Page 1 of 2 12 LastLast
Results 1 to 35 of 36
Like Tree7Likes

Thread: Phase Inverters with CCS

  1. #1
    Member
    Join Date
    May 2013
    Posts
    78

    Phase Inverters with CCS

    Hi all,

    I open this thread because I would like to talk about phase inverters with constant current sinks. Few designs apply them in the guitar world, and I think there can be more to squeeze from it with the partecipation of all of us.

    I've recently bought a Marshall 9200, and I re-jumped in the PIs with CCS world were I've been years ago with the Mesa Boogie 295 Simul Class.

    While fixing some weak points of the 9200 project (mainly bias and NPN for the CCS), I've done some simulations on the possibility to do some small improvements on the design of the PI.

    Well, the advantage of the CCS is (close to) infinite AC impedance (good for PI balance) while DC current flows as per the circuit setup. In the case of the Marshall 9200 circuit, the current is:
    0,7 V / 270 Ohm = 2,6 mA

    The concept is the same described here:



    Lets say the collector current tries to increase. The emitter current will also try to increase which makes the drop across the 1.6 k ohm resistor increase. Because the base voltage is held very constant by the voltage divider of the 3 k and 1.1 k ohm resistors the bias on the base-emitter junction will be decreased. This will decrease the base current and cancel out most of the collector current increase. The collector current has to change by a small amount to make the base current change, but the change is very small. The collector current is 3 mA and most 3904s have a current gain of more than 100 so the base current is about, or less than, 30 microamps. The divider current is 4.88 mA. The base current is 0.615% of the divider current.
    As you can see the distortion was very low, much lower than I had expected. Also the distortion at both outputs was very similar. When feedback was simulated the balance remained perfect. The distortion changed though. The plate on the left gave 0.09% while the right hand plate gave 0.175%. This low level of distortion means the distortion imbalance can be disregarded.
    Practical Phase Inverters.

    This configuration has in fact very low distortion, with a bit of a prevalence of 3rd harmonic.

    By looking at this other configuration by Crowhurst:



    Positive feedback can only be used over one stage in order to avoid oscil-
    lation. An easy way to accomplish this positive feedback is to couple cathode
    bias resistors of two consecutive stages in the earlier part of the amplifier,
    where the distortion is small.

    A momentary positive fluctuation at the grid of the first stage will pro-
    duce a momentary negative fluctuation at the plate, which is passed on to
    the grid of the following stage. This produces another positive fluctuation
    at the plate of the second tube. At the same time, the negative fluctuation
    at the plate of the first tube, resulting from increased plate current, will be
    accompanied by a positive fluctuation at its cathode. Similarly, a negative
    fluctuation appears at the cathode of the second tube. The fluctuation at
    the cathode of the second tube is much bigger than that at the first stage.
    Connecting the resistor between the cathodes will allow some of the fluctua-
    tion from the cathode of the second stage to cancel the fluctuation at the
    cathode of the first stage.
    https://archive.org/stream/Crowhurst...sic_3_djvu.txt

    So I came out with this simple modification on the 9200 to apply the following:
    - center the plates of the PI at 2/3 of B+
    - increase the current capabilities of the PI
    - apply some positive feedback between input stage and PI
    - give more dynamic range to the PI

    Anyone has some experience on them, from direct design or as a user of one amp with this configuration?

  2. #2
    Member
    Join Date
    May 2013
    Posts
    78
    Looking at similar configurations, I came into the 9005 poweramp:
    http://drtube.com/schematics/marshall/9005.gif

    Searching some more informations on that 9005 configuration and its origins, I've found that the Major had a similar configuration, with a concertina plus a differential amp based on a 12au7.

    Well, I've also found that the original schematics by Marshall are wrong, because the signal is taken from the cathode of the concertina, and the 1M resistor is also connected to the cathode of the concertina, even if the schematics say the opposite:
    http://drtube.com/schematics/marshall/1966u.gif
    http://drtube.com/schematics/marshall/1967u.gif
    http://drtube.com/schematics/marshall/1978u.gif

  3. #3
    Senior Member uneumann's Avatar
    Join Date
    Jun 2010
    Location
    Los Angeles
    Posts
    191
    Quote Originally Posted by Roberto View Post
    Hi all,

    I open this thread because I would like to talk about phase inverters with constant current sinks. Few designs apply them in the guitar world, and I think there can be more to squeeze from it with the partecipation of all of us. ...
    Anyone has some experience on them, from direct design or as a user of one amp with this configuration?
    I've built two amps with CCS in the PI. Full schemos are on my site, but these clips show the PI. The first (Princeton AOT) uses two EF86 tubes which is a bit crazy, but it was an experiment to see if screen voltage shifts could impact compression/tone at high drive levels. The CCS does balance things pretty well - the 500 ohm adjustment pot is helpful. However, tube variation with EF86's seems high so perfect matching is going to take some selection among a batch of tubes. The screen shifts do occur as hoped and the OD tone of this PI is good - although I can't say I've heard it in isolation. By the time the PI is in OD, there is distortion coming from other sources as well.

    p-aotv3.jpg

    The second figure is from my current project (BML) and it uses a CCS with two triodes. Again, very good results in terms of balance and gain and OD behavior. This design has a great clean tone so I can attest to the PI behavior under high, but still linear range signals.

    bml-pre1b.jpg

    A point you didn't raise, but it seems a virtue of the CCS approach, is that it presents a great tremolo opportunity. The BML schem shows this. The trem osc drives the CCS to vary the PI tube current. The trem tone is smooth and very "bias-y" sounding.
    Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
    Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

    https://sites.google.com/site/stringsandfrets/
    https://sites.google.com/site/ampaid16/home

  4. #4
    Senior Member
    Join Date
    Feb 2009
    Posts
    318
    I have also done a CSS tail, just with a normal 12ax7 LTP circuit. In my case, I was using a low B+ and wanted to do without the huge voltage drop (and loss of swing) of a tail resistor. It seemed to work quite well, but I didn't try any A-B comparisons.

  5. #5
    Lifetime Member
    Join Date
    May 2006
    Posts
    4,150
    A differential amplifier with only one input fed, the other held constant, amounts to a common cathode stage from the input to the first plate, but for the path from the input grid to the second tube, it is a common plate stage ( that is, a cathode follower ) feeding the second tube as a common grid stage with its output on the second plate.

    The low output impedance of the first tube stage's cathode pulls the second cathode up and down by nearly the same signal amount as the input grid moves, and since the second tube grid is held fixed, the second tube's cathode-to-grid voltage is wobbled about by the same signal as the first tube's grid-to-cathode, but the phase is inverted. So the plate signal of the second tube is **about** the same size/current as the first tube, but inverted in phase.

    That only holds true if the biasing elements on the shared cathodes don't eat any of the signal being pumped in through the first cathode. So the lower the impedance of the shared cathode biasing resistor, the more signal leaks out of the cathode1 to cathode2 path and the smaller the signal on the second tube's plate. Conversely, the higher the impedance of the biasing resistor on the two cathodes, the more the signal from the first cathode gets into the second cathode, and the closer the two signals on the plates are to the same magnitude.

    The classical diffamp PI in guitar amps usually tries to fix this up a bit by making the plate resistors of the two triodes different sizes to try to get the gains back to more similar sizes, even in the face of non-identical gains caused by low cathode biasing resistors.

    Biasing with a CCS is about as high an impedance as you can do for the shared cathodes without going to extreme measures. CCS biasing on the cathodes of a diffamp PI helps get the gains to be much closer. It can also make the usual tricks with feedback into the shared cathodes for "presence" and other stuff harder to do.
    Amazing!! Who would ever have guessed that someone who villified the evil rich people would begin happily accepting their millions in speaking fees!

    Oh, wait! That sounds familiar, somehow.

  6. #6
    Member
    Join Date
    May 2013
    Posts
    78
    Thanks uneumann,

    Quote Originally Posted by uneumann View Post
    I've built two amps with CCS in the PI. Full schemos are on my site, but these clips show the PI.
    can I kindly ask you to link them here? I cannot find them.

    The second figure is from my current project (BML) and it uses a CCS with two triodes. Again, very good results in terms of balance and gain and OD behavior.
    That's one of the point I was thinking at, the OD behaviour, thanks for this feedback.

    A point you didn't raise, but it seems a virtue of the CCS approach, is that it presents a great tremolo opportunity. The BML schem shows this. The trem osc drives the CCS to vary the PI tube current. The trem tone is smooth and very "bias-y" sounding.
    That's something I never thought about. I will develop it following your indications.

  7. #7
    Member
    Join Date
    May 2013
    Posts
    78
    Quote Originally Posted by octal View Post
    I have also done a CSS tail, just with a normal 12ax7 LTP circuit.
    Thanks octal, what about its OD performance? How was NFB connected?

  8. #8
    Member
    Join Date
    May 2013
    Posts
    78
    Quote Originally Posted by R.G. View Post
    Biasing with a CCS is about as high an impedance as you can do for the shared cathodes without going to extreme measures. CCS biasing on the cathodes of a diffamp PI helps get the gains to be much closer. It can also make the usual tricks with feedback into the shared cathodes for "presence" and other stuff harder to do.
    Thanks R.G., AFAIK the second input is used grounded as a NFB input, including the presence pot. But I think I've missed something on your post, because I know you are well knowledged.

  9. #9
    Lifetime Member
    Join Date
    May 2006
    Posts
    4,150
    I was speaking more generally than just your circuit. The common string of resistors from the joined cathodes to ground is often used as an input for feedback from the OT secondary, as well as a tinkering point to add "presence" and other things by feedback.
    Amazing!! Who would ever have guessed that someone who villified the evil rich people would begin happily accepting their millions in speaking fees!

    Oh, wait! That sounds familiar, somehow.

  10. #10
    Senior Member uneumann's Avatar
    Join Date
    Jun 2010
    Location
    Los Angeles
    Posts
    191
    Quote Originally Posted by Roberto View Post
    Thanks uneumann,

    can I kindly ask you to link them here? I cannot find them.

    That's one of the point I was thinking at, the OD behaviour, thanks for this feedback.

    That's something I never thought about. I will develop it following your indications.
    Sure - the first circuit fragment is from version3 of this amp (scroll towards the bottom of the page) ...
    https://sites.google.com/site/string.../princeton-aot

    The second circuit is from this amp ...
    https://sites.google.com/site/string...m-lite-project
    Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
    Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

    https://sites.google.com/site/stringsandfrets/
    https://sites.google.com/site/ampaid16/home

  11. #11
    Senior Member
    Join Date
    Feb 2009
    Posts
    318
    Quote Originally Posted by Roberto View Post
    Thanks octal, what about its OD performance? How was NFB connected?
    I didn't use any NFB. OD performance was good, I adjusted the current in the CSS with a trimpot for best sound. I used the "Valve wizard" circuit.

    The Valve Wizard (scroll down to the bottom, the very bottom circuit with the three biasing diodes in series is the one I used.) I substituted a trimpot for the 600 ohm emitter resistor. If I recall correctly, I had to go up in value a fair bit from what is shown in the schematic.

  12. #12
    Member
    Join Date
    May 2013
    Posts
    78
    Quote Originally Posted by octal View Post
    I adjusted the current in the CSS with a trimpot for best sound. I used the "Valve wizard" circuit.
    Is there any technical reason to use three 1n4148 instead of two, and just Iccs = 0,6 V / Rk ?

  13. #13
    Member
    Join Date
    May 2013
    Posts
    78
    Quote Originally Posted by uneumann View Post
    Sure - the first circuit fragment is from version3 of this amp (scroll towards the bottom of the page) ...
    https://sites.google.com/site/string.../princeton-aot
    https://1617738a-a-62cb3a1a-s-sites....-aot/P-AOT.png
    Thanks for sharing, I will take a deep look at it!

    The second circuit is from this amp ...
    https://sites.google.com/site/string...m-lite-project
    https://1617738a-a-62cb3a1a-s-sites....ML%20PRE1b.png
    This is more common to what I had in mind, and the bias shifting is really a nice plus!

  14. #14
    Old Timer
    Join Date
    Dec 2007
    Location
    Poland
    Posts
    1,456
    Quote Originally Posted by Roberto View Post
    Is there any technical reason to use three 1n4148 instead of two, and just Iccs = 0,6 V / Rk ?
    It seems that it is explained in the article your are referring (starting from: "Note, this circuit may not work if Re is less than about 400 ohms.....").

    Mark

  15. #15
    Senior Member uneumann's Avatar
    Join Date
    Jun 2010
    Location
    Los Angeles
    Posts
    191
    Quote Originally Posted by Roberto View Post
    Is there any technical reason to use three 1n4148 instead of two, and just Iccs = 0,6 V / Rk ?
    Just a comment about the valve wizard circuit. It works fine for small signal PI drive conditions. But - if the PI is fed large signals to overdrive it, the shared cathodes swing negative to the point where the CCS saturates and no longer behaves as a CCS. That will change OD waveforms - you'll have to try it to see if you like it. A small reference voltage can keep the CCS out of saturation. Your OP shows a -20v supply. In my circuits, I use a small regulator (12v or 15v) to accomplish the same goal. You just want to create some headroom for the cathodes to swing without saturating the CCS.
    Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
    Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

    https://sites.google.com/site/stringsandfrets/
    https://sites.google.com/site/ampaid16/home

  16. #16
    Lifetime Member
    Join Date
    May 2006
    Posts
    4,150
    Good point. You never want to starve that CCS for voltage.

    Since there are input DC blocking caps and a good 300V of plate supply, the simplest thing might be to just bias both grids up at about 50-75V.

    There are variants of CCS circuits that would let you run the CCS with very, very small reference currents. Using a MOSFET for the series pass device and then using a bipolar to "steal" gate voltage from the MOSFET based on the voltage across a MOSFET source resistor comes to mind.

    This setup would probably eliminate the need for a negative supply and the issues with input voltage size.
    Amazing!! Who would ever have guessed that someone who villified the evil rich people would begin happily accepting their millions in speaking fees!

    Oh, wait! That sounds familiar, somehow.

  17. #17
    Supporting Member
    Join Date
    Sep 2007
    Posts
    2,840
    Avoid fancy circuits in guitar amps. The guitarist always finds some way to make them misbehave. So if you want a current source, build a negative supply like your positive supply with a pair of diodes, and a couple of Cs and Rs. (You do not need that much current.) Then use a 100K resistor (or whatever your specific circuit requires) for the CCS. 100K is large compared to the inverse of the gm, and so the balance is good. Hard to make that circuit misbehave!

  18. #18
    Senior Member uneumann's Avatar
    Join Date
    Jun 2010
    Location
    Los Angeles
    Posts
    191
    Quote Originally Posted by Mike Sulzer View Post
    Avoid fancy circuits in guitar amps. The guitarist always finds some way to make them misbehave. So if you want a current source, build a negative supply like your positive supply with a pair of diodes, and a couple of Cs and Rs. (You do not need that much current.) Then use a 100K resistor (or whatever your specific circuit requires) for the CCS. 100K is large compared to the inverse of the gm, and so the balance is good. Hard to make that circuit misbehave!
    Not to be dismissive ... but building another power supply sounds "fancy" to me.
    Just how would you make a CCS "misbehave" if it's properly designed and built as some of the examples show?
    Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
    Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

    https://sites.google.com/site/stringsandfrets/
    https://sites.google.com/site/ampaid16/home

  19. #19
    Member
    Join Date
    May 2013
    Posts
    78
    Thanks uneumann!

    Thanks E.G.,
    Is this configuration better than a single 75 V zener instead of the 1n4148s, and then set the current with Re as before?
    EG for 7,5 mA would be around 10k.

  20. #20
    Member
    Join Date
    May 2013
    Posts
    78
    Hi Mike,

    Mesa Boogie 295 and 395 had this configuration, Coliseums as well.
    Marshall 9100, 9200 and relatives had this configuration.

    Tons of guitarists have enjoyed and still enjoy the CCS PI with no issues at all.
    I have a Marshall 9200 and the Mesa 295 is the best sounding power amp I've ever had (except mines! ).

    On top of that, the guitarist will be me.

    Not to be dismissive too, but I hope you don't feel outraged if I consider your post as the answer to the question:
    "you know the difference between a guitar and a piano?"


  21. #21
    Senior Member
    Join Date
    Mar 2008
    Posts
    368
    Quote Originally Posted by Mike Sulzer View Post
    Avoid fancy circuits in guitar amps. The guitarist always finds some way to make them misbehave. So if you want a current source, build a negative supply like your positive supply with a pair of diodes, and a couple of Cs and Rs. (You do not need that much current.) Then use a 100K resistor (or whatever your specific circuit requires) for the CCS. 100K is large compared to the inverse of the gm, and so the balance is good. Hard to make that circuit misbehave!
    Add some remarks: negative supply should be dimensioned by nominal current draw by cathodes and also the cathode resistor should support the power dissipation. In respect with current draw sometime can be pretty cumbersome. Also the supply should have very low ripple to not modulate the signal. Othervise it works pretty well in some project I did.
    And a question: Does it a danger to keep voltage over max. hk rating whilst the tube is heated and start to draw current, please? I don.t know how "healthy" is to keep -400v in cathode (I exagerated) expecting to drop whilst tube is heating. How much voltage should consider safe in respect with hk rating from this point ? If is a limit then tail resistor is limited by max. voltage admited by hk isolation aso. Not when running but when is cold instead. Thanks
    Last edited by catalin gramada; 04-11-2017 at 09:21 AM.

  22. #22
    Senior Member
    Join Date
    Mar 2008
    Posts
    368
    Quote Originally Posted by Roberto View Post
    Hi Mike,

    Mesa Boogie 295 and 395 had this configuration, Coliseums as well.
    Mesa used negative power supply without regulator in 400 series as well, best amps ever heard by others..
    Does it a guitar or a piano?
    Last edited by catalin gramada; 04-11-2017 at 12:23 PM.

  23. #23
    Member
    Join Date
    May 2013
    Posts
    78
    well, the 400 series is more FORTISSIMO than piano
    catalin gramada likes this.

  24. #24
    Supporting Member
    Join Date
    Sep 2007
    Posts
    2,840
    I apologize for my rude remark.

    cg: I have had no trouble with the brief over voltage, but I do not know why. I think there could be a problem; there just never has been!

    un: I do not think of the negative power supply as fancy. It is an additional simple circuit with components that interact with each other, but not so much with the components in the signal path circuit. It allows making the signal path circuit simpler.

    Perhaps I am misusing my experience as a designer of complicated signal processing software where you must modularize, then isolate and simplify the modules. It seems that the same rules work here.

    Roberto: Yes. A guitar is simple enough so that you can redesign it to achieve specific goals in a short enough time so that you can do a lot in a life time. A piano has so many interactions that even if you just try to copy a good design, you might end up with something very different. It is difficult to achieve much directed piano redesign in a life time.
    catalin gramada likes this.

  25. #25
    Senior Member uneumann's Avatar
    Join Date
    Jun 2010
    Location
    Los Angeles
    Posts
    191
    Quote Originally Posted by Mike Sulzer View Post
    I apologize for my rude remark.

    cg: I have had no trouble with the brief over voltage, but I do not know why. I think there could be a problem; there just never has been!

    un: I do not think of the negative power supply as fancy. It is an additional simple circuit with components that interact with each other, but not so much with the components in the signal path circuit. It allows making the signal path circuit simpler.

    Perhaps I am misusing my experience as a designer of complicated signal processing software where you must modularize, then isolate and simplify the modules. It seems that the same rules work here.

    Roberto: Yes. A guitar is simple enough so that you can redesign it to achieve specific goals in a short enough time so that you can do a lot in a life time. A piano has so many interactions that even if you just try to copy a good design, you might end up with something very different. It is difficult to achieve much directed piano redesign in a life time.

    No problem... I think you did make a valid point - which is that there are several ways to achieve a CCS. The "best" way for a given situation is often a matter of opinion and preference. I don't see any option we've discussed here as *always* better or worse. It depends on context too.
    That's part of the "art" part of circuit design -- considering all the options and arriving at an outcome that works (and sounds) satisfying.
    SoulFetish likes this.
    Old Tele man: Equations provide theoretical values, SPICE provides approximate values; but, the ears provide exact values.
    Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

    https://sites.google.com/site/stringsandfrets/
    https://sites.google.com/site/ampaid16/home

  26. #26
    Senior Member
    Join Date
    Jul 2011
    Location
    Adelaide, South Oz
    Posts
    634
    Looking into the cathodes of the PI you see a very low impedance, the CCS represents a very high impedance. Any residual negative rail ripple and noise will be divided by those impedances so the negative rail can be very simple.
    If using the negative rail for a bias supply for the output tubes then the rail needs to be more sophisticated and cleaner.
    Can you have the best of both worlds?
    YES.
    Run mosfet source followers AC coupled off the PI with output tube grids direct coupled to the mosfet sources (just the grid stop in between). Apply the bias (with additional filtering) to the mosfet gates AND use a current source load on the mosfet source followers back to the negative rail.
    Negative rail needs to be at least X3 the output tube bias voltage to guarantee drive to cutoff.
    I've done this with HIFi Amps but not tried it for guitar.
    Post #602 for schematic here:
    EL84 Amp - Baby Huey - Page 61 - diyAudio
    The feedback scheme is a bit different too.
    Cheers,
    Ian

  27. #27
    Member
    Join Date
    May 2013
    Posts
    78
    Thanks Gingertube,

    I've seen the schematic: diyAudio
    I will work on that. Maybe not for the Marshall 9200 (it would be complicated to implement it on the already crowded layout of that amp), but for the future.

    I also like the idea of a EL84 HiFi amp for the future!

  28. #28
    Senior Member
    Join Date
    Mar 2008
    Posts
    368
    Quote Originally Posted by Gingertube View Post
    Looking into the cathodes of the PI you see a very low impedance, the CCS represents a very high impedance. Any residual negative rail ripple and noise will be divided by those impedances so the negative rail can be very simple.
    If using the negative rail for a bias supply for the output tubes then the rail needs to be more sophisticated and cleaner.
    Can you have the best of both worlds?
    YES.
    Run mosfet source followers AC coupled off the PI with output tube grids direct coupled to the mosfet sources (just the grid stop in between). Apply the bias (with additional filtering) to the mosfet gates AND use a current source load on the mosfet source followers back to the negative rail.
    Negative rail needs to be at least X3 the output tube bias voltage to guarantee drive to cutoff.
    I've done this with HIFi Amps but not tried it for guitar.
    Post #602 for schematic here:
    EL84 Amp - Baby Huey - Page 61 - diyAudio
    The feedback scheme is a bit different too.
    Cheers,
    Ian
    Hey Gingertube. I used negative power suply to force a differential power stage driver into the balance. It works well till a certain signal amount when clipping put this stage into intermitent operation. Does a CCS the same thing, please ? Thanks
    Edit: this was never happen for the same amount with individual bypassed cathode resistors, and also my subjective impression it sound more "open, like that
    Last edited by catalin gramada; 04-19-2017 at 01:15 AM.

  29. #29
    Member
    Join Date
    May 2013
    Posts
    78
    Like blocking distortion?

  30. #30
    Senior Member
    Join Date
    Mar 2008
    Posts
    368
    Quote Originally Posted by Roberto View Post
    Like blocking distortion?
    yes

  31. #31
    Senior Member
    Join Date
    Jul 2011
    Location
    Adelaide, South Oz
    Posts
    634
    The distortion thing!!! Ian's rave - hopefully there is something useful in it.
    With a CCS in the tail then the total current thru' the 2 sides of the PI/driver are fixed. The CCS is set for 2 x the idle current of each triode or to think of it another way, to the idle current of the 2 triodes combined.
    If you drive one side to cut off then the other side can go to 2 x it's idle current - no more. You get a clamping function.
    When overdriven it is likely to clip really hard and that means lots of odd harmonics and higher order odd harmonics at that, the thing here is that the higher order odd harmonics are the ones which are musically "quint" (meaning no musical relationship to the fundamental note) and even small amounts of them are "finger nails down the blackboard" stuff. They just grate.

    That means that it (PI/Driver) has to run in Class A, that is not overdriven.

    That is why for a Class AB output stage where one side is regularly driven to cutoff then you need to use separate CCS for each of the push pull sides, and separate bypass caps.... BUT I abandoned this approach since it has the worst overload recovery of any output tube biasing methods and so is unsuitable for a Git Amp.

    The overdriven PI/Driver sound will be somewhat different from a resistor "tail" of lower AC impedance and unlike a lowish restive tail PI which has a fairly pleasant overdrive tone, it is likely to be far less pleasant and maybe quite awful unless you set the amp up such that the output tubes go into overdrive long before the PI/Driver stage.

    OFF THREAD ASIDE (FWIW):
    AS for that HIFI schematic I linked, the balanced shunt feedback from the output tube anodes does something special, it enforces balance for a start.
    Guitar guys who have listened to that amp with my corksniffer CD player and speakers have all commented on 2 things, its speed and the way it handles the attack of a guitar, and that it grunges pleasantly when driven hard. HiFi "dudes" talk about PRATT (Pace, Rhythm and Attack). This amp has great PRATT. Music like JJ Cale just is simply stunning with an easy relaxed loping pace.
    It is NOT a low distortion design but appears to be a benign distortion design. I've had 210W per Channel 0.001% distortion SS Power amps , JA80 4 x KT88 Jadis mono blocks, 845 Single Ended Triode Mono blocks and a heap of other high end HiFi Tube and SS Power Amps in my system and I always come back the amp to that design, currently my HiFi Amp is exactly that design but with 6SL7 diffamp and old coke bottle shape 6V6G output tubes.
    A glorious 12 Watts per Channel.

    I have had a try the Baby Huey Power Amp for Guitar on my to do list for far too long.
    For now I have to say for those interested try it at your own risk. What works quite delightfully for a HiFi Amp may not be ideal for a Git Amp.

    The Balance shunt feedback trades output tube gm for reduced internal impedance (lower rp). That means that it better drives the lowish Primary Inductance of Git Amp Output Trannies primary and results in improved bass, but it also better drives the leakage inductance and inter-winding capacitance of the the Output Tranny for improved top end. In a Stereo HiFi setup that also means improved stereo imaging (mostly phase related information) and you tend to hear a 3 dimensional wall of sound rather than 2 separate channels.

    So is this just some guy waving his primary sexual characteristic.
    Well I hope not, and the 60 or 70 guys around the world who have built one or more of these seem to agree.
    Is it all my own clever design - nah! like most things it is based on someone else's work, in this case Yves M's 6GW8/ECL86 design here:
    Push Pull ECL86/6GW8


    Now back to our regular programming.

    Cheers,
    Ian
    catalin gramada likes this.

  32. #32
    Member
    Join Date
    May 2013
    Posts
    78
    Thanks again Ian,

    Quote Originally Posted by Gingertube View Post
    AS for that HIFI schematic I linked, the balanced shunt feedback from the output tube anodes does something special, it enforces balance for a start.
    At the end you used the shunt feedback from the anodes or from the UL taps? I read you preferred the detail of the shunt from screens.

  33. #33
    Senior Member
    Join Date
    Jul 2011
    Location
    Adelaide, South Oz
    Posts
    634
    From the anodes - I tried it from the screens for a while before deciding it sounded better from the anodes. This seemed to be an impedance vs frequency of the speakers thing. With a Zobel (R +C) across the output tranny secondary to linearize impedance vs frequency then the screen sourced feedback seemed marginally better initially BUT long term listening and feedback from other guys who built it, suggested it was better with anode derived feedback. You will also note that Yves original did not have UL connections and so feedback was from anodes.
    Cheers,
    Ian

  34. #34
    Member
    Join Date
    May 2013
    Posts
    78
    Thanks Ian,

    a bit OT, but have you tried it also with cathode UL taps instead of screen ones?
    Or with the mixed fixed-cathode bias configuration?

    I will take care of the hint to let the PA saturates way before the PI.

  35. #35
    Senior Member
    Join Date
    Jul 2011
    Location
    Adelaide, South Oz
    Posts
    634
    Roberto,
    On my to do list is to try both together (menno's Supertriode configuration) !! I have a pair of Plitron VDV2100 CFB/H Toroidal Output Trannies and Matching Plitron Toroidal Power Trannies on the shelf for a pair of HiFi mono blocks, each with 4 x KT88. The Output trannies have 2 independent (one for each push pull side) 10% cathode feedback windings Plus 33% screen taps on the primary. The screen feedback voltage is from screen to cathode so the combination of 10% cathode feedback and 33% screen tap means we are running at the "usual" 43% Ultralinear point with Cathode feedback added. About 120 Watts per mono block. Have all the bits, just a matter of pulling finger!!
    Cheers,
    Ian

    P.S.
    Besides - I have 2 Guitar Amps to finish first, one with 6DQ6 Outputs (Va=300V, Vg2 = 150V for 30 Watts out) and one with a pair of 6v6, maybe I can try the BH output stage for that 6V6 Amp, it is for a friend and is already in the 'Impossible Amp" category as he wants a 2 channel amp, the clean channel to be also able to handle BASS and the lead channel to handle shreading , oh, and it has to have effects in and out to suitable impedance/voltage for his studio gear including digital reverb and it has to have independent effects MIX controls for each channel, a design which was semi out of control from the specification stage. The first Amp is just about done, the "Impossible" Amp is at Issue 5 of pencil schematics - not really impossible just difficult.
    - Back to our regular thread - sorry
    Last edited by Gingertube; 04-21-2017 at 08:55 AM.

Page 1 of 2 12 LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. In phase or out of phase for jumping channels
    By Dagan8 in forum Theory & Design
    Replies: 13
    Last Post: 12-23-2010, 02:05 PM
  2. car inverters as tube power supplies
    By gtrboy in forum Theory & Design
    Replies: 6
    Last Post: 10-12-2010, 10:11 PM
  3. questions about floating paraphase inverters
    By antieatingactivist in forum Maintenance, Troubleshooting & Repair
    Replies: 7
    Last Post: 11-11-2009, 10:26 PM
  4. Fender Deluxe Phase Inverters
    By Sir Cuitous in forum Theory & Design
    Replies: 14
    Last Post: 01-20-2009, 03:13 AM
  5. Phase Inverter and Reverb Phase
    By Daver in forum Theory & Design
    Replies: 5
    Last Post: 08-21-2006, 04:25 AM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •