Massively overbuilt 45 headphone amp.

Discussion in 'DIY' started by johnjen, Aug 4, 2018.

  1. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    So we have initiated a build project of a 1.5w/ch headphone amp using the venerable 45 output tube with 6J5 driver tubes.
    It will also be easily convertible to run 2A3 output tubes should the vintage (1938 RCA Radiotron globe/ballon) 45's fail.

    This thing will be monstrously massive with LOTS of BIG iiiirrrrnnnn.
    At last count there are 10 transformers of various types and sizes and a few more may be added depending upon our final configuration.

    In a word this is OLD Skewl tube design, pushed to work on high impedance headphones instead of low impedance speakers.

    We have already breadboarded the basic circuit and aside from the noisy makeshift power supply the 'magic' of tubes was obvious from the very start.

    This is the schematic we are basing this build upon and it will give you an idea of the analog section, but we have already considerably modified it.
    From here, https://www.dhtrob.com/schemas/#PowerAmps

    2A3 or 45 SE AMP.GIF

    The 'matching' power supply schematic will not be used, as a more 'robust' psu will be designed using 5 chokes, 5 filter caps, and separate filament psu's.
    We were originally thinking of monoblocks but the size of this thing was already getting out of hand :), so we 'scaled' back a bit, and we'll use just one mains xfmr with 2 secondary filter sections, one for each channel.

    One of the design goals is to be able to match the authority of The Rok amp in the extreme bottom end (below 20 Hz) with the same (or better) focus and finesse and LEDI (Leading Edge Dynamic Impact). And even in pre-prototype breadboard form, the bass was near the equal of the prodigious subsonic bass The Rok is capable of.

    This is why this will be a BIG iiiirrrrnnnnn amp, the plate chokes alone are rated at 64H which should enable the bottom end to reach down to the mid teens.

    But the Magic in the Mids (MitM) was clearly evident even from the crude breadboard, wired up in one day, pre-prototype.

    Once we get a 'proper' PS operating and use the oh so few tweako passive (non-iiiirrrrrnnnn) parts in the direct audio path (4/ch with 9/ch total, see the schemo above), things should start to blossom like crazy in terms of the overall SQ, especially with the MitM's.

    And yeah I know, boys and their toys and too much time on my hands etc, etc.

    This should be a kickass amp for my 800-jmod's and JggyB.

    JJ
     
    Last edited: Aug 5, 2018
    Jerry, atomicbob, highflyin9 and 6 others like this.
  2. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    So ordering BIG iiiirrrrnnnnnn is a PIA, especially when there is, say one or 2 pieces that are not 'in stock' and must be built (can you say up to 6 weeks).
    And granted this build is a tad unusual in how much BIG iiiiirrrrrnnnnn is being used given there are only 4 tubes, and that some of these xfmrs are build to order - only.

    Still the delay in actually being able to get all of the big ticket items together to be able to wire it all up will give us a chance to do layout and chassis work with less pressure to git'r done, at least quickly.

    And some of the resisterators and capacitators etc. are good to have on hand first so we can factor in their sizes etc.
    For instance, the 4.7µf 800volt caps are ≈75% the size of a hockey puck, not exactly a dainty item.

    But ordering them means that the selection of the big, heavy, large chunks needed for this amp are somewhat fixed now.
    Which in turn means we have the sizes of the xfmrs and can begin their layout, which in turn will determine the chassis size etc.

    It's sorta like dominoes all falling into place.

    JJ
     
  3. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    So here is an early component layout.
    And know it will change, especially during the build stage.

    JJ

    45 amp layout-6.pdf

    you'll probably have to zoom down the size, as it’s a big layout.
     

    Attached Files:

  4. highflyin9

    highflyin9 Almost "Made"

    Contributor
    Joined:
    Aug 3, 2017
    Likes Received:
    107
    Dislikes Received:
    0
    Trophy Points:
    33
    Location:
    DC/MD/VA Area
    Home Page:
    That's a big boy alright. I bet it'll be pushing 60-70 lbs. What AWG metal are you using for chassis?
     
    Phantaminum likes this.
  5. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    Thus far I figure I've got about 30lbs in BIG iiiiirrrrrnnnnnnn, so I figure, at least thus far, the amp will weigh in at ≈ 50lbs.

    Which is a bit over the top for 1.5watts/ch no matter how you slice it.

    JJ
     
    sodacose likes this.
  6. sodacose

    sodacose MOT: WTFAmps

    Joined:
    Feb 3, 2017
    Likes Received:
    34
    Dislikes Received:
    0
    Trophy Points:
    13
    Location:
    Wisconsin
    Home Page:
    Why the Lundahl 2743 plate load instead of the Sowter 8982 (or why not a LL2765 OPT)? Parts on hand?

    Looks like a monster in the works.
     
  7. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    Say there.
    We went with Lundahl because they are ≈ 1/2 the price, are (mostly) available from stock (faster delivery).
    Even so the Sowter parafeed xfmrs (8983) and the Lundahl Plate Chokes (LL2743) are still up to 6 weeks out.

    The LL2765 "headphone" output xfmr is a bit of an unknown and I figure a lowered output impedance will help provide more tightly defined bass.
    And the freq response of the parafeeds is -3dB at 10Hz which is more in keeping with our design goals.

    And those 100H 8982 plate chokes are roughly twice the inductance we figure we need, so we stuck with the design goals, instead of going completely off the deep end.

    We needed to show SOME kind of restraint… hahahahahahahahahahahahaha :)

    JJ
     
    Last edited: Aug 15, 2018 at 4:09 PM
  8. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    522
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Well Grounded
    Oh yeah I forgot the top plate gauge question.
    Really, I'm not sure yet.

    That part of the design process is still missing a few 'data points' to be able to 'properly' answer that question.
    But as a initial guess I'd say at least 3mm, (≈1/8") and probably thicker just due the span length of the entire top plate and how many holes and their size, like those 100µfd/440v caps need a 60+mm (2.35") hole, and there are 6 of them.

    JJ
     

Share This Page