Doug, Very good design work and excellent explaination. I am impressed with your conception, "if it ain't broke make it better". Yes, to my knowledge it is your invention. When I get my shop set up again (after a retirement ) I will incorporate it,with full credit to Bug Amps, into my "super Texan" guitar amp, and a redesgned SCA-35, A perfect example of a blocking amp! The very,very basic circuit is in an obscure book "all about zenner diodes" in my stored library if I can find it, I will get you a copy of the original circuit. A really perfect example is Jimi Hendrix in the '60's movie "Monterey pop" it will have a tech like yourself tearing your hair out! I will get one of your boards to check it out in the dynaco amp first. Dan Torres "Inside Tube Amps"
I've always assumed the reason it's called blocking is because if a large transient upsets the bias enough, a quiet sound that comes right after can be "blocked" from being heard because the average bias has the output tube fully cut off. You may hear just peaks with very bad crossover distortion, or if the sound is quiet at all, it might not punch through at all.
Good stuff.Reminds me of an article by Jim Carlyle (Clamped Bias, AB2,RC Coupled Amp). If you offer the bias kit I will buy one. I have quite a few projects to do but will put a push-pull to the forefront for this!
It looks like you've created a variable zener diode and this already comes in a single TO92 package LM431 voltage reference chip. have you heard of this? You might want to give it a try
Nice! I have been thinking on and off about blocking distortion for quite some time now, mostly about using a source follower in for single ended and class A amplifiers. The grid of a typical power tube is a pretty low impedance to be driven by a tube follower though. Here is something you can look into - Does this restorative clipping that discharges the coupling cap show up on the plate voltage of the previous tube? I would expect that the current might in the range of several mA depending on how hard you drive it - (Guessing from the hip here) so that would definitely show up on most phase inverter plates(12ax7's for sure and even 12at7's - maybe those CBS engineers were onto something with the 47k plate loads - Hersey - I know!!!) - and then also on the other plate of a long tail pair via quasi differential action. By this time though there is enough of a din going on that it might be lost in the Kerrrrannnng! Congratulations on the circuit! The closest I have seen of something like this might be a variable clipping circuit in one of the London Power books - but yours is applied to the bias circuit which is new to me! I am so glad to see this sort of video! I have been mulling over doing videos like this just because it seems like there are not many being made. - Leif!
Ok so, I was assuming you were guitar amp guys - I just checked out your site and it looks like High end audio, where the grid drive is not so much. So you should really look at the charging current effect on the previous stage - I would actually buffer it - direct coupled so you can really control that grid. I worked at a place that did some very high end stuff for a few years but we were driving triodes and the grid current could be overcome by a 6sn7 follower. The funny thing about high end audio is that after you leave you think that it stops being a thing ;-)
In a traditional fixed bias configuration, regulating the negative supply wouldn’t resolve blocking distortion. This is because the grid leak resistance would still allow the coupling capacitor to charge when current flows in the control grid.
Still trying to fully grasp the blocking phenomenon. But what an outstanding job you have done here. Curious- Would I need to make Gs in the bug bias circuit equivalent to the paralleled tube's (2 6L6s per side in push-pull) grid stoppers? Also, Would it be beneficial to have your circuit on each side of the two tube pairs? Thank you for giving me something that is well explained that I can study.
Having the R value in series with the diode be similar in value to the stopper value is just a place to start really. It should be adjusted so that the clipping character is similar and symmetrical. But yea, in parallel push pull it would be the stopper values in parallel. So half the value assuming they are the same value. In a parallel push pull situation only one BB circuit will be needed for each side and should run both tubes.
So essentially it's a dynamic bias shunt type of arrangement - useful in situations where you want to clean up the distortion caused by overdriving the output stage in an attempt to exceed its power output capability. I can't paste images here so excuse the link atrad-audio.co.nz/wp-content/uploads/2020/04/BugBias.png you mention the RGL resistor as being selected from the tube's datasheet. Typically in fixed-bias the datasheet will specify a maximum grid-to-cathode resistance. In this case it's the sum of the grid stopper, the "build-out" resistor (RGL in this case) and then teh resistance from there to ground. Which in your case has an extra 150K on it. So long story short: Grid-stopper PLUS RGL (on my screenshot of your vid) PLUS RGND (screenshot) need to not exceed tube's specified grid-to-cathode resistance.
You’re right, single stages suffer from blocking too. This solution will work there too but may change the distortion spectrum Because the signal clipping will not be during cut off anymore. This will probably result is more third order distortion in overload.
@@Edgarbopp I mention single ended because I have begun to notice some strange distortions in my homemade champ. I think initially I'm gonna change some decoupling cap values from overspec .1uf to vox level 0.01uf. also.. it's cathode biased anyway but I was watching this video more out of curiosity for the mechanisms than a solution. Thanks tho!
Damn, you explained this insanely well. I can't believe you only have 1.6k subs, you definitely deserve more!
That’s very kind of you to say. 🙂
Doug,
Very good design work and excellent explaination. I am impressed with your conception, "if it ain't broke make it better". Yes, to my knowledge it is your invention. When I get my shop set up again (after a retirement ) I will incorporate it,with full credit to Bug Amps, into my "super Texan" guitar amp, and a redesgned SCA-35, A perfect example of a blocking amp!
The very,very basic circuit is in an obscure book "all about zenner diodes" in my stored library if I can find it, I will get you a copy of the original circuit.
A really perfect example is Jimi Hendrix in the '60's movie "Monterey pop" it will have a tech like yourself tearing your hair out!
I will get one of your boards to check it out in the dynaco amp first.
Dan Torres "Inside Tube Amps"
Thanks Dan! 🙂
I've always assumed the reason it's called blocking is because if a large transient upsets the bias enough, a quiet sound that comes right after can be "blocked" from being heard because the average bias has the output tube fully cut off. You may hear just peaks with very bad crossover distortion, or if the sound is quiet at all, it might not punch through at all.
You might be right!
Good stuff.Reminds me of an article by Jim Carlyle (Clamped Bias, AB2,RC Coupled Amp). If you offer the bias kit I will buy one. I have quite a few projects to do but will put a push-pull to the forefront for this!
“Push pull, obviously”.
It looks like you've created a variable zener diode and this already comes in a single TO92 package LM431 voltage reference chip. have you heard of this? You might want to give it a try
Nice! I have been thinking on and off about blocking distortion for quite some time now, mostly about using a source follower in for single ended and class A amplifiers. The grid of a typical power tube is a pretty low impedance to be driven by a tube follower though. Here is something you can look into - Does this restorative clipping that discharges the coupling cap show up on the plate voltage of the previous tube? I would expect that the current might in the range of several mA depending on how hard you drive it - (Guessing from the hip here) so that would definitely show up on most phase inverter plates(12ax7's for sure and even 12at7's - maybe those CBS engineers were onto something with the 47k plate loads - Hersey - I know!!!) - and then also on the other plate of a long tail pair via quasi differential action. By this time though there is enough of a din going on that it might be lost in the Kerrrrannnng! Congratulations on the circuit! The closest I have seen of something like this might be a variable clipping circuit in one of the London Power books - but yours is applied to the bias circuit which is new to me! I am so glad to see this sort of video! I have been mulling over doing videos like this just because it seems like there are not many being made. - Leif!
Ok so, I was assuming you were guitar amp guys - I just checked out your site and it looks like High end audio, where the grid drive is not so much. So you should really look at the charging current effect on the previous stage - I would actually buffer it - direct coupled so you can really control that grid. I worked at a place that did some very high end stuff for a few years but we were driving triodes and the grid current could be overcome by a 6sn7 follower. The funny thing about high end audio is that after you leave you think that it stops being a thing ;-)
Congrats on the video Doug :D
Thanks for the encouragement!!
Would regulating the negative supply help resolve blocking distortion?
In a traditional fixed bias configuration, regulating the negative supply wouldn’t resolve blocking distortion. This is because the grid leak resistance would still allow the coupling capacitor to charge when current flows in the control grid.
Still trying to fully grasp the blocking phenomenon. But what an outstanding job you have done here.
Curious- Would I need to make Gs in the bug bias circuit equivalent to the paralleled tube's (2 6L6s per side in push-pull) grid stoppers? Also, Would it be beneficial to have your circuit on each side of the two tube pairs?
Thank you for giving me something that is well explained that I can study.
Having the R value in series with the diode be similar in value to the stopper value is just a place to start really. It should be adjusted so that the clipping character is similar and symmetrical. But yea, in parallel push pull it would be the stopper values in parallel. So half the value assuming they are the same value.
In a parallel push pull situation only one BB circuit will be needed for each side and should run both tubes.
A wish one demo of this sound please for my god !
So essentially it's a dynamic bias shunt type of arrangement - useful in situations where you want to clean up the distortion caused by overdriving the output stage in an attempt to exceed its power output capability. I can't paste images here so excuse the link atrad-audio.co.nz/wp-content/uploads/2020/04/BugBias.png you mention the RGL resistor as being selected from the tube's datasheet. Typically in fixed-bias the datasheet will specify a maximum grid-to-cathode resistance. In this case it's the sum of the grid stopper, the "build-out" resistor (RGL in this case) and then teh resistance from there to ground. Which in your case has an extra 150K on it. So long story short: Grid-stopper PLUS RGL (on my screenshot of your vid) PLUS RGND (screenshot) need to not exceed tube's specified grid-to-cathode resistance.
Adam Rosner the 150k pair will be in parallel to some extent depending on the source Z of the negative supply. But yea I agree!
Love the video Doug
Thanks!
You know what that does mean right ? GU 50 amp schematic update :P
"two output tubes obviously" where's the single ended amp love at
You’re right, single stages suffer from blocking too. This solution will work there too but may change the distortion spectrum Because the signal clipping will not be during cut off anymore. This will probably result is more third order distortion in overload.
@@Edgarbopp I mention single ended because I have begun to notice some strange distortions in my homemade champ. I think initially I'm gonna change some decoupling cap values from overspec .1uf to vox level 0.01uf. also.. it's cathode biased anyway but I was watching this video more out of curiosity for the mechanisms than a solution. Thanks tho!