Announcement

Collapse
No announcement yet.

Little theory please

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Little theory please

    Hi.
    I did some experiments reducing the gain of a input stage by a factor of 3 to acomodate into a preamp circuit. I did it by trial and error but at a point I.ll be happy to understand how to calculate the gain and output impedance of the circuit from the pic. The tube was a 12ax7 and amplification factor was 10 in example above. Can someone help me to understand how I can calculate the gain factor and output impedance of the circuit, please ? Thanks

    Click image for larger version

Name:	20190802_002039.jpg
Views:	1
Size:	2.30 MB
ID:	874800
    "If it measures good and sounds bad, it is bad. If it measures bad and sounds good, you are measuring the wrong things."

  • #2
    Here are the formulae:
    Click image for larger version

Name:	triode.png
Views:	4
Size:	1.06 MB
ID:	854417

    The values for µ, gm and rp are found in the datasheet. µ is fairly constant but gm and rp depend on plate current. RL is plate resistor in parallel with the load impedance (1M in your example).

    The Barkhausen equation for triodes states: µ = gm*rp.
    Last edited by Helmholtz; 08-01-2019, 10:56 PM.
    - Own Opinions Only -

    Comment


    • #3
      Thank You
      "If it measures good and sounds bad, it is bad. If it measures bad and sounds good, you are measuring the wrong things."

      Comment


      • #4
        I trying to understand the circuit amplification in terms of current. 0.4 mA looks pretty low even for 12ax7 but 190k impedance make it still usable in respect with 1M load. But still one more question please : except the voltage gain differences, the tube from circuit will perform as same (meant same gm ,same rp)if we get rid of 10k resistor and tied 2.2k resistor directly to the ground and using a bigger plate resistor, say 500k around, to ensure the same current over the tube, please ?
        Last edited by catalin gramada; 08-02-2019, 09:59 AM.
        "If it measures good and sounds bad, it is bad. If it measures bad and sounds good, you are measuring the wrong things."

        Comment


        • #5
          Originally posted by catalin gramada View Post
          I trying to understand the circuit amplification in terms of current. 0.4 mA looks pretty low even for 12ax7 but 190k impedance make it still usable in respect with 1M load. But still one more question please : except the voltage gain differences, the tube from circuit will perform as same (meant same gm ,same rp)if we get rid of 10k resistor and tied 2.2k resistor directly to the ground and using a bigger plate resistor, say 500k around, to ensure the same current over the tube, please ?
          Not sure if I understand. (We may have a double language barrier )

          From the voltages in your picture I calculate a plate current of around 0.66mA (voltage drop across plate resistor divided by 220k). The tube parameters gm and rp depend on plate current and to a lesser extend on plate voltage. If plate current and voltage stay the same, the outer resistances don't matter.

          Rk in the formulae above means the total cathode or tail resistance (12.2k in your circuit). If you reduce Rk, gain increases. No reason to increase the plate resistor. Plate current mainly depends on grid-cathode bias voltage and somewhat on plate voltage.

          So dropping the 10k tail resistor will mainly increase gain but will also reduce grid input impedance to 1M as you lose the bootstrapping.

          Did you find the µ/gm/rp graphs in the datasheet?

          Could you verify your measured gain with the formula?
          Last edited by Helmholtz; 08-02-2019, 03:52 PM.
          - Own Opinions Only -

          Comment


          • #6
            For your convenience the parameter charts over plate current for 2 different plate voltages:

            ECC83par.pdf

            Legend:
            Ua = Vp
            Ia = Ip
            s = gm
            ra = rp
            - Own Opinions Only -

            Comment


            • #7
              Originally posted by catalin gramada View Post
              if we get rid of 10k resistor and tied 2.2k resistor directly to the ground and using a bigger plate resistor, say 500k around, to ensure the same current over the tube, please ?
              I'm not sure of the question either
              If the plate resistor is changed to 500k the bias will be on the 'hot' side with a 2k2 cathode resistor. If you want it centre bias the cathode resistor should be increased to something like 4k7.

              Comment


              • #8
                BTW, 6.6mA plate current is huge for a 12AX7, max. value is 8mA, typical operation is around 1mA.
                The circuit shown exceeds the max. plate dissipation of 1W by 52%.

                Edit: Please ignore post. I used wrong plate current 6.6mA instead of 0.66mA (see post#5).
                Last edited by Helmholtz; 08-02-2019, 08:21 PM.
                - Own Opinions Only -

                Comment


                • #9
                  Originally posted by Helmholtz View Post
                  BTW, 6.6mA plate current is huge for a 12AX7, max. value is 8mA, typical operation is around 1mA.
                  The circuit shown exceeds the max. plate dissipation of 1W by 52%.
                  6.6mA? The circuit in the OP is 0.66mA, 0.15W

                  Comment


                  • #10
                    Originally posted by Dave H View Post
                    6.6mA? The circuit in the OP is 0.66mA, 0.15W
                    You're right, sorry and thanks. Mixed up my numbers.
                    Actually had calculated 0.66ma myself (post #5).
                    Too much jumping between different tasks.
                    Last edited by Helmholtz; 08-02-2019, 07:29 PM.
                    - Own Opinions Only -

                    Comment


                    • #11
                      No, I did it wrong, so I have to apologise, sorry guys, measured wrong things. I did it again and post the values in the attach. There are 0.78 mA, biased at 1.7v. Voltage gain almost 12.

                      Click image for larger version

Name:	20190802_224637.jpg
Views:	1
Size:	2.30 MB
ID:	854425
                      "If it measures good and sounds bad, it is bad. If it measures bad and sounds good, you are measuring the wrong things."

                      Comment


                      • #12
                        Fine. All questions answered?
                        - Own Opinions Only -

                        Comment


                        • #13
                          Originally posted by Helmholtz View Post
                          Fine. All questions answered?
                          Before this thread is ghosted, I'm wondering If Catalin's schematic looks like a very 'lop-sided' concertina phase splitter to anyone else but me. And if this design is significantly different from an engineering standpoint (noise, stability, parts count, etc.) than stacked plate resistors with a tap in between.
                          If it still won't get loud enough, it's probably broken. - Steve Conner
                          If the thing works, stop fixing it. - Enzo
                          We need more chaos in music, in art... I'm here to make it. - Justin Thomas
                          MANY things in human experience can be easily differentiated, yet *impossible* to express as a measurement. - Juan Fahey

                          Comment


                          • #14
                            With stacked plate resistors the gain of the tube unchanged compared to a single resistor of the same value so it will clip at a lower input voltage than the lop-sided concertina (I think)

                            Comment


                            • #15
                              Originally posted by eschertron View Post
                              Before this thread is ghosted, I'm wondering If Catalin's schematic looks like a very 'lop-sided' concertina phase splitter to anyone else but me. And if this design is significantly different from an engineering standpoint (noise, stability, parts count, etc.) than stacked plate resistors with a tap in between.
                              Yes, except the fact the outputs are not equally loaded so the gain difference will be greater than 1 as cathodyne have. An amplification factor of 10 around for input stage fit better in my project. Due to bigger cathode resistor the output impedance will be greater (almost same as Ra) but still usable in respect with 1M load

                              Click image for larger version

Name:	20190803_104456.jpg
Views:	1
Size:	2.02 MB
ID:	854431

                              My competences didn.t put me in situation to can do a complete stage analysis. But is something it works putting disparate infos head by head, as variable input gain stage in my project. I don't see any drawback reasons why not to use it.
                              Last edited by catalin gramada; 08-03-2019, 09:25 AM.
                              "If it measures good and sounds bad, it is bad. If it measures bad and sounds good, you are measuring the wrong things."

                              Comment

                              Working...
                              X