How often does the signal level exceed the bias voltage?

Discussion in 'Amp Tech Center' started by peteb, Mar 16, 2019 at 11:41 AM.

  1. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    And what does it sound like when it does?


    One of the fundamental laws of tube operation is that the signal exists within the boundary or limit set by the bias voltage.



    If the signal exceeds the bias, the grid goes positive, this is class AB2, A2, B2. There is a technical reason that limits Fender type tube amps to operate in class AB1, A1, B1.




    The goal here is to be able to gauge the correct signal level at the power tubes for any amp, based on the amp’s bias voltage.



    The second goal is to find out if Fenders go AB2, A2 or not.





    I did quite a bit of signal testing two years ago using a guitar to make the signal and quite a bit again last year using a signal generator.




    The tests with the guitars, mostly consisted of measuring the signal at different volumes with tone controls set in the middle. The results were that the signal rose linearly with the volume control setting, and that the max signal consistently ands up being a little bit more than half the bias.



    The most interesting results with the signal generator is that you can accurately measure both the input and the output anywhere and easily and accurately figure the gain. Another interesting result is that blackface tone controls affect the gain a lot and a simple tone control does not affect the gain at all. The other interesting result was that signal level anywhere in an amp strictly follows these two basic rules. If the path is not a path to ground, then there is no signal loss. If the path is a path to ground, then there is signal loss.






    Now it’s time to put the two tests together and see if a Fender amp will go class A2 or AB2.


    I think the answer is yes.





    Somewhere in the testing I have done, I think it happened and it wasn’t my focus at the time, so I disregarded it. But I think I can easily reproduce it and see if it is true.



    This is the test. One blackface champ. Measure the signal on the plate at extreme playing conditions.


    The bias is around 24 volts. I think the signal is going to go higher. Maybe 30. 30 but not a steady 30.
     
  2. D'tar

    D'tar Tele-Afflicted

    Posts:
    1,793
    Joined:
    Jan 11, 2013
    Location:
    WNY
    Extreme playing conditions.... I like that!

    Isn't this the point that overdrive distortion and clipping starts..
     
    Last edited: Mar 16, 2019 at 12:21 PM
    peteb likes this.
  3. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia


    Thanks D’tar




    I think overdrive starts earlier and the amp is really not meant to go to that point.
     
  4. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    https://www.aikenamps.com/index.php/the-last-word-on-class-a

    What about class A2, AB2, and B2?




    Thee numerical suffix appended to the class designation indicates whether or not grid current flows in any appreciable portion of the cycle. A "1" suffix indicates no grid current flows, while a "2" suffix indicates grid current flows for some part of the cycle. Class A2/AB2/or B2 requires a very low impedance, transformer-coupled or DC-coupled driver stage. The standard AC-coupled phase inverter or single-ended driver stages used in nearly all guitar amplifiers will not allow grid current flow, so they are class A1/AB1/B1 amplifiers.

    The advantage of class A2, AB2, or B2 is the complete lack of "blocking distortion", or transient intermodulation distortion. The disadvantage is the extra complexity of the output stage required to source current to drive the output tube grids into the positive region.
     
  5. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    This is the part of the Aiken page that links positive grid to class A2. The definition is grid current not positive grid, so it not entirely clear to me. This quote also alludes to the part about the fender amp that it is not equipped to drive the grid positive. But there has been a lot of discussion about gain lately and we know that there is a lot of gain available and one important job of the pre amp is to not provide too much signal in any place. And we know that there should be no practical reason that an amp could not be hopped up and make more volts on the grid.


    Aiken Amps:

    The disadvantage is the extra complexity of the output stage required to source current to drive the output tube grids into the positive region
     
    Last edited: Mar 16, 2019 at 3:41 PM
  6. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Test complete, I believe I made the grid go positive.


    The stock cathode bias is 25 VDC.


    As expected, the onset of overdrive was when the signal as a fraction of the bias, like 1/3 or 1/4



    The full potentiall of the amp is realized when the signal gets to half of the bias, but there is more after that, here’s the rough approximate numbers given as a high and a low


    Champ amp AA764 bias 25 VDC.

    Treble and bass on 10


    Guitar is a stock Tele on the bridge PU, max output signal considered to be about 100 mV.


    Volume setting / a large signal on the grid - a very large signal on the grid

    6 / 7 - 10V

    7 / 8 - 14V

    8 / 12 - 20V

    9 / 15 - 25V

    10 / 17 - 30V



    I figure the grid went positive in there. Then I thought hold on



    What is the cathode doing?


    I measured the DC bias on the cathode under heavy playing conditions. It was not affected very much.



    Then I thought I would show the grid showing positive by connecting the meter between the grid and the cathode. I saw -25 volts on the grid. Then I realized you can never show the grid going positive on a meter. I guess with two meters you could. Not on one meter because it involves reading and comparing AC and DC. I guess that’s one of the nice things about an oscilloscope, it shows the AC and DC together I think. Is that right?




    I conclude the grid was positive and the amp was operating class A2.



    What does it sound like?



    I would say there is no difference in sound as the amp transitions between class A1 and A2, other than just a little more of what was already happening.
     
    Last edited: Mar 16, 2019 at 5:15 PM
  7. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    And the last conclusion


    How much signal should one expect the power tubes to see on the typical amps?



    What I have found is that one would like to see the signal on the grid of the power tube to be at least half of the power tube bias and not more than the bias. Even though half of the bias is at the low end of the range, I say that an amp that the signall can only reach half of the bias, is adequate. The full potential of the amp is reached if the signal can be as high as half the bias.
     
  8. D'tar

    D'tar Tele-Afflicted

    Posts:
    1,793
    Joined:
    Jan 11, 2013
    Location:
    WNY
    Are you sure the grid went positive?

    My gibson (GA5)lives with the signal level greater than bias voltage with volume past 4. Darn near 2x. The grid begins to collect negative charge(-10v) to about half of idle bias voltage (22.5v.)

    Reaching halfway would surely keep things clean but thats where the fun stuff begins!
     
    Last edited: Mar 16, 2019 at 6:48 PM
    peteb and LudwigvonBirk like this.
  9. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Thank you D’tar for this info.

    I have posted on this before but no one has offered their own measurements. It is really useful to hear other people’s experiences.



    Didn’t it?


    Didn’t yourn?


    . “Darn near 2x”


    What power tube? One or a pair?

    What’s the bias?

    What’s the signal level?


    Would you say it is expected or an exception to the rule?

    A curve ball or a fastball?



    Much appreciated.
     
    Last edited: Mar 16, 2019 at 10:47 PM
  10. robrob

    robrob Poster Extraordinaire Ad Free Member

    Posts:
    7,260
    Joined:
    Dec 29, 2012
    Location:
    United States
    This is not a fundamental law. It is simply not true for cathode bias because the cathode resistor causes the bias to move with the grid signal. That's why we call them "fixed bias" and non-fixed "cathode bias".

    Remember the discussion of how a cathodyne phase inverter with a 1.5v bias can pass a 20+ volt AC signal? The cathodyne is an easy to understand extreme example of non-fixed bias.
     
    Bendyha likes this.
  11. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Thanks Rob, that’s a great point to discuss.




    You were the one that solved that mystery. At least around here. I am grateful to you because by solving that one, the world still has order. The cathodyne PI is a great example in that it looks like the signal exceeds the bias many times over, and it does, but to the tube itself, and that’s what we are taking about.


    The tube sees 1-2 volt difference in DC volts between the cathode and grid. This is the negative DC bias that keeps the current thru the tube from running out of control. Because it is a split load cathodyne phase inverter, the tube sees very little difference in AC volts between the grid and the cathode. This is the signal that gets amplified (small compared to the signal on the grid) that’s why there is no gain.






    I kept an eye on the signal voltage on the cathode of the power tube, as this also subtracts from the signal getting amplified. It seemed like most of the time it is small, like one volt so I disregarded it.
     
  12. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    D’tar’s results are a little different than mine. That’s why i would like to know the circuit and operating conditions.


    But I think the results are close. Did 4x get edited to 2x?


    My test showed that the signal could go higher than the bias, and nothing really happened as a result.

    But the vast majority of playing conditions occur when the signal is half or less of the bias.


    This is what I have found on the power tubes of different sized fender amps.



    2x6L6 amp: the bias is around -50 VDC and a good sized signal on the grid is 30+ VAC.

    2x6v6 amp: the bias is -35 to -40 VDC and a good sized signal on the grid is 20+ VAC

    1x6v6 amp: the bias is -25 VDC and a good sized sport signal on the grid is 12+ VAC.
     
  13. Sean Mac

    Sean Mac TDPRI Member Ad Free Member

    Posts:
    61
    Joined:
    Mar 2, 2016
    Location:
    Dublin. Ireland.
    I have always understood the difference between Class A and Class B to be related to the Plate current and whether it is driven into cutoff.

    The image below is three separate sections from The Radiotron Designers Handbook 4th edition.

    The middle section relating to cathode bias may not be totally on point but it seemed to me to be worth including in your discussion



    Amp Classes RDH4.jpeg

    I hope this is not an intrusion on your discussion...
     
    peteb likes this.
  14. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Thanks Sean,


    That’s helpful, I see class A versus class B described, and cathode bias but not anything about class A1 versus A2 or class AB1 versus AB2.


    1 and 2 are further subdivisions within A and AB. Notice they only say A1 and AB1 which is consistent with what Aiken says, that the amps we are used to operate class A1 and AB1 not A2 or AB2 and when some one says A or AB they really mean A1 and AB1.




    The definition for 1 versus 2 is grid current flows in 2 and not in 1.



    Aiken throws in that part about class 2 meaning the signal is more than the bias but there really isn’t any other documentation.
     
    Sean Mac likes this.
  15. Sean Mac

    Sean Mac TDPRI Member Ad Free Member

    Posts:
    61
    Joined:
    Mar 2, 2016
    Location:
    Dublin. Ireland.
    Hi Pete,

    I am not sure I have ever seen the A2 classification before. I suppose a transformer coupled single ended design might qualify but those kind of amps are outside my limited experience. Full frequency interstage transformers are expensive I imagine.

    I do remember reading about bifilar Macintosh designs somewhere....

    For cap coupled designs I have always seen it as a progression from A1 to AB1.

    The jump to B2 apparently brings a lot of higher harmonics into the distortion spectrum according to the book I quoted earlier.

    It is possible I'm completely confused about all this. I just remembered the bit about driving the plate into cutoff being the Class B definition.

    Edit to add

    A quick look at google confirms that Class A2 certainly exists and suggests either an interstage transformer or a relatively high current cathode follower is required to supply grid current.

    Very interesting...

    :)
     
    Last edited: Mar 17, 2019 at 6:57 PM
    peteb likes this.
  16. clintj

    clintj Friend of Leo's

    Posts:
    4,842
    Joined:
    Apr 4, 2015
    Location:
    Idaho
    The Class AB to B distinction is at idle. AB idles somewhere between cutoff and saturation, B idles right at cutoff. Under signal, AB passes between 180 and 360° of signal in each tube depending on input signal, while in B each will pass 180°.

    Peteb, you might like a copy of Richard Kuehnel's power amp book. I'm four chapters in currently and quite engrossed. I might just figure out why my oddball EL84 power amp in my Hammond sounds so good despite varying so far from the datasheet numbers by the time I'm done with reading.

    Sent from my Pixel 2 using Tapatalk
     
    peteb likes this.
  17. Sean Mac

    Sean Mac TDPRI Member Ad Free Member

    Posts:
    61
    Joined:
    Mar 2, 2016
    Location:
    Dublin. Ireland.
    Hi Clint,

    when I quoted the bits from RDH4 I was thinking about Pete's initial question. Signal voltage and bias voltage.

    The point of Class A becoming Class AB being defined by the plate current reaching cutoff for some part of it's cycle.

    I will have to learn to proof read my posts more carefully :rolleyes:

    Class A2 is new to me, so I learnt something....

    :)
     
    clintj and peteb like this.
  18. LudwigvonBirk

    LudwigvonBirk Tele-Holic

    Age:
    117
    Posts:
    687
    Joined:
    Aug 26, 2017
    Location:
    Madison
    Jfyi, Robrob, bless his kind knowledable soul, didn't "solve that mystery", he clarified a known and well-documented tube circuit paradigm. And it isn't controversial or something only secretly now understood only "around here".

    Anyway... the subject of the thread is interesting, let's attempt to keep the s/n content ratio on the high side (I'll try....)
     
    peteb likes this.
  19. peteb

    peteb Friend of Leo's

    Posts:
    3,750
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Keeping the signal below the bias and keeping the grid negative may be more about ‘design center’ than an absolute rule but we have yet to see evidence of the signal significantly going over the bias level.


    And Rob did make order out of disorder.



    Where else is that documented?
     
  20. Bendyha

    Bendyha Tele-Afflicted Silver Supporter

    Posts:
    1,780
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    :confused::(:cry:
    other emoji where censored out
     
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.