole. where the N-device is on the high-side and P-device is on the low-side.

EDIT: Wait, NVM, that doesn't invert the signal. Having the N-device on the low-side and P-device on the high-side would...but then there's some risk of shootthrough-like behaviour if the devices aren't matched properly.

 

[MikeMl]

Sorry, that post seemed a little incomprehensible to me. Sort of like it was typed on a phone.

That's ok. English is not my native language...

Here is a picture that has serious problems. V(out1) vs V(in) at three different values of load resistance.

Note the skewed threshold.

Note the output voltage when the load is 100Ω.

Note the power dissipation in R1.

 

[earckens]

Harkens back to day of RTL.

About your current sink, current source requirement? Do you mean you want a single circuit that can both source 12V/100Ω = 120ma and sink 120mA. Or do you mean that sometimes you would like to source 120mA into a grounded load (using circuit A) and other times you would like the circuit to sink 120mA from a load returned to +12V, and you would be willing to change (to circuit B) depending on the nature of the load?

The load (in my case here) is high impedance.

 

[MikeMl]

The load (in my case here) is high impedance.

That is not what you said here:
source = sink, impedance about 100Ohm

 

[dknguyen]

That's ok. English is not my native language...

Here is a picture that has serious problems. V(out1) vs V(in) at three different values of load resistance.

Note the skewed threshold.

Note the output voltage when the load is 100Ω.

Note the power dissipation in R1.

View attachment 111331

Ah, I see. THat clears things up.

 

[earckens]

That is not what you said here:
source = sink, impedance about 100Ohm

My apologies! I re-read that post, the other answers are ok, sorry, impedance should be high

 

[ChrisP58]

One of my favorite little utility components is the pre-biased BJT transistor, sometimes listed as a digital transistor. Here's one example. https://www.diodes.com/assets/Datasheets/ds30384.pdf

Very nice for inverting signals and for LED current buffers.

There are many more from multiple vendors with different resistor ratios, collector current ratings, and in both polarities.