MagneQuest/Peerless Forum

Welcome! Need support, you got it. Or share you ideas and experiences.

Return to MagneQuest/Peerless Forum

6EW7_2A3 Parafeed PP

208.58.148.226
Posted on October 6, 2009 at 10:50:09
nullspace
Audiophile

Posts: 203
Location: Philly, PA
Joined: June 18, 2004
 Hi All,

I've been talking to Mike about this amp for a while now, and it's finally done.

It's pretty straightforward: input transformer-buffered volume control (Penny&Giles RF11 in this instance), 6EW7 twin triode doing both input and driver duty with the little triode DC-coupled to the big triode, MQ autoformer as phase splitter, and 2A3 parafeed PP finals, with cathode biasing all around. The DC side of things is full-wave rectification with 6CJ3 TV damper diodes into a choke-input filter.

Results are preliminary, but suffice to say it's a much different amplifier than the 27 preamp/ 56->45 amp combination. Very cliche, but with the PP amp I'm hearing all kinds of details/ articulation that were lost in the previous setup. But at the same time, I spend a whole lot less time singing along (which may be a good thing) as there seems to be less emotional content. If I could just get the transparancy of the PP amp along with the emotional connectedness of the SE amp I'd be all set...

  • Schematic
  • JPG of signal chasis
  • JPG of PS chasis


    Regards,
    John


    		•
    Slight correction, posted on October 6, 2009 at 19:00:49
    danlaudionut
    Audiophile

    Posts: 3564
    Location: Upstate NY
    Joined: June 6, 2002
    		 John

    The OPT should be connected on the
    other side of the plate chokes.
    The way it is drawn you will
    not hear anything.

    DanL
    Geez..., posted on October 6, 2009 at 19:18:44
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 I wouldn't be so embarrassed if Allen Wright hadn't upbraided the TubeDIY crowd for just that mistake about a month ago...

    Thanks for pointing that out, Dan. I've fixed the schematic and uploaded it, so it should be correct now.

    Regards,
    John

    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 14:13:59
    Caucasian Blackplate
    Industry Professional

    Posts: 4445
    Location: Seattke
    Joined: June 18, 2004
    		 Have you compared the parallel feed push-pull to just push-pull?

    I have a pair of S-240's laying around, I suppose a similar amp may be somewhere in my future...

    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 15:48:26
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 It's on the list of things to try. I just got it up and running aout a week and half ago, so I'm still getting used to the sound as-is.

    Regards,
    John

    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 11:51:28
    mwiebe
    Audiophile

    Posts: 182
    Joined: November 3, 2001
    Contributor
      Since:     January 5, 2006
    		 John,
    I’ve used a few 6EW7s, but none as a P-P driver like you have. What is your plate to cathode voltage on the high mu triode and the lo-mu triode? I expect there is a bit of voltage drop over your CCS.

    You do not show a grid bias on the 2A3, I assume you have cathode resistors, are they bypassed? Why the 8.2uF blocking capacitors between the 2A3 and outputs? If those are stock S-240s you could try pulling those out. And then at some point you could pull out the 2A3 plate chokes and see what that does to the music.

    You could try to go straight from the plate of your hi-mu driver to the splitting choke and see what you think. I don’t know what you have as sources but the gain should be enough without the lo-mu triode. If you have a suitable choke, EXO-099, you could pull the CCS and choke load the hi-mu then straight to the splitting choke. You could put the BCP-15 below the lo-mu section and drive the splitting choke from the lo-mu cathode. These are all pretty easy things to do and listen to see if any get you closer to what you are looking for.
    Matt
    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 12:20:43
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Hi Matt,

    Operating points for the 6EW7, the high-mu triode is about 4.8mA, 121.8Vpk, -3.9Vgk and the low-mu triode is about 38.1mA, 152.4Vpk, -17.7Vgk. Those are average numbers, btw. The two examples I have aren't what you call matched. I trimmed the CCS (one is at 4.6mA and the other at 5.0mA) so that the operating points were, broadly speaking, comparable.

    Rk and Ck for the 2A3 are on the centertap of the filament transformer. At some point I'll be converting from parafeed to series feed to see what's what, and I tried to wire the signal chasis in such a way that it wouldn't be too onerous to implement. Phil Sieg did a parafeed PP amp some years ago, Marcel & Clovis in VALVE, where he ditched the parafeed DC-blocking caps in favor of a trimpot between the output tubes' cathodes with Rk and Ck off the center, keeping any DC imbalance off the primary of the output transformer.

    I do have enough gain to not need voltage gain from the low-mu high of the 6EW7, but I don't know that the high-mu section has the oomph to drive the PP finals directly; thoughts?. Setting the low-mu section up as a CF sounds very interesting, though. I'll have to give that some serious thought.

    Thanks for your comments.

    Regards,
    John

    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 13:46:47
    mwiebe
    Audiophile

    Posts: 182
    Joined: November 3, 2001
    Contributor
      Since:     January 5, 2006
    		 John,
    Since you want to listen to a few options with this topology you are limited on the hi-mu plate voltage. I usually run this up around 180-200Vpk around 5ma but I’m not sure how to do this simply with your design. I’ve also seen designs that run the 6EW7 hi-mu at 60V and 2ma. Your lo-mu section is what I usually end up at.

    I guess the first thing I would do is strip all the parallel feed stuff off the 2A3s and just see if you like what you hear. Should be easy enough to listen back and forth. You could also put adjustable CCS in the 2A3 tails to control DC imbalance on the outputs or you may find you don’t mind a little DC imbalance.

    Well the hi-mu section is not really high mu is it at 17 or so. So my suggestion of going direct was not very good unless you are using an active preamp with gain. You could swap in a 6FD7, same pinout I think, the low mu section is the same but the hi-mu section is 64, that may give you a few more options as you tune things in.

    Just some thoughts.
    Matt
    CCS on tails, posted on October 7, 2009 at 21:03:31
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005
    		 Well ... actually with just one CCS under the 2A3s tails you cannot adjust for any DC imbalance. But if you tie the 2A3s cathodes and use a CCS to bias the pair you get a differential amp that should sound better (IMHO).

    Nice looking wiring btw.

    Gianluca

    RE: CCS on tails, posted on October 7, 2009 at 22:52:52
    mqracing
    Manufacturer

    Posts: 3314
    Joined: June 29, 2000
    Contributor
      Since:     May 7, 2004
    		 Hi GL:

    Can you show this to me schematically???

    thanks,

    MSL



    Builder of MagneQuest™ & Peerless™ transformers since 1989
    RE: CCS on tails, posted on October 8, 2009 at 07:27:39
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004

    Hey Mike,

    As I see it, there's a couple way to use a CCS and ditch the parafeed DC-blocking caps.

    'A' is along the lines of how Phil Sieg did it with Marcel & Clovis, using a 50ohm pot to balance the DC current.

    'B' is from the May '06 issue of AudioXpress, a 6AS7 amp by Richard Sears. The cathodes are AC-coupled with the a decent-sized cap, and then individual CCSs provide the balanced DC current.

    I will say that one thing I'm kicking myself for is not having ordered a couple 2.5V filament windings on the power transformers. When I go to try this I'm going to be stuck picking up some Hammond filament transformers just for the experiement.

    Regards,
    John

    Missing parafeed caps in the schems, posted on October 8, 2009 at 08:20:52
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005
    		 I'd stick with a series feed transformer in your PP amp: the LCL combo at the output is anyway a resonant cell (that follows the LCL before the 2A3s) which introduces further phase shifts and blah blah blah

    Is the amp a bit "boomy" with a strong bass end?

    Gianluca

    RE: Missing parafeed caps in the schems, posted on October 8, 2009 at 08:34:55
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Hi Gianluca,

    Unless I'm mistaken (50-50 chance, I suppose), getting rid of the parafeed caps is the whole point of ensuring that both tubes draw the same DC current (in this case, 60mA for the 2A3s), so they are purposefully left out of the schematic.

    It's difficult to say; my mains only go down to ~120hz. (Sub)woofers are driven by a SS pro amp, with the signal coming from the speaker posts of the main amp. There is definitely a stronger low end with the PP amp vs. my 45 SE amp, but any 'boom' would be hidden by the relatively early roll-off of my OB woofers -- I suspect that I'm having some phase cancellation between the mains & (sub)woofers, and I've been unable to fix it as of yet. I'm still working on it...

    Regards,
    John

    RE: Missing parafeed caps in the schems, posted on October 8, 2009 at 10:10:52
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005
    		 >Unless I'm mistaken (50-50 chance, I suppose), getting rid of the >parafeed caps is the whole point of ensuring that both tubes draw the >same DC current (in this case, 60mA for the 2A3s), so they are >purposefully left out of the schematic.


    I don't think that is correct. Where plate, choke and transfomers met there are 300V or so: there is a path to ground through the transformer windings which is low DCR... I assume a couple of amperes would be flowing in each coil of the OPT and burn them. AC world gets even more complex.

    Again, I am much impressed by simple circuits and I like yours and I'd make it even simpler.

    A direct comparison parafeed/series feed would be a great experience to share. Please do keep us posted.

    I will somehow recycle your driver stage for my DAC ... I got 5 6EW7s on the bay last night after your posts, would you mind givind details on plate voltage you are running the 6EW7 at?

    OH! I love my beeseven in nickel ... they need special care being sensitive to pick up but those are the input tranny I prefer (I tried sowter, cinemag which are great too, lundhal, vintage altecs). All were nickel/amorphous (not sure about altecs) ... I am trying to put my hands on some UTCs or Peerless but they are utterly expensive

    RE: Missing parafeed caps in the schems, posted on October 8, 2009 at 12:22:03
    mqracing
    Manufacturer

    Posts: 3314
    Joined: June 29, 2000
    Contributor
      Since:     May 7, 2004
    		 Hi GL:

    you wrote;

    :::Where plate, choke and transfomers met there are 300V or so: there is a path to ground through the transformer windings which is low DCR... I assume a couple of amperes would be flowing in each coil of the OPT and burn them. AC world gets even more complex.:::

    I may have missed your point here or simply misunderstand...

    but... as I see it... the plate voltage of say 300vdc is present at the "junction" of the where the plate, choke and output trans meet... correct... but this would also be the case if we were using a series feed arrangement (i.e., it would have the same dc potential at the juntion of the plate and the output transformer).

    An important difference is that (depending on the tailoring of the impedances and etc) in the parafeed arrangement we are likely to have less "backwash" into and through the power supply due to the higher PSRR of the parafeed arrangement.

    300vdc across the ouput tranneys primary or across (or through) the plate choke is of no consequence... likewise the 60 mils of plate current is of no consequence from either the output tranney or plate chokes point of view...

    The magnitude of the ac signal currents will be determined by the effective impedance load that the tube is working into... but this current will only bee a fraction of the capacity of either transformer from a copper current density point of view... or a heat loss point of view.

    so... I'm not seeing a problem... am I missing something???


    Mike




    Builder of MagneQuest™ & Peerless™ transformers since 1989
    DC problem, posted on October 8, 2009 at 12:28:38
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005
    		 I mean, in the schematic posted above the coil of the OPT will be tied at 300V at one end and ground at the other so a net current will flow which is 300V/whatever_the_DCR_of_the_coil_is. In the series feed the coil is hanging beetwen the B+ and the plate which is not ground.

    That's why John needs a cap beetwen the two OPTs coils to block the DC.
    Ciao

    RE: Missing parafeed caps in the schems, posted on October 8, 2009 at 10:25:25
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Thanks for setting me straight Gianluca; if I moved the parafeed cap to the center tap, would that suffice?

    Thanks for the compliment. I will be sure to keep everyone abreast of my series vs. parallel feed experiments.

    Operating points for the 6EW7, the high-mu triode is about 4.8mA, 121.8Vpk, -3.9Vgk and the low-mu triode is about 38.1mA, 152.4Vpk, -17.7Vgk. Those are average numbers, btw. The two examples I have aren't what you call matched. I trimmed the CCS (one is at 4.6mA and the other at 5.0mA) so that the operating points were, broadly speaking, comparable. Matt suggested that he runs the low-mu triode about where I have it, but the high-mu triode around 180-200Vpk and 5mA, so there's another possibility.

    Regards,
    John

    It would work. Schematic updated., posted on October 8, 2009 at 11:27:08
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005
    		 Thanks for the hints on 6EW7, I'll have 400V B+ so there will be room for a DC coupled stage.

    PS B) doesn't make that much sense to me.

    EDIT: I would make it even simpler. Get rid of the CCS under the 2A3s and ground the filaments (via the filament winding center tap of the power tranny) and apply the bias at the EXO173PPS.

    Gianluca

    more, posted on October 8, 2009 at 12:20:08
    gluca
    Audiophile

    Posts: 749
    Location: SomeWhere Islands
    Joined: February 17, 2005

    @ John

    Yes the -45V (that is 5*9, I mean do use 5 batteries to make it simple and dirt) goes to the PPS center tap. I used the EXO173 that way a number of times. And you will have no way to balance the currents as Mike suggested but the OPT is parafed so ... DC balance doesn't matter the OPT that much.

    @ Mike

    Definitively, I am more than curious to hear about differences between parafeed and series feed in PP. You know I like compact amps ... if I had the time ... way too many ideas.


    EDIT: added again the schematic for parafeed cap ... I lost it along the way

    RE: Adjustable balance differential CCS, posted on October 9, 2009 at 10:56:22
    John Swenson
    Audiophile

    Posts: 2013
    Location: No. California
    Joined: October 13, 2002
    		 I have yet another variant of this, last night I worked out a CCS design based on Gary Pimm's CCS that can be used for a differential amp but still allows for adjusting the current between tubes. With this you can put the output pair in true differential mode AND balance the voltages across the OPT so no need for either the cap between windings on the OPT or the CAP between the CCSs as shown in #B.

    As Mike mentioned at that point you don't need a separate phase splitter, the output stage IS the phase splitter, one tube's grid gets the signal, the other gets DC. As Mike mentioned the whole thing could be DC coupled so you have no caps anywhere in the signal path. And its still parafeed.

    I don't have a schematic program right this second to draw up the CCS circuit, so it will have to wait till tonight.

    John S.

    Diff CCS schematic, posted on October 18, 2009 at 23:42:43
    John Swenson
    Audiophile

    Posts: 2013
    Location: No. California
    Joined: October 13, 2002
    		 OK, I finally drew it up. Here is the schematic for the adjustable differential CCS.

    Its essentially a standard Gary Pimm SBCCS with an extra differential cascode. The voltage offset for each cathode output is adjustable so you can tweak the idle current through each tube.

    The voltage shown as 150V in the schematic feeds the voltage references, it should be between 150V and 400V max. If your B+ meets that requirement you can connect direct to it. Otherwise use a resistive divider to get it down low enough.

    The cathodes will HAVE to be at least 35V above GND at all times, note that is NOT the quiescent point but the lowest the cathode will ever go under geatest signal swing. If this is too much you will have to either run the GND terminal from a negative supply or bias the input at a positive voltage.

    John S.

    RE: Diff CCS schematic, posted on October 22, 2009 at 07:16:14
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Hi John,

    Sorry for the delay; I just saw your post w/ schematic. Thanks a lot for drawing that up.

    As a first pass, I was thinking of just using a single GP SBCCS in place of the Rk/Ck on the center tap of the filament transformer. I'm a little hazy reading datasheets, but it seems that the IRF820B should be okay with sinking 120mA as long as I use a hefty heatsink -- I was thinking of this one and mounting both IRF820B on the same hs. Thoughts? Am I missing something?

    Regards,
    John

    RE: Diff CCS schematic, posted on October 22, 2009 at 16:32:02
    John Swenson
    Audiophile

    Posts: 2013
    Location: No. California
    Joined: October 13, 2002
    		 Hi John,
    the idea here is to use the tradition differential topology which has a single CCS connected to both cathodes. The circuit I came up with gives you that but lets you adjust the voltage offset for each cathode separately so you can tweak the stage to get the same voltage on each plate. This lets you do a PP parafeed without ANY capacitors in the output stage AT ALL. The PP OTP connects directly to the plates, the primaries are connected directly to each other, no cap. This "center tap" is not connected to anything either. Because the idle plate voltages have been adjusted so they are the same there will be no DC current flowing through the OPT. (well nothing is ever perfect, there will be SOME current through it but it should be less than 1mA and Mike has stated that this should be no problem)

    It gives you the advantage of a true differential stage without any capacitors either in the cathode circuit or the plate circuit.

    With a true differential stage you don't need a phase splitter, the differential topology will do that just fine.

    John S.

    RE: Diff CCS schematic, posted on October 23, 2009 at 07:05:07
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Thanks John. I see what you're getting at and I'm sure as the project progresses I'll give this a go.

    I didn't think that the single SBCSS was a substitute for your proposal, just an easy first-pass on some differential action with a simple mod to the amp as-is. With the CCS you're suggesting, I'd need to generate a negative bias and I'd ditch the phase splitter to go 'compact', and that to me is more like a major rework, which isn't where I'm at right now. Give me a couple months, though. I'll keep you posted.

    Regards,
    John

    RE: Adjustable balance differential CCS, posted on October 9, 2009 at 11:55:50
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 I'd love to see that, John. Thanks very much for your comments.

    Regards,
    John

    RE: It would work. Schematic updated., posted on October 8, 2009 at 12:08:41
    mqracing
    Manufacturer

    Posts: 3314
    Joined: June 29, 2000
    Contributor
      Since:     May 7, 2004
    		 Hi GL:

    you wrote;

    :::I would make it even simpler. Get rid of the CCS under the 2A3s and ground the filaments (via the filament winding center tap of the power tranny) and apply the bias at the EXO173PPS.:::


    If we were aiming at simplicity as a virtue (which is prima facie a good first hunch imo) then I would be tempted to modify schema B as drawn by JB to the "compact" circuit... wherein the phase inversion is done in the output stage itself... tie the bottom triode's grid to ground...

    now we could dispense with the 173 phase invertor... and use a simple voltage amp in the front to drive the 2A3's... either via RC coupling, LC coupling or even direct coupling (i.e., see Keto's direct coupled compact).

    I like the concept\direction of schema B... and on paper parafeed PP does offer some advantages... advantages which imo are worth exploring. So I would not be overly anxious to abandon the parafeed strategy to go with a conventional series feed approach.

    Schema B as drawn (with proper implementation of the CCS's) should be able to keep the magnitude of the unbal dc plate current well below 1 madc. The S-240 output trans won't even blink an eyelid with this minimal amount of unbal plate current flowing through it's primary windings. It'll be fine.

    If I were building this amp for full range service... I would perhaps use a SUPER plate choke.... some guy with a greater amount of L.

    But least we have too many cooks in the kitchen...

    MSL





    Builder of MagneQuest™ & Peerless™ transformers since 1989
    RE: It would work. Schematic updated., posted on October 8, 2009 at 11:54:22
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Thanks for the suggestion, Gianluca -- that definitely sounds like something to try. The -45V bias supply would go to the center tap of the EXO173PPS; is that correct?

    Regards,
    John

    RE: It would work. Schematic updated., posted on October 8, 2009 at 12:13:22
    mqracing
    Manufacturer

    Posts: 3314
    Joined: June 29, 2000
    Contributor
      Since:     May 7, 2004
    		 but wouldn't you then lose the policing action of the ccs's command of enforcing a dc balance btwn the tubes???

    MSL



    Builder of MagneQuest™ & Peerless™ transformers since 1989
    RE: 6EW7_2A3 Parafeed PP, posted on October 6, 2009 at 14:02:58
    nullspace
    Audiophile

    Posts: 203
    Location: Philly, PA
    Joined: June 18, 2004
    		 Thanks for the comments Matt. I'll keep them in mind as I fiddle with this or that.

    Regards,
    John

    Page processed in 0.093 seconds.

    [ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2009, Audio Asylum, all rights reserved ] [ Go Green ] [ Take5 ]
    [ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]