1. Win a Broadcaster or one of 3 Teles! The annual Supporting Member Giveaway is on. To enter Click Here. To see all the prizes and full details Click Here. To view the thread about the giveaway Click Here.

Trying to make a more accurate bias chart

Discussion in 'Amp Tech Center' started by peteb, Jul 17, 2017.

  1. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Thank you old man tele, I read thru it and will look at it closer.




    I don't think I'm trying to prove anything, I'm just trying to relate as closely as I can bias percentage to tube operating condition.


    And I don't think it is a stretch to generalize the low power Amps are going to have a lower bias. I am also looking to understand how a cold biased amp can sound hot.


    I need to check your last post, thanks but gotta go
     
  2. Old Tele man

    Old Tele man Friend of Leo's

    Posts:
    3,941
    Joined:
    May 10, 2017
    Location:
    Once banned always banned
    Sorry: BIAS determines/sets the tubes "operating Class (A,AB,B,C, etc.)." Can it (bias) be adjusted up/down? Yes, but then you're ALSO changing its Class-operation.

    It's (1) the AC-drive signal riding "on" that DC-bias voltage and (2) the tubes' effective transconductance (gm') value and (3) the plate load resistance/impedance that determine (4) the OUTPUT signal/power. And it's ALL limited by the B+ available and the tube's diode-line.
     
    Last edited: Jul 19, 2017
    peteb likes this.
  3. Bendyha

    Bendyha Friend of Leo's Silver Supporter

    Posts:
    2,553
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    Here is something to try to apply your theories too.
    Take the 6V6 GE data for a single tube.
    Being one tube, it will have to deal with all the amplification itself, no matter how you bias it, it will be Class A, maybe not always good Class A, but it cannot be anything else.
    Here is the standard characteristic at typical operation.
    upload_2017-7-19_23-1-33.png

    Lets look at this setup closer. 250V on the plate & screen, a 5K load from the transformer primary, and 45mA of bias at idle.
    Okay....lets take a chart for those conditions, first of all the 250V screen curves, I will add the 100% bias at 14W in red, and the 50% 7W dissipation in green.
    Then draw in a loadline for 5K at 250V plate. this is with a 0% bias.
    upload_2017-7-19_23-8-37.png
    Okay, this has 0% bias, it is at cut-off at idle, but the signal can still swing down to about 20V if you put a signal in to shift the -35V bias to about the -5 volt line, so an symetrical imput of 60V p-p. half the signal will be clipped off and the other half cleanish Class A. If you imput more signal, the swing can't go further...this is saturation.

    Next, we will bias the tube by sliding this 5k line upwards along the 250 plate line untill we reach the suggested "no signal plate current" bias of 45mA

    upload_2017-7-19_23-18-59.png

    Now the can swing down 225V before reaching the grid limiting diode line right below the knee (PERFECT TEXT BOOK POINT), and can swing up 225V to 475V before cut-off. so it is perfectly in the middle with about 80% of maximum alowable heat tolarating bias. but 100% perfect!
    Here we have the control grid sitting at -12.5V negative to the cathode - this is the bias voltage. So for maximum clean output we can swing down to -0V so an imput much smaller than the example above, a 25V p-p is the limit before saturation starts to set in, although swinging downwards to the cut-off point, this is at -35, 22.5V away, so we need a 45V p-p imput before that sets in, by which point the asymetric distortion is setting in nicely. So are we central after all? Of the maximum voltage swing, yes.

    You might argue that this is not push-pull conditions, and everything is different. But untill you can grasp Single End Class A, there is no way you can cope with Push-pull Class AB1.

    Next I will add a 50% bias, and a 100% bias as allways with the bias point as a pink star on the 250V plate line. These are for you to try to draw conclusions about, and explain your theories with.

    upload_2017-7-19_23-44-52.png

    And the 100%

    upload_2017-7-19_23-45-47.png
     
    peteb and RLee77 like this.
  4. Bendyha

    Bendyha Friend of Leo's Silver Supporter

    Posts:
    2,553
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    One more. you have mentioned biasing to saturation a few times, this is where the loadline would be, at about 200% dissipation!


    upload_2017-7-20_0-15-9.png
     
    peteb and RLee77 like this.
  5. Old Tele man

    Old Tele man Friend of Leo's

    Posts:
    3,941
    Joined:
    May 10, 2017
    Location:
    Once banned always banned
    Bendyha, excellent "show & tell" illustrations.
     
    Last edited: Jul 19, 2017
    peteb likes this.
  6. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    I read thru bendyhas posts, but need to go back thru them. Thank you


    I read old tele mans paper and found it very interesting. Thank you


    It says


    Class of operation, measured by angle of conductance, is dependent on the ratio of idle current divided by the plate maximum power peak current.


    I don't know where the plate maximum power peak current comes from. I don't know if they measured it or if it is a tube spec. I don't see that spec listed. They use 297 and 315 mA for two different amps.



    I find it interesting that class of operation depends on plate current but not the idle power dissipation.
     
  7. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    More on the paper provided by old tele man

    "Plate Conduction Angle...by Earles L. McCaul • Relationship to Plate Dissipation"


    It says RCA specifies, and the their two sample amps follow the spec including one amp with RCA data, they expect the plate conduction angle to be between 191 and 203 degrees, not very far from class B (180 degrees) and far from class A (360 degrees)

    Also, from their two Amps, these are both 2x6L6 Amps, different power, different bias points, the bias points can be compared to the plate conductance angle!

    A sample of two:


    Table 1 – Example RCA 55W amplifier vs. typical Fender 45W amplifier.


    RCA amp: plate conductance angle = 203, bias = 87%

    Typical amp: plate conduction angle = 193, bias = 55%


    This shows the larger bias with a larger plate conduction angle, (and the higher power amp with a higher bias)






    What I get out of this is that in all cases the plate conduction angle is small and much closer to class B than class A and a rebias is not going to change the class of operation, nor will a rebias change the sound just because of the change in plate conduction angle.



    So then it is back to how do the bias points relate to tube saturation, to affect the sound. It looks difficult to relate bias points directly to tube saturation.


    But this:


    Plate conduction angle (and operating class)goes up with Increased current.


    The risk of redplate goes up with increased idle dissipation power,and can be accurately judged using the 50, 60, 70 rule
    [​IMG]

    Saturation will be increased with increased idle dissipation power but it may not make very much difference. Is saturation more dependent on the AC signal superimposed on the idle current? Is it dependent on the size of the signal to reach saturation? Is the size of the signal the voltage of the signal?



    Do we have the trifecta of electricity?


    Plate dissipation angle and class of operation depend on plate idle CURRENT

    Redplating depends on plate dissipation POWER

    And saturation depends on plate signal VOLTAGE




    Follow up reasoning for why the plate voltage (AC) is what causes saturation:


    You can set the idle to class A, AB, or B.


    You can set the bias to make the tube redplate.


    But no matter where you set the bias, you can't make the amp saturate without the Ac signal.





    Now I'm ready to add and subtract AC from DC, which is interesting, and even more inside the tube, I saw some of that in bendyha' write up.
     
    Last edited: Jul 20, 2017
  8. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia

    OK, I'll see what I can do with this.




    This is the stuff I'm looking for. I'm not following what sets the voltage level at which "the swing can't go further".

    Also, this one is class A? With half the signal getting cut off?




    Ok on to my theory,




    My theory is that under powered amps can get similar tube operating conditions as higher powered amp that use the same tubes, by using a (proportionally) lower bias dissipation power.

    If I apply that theory to the graph, I see a 14 W class A amp at 100% bias, and I see a 7 W class A amp with a 50 % bias, both amps use the same tube type.

    My theory holds up if the 14 watt amp with the 100 % bias has similar tube operating conditions as the 7 watt amp with the 50 % bias.

    Does the graph show that the higher watt, higher bias amp is closer to saturation? I would think it would.




    I amp not sure if this is a good test for my theory. Is this a higher and a lower power amp, or is this the same amp at two different bias points?




    Where I am with my theory after reading all those graphs is:



    I see the point that a 6v6 is a 6v6 and it doesn't matter if it is in a 10 or 20 watt amp, what goes on inside the tube is going to be similar in both. If you want the amp to saturate you gotta set it up so it can saturate. I think a lower power amp can get away with less bias, it's going to lead to less power just like in your example above. But either amp with low or high bias won't saturate unless the AC signal takes it there.



    It gets back to the bias guideline as being only an upper limit.



    Thanks all
     
  9. Bendyha

    Bendyha Friend of Leo's Silver Supporter

    Posts:
    2,553
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    The premise of the paper is to prove that you can't draw relationship conclusions from features that are not related, and that "conduction angle is solely determined by the plate DC-to-PK current ratio" the two things that determin current ratio...nothing more.

    When amplifying symetrical sinewaves within the limits of no clipping or saturation,
    do you prefer the sound of 237° to 194° ?

    As far as improving the sound of your amplifier goes, this other paper from the same author is maybe just as helpfull -ET&MPH-Equations. and stanglely reaches a similar conclusion.
     
    Last edited: Jul 20, 2017
    peteb likes this.
  10. robrob

    robrob Poster Extraordinaire Ad Free Member

    Posts:
    8,138
    Joined:
    Dec 29, 2012
    Location:
    United States
    peteb and Bendyha like this.
  11. Bendyha

    Bendyha Friend of Leo's Silver Supporter

    Posts:
    2,553
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    Yes, but the label doesn't define the character or behavior, the classes are all overlapping with malleable limits, susceptible to change; through input voltage, the current supply and demands, and the handling capabilities of surrounding components.

    One can set the idle current low or high, or cool or hot, or the Q point bias voltage high or low, or bias to 30% or 100% of maximum dissipation.......all different ways of saying the same thing.

    But say 72%, and everyone knows exactly what you mean.

    Say middle high AB1, and we have at the most a vaguely definable range.

    An amp designed and built with the plate and screen voltages set to a fixed level, the speaker reflecting through the out-put transformer a predetermined impedance and the right tubes stuck in the sockets, can have its bias raised or lowered.
    But if it was designed to be a Class AB1 push-pull amp, you can lower that bias right down to near cut-off, or raise it up to a state where your plates start to glow with current excess, and the whole time you won't be able to leave the realms of Class AB1.
    If the control grid was not designed to deliver current, it won't, it will block, and Class B2 can not be reached, bias to cut-off and the cross-over distortion will soon make you move back up to AB1
    Class A.... well when playing at lower volumes, you will be in Class A anyway, it is part of AB after all.....bias hotter and you might have a tiny bit more, but not much, bias to maximum dissipation, you still won't have that much more, when you play louder you will again be in the B part of the swing. If you changed the out-put transformer to an impedance about 2/3 higher and increaced its physical weight by about the same amount, then you might well get a true Class A amplifier if you biased up to the right level (that might only be 80% max like in the post #43 on the second graph.
     
    Last edited: Jul 20, 2017
    peteb likes this.
  12. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Bendyha and Rob, I really appreciate the help. Thank you


    I hope you guys both understand I'm just messing around, doing some research, learning about tubes. It keeps me off the streets.



    And I'm not trying to make up any new theories, just trying to connect the dots, and fill in some gaps.

    I will go over your posts in detail but I'm about to perform a test. I need to go into these situations with a clean slate and build my own knowledge, then I will look at your posts.


    But first a sanity check:



    A 6v6 has max plate dissipation Of 14 W per tube.


    What is the max power out of the amp per 6v6 tube?


    Isn't it also 14 w per tube? Or is it 11 like a 22 w deluxe reverb?



    It has been pointed out to me that plate dissipation is not the same thing as power out of the amp, and I can see that they are not, but why are the numbers so close?



    I thought of an example to show that plate dissipation is not the power out and the numbers are not close.


    A champ, class A, will dissipate 12-20+ W but the out put is more like 6.




    The test I'm about to do:


    Measure and compare grid signal voltages on cold and hot biased 6v6 amps.



    What should the grid signal voltage typically be?
     
  13. Old Tele man

    Old Tele man Friend of Leo's

    Posts:
    3,941
    Joined:
    May 10, 2017
    Location:
    Once banned always banned
    peteb - if you want to learn more about Class-AB1 operation, here's a paper that dissects and expands upon data sheet numbers and how they can be calculated:

    http://thermionic.info/mccaul/McCaul_GettingMoreFromTubeDataSheets_2008.pdf

    P.S. -- "Class-A" is 360-degree conduction, ie: a full sinewave, otherwise it's NOT Class-A. Also, the "1" subscript in Class-AB1 denotes "NO grid current flows...ever." That's what robrob is referring to above when he talks about power output tubes NOT typically going into saturation because their Class-AB1 operation keeps them from being driven into saturation. The Class-AB1 process is: normal drive signal NEVER exceeds the bias voltage--otherwise it's no longer Class-AB1, but rather Class-AB2, and the change from AB1 to AB2 greatly affects the loading back (lower Z) onto the phase-splitter because NOW the grids are being driven hard enough to make them draw current--just like a screen grid drawing current. There's a LOT more to the process than simply tweaking the bias...many OTHER aspects of associated circuitry (phase-splitter, OT, etc.) must be considered too.
     
    Last edited: Jul 20, 2017
    peteb likes this.
  14. Snfoilhat

    Snfoilhat Tele-Afflicted

    Age:
    39
    Posts:
    1,396
    Joined:
    Apr 8, 2016
    Location:
    Oakland, CA
    Seems to depend on the 'recipe' of load and plate voltage and whatever else: see the three example single-ended 6V6 amplifiers in this table, below. Three different maximum signal power outputs, depending on the recipe. 2W, 4.5W, or 5.5W.

    Also, it seems like operation class can't be ignored. 2 6V6s in push-pull AB can make 22W in a Deluxe Reverb, but there is no such thing as half of that, one 6V6 in push pull. Like Bendyha said above, with one tube class A is all you can get. Two tubes in p-p are potentially greater than the sum of their single-ended parts? I think someone earlier wrote you'd be unlikely to see a single 6V6 making more than 6ishW?
    6V6GT-GE_load_lines.png
    These data (in color) are from my amp. I didn't know how to do this until reading the thread, so thank you all and any errors are my own. Wanted to understand what's happening when I use the different OT secondaries with the different speaker combinations in my room. Seems like if the reflected impedance got much smaller, the line would rotate clockwise right into the over-dissipation zone.
     
    Last edited: Jul 21, 2017
  15. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Thanks bendyha, but I always thougt engine idle as an imperfect analogy to amp idle. And my question of bias in low powered amps vs high powered amps clarified something.

    The slug and the hot rod may have the same idle but it has nothing to do with max power.

    I think idle in amps affects power, increase idle, and you increase power out.
     
  16. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia

    Thanks Rob, I think it all sounds right. I am trying to identify tube operation points, I like what you have put together, I'm looking more for numbers to define tube operation points.
     
  17. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Old tele man,


    I looked it over, I may go back for more info and data.


    This is very much of interest, I was just measuring signal voltages and comparing them to bias voltages, it was very interesting and I did not know the significance of ab1 vs ab2 before, I think I see it now.
     
  18. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    Agreed


    Class of operation dictates that class ab is superior to class A in power output per tube.

    Known amps:

    Class A: champ: 6 W per a 6v6, dissipation 12 w

    Class Ab: deluxe reverb, 22 w, 11 watts per tube, dissipation 14 w per tube


    note: the power dissipation per tube is slightly higher in class AB, but the power out is almost double.
     
  19. peteb

    peteb Friend of Leo's

    Posts:
    4,208
    Joined:
    Apr 25, 2003
    Location:
    Cascadia
    I did some testing, here's what I found out.

    The purpose of this thread is to try and relate bias points, plate dissipation %, to tube operating points. The effort is showing that there is difficult to directly relate bias points to tube operating points such as cutoff and saturation.

    I have learned that the best measure of tube operating point is how wide open or shut the tube or valve is by comparing the negative DC bias voltage that closes the valve to the AC signal voltage that counter acts the bias voltage and opens the valve.

    Even though the signal voltage gets subtracted from the bias voltage, There is probably no better way of comparing the signal voltage to the bias voltage then to express the ratio as what percentage the signal voltage is of the bias voltage. The bias voltage represents a closed window and signal V divided by the bias V expressed as a percentage says how much of the window is open.

    for lack of a better term, let's call it the signal voltage as a percentage of the bias voltage. What's normal? I did some measuring.

    Samples:
    1x6v6 amp, VACmax/Vbias = 50+%, bias = 124%
    1x6v6 amp, VACmax/Vbias = 50 +%, bias = 137%
    2x6v6 amp, VACmax/Vbias = 55%, bias = 31%
    2x6L6 amp, VACmax/Vbias = 58 %, bias = 55 %

    I consider all of the amps biased in the middle range, the class A amps are over 100% because they are class A.


    They all seem to have a VACmax/Vbias in a similar range, which defines the tube operating point, even though the biases are different. The class A biases will be high automatically, so directly comparing biases may not be valid across classes, but we see different biases all with similar operating points. So this makes directly relating bias Point to operating point more challenging, and possibly showing it can't be done.



    Not only was I trying to relate bias point to operating point, but I was interested in seeing if a lower powered amp could be biased lower then a higher power amp and still have the same tube operating conditions. The small data sample above supports but does not prove this. The low biased amp has a healthy VACMax/Vbias percentage despite the low bias.
     
  20. robrob

    robrob Poster Extraordinaire Ad Free Member

    Posts:
    8,138
    Joined:
    Dec 29, 2012
    Location:
    United States
    A low powered amp can be biased in the same range as a high powered amp. The question you should be asking is, "Is the tone better in a low powered amp with a cooler bias and is the tone better in a high powered amp with hotter bias?"

    All the Fender tweeds were biased cool from the factory roughly around 55% so it seems Leo didn't think so.

    I don't think so either. In my experience high and low powered amps sound best when biased between 60 to 70% (70% is my personal max for Class AB amps).
     
    peteb likes this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.