Amplifier using LM324n

[Mafura]

Hi everyone. I have just started a project where I have to build an amplifier for a condenser microphone. I have to use an operational amplifier LM324n. I need to read in a signal from the microphone, amplify it and then send it to a daq card. I am stuck in the amplification stage. Can anyone help me with a basic circuit that I can start with. I have never worked with audio signals before. Where do I start?
Thank you

 

[betwixt]

The data sheet for the LM324 would be a good place. Usually they have lots of application notes, including their use as audio amplifiers which you can work from. The formulas for calculating the feedback resistors are also in the data sheet.

Brian

 

[Mafura]

Hi Brian, I have checked out some of the data sheets. What I need to know is why do I need a capacitor after the microphone when I am only monitoring signals. Because for this project I need to monitor exactly what is going on in a different system and I want that signal to be displayed on a computer. I am fairly confused about the issue of the capacitor, I mean what role does it play and what significance will it have on my system results?
Thank you

 

[betwixt]

The input capacitor allows the input source to (the microphone in your case) to have a different voltage across it than the LM324. Consider that microphones basically come in two flavors, dynamic types and capacitive types. Dynamic one use a magnet and coil to convert the sound waves to a voltage so there is a low resistance across the coil connections hat could upset DC conditions in the amplifier. Capacitive types have a very high impedance pick-up (tens or hundreds of Megohms) so they have an internal FET amplifier so they can present something more user-friendly to the outside world. The internal amplifier has to be powered and the output pin will have a voltage on it which would also upset the LM324. The capacitor blocks the DC while allowing the audio through in both cases.

The capacitor value has an effect on the low frequency response of the circuit so it's value should be chosen so the lowest frequency you are using still passes adequately. If you use a larger value than necessary, it will still work but the time taken to charge it via the feedback loop might become excessive. It's a compromise I'm afraid.

Brian.

 

[Mafura]

Thanks for the reply. I have been playing around with some circuits and the one that nearly managed to do what I wanted is the one I have attached. I want to amplify a voltage of 3mV to about 3V. When I check the result from the oscilloscope I notice that the two signals are not in phase with each other. I have not been able figure out why and I am not really sure if the circuit I have is a complete one either. Can you please advise....I want to simulate a circuit and then build it immediately but I want to be on the safe side and not blow up my components and I also want to make sure that what I have will give me the best results.
Thanks

 

[FvM]

The circuit can basically work, but some points should be considered.

LM324 has a GBW (gain-bandwidth product) of about 1 MHz, setting a gain of 1000 results in a signal bandwidth of 1 kHZ, I guess, that's not what you intend for an audio amplifier.

Furthermore, LM324 has a class B output stage, that causes really bad crossover distortions, unless biased for an unidirectional output current. You can overcome the problem by placing a pull-up resistor between OP output and VCC. Or use a higher GBW and class AB output standard OP. Achieving a gain of 1000 in a single stage may be still unsuitable, also it's possibly too much for the expectable input level.

 

[Mafura]

Hi, Now I'm totally clueless. Can you give me an example of what I am supposed to do becuase now I am stuck and also the signal gets cut off at the peak of the negative values. I have tried using a 2 staged amplifier but even that one does not work. In that one the negative values are the ones that get cutt off and the gain is around 500.

 

[betwixt]

Do you want to explain FvM or should I ?

Brian.

 

[betwixt]

FvM hasn't responded yet so I'll try to answer.

In amplifiers of that kind there are several factors that set the maximum gain it can achieve, normally, the feedback resistors you choose are used to set the gain below the maximum level. One of the factors is called GBW or "Gain BandWidth" and it basically tells us the highest frequency the amplifier can give full performance when it's gain is 1. Note that not all amplifiers will work with a gain of 1 in the circuit you used but bear with me...
For the LM324 the GBW is 1MHz. At a gain of 10, the bandwidth is reduced by a factor of 10, at a gain of 100, it is reduced by a factor of 100 and so on. As you have set the gain to 1,000 the bandwidth you can expect is 1MHz/1000 or 1KHz which is grossly inadequate for audio use.
You would get far better results by using two stages to reach a gain of 1000, perhaps two stages with a gain of 32 would be better and the bandwidth would be more than 30KHz.
The other problem you have is the LM324 does not have a particularly good output stage and you compound it's problems by using a low 5V supply. It tends to produce distortion around the point where one output transistor cuts off and the other starts to conduct, this causes what we call "cross over distortion". It may not be severe enough to cause you problems but be aware that it tends to happen if you notice distortion at low signal levels that seems to diminish as the volume increases. There are many better devices for audio use.

Brian.

 

[FvM]

I simply understand, that the signal is cut off because the gain is to high.

 

[Mafura]

Thanks for the responce. I will try using 2 stages but I have already tried using 3 stages each with a gain of 10 so that I have 10*10*10 for the gain. What I do not understand is that when I am calculating the bandwidth from GBW do I use only the first stage so I have 1MHz/10 or will I do that for all the different stages?

 

[Mafura]

Hey Brian, can you check if it is safe to build the circuit which I have uploaded. The signal is amplified to 3 in the begining but it decreases and reaches 2 and then it remains constant at that value. Will this not affect my system once I am doing the monitoring of the signal coming from the microphone once I have sent it to LabView? I want to build a circuit and then test it with an oscilloscope then if it working I want to do the monitoring using a LabView program. Please advise.

 

[FvM]

Why did you omit the bias voltage divider present in your original circuit. The circuit can't work without it.

In addition, 3 mVpk with 1000x will necessarily cause clipping of the output voltage. You have to reduce gain or input voltage to 50% or less.

 

[Mafura]

In addition, 3 mVpk with 1000x will necessarily cause clipping of the output voltage. You have to reduce gain or input voltage to 50% or less.

I will try to reduce the gain because I need to keep the input voltage as 3mV and then I will add that resistor which I omitted. I will test it now and see what happens, if it does not clip off the output then I will let you know but in the end I want to build it. Thanks

 

[betwixt]

It won't work I'm afraid. Aside from the problems mentioned earlier about the LM324 not being a good device for this application, you have the DC levels completely wrong.

This isn't a 'scientific' explanation but I hope you will understand: when you connect an amplifier with a feedback path between it's output and inverting (-) input, it adjusts it's output voltage to try to make the voltage between it's inputs (+ and -) zero. It is by controlling the amount of feedback that you force the amplifier to produce more voltage to compensate for the amount it is reduced in the feedback resistors.

In your design, the + input of the first stage is referenced to ground by R10 so it has zero volts on it, as a consequence the output pin will also go to zero volts (or as close as it can get) in an attempt to make the - input zero as well. With all the stages connected as you have done, all the outputs will be close to ground voltage.

You can try fixing it by adding another 10K resistor between the + input of the first stage and the supply voltage, the voltage at the input pin will then be half the supply voltage and all the amplifier outputs will then be at about 2.5V, giving their output voltages a chance to rise and fall around center point. You should be cautious about slight discrepancies in the voltages, the LM324 output an inputs are not guaranteed to track exactly, there is an additional factor caused by imbalance in their input stages that might add or subtract a little from the expected output voltage, bear in mind this difference is amplified in the following stages!

Another thing to watch out for is the low frequency response. You have calculated the gain by using the ratio of 100k/10K for the feedback resistors but ignored the effect of the 100nF capacitors in series with the 10K. The capacitor has to be there to satisfy the DC conditions but it makes the 10K appear higher in value as the frequency drops. The actual value isn't just 10K, it is 10K + 1/(2 * pi * F * C) where F is in Hz and C is in Farads. So for example at 10Hz, instead of 10K, it appears to be 10.157K and the gain is accordingly a little lower. At lower frequencies, the capacitor has more effect on the value (it has higher reactance Xc). You can overcome this by making the capacitor value higher but at the expense of it taking longer to "wake up" when the power is turned on because of the time it takes to charge up.

Brian.

 

[Mafura]

Hi, I have attached a modified circuit but I'm still not sure if it will work when i build it. the problem or something that i noticed is that when i measure the output from the first amplifier the voltage is amplified by 2000. it comes out as a DC voltage of 6..V. I don't know how that value is calculated or where it comes from. I also want if I can be able to use 9V as the supply voltage to the amplifier. Will it cause damage to it or will it work fine? Please find the attached documents and advice as soon as possible. I am running out of time because I need to build this by the end of tomorrow. If what I have will not work please someone/anyone just give me a proper circuit that I can work with.
thanks

 

[Mafura]

here is another one which i have tried out. it works wellwhen i am using 9V supply but it does not work so well when it is supplied by 5V and i think its because in the first stage i'm amplifying with a gain of 2000 but the problem is that i don't know how that happens. please find the attached document below and please give me some more advise on this audio subject as it is fairly new to me.

 

[betwixt]

I am suspicious of the simulation, to be honest, I never trust simulations!
The actual voltage gain, using your second schematic, should be (1 + 320/10) * (1 + 310/11) = 963.

The LM324 is good up to about 30V supply and it gives best performance when the voltage is more than 15V so 9V is still below optimum.

The best book I have on audio amplifiers is "Audio/Radio Handbook" by National Semiconductors but it was was published in 1980 so I doubt it is available now. It goes into depth on amplifier design and gives lots of practical examples and calculations for lots of different schematics. I'm sure more recent books are available but I will have to leave it to others to advise on them.

Brian.

 

[FvM]

I am suspicious of the simulation, to be honest, I never trust simulations!

It usually can keep up with hand calculations.

I don't understand about the "gain of 2000" point, I guess, you're confusing Vpk and Vpp. 3mVpk * 1000 gives 3Vpk respectively 6Vpp. I also don't understand about DC in this respect, because the circuit doesn't provide a DC gain above unity. However, the 2 and the 3 stage circuit can work both as well. The 3 stage amplifier has a gain of 1300 more exactly (11 per stage). 10k and 100 nF gives 1 ms high-pass constant, respectively 160 Hz low cut-off, which seems O.K. at first sight. I would like to remember what I said about LM324 cross-over distortions in a previous statement, I also made a suggestion to reduce it.

 

[Mafura]

I think I was confusing Vpk with Vpp and I am aware that the gain in the first and second stage is 11*11 that is why I changed the resistor in the third stage to 14 so that I have a gain of 11*11*8. Which is still fine for the application that I will be using the amp in. I will look into cross-over distortions as I am not really aware what it is and does. All in all this has been very helpful an even bigger challenge which I am now faced with is getting the labview program from reading in the rms voltage from the amp against the frequency to do the monitoring to work. I am now busy building the circuit and will test it when our labs are open just to make sure before I put in the mic and acquisition board. I will post the results of the test soon.

Thanks all