1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

The DIY'rs Cookbook

84 85 86 87 88 89 90 91 92 93
  1. johnjen
    So an update.

    The BIG 45 amp has just passed the 200hr mark and has settled in quite nicely with the SQ seemingly not calling attention to itself in anyway…

    Except when it should, like when I crank on the MOAR knob.
    Those are the moments when I pause what I am doing and look over at the amp and go, 'did you just do that?'
    And as I finish the Danse Boheme from Carmen where it concludes with a thunderous drum hit, I just shake my head, smile and enjoy the amazement.

    And news from the HP/pre-amp is that this purple heart wood is hard as a rock, as in, it chews up drill bits like we snarf on popcorn, fast an furious.
    It seems it comes from the same family of woods as iron wood, for those that know about woods.

    It is being assembled and wired in it's purdy new chassis with operational functionality soonish…

    Yet another teaser…

    HP_pre-amp w_top plate.JPG

    gefski and Luckbad like this.
  2. johnjen
    Update - upgrade - report - status - news at 11.

    So I've been fussing with the BIG 45 amp a bit more.
    I added 15Ω resistorators to the output xfmrs.
    This effectively increased the load the output tubes and output xfmr 'sees' because instead of the spread of the load from ≈ 330 to 680Ω that the 800's present to the amp, the spread is now 14.4 ±0.1Ω, which is not only closer to the 16Ω load the xfmr 'expects' to see but also the Ω curve is MUCH flatter.
    This made a BIG (in keeping with the theme for this BIG 45 amp) change for the better in the SQ of the amp.

    The Magic in the Mids has taken a decided step up along with just about every other acoustic and sonic attribute I look for (see the 1st post for the list, hahahahahaha).
    I figure that adding additional load to the output helps in a number of ways beyond what has been mentioned above.

    I also have changed the output wiring from the 45 tube to the output xfmr with indeterminate results thus far.
    I can say, it hasn't resulted in a SQ reduction, but that may not be 'enough' since the aim is to achieve 'Better'.
    I'll need to give it ≈200hrs+ before I'll know if this last change is truly 'Better' or not.

    This 200hr runtime mark has been repeated numerous times and was again reinforced after I tweaked the Bias voltages and current on all 4 tubes.
    I figure that the 2 tweaks above will also 'blossom' after this 200hr mark is reached.
    And there may well be additional SQ improvements at greater runtime amounts, but I keep fussing with the circuit and so I never reach much beyond 2-300hrs of continuous , unmolested, operation.

    But the SQ at this point rivals the best this amp has ever produced.
    I feel my skull getting thumped when the very bottom end gets triggered by some extended bass note or concussive effect (think cannons, or those BIG 6' stand up drums).
    Now I just have to wait for ≈ 2.5 weeks at ≈ 12hrs/day, and this is the kicker, LEAVE IT ALONE, as in NO changes.
    Of course my will power only can be stretched so far, as additional parts ARE inbound,
    and, well, see, it's like this…

    There is only 1 set of caps directly in the audio path, in the entire analog section of this amp, which means they ARE the single weakest point in terms of SQ.
    And if we keep these parafeed output xfmrs we have to keep ALL the DC voltage from reaching them, but at the same time let ALL the music thru.
    Which translated means, I have another set of output caps inbound and they are much more closely matched to the desired spec of 5.7µf ± 0.2µf.

    And for those that know about tweako caps, they can all to quickly, get WAY beyond silly expensive, sorta like tubes and xfmrs etc.

    And in other news…
    The HP/pre-amp is moving closer, ever closer, to my domicile, literally, as the base of operations for it's build is 'on the move'.
    Sorry, there is no film at 11.

  3. johnjen
    I truly must be weak willed when it comes to tweaking my tunes.
    But ya know I think it's a good thing because it speeds up the forward progress, and significantly so.

    The Audyn 5.6µf @600vdc caps showed up today, and of course i HAD to install them, well, just because.
    The SQ immediately took a step up, and in unexpected ways.
    And that was after only ≈ 1hr of playing.

    What was most immediately apparent was an increase in the acoustic power delivered from cymbals and other high freq percussive instruments.
    Additionally horns and the like have also become more 'powerful' in their acoustic contribution.
    It isn't that they are louder, or their basic character changes, well other than saying it's more like the real instrument sounds.

    But it is that, there is more there, there, and viscerally so.
    And I see this as the result that more of the generated acoustic power is properly delivered to my ears, which in turn means less of that acoustic power is not focused where it shouldn't be, in the first place.
    Think about, that the 'set amount' of energy is being smeared thru time, which has now been reduced and instead has been delivered when and as it is supposed to be.

    Cymbals, like in a drum kit, are powerful.
    Think about it, you have a 12" or 16" round tuned plate of metal that is designed to ring from it's entire surface.
    And when whacked can fill a room.

    That presence, that visceral power just got turned up a notch or 2, with this new cap.
    And I still need to give it 200+hrs, but if this cap upgrade follows the same 'settling in' pattern as other tweaks with this type of immediate improvement right out of the gate, well…
    But there are at least 3 more 'refinement' experiments to follow, all centered around this cap change.

    And these results lend even more weight to my experiments dealing with the 'optimal' orientation of non-polarized film caps along with alternative placement of these caps within the circuit.

    So as the SQ continues to climb to new peaks, using all of the descriptors I have come up with, which individually and collectively all add up to 'Better', the net effect thus far, is that this amp, is the single greatest step up in SQ I have yet made.
    However it is standing upon ALL of the previous tweaks and improvements made thus far, which tends to skew this title of being the mostest, goodest, not to mention the amount and number of tweaks it has been the recipient of itself…

    And the one that is the most satisfying, strangely enough, is I5 (Intelligibility) where the words sung are more obvious and clear as to what the actual words are.
    So instead of trying to choose from possibilities, due to being able to hear the inflection it becomes more obvious what the meaning actually is.

    This in turn provides a greater depth to the appreciation behind the lyrics and indeed the entire song.

    Last edited: Mar 20, 2019
  4. gefski
    Thanks JJ, for nailing, once again, why we spend time and $ on this stuff.

    Throughout Head-Fi there are sometimes soooo many posts just talking about frequency response comparisons -- bass-mid-treble, bass-mid-treble, on and on, that the joy of music seems far off.

    Right now I'm listening to Ralph Towner guitar & Gary Peacock bass, the guitar string attacks and KNOCK on the body, supported by rhythmic power and presence of Peacock's foundation, has me, as you suggest, feeling I'm sharing exactly the alive interaction between these two. (And these aren't "phoned in" performances, they are IN that studio)
  5. johnjen
    Yeah many are drawn to the specs as THE major focal point in our hobby.
    But until there is the direct experience with SQ that eclipses those numbers and graphs, they tend to remain as the final arbiter of what is important, at least for them.

    And while numbers and graphs are helpful, for me, they are but the initial starting point, in considering to acquire a piece of gear, or not.
    And that's about where these numbers and graphs start and stop.

    For me it's the experience of listening to and getting swept up in the music, which has little to nothing to do with measurements, nor the technical side of all of this.
    And there are countless examples of gear that measure superbly yet fail to grab me by the ears and won't let go.

    But then we all have our own sets of proclivities and desired aspects we seek in listening to music.
    And of course since we as a group listen to music for different reasons, i.e. some listen TO the music while some listen FOR the music, this distinction alone helps us refine those experiences we seek when we have our systems play music for us.

    This in my mind is also directly related to the degree of 'Calibration' we have experienced, as in what is the 'best' we have ever heard and to what extent is attaining/maintaining that level of SQ (or exceeding it) important?
    We each will have a different answer, and I for one like to push the edges of the envelope from the technical side, mostly because I have the experience and understanding to do so.

    But the final decision to keep any tweak is solely based upon the acoustic results, not what the specs say.
    And for me the specs are a form of feedback of what I have changed, but not what I have achieved.

    My BIG 45 amp is prime example of this.
    It has eclipsed and set a new level of peak SQ, and I don't have a clue what it's freq response is.
    I have yet to measure it, as it isn't important at this point in the development of this amp, since it would be devoid of any useful significance, nor would it lead to any actionable steps to take.

    And as I sit here and listen to Eric Clapton's Unplugged album there is absolutely NO desire to want to know what ANY of the 'normal' specifications are for this amp.
    They just have no meaning while listening.
    And I'm anticipating when I play Walkin' Blues, because the stage riser acts like a drum head with his footfalls being the 'kicker'…

    Last edited: Mar 21, 2019
  6. johnjen
    So this cap change has delivered, BIG time (I couldn't resist :ksc75smile:), new peak levels of performance.
    So instead of waiting for those 200hrs to accumulate, I've ordered 2 sets of Hovland SuperCap bypass caps to push the total capacitance closer to 6µf to find out where the true sweet spot actually is.

    But at this point the dynamic quotent of the LEDI (Leading Edge Dynamic Impact) has risen yet again to new levels of visceral impact on all 'voices'.
    But where it's most noticeable is in the low end as all manner of bass has attained a degree of nuance and inner focus I have never heard before, at all, ever.
    It isn't how loud or powerful the low end is, rather it's how the related harmonics 'flesh out' the bass lines, as in how much more there is, there.

    But the mids are also improved and becoming more resolving, which among other aspects yields an increase in the perceived contribution the soundstage makes to the overall sound.
    This is related to both C3 and I5 (Cohesive, Coherent, Coupled & Intelligibility).
    Such studio aspects as the use of background echo, and other 'effects' are more easily heard as and what they truly are, all the while enhancing and melding with their 'parents' 'Voice' (the instrument that the effect is enhancing).
    And this is a wonderful addition to orchestral music as the hall now contributes more of it's share to the ambiance which adds richness and REALNESS to the music.

    And the tinkle tinkles on top are also more refined and REAL as that same increase in the degree of inner detail and focus mentioned above results in being able to more easily make out those subtle cues that reveal the true nature of the guitar plucks, cymbal hits, snare drum brush strokes, etc.
    Like whether it’s the plastic nubbin on the end of the drum stick or the wooden end that is tapping on the bell of the high hat.
    The 'acid test' will be if I can hear all of those little rivets rattling on a sizzle cymbal, AS rivets (and perhaps even be able to count them :ksc75smile:).

    So I'm rather pleased and await the results of the next few experiments (adding just a touch more capacitance to dial in this portion of the sweet spot) and get the new caps fully 'settled in' etc.

    And the re-design of this BIG 45 amp has some MAJOR design advances in store for its final build.

    I'll write up what is involved in separate posts, as these design aspects, while they are not new, except in how and where they are going to be applied, and what they will accomplish, are somewhat novel.
    AND they are in keeping with the overall design criteria of using old school techniques, simplicity, low parts count, no resistors nor caps (well, maybe 2 caps) in the signal path, and NO regulators (in the strictest sense of the term).

    And since I still need to wrap my head around some of the design and functional niftyness that this direct coupled, all single triode, BIG 45 amp, with all of ≈1.5 watts/channel of earthquake pounding power, will deliver, it might take me a bit to be able to write up an adequate description of what is being accomplished.

  7. johnjen
    So tonight I'm playing with how much is too much, vs. is enough, vs. is the sweet spot, on the cap values for the one cap in the audio path.
    5.7µf is enough, 5.9µf was too much, 5.8µf seems to be somewhere on or just over the peak, time will tell.

    And the difference between L & R runs from 0.01 to 0.0075µf, and so once again I lucked out in being able to match them this closely.

    The SQ is morphing with peaks from around the corner of major blossom action ahead.
    And there is this new degree of 'clarity' to the acoustic 'image' that I/we create in our minds eye.

    It's like another one of those veils was removed, ya know the one you never saw before, until after it's gone.

    Last edited: Mar 27, 2019
    gefski likes this.
  8. johnjen
    So I've settled down on 5.83µf ±.01µf as the sweet spot on these parafeed output xfmr dc blocking caps.

    It's a 3 cap stack made of an Audyn 5.71µf + Hovland SuperCap 0.9µf + a Hovland Musicap 0.022µf (all as measured, not rated).

    This is where I was back on the 3/27, only now I have added ≈60hrs and the SQ blossom factor is slowly and steadily increasing.
    And reaching yet greater levels of SQ as the hrs accumulate.

    And yet again I5 (Intelligibility) is the single most welcome improvement as lyrics become all the more clear and 'meaningful'.

    And so far I have 3 pages of written material on the designs for final build of the BIG 45 amp.
    I have yet to figure the organization and how much background detail to include in this write up so it will make sense and be comprehensible to those who aren't familiar with the inner details of tube amp design.

    And it appears at this point we will depart from a 'strictly' old skewl design approach and use some sillycone based devices, along with even more TUBZ.
    And it will be a directly coupled amp, which if you know what that means/implies is somewhat unique, except for DIY custom one off designs, which this is but one example.

    Generally 'mass produced' tube amps don't dare use a direct coupled design due to their ability, under some conditions, to fail catastrophically and in a cascade mode where when a 'stage' goes sideways, it trips the next stage, or the power supply or, or…
    This is mostly a design issue, or should be.

    So next I'm gunna play with the resistance I added to the output xfmrs and see where the sweet spot is.

    Last edited: Apr 4, 2019
  9. johnjen
    Oh yeah, I meant to add that those 0.022µf Hovland Musicaps are 'Very Special' in that they really help the top end all the way down into the upper mids.
    I was surprised at how much additional detail and visceral impact they contributed to such 'voices' as cymbals and even picolo's etc.
    And that they 'blended' all of this nuance and inner focus seamlessly into the entire bandwidth was unexpected as well.

    And the proclivity for tube amps to be able to self destruct is embedded in the basic's of what is needed to be able to run tubz, not to mention how they can operate.
    I mean when you have several farads of stored electrical energy at elevated voltages, which all adds up to a substantial amount of potential energy, and then if a component fails (which fortunately doesn't happen very often) there has to be a place for all that energy to go, or keep from going, as the case may be.

    This is all part of the designers art, to be able to accommodate any and all contingencies, regardless.
    These issues, from a manufacturing and support POV, HAVE to be handled, and the design HAS to "accommodate any and all contingencies, regardless."
    Or the business won't survive, for the obvious reasons.

    And direct coupled amps pose additional challenges in that there is nothing to block the output from one tube that then feeds directly to the input to the next tube, and so if the upstream output voltage should 'exceed the need', the next stage can also 'exceed the need' and that's when melt down can become 'spectacular'.

    The thing is, direct coupled amps are essentially the purest types of designs due to their simplicity and minimal parts count in the audio path.
    They also tend to minimize the phase shift as the signal passes thru the amp (no caps).
    And since direct coupled amps don't let the audio signal pass thru any interstage caps, (nor are there and bleed resistors), the music signal is 'altered' as little as possible since it's only traveling thru wire and tubes, with but one exception in this case.

    That being the parafeed DC blocking and coupling cap.
    Which is why I was working at optimizing this one critical component's placement in the audio circuitry, and dialing in the amount of capacitance to tweak the SQ.

    Some may know that while tubz are important in terms of the overall SQ of any tube amp, the output xfmr is actually more important and has a greater effect on the actual performance and SQ of the amp.

    But unfortunately, rolling xfmrs just isn't 'a thing'.
    And thankfully they just aren't 'plug and play', not by any stretch, (can you imagine what COULD happen if…?)

    And we chose parafeed output xfmrs because they are capable of extended low frequency response (these Sowter 8995's have a power bandwidth of -3dB at 5hz & 60KHz).
    And a direct coupled amp should be able to reach those power bandwidth numbers far easier than any other design, and this 2 single triodes (per channel) design is not limiting this bandwidth either.
    As such this amp is, even now as I type this, able to reach our original design goals, and then some.

    I keep looking over at it and saying "did you just do that?" and then shake my head in amazement as I listen to a Cat Stevens track from 1971…

    And for those who are familiar with my 'special re-purposed terms' (see the 1st post for the list) the amp is now getting S/S (Spooky/Scary) with a full helping of I HEARD thrown in for good measure.

    And I've only reached ≈ 80hrs, which isn't even half way to the magic number of 200hrs.

    Last edited: Apr 4, 2019
  10. johnjen
    So I'm hot on the trail of dialing in the added 15Ω load resistors to the secondaries of the parafeed output xfmrs.
    These 15Ω resistors which are in parallel with the 800's (≈350-650Ω impedance), which results in a 'reflected load' of 14.5Ω ±0.2 instead of 16Ω.

    Yeah this certainly seems to qualify as an exercise in the picking of nits, however…
    This relates directly to the load the tube 'sees', so optimizing the voltage and current it operates with, tweaks its 'environment' so to speak.
    Which in turn means it's operating in it's 'happy zone' and not being forced outside of that optimal envelope.

    At least that's the course of action I'm experimenting with to find out if these numbers will reflect in SQ improvements, or not.

  11. Royaume
    Help!!! I am buying an Yggy with 110V transformer. I need to replace it with 230V transformer. Can you offer advice and recommend good audiophile-grade replacement transformers? Am I even taking the right approach? All help greatly appreciated.
  12. johnjen
  13. johnjen
    So during my output impedance fussings, I have tried using 15Ω and 20Ω and 23Ω load resistors and determined that,
    #1 this tweak has a much bigger impact than I would have thought.
    #2 even tenths of an Ω can make a difference.
    #3 nailing the final Ω raises the SQ bar yet again.

    These small changes from 15Ω to 20Ω to 23Ω clearly make an obvious difference and it is relatively easy to know fairly quickly, as in within 1 hr of playing time.

    Due to the math, 0.1Ω changes to the secondary of the output xfmr, at these low values (≈20Ω), makes a reflected load change of ≈15Ω, which the tube sees directly.

    So it looks like a 16.9Ω to 17.0Ω is the optimized resistor value, which I have concocted using 3 resistors.
    And the SQ has peaked yet again, with only ≈5hrs of playing time.

    I'm hoping to find a matched pair of single resistors that meet this spec., but 17Ω is an 'uncommon' value to say the least.

  14. johnjen
    Back in Jan of 2016 I posted about
    Phase Δ Reduction,
    A Thought Experiment…


    And in it I used the analogy of how a moiré pattern is created and how as layers of the same pattern were rotated wrt each other, complex moiré patterns would emerge then simplify, then morph into complex patterns and so on as the layers continued to rotate towards perfect alignment, where all that remained was the original pattern.

    Then take this analogy one step further and imagine the original pattern is the music signal, (with it's layers and their degree of alignment), and as it passes thru the circuit, with various degrees of being 'unmolested', so to speak, we can approach ever closer to just the original musical signal as the alignment approaches perfect alignment.

    I bring this up again because the BIG 45 amp (which currently has 12 xfmrs, 10 PSU filter caps, 4 battery packs, 4 tubes, and only 2 caps and no resistors in the signal path) is demonstrating this moiré pattern as the amp continues to settle in and as the soundstage and all 'Voices' within it, come into tighter focus and greater resolution, the SQ continues to raise the bar on more or less ALL of the sonic and acoustic attributes I have identified and use to determine what, in fact, truly is 'Better'.

    It also reinforces my idea about how, as the number of CP's (Choke Points) are reduced, those that remain, when ameliorated, yield much greater than expected results.

    This last tweak of dialing in the Ω on the output xfmrs has accentuated this process, but again in unexpected ways.

    For instance the improvement to the overall SQ took a major step up with the 15Ω load resistors, and stayed, but changed slightly, with the 20Ω loads, and then the SQ fell off with the 23Ω loads.
    The peak SQ returned and has reached new heights after I changed to 17Ω loads, where it remains as the hrs. accumulate (now at ≈ 50hrs)

    The moiré patterns are still coming and going but the range of complexity (the degree of being out of focus vs, being in focus) has now narrowed considerably.
    In that now when the complexity is at max levels, the SQ is still better than in the past when at max complexity.
    And when the original pattern (music signal) is molested the least, new peak levels of SQ are set (the performance bar has risen yet again to new heights).

    IOW the max complexity has reduced max levels, along with the max alignment also gaining tighter and tighter focus with greater resolution for all aspects of the soundstage.

    And right now as I listen to Eric Clapton's Pilgrim album and as the amp is transitioning from complexity to alignment, as the SQ is blossoming,
    I can't take my 800's off, yet again.

    And additional developments are brewing both for the BIG 45 amp final design and build, along with the HP/pre-amp.
    And as each step in the design process results in fewer and fewer parts, every portion of the circuit performs even more functions, all at the same time.
    This is a 'sideways' way of saying as the circuit design is more and more refined fewer and fewer parts remain, at least in the direct audio path.
    The power supply on the other hand is getting more sophisticated (read as complex), yet keeping with the same multiple functions for each portion of the circuit.

    Yeah we're having some fun now… :ksc75smile:

    Last edited: Apr 12, 2019
  15. johnjen
    So a ways back the term REALNESS emerged and was associated with using the AOIP method of delivering the digital audio signal to the dac.

    My BIG 45 amp is now passing into its second half of settling in, on its way to full Blossom Action.
    IOW I'm at ≈100hrs where usually between 100-125hrs the SQ blossoms and in wonderful ways.
    And then does so again after ≈200hrs as I have mentioned before.

    But what is not just peeking from around the corner, but is running straight for me is, REALNESS^4 (REALNESS to the 4th power).

    For me REALNESS began with the enhanced sense of 'being there', of hearing deeply into the music as it is being played, as if I was wired into the master control mixer, like listening to the master recording.

    Then there was REALNESS^2, which brought the left to right, and up to down sense of placement, which resulted in being able to locate each 'Voice' in 2d space.
    Then there was REALNESS^3, which added front to back, and with this added sense of depth, how the acoustic environment 'held onto', and then released, this 'stored' acoustic energy.

    Tonight REALNESS^4 is coming to the fore where the entire soundstage is being presented with a new sense of wholeness along with a duration thru time.
    This ability to deliver the whole acoustic soundstage and maintain it with stability and cohesiveness, yields a new degree (REALNESS^4) to the ongoing 'stream' of the music, as it is being heard.

    It's like there isn't a drum hit over there, or a guitar over here, etc.
    Rather it's music, and as it is being played before me in a REAL way, it isn't disjointed nor segmented into this 'Voice' and that 'Voice', but as a whole where all 'Voices' are seamlessly integrated into the soundstage AND where the sense of dimensional placement and the full acoustical context is 'built in, not added on'.
    AND where all of the acoustical energy being created 'comes from' it's parent 'Voice', thru time, which is tied into tLFF and less smearing of the creation of the acoustic energy thru time.

    You could also describe this experience as, the music is playing in this ongoing continual NOW.
    And this applies to ALL of the music, not just the 4 and 5 star rated tracks.

    Which yields hearing my music as if being heard anew, yet again,
    one moar time with FEELING…!


84 85 86 87 88 89 90 91 92 93

Share This Page