RC LPF question

Discussion in 'General Electronics Chat' started by samy555, Aug 20, 2015.

  1. samy555

    Thread Starter Active Member

    May 24, 2010
    116
    3
  2. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    The two circuits you indicate are power supply decoupling circuits intended to lower the power supply impedance for proper circuit oeration and prevent unintended coupling between stages through the power connections. Specifically, they prevent motor noise from getting into the receiver rf stage.

    ak
     
  3. samy555

    Thread Starter Active Member

    May 24, 2010
    116
    3
    Can you link the frequency 5Hz with anything as well as for frequency 3.3KHz.
    What is the thing the circuit designer fear of arrival to battery?
     
  4. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    Notice the direction of the two filters. They are filtering things *from* the battery, not to the battery. These filters are not designed for a specific cutoff frequency.

    ak
     
  5. samy555

    Thread Starter Active Member

    May 24, 2010
    116
    3
    No
    I can not imagine that the circuit designer chose those values randomly
    Sure he is intended that particular values!!!
    thank you AnalogKid
     
  6. #12

    Expert

    Nov 30, 2010
    16,271
    6,788
    Here are a couple of ideas:

    1) I can't imagine that the circuit designer of my car randomly chose a DC (zero Hz) battery when we all know the engine never operates at zero RPM. Even the starter motor does not operate at zero RPM!!! The light bulbs can operate at zero Hz, but they would work just as well at 60 Hz or 400 Hz. There must be a reason.

    2) Attenuation is always a matter of proportions. When you have 2 transistors about a centimeter away from each other, operating in the MHz range, by how many db do you want their load current isolated from each other? If you want to find the limit, remove the offending components and replace them with values that are less effective until he circuit doesn't work correctly.
     
  7. Robin Mitchell

    Well-Known Member

    Oct 25, 2009
    732
    199
    When decoupling power supplies it's good practice to use both an electrolytic and ceramic capacitor. For example, a 100uF electrolytic is very good at removing large ripples (low frequencies: such as those found in an AC supply). The ceramic capacitor (22nF) however has a good response to sudden changes in voltage (higher frequencies).

    These figures are just ball park number but personally if I am decoupling a power supply I use a 100uF electrolytic and a 100nF ceramic. Then you can use lower values such as 22nF or even 10nF for places near sources of electromagnetic interference (like the antenna of an FM transmitter).
     
  8. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,782
    943
    I see the 330 ohm resistor as more of a current limit element for initial power on with an uncharged 100 uf cap. A battery powered receiver does not need a low pass filter even though that combo of resistor and capacitor can be used for that purpose it is not functioning that way in this circuit because there is no low pass AC information being generated or used here.
     
  9. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    No, he didn't, and that is not going to change no matter how many times you ask the same question.

    I did not say the values were chosen randomly, I said "These filters are not designed for a specific cutoff frequency." The component values were chosen based on the circuit operating current, traditional and available values for capacitors in this type of application, physical size, existing inventory, cost, experience, and many other factors that have nothing to do with a specific cutoff frequency. The filters are in the power parts of the circuit, not the signal parts. They are power filters, not signal filters, and in this circuit there are no rigid design parameters for power filters. They are there to remove noise produced by the motor and to correct the output impedance of the battery at very high frequencies. Just because it looks like an audio filter, that does not mean that it actually is one.

    And as pointed out in post #7, one of the filters could be improved with an additional capacitor. Again, this has nothing to do with any specific cutoff frequency. In fact, the traditional capacitor he mentions (100 nF) is only 0.1% of the capacitor already there (100 uF). But it has a completely different performance at high frequencies, where the 100 uF capacitor is basically useless. In this case, it is the construction of the capacitor that is important, not its value.

    ak
     
    Robin Mitchell likes this.
  10. Robin Mitchell

    Well-Known Member

    Oct 25, 2009
    732
    199
    That never crossed my mind and makes sense! A 100uF CERAMIC capacitor would be both a great low frequency de-coupler as well as good high frequency AC filtering characteristics!
     
  11. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    Not really. A 100 uF ceramic would be larger than the aluminum/ceramic combination, with much higher internal inductance and a much lower self-resonant frequency. In the MHz region it would not work as well as caps 1% the value. And cost a bundle.

    ak
     
    Robin Mitchell likes this.
  12. samy555

    Thread Starter Active Member

    May 24, 2010
    116
    3
    Thank you all for trying to help

    Unfortunately, not much convinced
    I understand completely the use of several capacitors
    I understand that it must be around 1n, 100n and 10u
    But no one talked about the values of resistors
    Why 330 ohms and 2200 ohms?
     
  13. #12

    Expert

    Nov 30, 2010
    16,271
    6,788
    Because the combination of the resistors and the capacitors create the time constants.
    What did you want to know about the resistors, beside the fact that they form time constants with the capacitors?

    How about this: Attenuation of the higher frequencies happens because the Xc forms a voltage divider with the resistance.
     
  14. samy555

    Thread Starter Active Member

    May 24, 2010
    116
    3
    I do not understand this way
    I understand that, for example, the presence of a HPF of cutoff freq = 300Hz before an audio preamplifier because the lower frequency of the human voice = 300 Hz.
    You explained to me the reason for using the fc1=5 Hertz (motor noise)
    But you could not provide a convincing answer regarding the frequency 3.3kHz.
    Talking about time constants not much different and not far from cutoff frequency of a LPF.
    Thank you very much
     
  15. Robin Mitchell

    Well-Known Member

    Oct 25, 2009
    732
    199
    Again, you just educated me some more :)!
     
  16. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    Not me.
     
  17. #12

    Expert

    Nov 30, 2010
    16,271
    6,788
    I'm surprised Mitch hadn't considered the problem from that point of view, but not surprised at AK. He was here only a few weeks when I noticed he could keep up with me on any subject (that had happened so far). We may yet find an alley I have been down and AK hasn't, but that's why the sum of several of us make a good website. ;)
     
  18. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,531
    1,248
    I am reporting post #17 to the moderators because #12 has been taken over by aliens.
     
    #12 likes this.
  19. Robin Mitchell

    Well-Known Member

    Oct 25, 2009
    732
    199
    Well I just did not consider the problems with a practical 100uF ceramic capacitor. Everyone at some point forgets to take something trivial into account. I have seen experts (like my old boss), who would make some of the silliest mistakes yet could tell you everything about the most complex stuff. Plus my digital ability dwarfs by analogue ability (just comes more naturally).

    Of course the strange thing is that I compared a practical electrolytic capacitor to a theoretical ceramic device...dooooiinng :S
     
  20. MrAl

    Well-Known Member

    Jun 17, 2014
    2,425
    490
    Hello there,

    After a quick glance at this circuit and after reading some of the posts that follow is it very clear that the designer with a blindfold on dipped his hand into a barrel of small caps of various values and pulled out a handful, then loaded them into a dart gun, then took aim at the circuit board, then fired repeatedly until all the caps were used up, then soldered them in place wherever they happened to have landed. You didnt see that? <chuckle> :)

    Ok, the 330 ohm is there to help isolate the positive power supply rail (battery) from the input stage. The 330 ohm limits the peak current draw from the battery so the current dips on the left side wont change the battery voltage too much even at somewhat high frequency, which might interfere with the circuit on the right side of that resistor. Curiously however, there is no bypass cap across the power supply rails which would have helped this situation too.
    Looking in the other direction, the 330 ohm also helps to keep variations due to the right side circuit from affecting the left side of the circuit which could interfere with reception. The 100uf cap helps this situation because it keeps the voltage across it more constant than at the right side of the 330. That's your typical LP filter, but the exact cutoff frequency wont be that important as long as it can filter out any lower frequency components that originate in the right side circuit. If that is audio, then the 330 ohm and 100uf should have a low enough cutoff to keep the lowest expected audio frequency from affecting the front end (left side circuit). You have to also realize that the designer could have made a mistake in the choice of filter components too, so some modification may improve performance. For this filter the lower the cutoff the better although we dont want to get into a cap that is too large either. The 330 ohm needs to supply the right amount of average current, so we should not change that unless the value is a mistake. It has to allow enough current for the front end voltage requirement yet still act as part of a filter. Assuming the cap value was a decent choice to begin with, then anything above 100uf would work too, and maybe even better.

    The front end resistor 2.2k and cap 22nf seem to act in a similar manner. the 2.2k appears to be there for biasing and proper operation while the 22nf is there for keeping the voltage supply to the collector smooth. Probably anything over 22nf would work too, but the 2.2k resistor value probably should not be changed as that is part of the DC bias scheme.

    So you see it is mostly about DC bias for the resistors, and once we have the right DC bias then we select the associated cap to provide enough filtering action without going too large which would just be a waste if a smaller value would work ok. We usually like to stay small on the resistor values, but the smaller the resistor value the bigger the cap has to be, so we try to pick a resistor value that is large enough to get some filtering action yet not too large such that we cant get enough DC current through to supply the next stage.
    If for example we changed the 330 ohm to 33 ohms, then we would have to change the cap from 100uf to 1000uf. We would then have more voltage for the front end, but if we dont need that then it is a big waste to have to use a 1000uf cap when a 100uf cap works fine with a 330 ohm resistor. If we did in fact need that extra voltage though then we would be forced to use the 33 ohm and 1000uf. If on the other hand we could get away with less voltage, then we might be able to go to a 3.3k and 10uf capacitor and that would make the cap cheaper and smaller.

    I believe you could gain some insight by studying a resistive voltage divider with a cap for filtering on the output. You will see that as you increase the resistor values the cap value can be smaller for the same degree of filtering. The downside is that the resistive loading of the next stage will pull the divider voltage down more with larger divider resistor values. So it's a matter of choosing the right resistor values first, then finding the cap value that will provide just enough filtering.
     
    Last edited: Aug 22, 2015
Loading...