op amp: Voltage follower (unity gain buffer)

I have a voltage follower (aka unity gain buffer) set up on a breadboard.

The op amp is an MC33078N: http://www.digikey.com/product-detail/en/MC33078N/497-1953-5-ND/599527 (datasheet)
To set up the voltage follower, I connected the inverting input to the output.
My input signal is a 32k Hz square wave with a peak voltage of 5 volts and a min voltage of 0 volts.

Orignally, I had a VCC+ of 5 volts on the op amp and VCC- of 0 volts. My out put signal had a peak-to-peak voltage of only 4 volts (1 volt signal loss).
I attempted to fix the problem by changing the VCC- to -5 volts, same problem.
I tried increasing the range but the signal still showed a 1 volt loss and some added distortion at the mins.

I believe some characteristic of the op amps I am using is not suited for what I am doing but I am not sure what that is. If necessary, I can include a video of the output behavior of the op amp vs the input signal as a function of varying VCC voltages

Not all op-amps can allow their output signal to go all the way to either their positive or negative voltage rails no matter what the input signal and circuit gain is designed for. There are so called 'rail-to-rail' op-amps available that allow their output voltage swing to come very close to their power rails. So you need to use either such a rail to rail op-amp or just use your existing opamp but run the Vcc at 6vdc or higher.


I'm curious , why do you think you need OPA as a buffer for square wave? Output current of IC is even less than max output current you may get from arduino. To boost higher up in amps you may use logic IC, TC4424/4427/4428 or TC426

Lefty: Usually it would be a lot closer to the rail voltage
Magician: I am trying to "split" the square wave; I want to use 1 PWM pin and split the signal using unity gain buffers so that I can have four channels each with a Square wave, from the original PWM, driving its own piezotransducer

I am starting to think that maybe my Slew rate is too low, any thoughts?

I want to use 1 PWM pin and split the signal using unity gain buffers

I still don't get it, why can't you use normal buffers or for that matter as the signals each piezotransducer receives are the same just one buffer that can drive them all.


I probably didn't fully comprehend the response, what is a "normal buffer" i.e. how does it differ?

Your choice of OPA as a driver seems absolutely "off" in first, it's not rail-to-rail (pointed above) and secondly it's can't drive high capacitance load (piezo) - IC wasn't design for such task, if you look in data sheet everything specified for 100 pF. Look at microchip TC426, or if you persist on using OPA, you can find one 80 mA output current at adafruit

Oh I see, thank you for the clarification. I will look up the TC426 and give that a try.
Out of curiosity, I have two questions regarding your last post:

  1. how do I know if its a rail-to-rail; will it explicitly be stated? for example if i go on digikey and filter op amps, what criteria can I use to specify that I want rail-to-rail op amps?
  2. I do not see how the input capacitance plays a role or what its significance is in op amp designs? Like you said the datasheet lists a 12 pF input but I don't understand why that effects the op amp at all let alone what it is driving a PZT at its output

I found the rail-to-rail option, still not sure about the capacitance though

I do not see how the input capacitance plays a role or what its significance is in op amp designs?

I was referring to OUTPUT driving capability, (if my post is not altered, you never know what some regime does with a freedom of speech)

Slew rate and GBP specified for 100 pF, same with figure 8 in data sheet. FUI, piezo transducers usually have a 10-100 nF capacitance