Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Decoupling and projects

  1. Apr 18, 2014 #1
    Hello, I have two questions today

    Reading ahead of my class, I noticed that we were coming up on frequency response. My book kept discussing midrange gain, but it doesnt define it anywhere.

    I tried looking online and I couldn't find a good definition there either. The only thing I could find that was similar was about midrange frequencies (not midrange gain). From the picture in the book, it looks like midrange is defined as 0Hz to critical frequency which is about 3dB less than Aol. But that doesnt really help me with midrange gain. In my research I did noticed that midrange gain is the range of frequencies in which I CAN use the formulas I have learned (such as -Rf/Ri, or Rf/Ri + 1). I wasn't sure why they didnt work for frequencies passed this "critical frequency" which brings me to my first question.

    This got me thinking about how Op Amps can amplify DC (since this wasn't explicitly discussed in my text either). So doing research I found that Op Amps are decoupled.

    This confused me because I remember back from my freshmen electronics classes that decoupling meant to un-couple/bypass... to eliminate (though not completely) an AC signal while passes DC. Or I always remembered it as eliminate voltage spikes.

    So my question comes out of the confusion in that the decoupled Op Amp amplifies both DC AND AC signals. With that said, I know that Op Amps have a bandwidth after which they are not as efficient at amplifying (the gain changes).

    Is this because...
    Since the Op Amp is decoupled, it can amplify low frequencies (0Hz) and lower AC frequencies (relatively, based on the capacitor due to Xc). Once a certain frequency is reached, the decoupling capacitors begin to filter/begin working as a rectifier?

    I guess I am confused as to how a dc coupled capacitor can allow ac through (unless my explanation above is somewhat correct... that there is a range from 0Hz to some relatively low frequency that can pass while higher frequencies begin to get blocked more and more). I understand how rectifiers work, so this makes sense to me (I just cant find a website or text in my book to verify that my logic is ok)

    Also, I was confused why midrange is from 0Hz to the critical frequency when this range is clearly not the midrange of the range of frequencies that exist. Seemed a bit confusing. Seems like it should be called "low range"

    Finally, in lab, we are doing things like measuring slew rate or voltages at various pins. We built a diff Amp, saw how an inverting amp inverted a signal... etc

    Does anyone know of anything cool I can do with my op amp? Are there any cool labs/ projects I can build for my own knowledge/curiosity? I tried looking on youtube, but a lot of the labs there are the same thing... proving theory. "This is a slew rate now go look at what a slew rate looks like on a scope in real life". "This is what a buffer (VF) does. No go look at what a buffer does in real life."

    We keep putting the Op Amp in a circuit by itself and putting jumpers in at different points to see its affects. I actually want to build a circuit where my 741 does something (rather than constantly, week after week, prove quantities that are in the data sheet). I am very curious!!

    Thank you so much! As always!
     
  2. jcsd
  3. Apr 18, 2014 #2
    also, is decoupling and dc coupling the same thing?
     
  4. Apr 18, 2014 #3
    You may be overthinking some terms. In power regulation there are diodes that provide a path for transients that occur when the transistors switch on and off. I've seen them called "freewheeling diodes", "flyback diodes", "body diodes", and "transient diodes". The point is that there really isn't an official dictionary for EE's.

    I think you hit the nail on the head that "midrange gain" is where the op amp operates in a way that you can use simple op amp analysis. This is basically the region of operation where the two inputs are locked at the same voltage. This gives you linear operation in the form (Output) = (Gain)x(Input).

    Have you noticed that most op amp circuits have some route from its output back to its negative input? This is providing feedback. If the output voltage drifts a little bit and gets higher than it should be, the voltage at the negative input should go up a bit too. The op amp detects this on its negative input and drops the output. The opposite happens if the voltage slips and goes below what it should be. This provides negative feedback and it locks the output to whatever gain you designed in your op amp analysis.

    In fact, this the whole point of op amps. Early engineers could easily make amplifiers to send phone signals but the gain of their amps was unpredictable. Op amps were invented using tube transistors that could provide stable and predictable gain.

    Op amps have a frequency limit as you know. As the frequency increases the op amp cannot keep up with the changes in voltage. The output falls behind and you loose negative feedback. The op amp becomes unstable. The more gain you try to attain, the harder it is for the amp to keep up as the frequency increases. This is why you have a gain-bandwidth product. More gain equals less bandwidth and vice versa.

    In addition, the phase can shift too. At a certain frequency, the phase at the output will be 180 degrees different from the phase at the input. Now the negative feedback becomes positive feedback and there's no chance for stability.

    Circuits tend to have noise from harmonics and other sources. If that noise is at or near the frequency where the phase shifts 180 degrees the output will get blown up with amplified noise.

    Capacitors tend to act like high-pass filters. The cap looks somewhat like a short circuit (a wire) to any sufficiently high frequency signal. When you talk about decoupling caps there are two ways to use them. If you place a cap in parallel with a signal it will short out the AC components and leave the DC component. If you place a cap in series with a signal it will pass the AC components and block the DC component.

    I'm guessing by your questions that you haven't taken a microelectronics class yet. I'll wager that coming up soon your going to be looking at the inner workings of an op amp.
     
  5. Apr 18, 2014 #4
    When you say the "frequency cannot keep up" is this because of the slew rate? If the frequency is too high, the op amp literally cannot linearly "mimic" the signal fast enough? Or is this because of something else.

    Also, is there a difference between decoupling and DC coupling? or do they mean the same thing. I wasn't sure if my own definitions were correct:
    Decoupling = Blocks AC (eliminates spikes) and passes DC
    DC coupling = passes both AC and DC (as in an AC signal riding a DC signal)

    or do they both mean the same thing - in which case decoupling passes AC signals but only up to a certain frequency.


    And yes, I have not had microelectronics yet. It sounds like something that would benefit me greatly though. It is actually not in my roster at all. What it be called just that?
     
  6. Apr 18, 2014 #5

    analogdesign

    User Avatar
    Science Advisor

    When the frequency gets too high the opamp gain decreases because it cannot "mimic the signal fast enough" as you say. That's a decent way to put it. If you need gain at a higher frequency you either need a faster op amp or you need to not use an op amp.

    Your definitions for decoupling and coupling are decent. Coupling just means connecting one system to another. DC coupling means they are connected at DC (so their bias points must be compatible). AC coupling means they are coupled at some band of frequencies above DC (depends on the nature of the coupling network) but they are decoupled at DC so you can bias them independently.
     
  7. Apr 18, 2014 #6
    Thank you so much! Now I see the difference between decoupling and Direct coupling! phew!

    Does anyone know why the midrange gain frequencies are called "mid" range if they start at 0 and end around 10 Hz (for the 741)?

    And lastly, is there anything interesting I can do with my 741? For the past 6 weeks, I have just been proving specs in the data sheet for it (slew rates, gain, how to account for offset...etc) but I haven't gotten to USE my op amp in anything. I looked online, but only found a few, simple, audio amplifier circuits... which wasn't two interesting to me cause I was already building those with class AB power amps (though I can appreciate the greater simplicity of just using a single IC chip now)

    I was wondering if anyone knew of a cool project that I could use my op amp with.
     
  8. Apr 18, 2014 #7

    analogdesign

    User Avatar
    Science Advisor

    The 3dB frequency of a 741 is quite low so that is why "mid-band" is also low. If you use negative feedback to get a smaller closed loop gain the closed-loop 3dB frequency will increase a lot and be more useful.

    You can do all kinds of cool things with a 741. When I was in school we designed an analog "Dolby" system. Basically we put high-pass filters to do pre-emphasis on an analog music signal and recorded that to tape. We then made de-emphasis (low-pass) filters to read the music back off the tape and demonstrated noise reduction. It was pretty cool.

    Another project you can do is analog fiber-optic communications (did that in a lab in college as well). Basically you connect the 741 as an audio amp to modulate an LED. Then this goes down a fiber and on the other end you module a photo-diode. You connect another 741 configured as a transimpedance amplifier to read out the photodiode. In our case the fidelity wasn't amazing but it worked all the way down the hall and was a lot of fun.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Decoupling and projects
  1. Decoupling capacitors (Replies: 1)

  2. Decoupling capacitors (Replies: 2)

  3. Decoupling filters (Replies: 19)

Loading...