Op-Amp Problems, constant output voltage

[jthistle]

I am having problems with a basic op-amp circuit. I am use an LM324 quad op-amp powered by a clean 5v supply. I have attached a schematic of the circuit I am using. I would expect to get a result like the graph I attached. However, no matter what I put on the input pin I am getting a continuous 3.81v from the output pin, even if I set the input pin to ground. Any ideas? It's been a few years since I have messed with electronics and I am trying to get back into it...

Thanks,
Jordan

 

[(*steve*)]

Is this in real life, or in a simulation?

 

[KrisBlueNZ]

That circuit is designed to use split supply rails. The positive supply pin on the LM324 should be positive relative to the 0V "common" rail (the earth symbol in the schematic) and the negative supply pin should be negative relative to the 0V rail. The LM324 is a "single supply op-amp" but that doesn't mean it will automatically work with a single supply in circuits designed for split supplies.

 

[jthistle]

@(*steve*)
The circuit that doesn't work is a real circuit on my bread board. the simulation was for comparison.

@KrisBlueNZ
I apologize the schematic I posted isn't 100% accurate I through it together trying to get this posted. The input signal to the opamp is positive only varying between 0 and 0.01 V with a pseudo sine wave. Do I still need a split rail supply? I was under the impression that I only needed a single supply for this. If this is not possible do you have any suggestions for a layout? I would prefer to use a single supply and ultimately I'm looking for a gain of approximately 100.

Thanks again,
Jordan

 

[(*steve*)]

I apologize the schematic I posted isn't 100% accurate

Let's fix that up first or we'll be trying to solve a different problem to the one you have.

 

[jthistle]

The attached schematic is what I am doing eventually, but I am modifying it to have 2 amplification stages instead of one.. I was trying to test just a single stage of amplification to start with, however, I get a constant voltage output. I also tried setting up the op amp with different gains and I still get the same result. I apologize for the confusion.

 

[(*steve*)]

Several questions.

1) Are LDR's sensitive to IR light?
2) Is the response time of the LDR fast enough for your purposes?
3) What are your purposes?

I could look some of these things up, but I have to run...

 

[jthistle]

1) Yes, the LDR is sensitive to IR light.
2) It should be fast enough.
3) It is a pulse sensor. The schematic is one that has been used and shown to work in the past. The IR light penetrates the ear or finger and the blood flow causes minute changes in the light captured by the LDR. The output from the op amp is fed into a microcontroller for processing and display. The microcontroller side is no problem, I program for a living. The max heart rate I am interested in is 180 bpm which would mean approx 1/3 of a second response time.

The circuit also isn't working with a red light source. Red is also supposed to work for this purpose but gives more noise in the signal. I tried a visible light to make sure the IR wasn't the problem.

 

[KrisBlueNZ]

OK, you need to modify that circuit to work from a single supply. You need to create a "virtual" or "derived" ground rail at around half the supply voltage. Then you need to return R2 and R3 to this rail instead of the 0V rail.

A derived ground rail is typically created using two equal resistors (typically around 10k) in series, connected from +5V to 0V, so the centre point (the derived ground) will sit at around 2.5V. You also decouple the derived ground to 0V with a capacitor so it is "firm" to AC - i.e. not affected by AC currents from the circuitry that is using the rail. In this application though, these AC currents will be at a very low frequency, so a very large capacitor would be needed. So it's impractical to make the rail "firm" with a capacitor. You can use another element of the LM324 as a voltage follower to buffer the 2.5V from the divider and provide a firm derived ground rail, or you can use a TLE2426 (Texas Instruments), a TO92-packaged voltage rail splitter.

These changes will shift the DC voltages on all pins of the op-amp up to 2.5V. Assuming your microcontroller's ADC uses a 5V reference, it will read nominally half scale (128, 512, 2048 or whatever) when there is no signal.

The LM324 is not a very good op-amp in many respects. In this case, you will lose quite a bit of the useful range of your ADC because the LM324's outputs cannot swing very close to the positive rail. There are many op-amps available with "rail-to-rail outputs" from various companies. Search on Google, or Digikey, Mouser etc. Make sure the op-amp has rail-to-rail OUTPUTS; rail-to-rail inputs are more common but aren't needed here because the input signal is so small.

I would also suggest using a variable gain amplifier, so the microcontroller can adjust the gain to keep the ADC input voltage high enough to measure clearly without clipping. That is unless you discover that the heartbeat signal from the LDR has a fairly constant amplitude regardless of measuring conditions (thick skin, sensor not positioned quite right, etc).

 

[jthistle]

Alright, I appreciate the help. I ordered a couple tlv2473's, they seem to fit the bill. They are a single supply op amp with rail to rail input/output. Hopefully that will do the trick.

http://www.ti.com/lit/ds/symlink/tlv2473.pdf

 

[KrisBlueNZ]

Yes, the TLV247x looks very nice and ideal for your application. I'm really impressed with TI's op-amp range, and they're not expensive.