Question for Dave G. regarding 330K resistors...

Discussion in 'Fisher' started by gkargreen, Aug 2, 2017.

  1. larryderouin

    larryderouin Do I get Food, Med's, or more gear this Month? Subscriber

    Messages:
    19,509
    Location:
    Glen Burnie Md.
    The old standard was 220K with a .1uf. It still works and can be used with no problems. However he cautioned against going higher due to possible LF oscillations. The time constant is a little larger but you can't hear the difference unless you get LF Oscillations. Dave did some calculations a few years ago and found .72uf or .73uf (Those numbers stuck for some reason but I can't remember which one. However it's irrelevant to this purpose as no one makes a cap that size.) was the ideal. But due to lack of a cap that size he recommended a 0.68uf or .082uf as the next best value.(At the time both values were fairly new and not widely available and hard to find). As .068uf is closer than .082uf it's a better value to use, although the .082 is good too. No audible difference is discerned, but it will show up on a scope that is sensitive enough.

    I don't remember the exact calculation but the resistance multiplied by the cap value sounds like it might be the correct . Lets see if this works.
    330K x .047 = 15510 Original Factory Values.
    220K x .071= 15620 Ideal replacement but not feasible due to lack of cap value.
    220K x .068 = 14960 Closest replacement to the 330K/.047uf. Parts available.
    220K x .082 = 18040 Last two values are useable if .068uf aren't available.
    220K x .1uf = 22000
     
  2. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    Thanks, Dave. Is there a formula you use to calculate the values for the time constant, maybe like 1/RC or something?
     
  3. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    8,676
    Location:
    Ball Ground, GA
    Within the power amplifier, there are a number of LF time constants -- the OPT itself being one of them -- so it is something that must be determined after a basic amplifier design is established, and the stability testing portion of R&D begins. Basic formulas will determine the response of a given individual TC, but when multiple TCs exists, then it begins with measurements for gain and phase shift in the frequencies of interest and altering the available TCs to suit. It's just not as cut and dried as I know you'd like it be. That's why it's best to maintain the original constants if any values within them are altered.

    Dave
     
  4. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    Yes, that makes sense. The OPT would certainly have to be considered as there would be no sense in making the circuit have a much lower response that the transformer could not itself pass, and making other circuits also be beyond that capability as well. BTW Dave, I reworked my W4-AM amps using your suggestions, they are sounding real good in my system right now, thanks for all the good suggestions!
     
  5. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    Using the values that Larry provided = 330,000 x 0.047 x10*(-6) = 0.01551, reciprocal is 64.5 hz
    the Scott 330,000 x 0.1 x10*(-6) = 0.033, reciprocal is = 30.0 hz
    so the proper value for c (using a value of 220,000 for R) is 0.15 ufd, so I should up the coupling caps from 0.1 to 0.15 ufd to keep the same time constant. Well, this whole thread has been an enlightening and good teaching moment for me, and I think for others as well. Thanks to all for helping me puzzle this out!
     
  6. heyraz

    heyraz Super Member

    Messages:
    1,319
    Location:
    Long Branch, NJ
    If available capacitor value choices are limited, and the design Time Constant wants to be maintained...why can't the resistor value be altered? Why not parallel two resistors together to get the closest value to work with 0.068uf?
    Modern resistors are smaller too.
     
  7. gadget73

    gadget73 junk junkie Subscriber

    Messages:
    31,767
    Location:
    Southern NJ
    you can, but that gets back to considering the drive abilities of the phase inverter. If you make the resistor too low, the inverter might not be able to generate enough voltage to drive the output stage to full power.
     
  8. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    Thanks, gadget, this keeps getting better and better, I'm learning here!
     
  9. tube-a-lou

    tube-a-lou Super Member

    Messages:
    4,219
    This is what I finally did to my Fisher KX200, I was having a bit of a problem with bloated LF in my amp but when I change
    to the .068uf caps it went away. Great thread guy's Thank You.
     
  10. Patrice B

    Patrice B AK Subscriber Subscriber

    Messages:
    1,985
    Location:
    Montréal, Canada
    Sorry to chime in and hope my intervention is ok: I asked a similar question back in 2009 when restoring my LK-72. The answer was, at the time, that the change wasn't necessary because the resistance seen in the circuit wasn't really 330k.

    I think it was in part because of the interaction of the D.C. Balance circuit just ahead. Mr Gillespie, please correct me if I'm wrong.

    http://hhscott.com/pdf/340A.pdf
     
  11. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    Thanks for chiming in, Patrice! In looking at the circuit, I see what you mean, the DC balance control is in parallel with the grid bias resistors in what appears to be a resistor divider network, nice catch!
     
  12. Patrice B

    Patrice B AK Subscriber Subscriber

    Messages:
    1,985
    Location:
    Montréal, Canada
    Found the post by GordonW: http://audiokarma.org/forums/index....ring-after-a-while.221779/page-2#post-2648959

    "Actually, the LK72 does NOT need the 330K grid resistors changed.

    It's funny to think about at first... but through Thevenin's and Ohm's laws, you can determine that, at DC, the grid actually sees the PARALLEL combination of the voltage divider created by the 330K resistor to ground and the roughly 520K combination of half the bias pot (500K divided by two) and the 270K resistor from the bias supply, as its DC feed resistance. That combination (330K || 520K) gives roughly 202K ohms. Well within spec for the 7591.

    I've never seen a Scott with this setup have problems maintaining grid bias...

    Regards,
    Gordon."
     
    Last edited: Aug 7, 2017
  13. Patrice B

    Patrice B AK Subscriber Subscriber

    Messages:
    1,985
    Location:
    Montréal, Canada
    However, the 340A has 100k as D.C. Balance and not 500k like the LK-72...
     
  14. Patrice B

    Patrice B AK Subscriber Subscriber

    Messages:
    1,985
    Location:
    Montréal, Canada
    Ok, parallel resistance gives around 160k if I follow correctly what Gordon said?
     
  15. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    8,676
    Location:
    Ball Ground, GA
    Absolutely correct. The total DC grid resistance is about 163.5K in the 340A, so as previously stated, there is no reason to adjust any of the grid resistances in the Scott design.

    Dave
     
  16. gkargreen

    gkargreen Active Member

    Messages:
    326
    Location:
    Fairfax, VA
    See, Patrice, you moved the needle in the discussion and we learned even more, thanks for your contribution as well as Dave's confirmation, that makes good electrical sense!
     

Share This Page