# Bias voltage of non-inverting op amplifier drops to 0 when input signal connected

I am working on a non-inverting opamp amplifier, using OPA2134, to amplify the 150mV peak signal (1 kHz sine wave playbacked) from a phone audio jack to ~4V peak sine wave, that drives a LED using a N-channel MOSFET. My goal is to turn on the LED by playing music from the phone.

I have biased the non-inverting input of the opamp to 2.5V using a simple voltage divider from 5V. I want the opamp to work with a single supply (+5V).

List of signals:

• OPA2134 V+ - +5V
• OPA2134 V- - 0V (GND)
• Input Vin- 150mV peak, 1kHz sine wave.
• Desired output - ~4V peak, 1kHz sine wave.
• Bias voltage - 2.5V

The input is ac-coupled, gain is set to unity at dc, and the output is also ac-coupled. I am currently not concerned with the noise from bias divider.

If the input is not connected to the phone audio jack, the inpt is floating, and the non-inverting input of the opamp is at 2.5V, the output of the opamp is also at 2.5V and the output after the output capacitor is 0V.

If however i connect the input to the audio jack the bias voltage at the non-inverting input drops to 0V. If then i play audio from the phone, the noninverting input has a signal of 1kHz 150mV peak wave instead of 2.35V to 2.65V and there is nothing at the output, since the opamp can not work with an input signal so close to it's rail.

Here is my circuit, that simulated in falstad operates as desired. For simplicity I marked the opamp max voltage to 5V and min voltage to 0V in falstad. I know that in real life, the OPA2134 can not operate rail-to-rail. The signal on the left is the input (-150mV to 150mV), signal on the right is the output (0V - 5V)

• Firstly, 1 KHz signal turning the LED on-off would make no sense I guess. Too fast to allow a LED to turn-off and perceive it. Commented Nov 28, 2016 at 10:19
• You are right in the way, that the LED can't be perceived by the naked eye. Not what i am trying to do and not the problem behind the question. Commented Nov 28, 2016 at 10:33
• The instant you connect the signal source sure, the bias of 2.5 volts drops to 0 V then, after a few tens of milliseconds (due to the 100 nF capacitor charging from the two 150 k resistors it'll rise back to 2.5 volts. Is this the problem you refer to? Have you in fact just simulated for a few milliseconds and not given the sim long enough to stabilize? Commented Nov 28, 2016 at 10:48
• No, it does not drop to 0 in the simulator, it drops to 0 in real world, and does not rise up to 2.5V afterwards. Tested and measured with an oscilloscope. Commented Nov 28, 2016 at 10:50
• Are you sure that the 100nF cap isn't short circuit? try replacing it. Commented Nov 28, 2016 at 10:53

## 2 Answers

Are you sure that the 100nF cap isn't short circuit? Try replacing it.

I see you had the input capacitor shorted but you will likely hit another problem with this op-amp.

There is a significant problem in running your amplifier at a total power rail of only 5 volts and it is listed on page 5 of the data sheet. The table on page 5 states that the input common mode range is (V-) + 2.5 volts to (V+) - 2.5 volts and this leaves you with an input range of nothing when running from a 5 volt supply.

If you are just building a one-off you may get lucky but, if you are building for production then either increase the supply rails or choose a different op-amp.

• Thank you for the input, I was already aware of that, this is just a prototype from components i had lying around. But a good note nevertheless. Commented Nov 28, 2016 at 11:37