How can I reduce levels of a pulse?

[hannover90]

Hello,

I want to reduce the level of a pulse from 2.5V/5.0V to 0V/2.5V.

The common levelshifters rise the levels of the pulse (from 2.5V/5V to 5V/7.5V).

Does anyone has an idea, how can I reduce the level of a pulse?

Many thanks in advance.

 

[gag2000]

Hi,
Maybe a pulse tranformer could to the job.
Alternatively, you can use a comparator, supplied with 2.5V and with 5V tolerant inputs, put the threshold at (5-2.5)/2, it should work.
Or, you may use a serial capacitor to get rid of the 2.5 DC voltage, and use a comparator with 2.5V supply.

Hope this helps.

 

[erikl]

Use a series decoupling capacitor and then a resistor to GND (CR differentiation element).

 

[hannover90]

gag2000 and erikl, thank you very much for so many helpful tips.

 

[BradtheRad]

The right combination of led/ zener/ diodes, will drop 2.5V from your signal.

You may have to accept some amplitude reduction.
The potentiometer may or may not be needed.

 

[crutschow]

Use a series decoupling capacitor and then a resistor to GND (CR differentiation element).

For that technique to work you will likely need to add a Schottky diode clamp (diode across resistor, anode to ground) to avoid level shifting as determined by the duty-cycle of the pulse. This will cause the bottom of the pulse to be about 0.5V below ground, which can't be readily avoided.

 

[Borber]

What is Differential Amplifier or The Voltage Subtractor?

 

[crutschow]

What is the pulse width and frequency?

 

[erikl]

Use a series decoupling capacitor and then a resistor to GND (CR differentiation element).

For that technique to work you will likely need to add a Schottky diode clamp (diode across resistor, anode to ground) to avoid level shifting as determined by the duty-cycle of the pulse. This will cause the bottom of the pulse to be about 0.5V below ground, which can't be readily avoided.

You're totally right, crutschow, I forgot this. But even adding the Schottky diode clamp doesn't seem to help a lot: View attachment CR-lvl-shft.pdf

But I don't really trust this simulation result. Why wouldn't it clamp?

 

[crutschow]

..............................
But I don't really trust this simulation result. Why wouldn't it clamp?

Dunno. My simulation (below) works fine. It would appear your diode model is not working, since obviously you can't have that much forward voltage across a normal diode without many amps of current.

 

[erikl]

It would appear your diode model is not working

You're right again, crutschow: The BAS40 Schottky diode is too "weak", i.e. has too much series resistance. The stronger 1N5817 actually clamps correctly (even if my 1N5817 model seems inferior to yours) ;-) :