Announcement

Collapse
No announcement yet.

Marshall 2103 Cold Maximum Bias Mystery

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

  • Marshall 2103 Cold Maximum Bias Mystery

    It's a 1977 2103. Anode voltage measures 475. I'm out of adjustment on the trim pot at around about a maximum 16ma of anode current. The amp is a UK spec model, running EL34s. The previous owner had these changed to a new set of JJs before selling the amp to me. The grid leak resistors are 220k (and measure 222k), the resistor running into the bias trim pot is 56k (and measures 55k). The trim pot has a clean sweep from 0k up to 23 and a bit k. My meter and bias probe work fine (I checked them on another amp to be sure).

    The only slightly odd thing I've found is that the speaker impedance was set to 8k but the twin McKenzie speakers measure 4k. It could well have been set like this since it was built for all I know. I corrected it. I am now at the limit of my knowledge/diagnostic abilities. I don't really want to go piggybacking resistors as that would be addressing the symptoms rather than the cause. Can anybody advise me on what to do next?
    Last edited by greengriff; 06-05-2020, 01:18 PM. Reason: typo

  • #2
    What are the voltages at the screen grids pin 4? And control grids pin 5 (to avoid oscillation it may be bet to take those measurements with the ECC83 in the phase splitter socket removed)?
    And the bias supply output, at the node of the 220k bias supply output and grid leak resistors?
    Maybe its just a matched set of weak EL34s, any others to try in there?
    My band:- http://www.youtube.com/user/RedwingBand

    Comment


    • #3
      There is another resistor for the bias supply just off the rectifier (schematically speaking) that forms a voltage division with the bias trim pot. In the 6550 model it's 120k. In the EL34 model it's 220k. It's possible a mistake was made from the factory or that this amp WAS actually a USA destined model at some point. Check the value of that series resistor for the bias circuit voltage divider.
      "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

      Comment


      • #4
        I noticed looking at the schematics that the reissue 2203 has a 47k in place of the 56k that's in mine. Piggybacking in another resistor with croc clips to bring the value down to approx 47k does allow me to bias properly. The 56k appears right for the era. I wonder what the thinking behind that was? Trying to prolong valve life? So I guess I will have to piggyback as a permanent solution. I don't want to start messing with the PCB on such an original amp.

        Comment


        • #5
          Originally posted by pdf64 View Post
          What are the voltages at the screen grids pin 4? And control grids pin 5 (to avoid oscillation it may be bet to take those measurements with the ECC83 in the phase splitter socket removed)?
          And the bias supply output, at the node of the 220k bias supply output and grid leak resistors?
          Maybe its just a matched set of weak EL34s, any others to try in there?
          Oddly I didn't see your reply before posting my first reply. I don't know why. I've now packed all my kit away to go to work, so I'll have to attend to this later. I dug out a the schematic for the 70's 100w power sections (page 245 of the Doyle book if anyone is interested), and it seems 56k is correct for UK models of the time (it shows 47k as the value for US models). Changing the value to 47k-ish by temporarily piggybacking another resistor does allow me to bias properly, but that suggests that all late 70s UK 100 watters were biased stone cold by design. I find that hard to believe so maybe there is another issue. I don't have another full set of EL34s kicking around (only 2 in my 4010) but I'll run through your other suggestions ASAP. Cheers!

          Comment


          • #6
            Originally posted by Chuck H View Post
            There is another resistor for the bias supply just off the rectifier (schematically speaking) that forms a voltage division with the bias trim pot. In the 6550 model it's 120k. In the EL34 model it's 220k. It's possible a mistake was made from the factory or that this amp WAS actually a USA destined model at some point. Check the value of that series resistor for the bias circuit voltage divider.
            It's not leaping out at me. Do you mean the grid leak resistors? They are 220k on my amp.

            Comment


            • #7
              The 56k appears right for the era. I wonder what the thinking behind that was? Trying to prolong valve life?
              I guess 56k works/worked fine with tubes that actually conform to the EL34 datasheet.

              I would bias to around 34mA cathode current. With good tubes this should be achievable with a grid bias voltage of -38V +/-1o%.
              Last edited by Helmholtz; 06-05-2020, 03:46 PM.
              - Own Opinions Only -

              Comment


              • #8
                Originally posted by Helmholtz View Post
                I guess 56k works/worked fine with tubes that actually conform to the EL34 datasheet.

                I would bias to around 34mA cathode current.
                Yeah, maybe that is it: valves just ain't what they used to be! For an easy life I have a little gadget that goes in between the valve and the base, breaking the circuit so I can insert my meter and measure anode current. I find that an anode current of around 28-30ma as a starting point sounds good, but I've no objection to going higher or lower. Is there a formula I can use to work out what your suggested cathode current means in anode current terms?

                Comment


                • #9
                  Chuck may have been referring to the resistor between the bias winding and the bias rectifier diode, 10k on this North America 6550 version https://drtube.com/schematics/marshall/2203u.gif
                  The EL34 spec version may have had a higher value resistor there, eg 18k.
                  Check out the various JMP 100W models, all will probably have had the same power supply spec as their contemporarys https://drtube.com/library/schematic...ll-schemas#JMP
                  Last edited by pdf64; 06-05-2020, 03:54 PM.
                  My band:- http://www.youtube.com/user/RedwingBand

                  Comment


                  • #10
                    The JCM800 versions get their bias feed from the HT winding rather than a bias winding. So the resistor on the cathode end of the rectifier diode has a bigger resistor. 220K for EL34, 150K for 6550. I think this is what Chuck was referring to.
                    https://drtube.com/schematics/marshall/jcm800pw.gif

                    (I think we're all assuming there is a typo in the thread title and this is a 2203, not 2103).
                    "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

                    Comment


                    • #11
                      A lot of guessing and misunderstanding could be avoided if the OP finds and posts the appropiate schematic.

                      Still wondering about the actual bias voltage.
                      - Own Opinions Only -

                      Comment


                      • #12
                        Originally posted by pdf64 View Post
                        Chuck may have been referring to the resistor between the bias winding and the bias rectifier diode, 10k on this North America 6550 version https://drtube.com/schematics/marshall/2203u.gif
                        The EL34 spec version may have had a higher value resistor there, eg 18k.
                        Check out the various JMP 100W models, all will probably have had the same power supply spec as their contemporarys https://drtube.com/library/schematic...ll-schemas#JMP
                        That resistor is 27k on UK models.

                        Comment


                        • #13
                          I can't find the schematic online. I have a paper copy in the Doyle book, but it's low enough resolution that it wouldn't scan well. Everything I can find online shows a 47k in that position.

                          Comment


                          • #14
                            Originally posted by g1 View Post

                            (I think we're all assuming there is a typo in the thread title and this is a 2203, not 2103).
                            No, it's a 2103 (2x12 combo)

                            Comment


                            • #15
                              I can't find the schematic online.
                              So the circuit is different from this: https://drtube.com/schematics/marshall/78331-3.gif ?
                              - Own Opinions Only -

                              Comment

                              Working...
                              X