Hi guys,

I wonder if any of you guys have accoutered this before. I built an AC coupled non-inverting amplifier using an LM358 as in:


The problem I'm facing is output latch-up at the maximum output voltage upon power up. Only after a moment like in 10-15 seconds, the voltage gradually reduces to the output offset value I set by the voltage dividers at the non-inverting input.

I tried connecting the voltage reference later a moment after power up after reading http://www.allaboutcircuits.com/textbook/semiconductors/chpt-8/op-amp-practical-considerations/ . It says that I need to fire the inputs only after the main power has charged it's internal capacitors. However, upon connection, the output shoots to maximum and latches and only after 10-15 seconds it 'normalizes' just like before.

I don't wan't the user of my device waiting for this time period. How can the latchups be prevented?

Help please..thanks!

 

Post a schematic for your circuit, and the waveforms if you're able.

 

Hi,
Thanks in advance!
Well here is the actual circuit I'm using:

System voltage is 3.3V.
You could also appreciate a picture of my breadboard prototype:

Using the MCU's ADC to sample the output of the op amp, I was able plot this waveform:

Can you make it out from these?
Thanks in advance!

 

What is the supply voltage?

 

The system voltage is 3.3V.

 

An LM358 can barely run on only 3.3 V. Look at the input common mode range and output voltage swing limits on the data sheet.
You are asking for an AC gain of 4,300. That's 0ver 72 dB. Very few new opamps can do that. The 358 isn't new, and can't.
The amp is saturating because of the enormous DC error gain before the 22 uF cap can charge up to the input offset voltage. The charging current comes from the output through the 4.3M resistor. It's a simple R-C time constant adjusted for the fact that the output cannot swing to the positive rail.

Use a higher power supply voltage or change to a newer, rail-to-rail opamp rated for 3.3 V operation.
Reduce the single-stage gain so that you have at least 20 dB of negative feedback *at the highest frequency of interest*.
If you really need 73 dB of gain, you probably will have to have two amp stages in series.

ak

 

Hi,
Thanks in advance!
Well here is the actual circuit I'm using:
View attachment 98326
System voltage is 3.3V.
You could also appreciate a picture of my breadboard prototype:
View attachment 98327
Using the MCU's ADC to sample the output of the op amp, I was able plot this waveform:
View attachment 98329
Can you make it out from these?
Thanks in advance!

Because it take about 15 seconds (or more) to charge the 22uF capacitor to the level of the voltage divider. You are charging it with a 4.3M ohm resistor so that is a time constant of 95 seconds. Luckily, your voltage divider on the non-inverting input is less than 63% of Vcc so you don't have to wait so long. Also, there is probably some residual charge in that 22uF cap from your previous use.

I would dump that 22 uF cap and add a voltage divider there at your desired virtual ground. And, a cap from true ground to the virtual ground we are making with this voltage divider.

EDIT: this time AK beat me!

 

Thx guys. The project is not finished but this chapter is.
Well, did capacitor sizing, double stage amplifying (four actually), precision 5.5V opAmp (TLV2462). Now got clean signals. Please admire.