0
\$\begingroup\$

I have a sensor which outputs a voltage signal alternating a few millivolts around zero volts and want to design a circuit which prepares this signal for processing by a microcontroller like in the picture below:

enter image description here Now the problem is that i only have a supply voltage range between 0V and 3.3V and therefore need to add some DC offset to my signal and then amplify the alternating part.

I think that op amps are not suited for my application because they amplify the DC signal as well as the AC Signal and so I thought of a class-A-amplifier circuit like this:

schematic

simulate this circuit – Schematic created using CircuitLab

Now the question: Is it possible with the class-A-amplifier to process my millivolt input signal to the larger one in the picture at the top or are there any better methods?

\$\endgroup\$
  • 2
    \$\begingroup\$ I think I need to know more about the width of those pulses and the likely span between one and the next one. You've provided no information I can see about that. \$\endgroup\$ – jonk Sep 1 at 18:18
  • \$\begingroup\$ If R2 is not zero-ohms, the transistor cannot be turned on enough to pull a pulse to ground. Even when it is turned on hard, it may only get to +0.1V above ground. For this circuit to be design-able, you'd likely need a negative DC supply. \$\endgroup\$ – glen_geek Sep 1 at 19:03
  • 1
    \$\begingroup\$ The selection of an opamp it is not follow just the general equation as presented in text books. That’s why there are so many types! For your particular application, a close look at input offset and circuitry, rail supply topology and output circuit (to name some), will help you to find solution. \$\endgroup\$ – GR Tech Sep 1 at 20:06
  • \$\begingroup\$ @GRTech Thanks for the tips, do you have any specific recommendations on further readings (books, links..)? \$\endgroup\$ – Thauer Sep 1 at 20:22
  • \$\begingroup\$ Unfortunatelly I do not have a single source. Try to read any parameter of specs from companies data book for bjt, fet or bfet trechnology opamp, then search do demistify. A good start nutsvolts.com/magazine/article/… \$\endgroup\$ – GR Tech Sep 1 at 20:32
1
\$\begingroup\$

Use an opamp circuit, non-inverting config, with a gain of 30. I. E. R2/R1 = 29

Connect R1 to gnd through a large cap.

Make a resistive divider across your Vsupply to get midpoint of 1.5v

Connect your signal through sufficiently large cap to non-inverting pin.

Connect resistive divider centre point also to inverting terminal through large resistor (say 1Meg)

That's your circuit.

Choice of opamp and of caps depends on your pulse width and spacing etc etc. Also opamp power supply should be properly selected for opamp chosen.

\$\endgroup\$
1
\$\begingroup\$

Your circuit is Ok from the point of view of the input signal polarity since it adds an offset to the input in order to avoid its saturation when \$V_\mathrm{in}\$ goes below zero: note however, from the picture you show, it seems that you need a voltage gain of at least $$ A_v=\frac{V_\mathrm{out}}{V_\mathrm{in}}\approx\frac{1.5\mathrm{V}}{50\cdot 10^{-3}\mathrm{V}} $$ which is readily obtainable from the circuit you propose, provided you put some care in its design. Also you must consider carefully the time duration of input pulse, in order to design your amplifier with a bandwidth sufficiently larger in order to avoid excessive "linear distortion" in the output signal. I propose a design with an OpAmp where all the parameters above can be changed without too much calculations

schematic

simulate this circuit – Schematic created using CircuitLab

When \$V_{in}=0\$, this amplifier is characterized by the following offset voltages $$ V_+=V_-=\frac{V_{CC}}{2}=1.65\mathrm{V} $$ thus the given negative \$V_\mathrm{in}\$ cannot saturate its input. Every parameter, from the gain \$A_v\$ to the cutoff frequencies \$f_H\$ and \$f_L\$ can be easily changed by changed the circuit parameters according the known "calculation rules" for the OpAmp.
There is only a particularity: the bias network I've used to bias its \$V_+\$input is called noiseless because it avoids the injection of (too much of) the shot noise due to the current flowing in the \$R_b\$ resistors. It is the only part I advise you to use, even in your BJT amplifier, if you decide to go for it: \$50\mathrm{mV_{pk}}\$ may not be the lowest voltage, but it is a sufficiently low value in order to start to worry for the input signal to noise ratio.

\$\endgroup\$
  • 1
    \$\begingroup\$ Thanks for going into Detail! Do you have any recommendations on further readings in designing this kind of op amp circuits? None of the books i read so far presented such circuits. \$\endgroup\$ – Thauer Sep 1 at 20:25
  • 1
    \$\begingroup\$ Offset circuit are seldom described in textbooks: however you can find something in the application notes of Analog Devices, Texas Instruments and Maxim pertaining low voltage CMOS OpAmps (\$V_{CC+}-V_{CC-}\le 5\mathrm{V}\$). The noiseless bias network (and many more methods) is briefly described in the classical book by Motcenbacher & Fitchen "Low Noise Electronic Design" (or the later edition written only by the first author), which I advise you to have a look to, if you deal with low level signals. \$\endgroup\$ – Daniele Tampieri Sep 1 at 20:36
0
\$\begingroup\$

You have some kind of misconception about op-amps. You just have to find a correct op-amp configuration to do it. They can add output DC offset and made not to amplify input DC component.

\$\endgroup\$
0
\$\begingroup\$

Here are two solutions; first is discrete bipolar transistor circuit; 2nd is opamp circuit.

You could use a PNP emitter follower into a NPN common-emitter. Calibration and offset drift may be a bother. And the output will not give a solid 0.0 volts. But the opamp circuits cannot give a solid 0.0 volts output either.

schematic

simulate this circuit – Schematic created using CircuitLab

The RC filter isolates the opamp from the several volts spikes of the ADC as it takes samples.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.