No announcement yet.

Capacitor Time Constant

  • Filter
  • Time
  • Show
Clear All
new posts

  • Capacitor Time Constant

    It has been 10 years since i took a basic AC/DC Class at my local college.
    We did have to figure time constant answers on a few tests, but i am sure they were pretty basic questions.

    JUST Curious is all.
    I am j wondering now, all these years later, how it applies to guitar tube amps.?
    Is it something a designer would consider when Designing/Building the amp for production, i guess.?

    I am NOT an amp tech, but i HAVE worked on quite a few.
    Not sure i ever had to consider a Time Constant for you typical amp repair.
    So when does the Time Constant become am important consideration.?
    Thank You

    BTW........i did a basic search, and i mostly got info on what a Time Constant is and how to calculate it, but not why...generally becomes an important tool or consideration.
    Thanks Again

  • #2
    Hi! I know very little, but may help point you in the right direction. Time constant equations require a cap AND a resistor, thus RC time constants. In tube amps, the bias circuit is one consideration--if you have too much filtering, the bias voltage takes too long to rise to nominal, and too long to change values as you adjust the bias pot.
    I build and repair guitar amps


    • #3
      Xtian provided an excellent example. Read these MEF threads for more info...

      Optimizing RC time constants in the filter cap supply


      Sag effect, what time constant?


      • #4
        So a time constant is a property of an RC circuit. It's just a number based on the R and C values. That number is useful for describing some circuit behaviors - as the above post describes for power supplies. RC circuits also are useful for doing other things like high-pass filters or low-pass filters. When RC circuits are used for those purposes, the time constant is not that useful a number, instead it's more useful to compute a cut-off frequency, which is also a number based on the R and C values. So that's just an example of how context matter in terms of what measure or computation you want to do with R and C values. So, yes, R C values matter anywhere they're used in amp designs, but you don't always think of their values in terms of their time constant. Try reading a bit about High-Pass (HP) and Low-Pass (LP) filters. You'll see how those RC circuits are described with different calculations based on R and C values.
        Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
        Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.


        • #5
          It depends on what you are doing. In my career as a bench tech, I almost never need a time constant. If I am updating old caps, I replace them, but I don't need to calculate time constants for that. Same with a coupling cap or a cathode bypass cap. If I have an open or shorted 22uf cathode cap, I don't care what frequency it represents, All have to do is replace the bad part.

          If you intend to CHANGE the amp, then you might. Maybe it is too bottom heavy so you may calculate a frequency change and the cap value that represents that. That sort of thing. Changing tone circuits. Maybe you might get curious about the low end cutoff of a triode stage, and you can calculate that. But you don't need to do that for repair and maintenance, RC constants come up in design efforts..
          Education is what you're left with after you have forgotten what you have learned.


          • #6
            Not much to add. I'll back up Enzo's indication that the issue is more likely to come up in design or modification. That said, and as a designer, other than the bias circuit (which is most important as it's a function parameter the tubes need to operate safely) about the only other area where it comes up for me is the power tube grid circuit. Too long a time constant there can exacerbate grid loading. This is really only an issue for amps that will drive the power tubes into clipping. As guitar amps typically do. There are some mods for existing models that address this to improve the clipped tone and reduce grid blocking and crossover distortions that can be consequences of grid loading.
            "Take two placebos, works twice as well." Enzo

            "Now get off my lawn with your silicooties and boom-chucka speakers and computers masquerading as amplifiers" Justin Thomas

            "Being born on third base and thinking that you must have hit a triple is pure delusion!" Steve A


            • #7
              If I remember anythng of my early electronics courses, time constant of an RC network is the time it takes for a circuit to respond to a step change in input voltage. Not the 100% complete change in state, but 1/2 the square root of 2, 0.707*. And the "corner frequency" is simply the inverse of that. We'll call time constant Tc, and corner frequency Fc. So Tc = 1/Fc and Fc = 1/Tc. I'm sure the pro engineers will jump all over this one, so let loose guys! Is my memory faulty or what...

              For the example of initial charging of a power supply filter, the "time constant" is the time it takes to charge the filter cap to 0.707 of its eventual fully charged state. In play are the capacitance value, internal resistance of transformer winding, forward resistance of rectifier, any resistors that pass charging current, maybe a couple more factors.
              This isn't the future I signed up for.


              • #8
                Cutoff frequency (either highpass or lowpass) and time constant are actually the exact same thing, both involving chraging/discharging a capacitor through a resistor, to a certain voltage.

                Only that there´s 2 ways to say the same thing, we use one or the other depending on application.

                In Audio, we typically deal in Frequencies, so we say an amp or preamp with 100k input impedance fed through a .1uF cap will rolloff below 16 Hz (15.9 for nitpickers )

                In any timing circuit, which are the rule in digital stuff, we say its time constant is 0.01 seconds .... exact same thing.

                So much so that calculators give both results at the same time, go figure, you use the one which fits you best at the moment:
                Juan Manuel Fahey


                • #9
                  Thanks for the replies.
                  Not a super exciting topic i know.

                  I suppose it was sort of what i thought........not very relevant for the average repair, more a designers consideration.
                  Yeah i guess if it is important anywhere, the bias circuit would be it.

                  Thanks Again


                  • #10
                    As has been mentioned before, time constants determine things like filter frequency response, charge/discharge (resp. rise/fall) times, phase shift, PS smoothing properties and more.
                    Exact formulas for each effect can be found in literature if needed.

                    I think the most important thing to remember about RC time constants is that relevant circuit resistance is just as influential as the capacitance value.

                    - Own Opinions Only -


                    • #11
                      Another more obvious example of time constants is found in muting and ducking circuits. For a 'power-on' mute circuit, the delay of the signal is usually determined by an RC circuit.
                      "Everything is better with a tube. I have a customer with an all-tube pacemaker. His heartbeat is steady, reassuring and dependable, not like a modern heartbeat. And if it goes wrong he can fix it himself. You can't do that with SMD." - Mick Bailey


                      • #12
                        I've been thinking about this stuff lately with regards to supply filters . Is ripple rejection simply a matter of time constant and can keep the voltage up by using a smaller resister and larger capacitor or do you need a minimum series resister to isolate the different nodes ?


                        • #13
                          Originally posted by 35L6 View Post
                          Is ripple rejection simply a matter of time constant and can keep the voltage up by using a smaller resister and larger capacitor or do you need a minimum series resister to isolate the different nodes ?
                          Ripple rejection depends on the time constant(s) but in a power supply you don't have an isolated RC filter.
                          Non-zero source impedance and load current affect time constants and filtering - as well as other circuit properties.

                          E.g. if you replace the series resistor with a lower value and increase capacitance by the same factor, you get a lower impedance circuit.
                          That results in a stiffer power supply, likely to change amp dynamics.

                          Also the smaller series resistor will drop less DCV.

                          As power supplies are complex, it's best to use something like PSUD2 to see how different component values change the performance.
                          But that will only show steady state results - not dynamics.

                          - Own Opinions Only -


                          • #14
                            Originally posted by Leo_Gnardo View Post
                            the "time constant" is the time it takes to charge the filter cap to 0.707 of its eventual fully charged state.
                            I remember one time constant being 63% and the 10% to 90% rise time being 2.2 time constants.


                            • #15
                              I have a line voltage radio conversion . With tube rectifier HT1 is about 135 volts so I don't have much voltage to give away . Single ended hums more too . I hadn't even considered dynamics . As usual each answer raises new questions . I guess I need to learn simulation .