# High pass filter with triangle wave output

I'm having a problem with a high pass filter, maybe someone has an idea how to fix it.

Initially I have an IC that provides a sinusoidal signal of approximately 0 to 3V positive with a frequency varying between 50KHz and 100KHz (Figure 1).

Figure 1

The purpose of my high pass filter is to remove the DC component, leaving at the end the signal varying between approximately + 1.5V and -1.5V with the same frequency range (Figure 2).

Figure 2

I am using a Basic RC circuit with an op amp (TL081) as a buffer, and with symmetrical power supply of + 5V and -5V. The cutoff frequency is 100Hz (Figure 3).

Figure 3

My problem is that on the output of my op amp (Vout) I have a triangle wave (Figure 4).

Figure 4

The signal in "Vout" in the Figure 3 is a triangle wave with Vpp ranging from 0 to + - 1.5V and with the desired frequency (50KHz to 100KHz).

An interesting detail is that before the opamp ("Vt" in the Figure 3) the signal is correct, ie in "Vt" the signal is a sine wave of 1.5Vpp with frequency from 50KHz to 100KHz.

All observations were made on a real circuit with an oscilloscope.

Anyone have any idea what happens to my Op Amp?

* UPDATE

I did some tests with a power supply of + -15V. And the results were the same, according to the oscilloscope images below:

at "Vt" in Image 3:

at "Vout" in Image 3:

• VF, I have done what you have drawn in your schematic a million times. It works exactly as you want it to - sinusoid referenced to ground. My guess: Either you have a hook-up error or there is a bandwidth limitation with the op-amp. – FiddyOhm Aug 8 '17 at 23:34
• Try shorting out the capacitor. If you're still getting a triangle, it says that the op amp you are using is not up to spec, and its bandwidth is too low. If this happens, check by reducing the input frequency to 5 to 10 kHz. – WhatRoughBeast Aug 8 '17 at 23:36
• Pin 4 not connected to -5V? That's one of the reason I can think of since you still getting DC at output. – Jason Han Aug 9 '17 at 1:37

Have a look at the TL081's slew rate - the maximum rate of change of the output or $\frac {dV}{dt}$ but it is 13 V/µs quoted at +/-15 V supply.

Now have a look at your output. Your $dV$ = 3 V and at 50 kHz one half cycle will be $\frac {1}{2} \frac {1}{50k} = 10 \mu s$ giving you $\frac {dV}{dt} = \frac {3}{10µ} = 0.3 \; V/µs$ and 0.6 V/µs at 100 kHz.

The figures indicate you should be OK at +/-15 V but maybe not at +/-5 V. If you can drop your frequency by a factor of 10 and see if that solves the problem you might start to get better insight.

• I did the test with +/-15 V supply, and the result remains the same. Is it possible for my op amp to be fake and display this result? – VF35468 Aug 10 '17 at 9:46
• I see your update with photos. You seem to have done everything correctly. Add a photo of your setup. Someone might spot an error between your schematic and the physical and your measurement. – Transistor Aug 10 '17 at 10:42
• The problem was with my IC. I replaced it with a new one and it worked. Thank you! – VF35468 Aug 13 '17 at 12:08
• Thanks for letting me know. I wonder what defect would cause that problem. – Transistor Aug 13 '17 at 15:49

A quick look at your design and the IC specs tells me this chip is not designed for single supply 5V. This is your problem with under-biasing for the internal FETs

My guess is inadequate Vgs causing FET saturation and making your differential mV saturate internal BJT stages to a square-wave and then with slew rate limiting , the output looks like a triangle.

There are many OA's that are rated for 5V, 0V your BW, Slew rate and with CMOS outputs you can get rail to rail but at lower output current. (10K load)

From the graph below Vcc=7V is on the "hairy edge" for Vcc min. meaning 5V is not reliable.

• I did the test with +/-15 V supply, and the result remains the same. Is it possible for my op amp to be fake and display this result? – VF35468 Aug 10 '17 at 9:46