Fresh build ... first [and second] impression[s]

Guest · 2199

0 Members and 1 Guest are viewing this topic.

Deke609

  • Guest
on: August 03, 2019, 11:49:13 AM
After several marathon assembly sessions, I finished my Kaiju w/ DCF upgrade last night (base is still unfinished wood). Building this beast is a pretty good workout - it is remarkably heavy. When building I like to periodically raise the chassis overhead and shake loose any stray lead and wire clippings.  Hoisting the Kaiju overhead took some effort, with a death grip on the chassis motivated by equal parts fear of dropping and damaging the amp and fear of dropping it on my head and knocking myself out. No joke!

I added flux bands to the PT and OTs, and copper tubing/braid shielding to pretty much any wiring longer than 2 inches carrying signal or high AC voltage.  All shielding/flux bands connected with a single wire to chassis ground.  Film caps: 3 X 3.3 uF V-Cap ODAMs paralleled with 0.1 V-Cap CuTF bypass (10 uF total) as parafeed, and 0.1 uF CuTF for interstage coupling.  Tubes: stock GE Jan 5670s and fresh EML 300Bs.

I'm glad I went with the ODAMs instead of an expanded chassis to fit the gigantic Miflex 10uF caps.  Even an additional 4 inches of width (the max I could have accommodated on my audio rack) wouldn't have been enough.  With the ODAMs installed on the stock chassis I think I still have room to squeeze in a film cap as cathode resistor bypass down the line.

Based on past experience, the caps and EML tubes will take 150+ hours to burn-in, so I won't have a sense of the final sound for a few weeks.

Nevertheless, one thing is clear: this thing is RESOLVING. Paired with my Beepre, and feeding my Audeze LCD4 headphones, the Kaiju's imaging and focus is an obvious step up from my SII-45 and -2A3 amps. Even with no burn-in it is stunning. I'd assumed that I'd pretty much gone as far with resolution as I reasonably could with my setup - not so! 

This has me wondering what accounts for the greater resolution.  As recently confirmed by PJ and PB, the circuit of the SII w/ SR and DCF upgrades is essentially identical to the circuit of the Kaiju with DCF.  I see three possibilities: (1) better output iron; (2) the 5670 driver tubes; and (3) the spacious component layout and separation.  I will explore the first two in my SII-45 sometime soon.

Now it's time to catch up on all the pay-the-bills work I put off while working on the Beepre filament voltage and Kaiju build  ;D  This stuff really is addicting. 

cheers and happy building/listening, Derek  [edited to correct typos] [Edit 2: changed thread title]
« Last Edit: October 10, 2019, 02:34:45 AM by Deke609 »



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #1 on: August 03, 2019, 01:18:30 PM
(1) better output iron
I've had the experience of listening to a Kaiju with Magnequest TFA-2004s against mine with OT-5s, and indeed the stock iron is very good.  The Stereomour transformers are quite nice and represent a step-up from what you'd get from a $300-500 pair of series feed output transformers, while the Kaiju iron presses beyond the performance of iron at any price point.

(2) the 5670 driver tubes
They are empirically a step up from the 12AT7s.

(3) the spacious component layout and separation.
This is least likely to exert much influence on the final results IMO.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #2 on: August 03, 2019, 03:03:11 PM
Many thanks PB.


I'd like to explore upgrading the iron on the SII-45 and reconfiguring it to use 5670 drivers (understanding that'll lose a bit of power doing so). But I'll wait until after completing my project of designing/building a super simple parafeed amp. In the process I hope to learn enough to either figure out how to do the switch to 5670 myself, or at least enough to ask the right questions about how to do it.


cheers, Derek
« Last Edit: August 03, 2019, 03:14:32 PM by Deke609 »



Deke609

  • Guest
Reply #3 on: August 03, 2019, 03:13:58 PM
@PB: In the Kaiju there is a 5 terminal solder strip connected to one of the PT's mounting bolts. Only 3 terminals are used - one for ground connection of the PT, and 2 as wiring junctions for AC line voltage in. Are the 2 unused solder terminals being saved for something, or can I clip them? They take up valuable real estate that I could use for my parafeed caps.


Many thanks, Derek



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #4 on: August 03, 2019, 05:43:55 PM
We cut those two terminals off in some other kits to get the power transformer close to the wood base.

For the Stereomour running 45s, the 5670 is a more viable option than it would be with the 2A3, but you'll probably want to use a bypassed cathode resistor or the TL431 biasing like we use in the Kaiju to avoid losing too much gain.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #5 on: August 04, 2019, 04:07:59 AM
Thanks PB. I will clip off the extra 2 terminals. In retrospect, I'm embarrassed for even asking whether I could. If in the future I need the extra 2 terminals, it will take all of maybe 10 minutes to install a new 5 terminal strip.



Deke609

  • Guest
Reply #6 on: October 10, 2019, 03:51:48 AM
I recently listened to both the Kaiju and my SII-45 fed by my Beepre. The Kaiju had more than 2 months of regular use, so I think it was pretty well settled in. Both amps were configured for 16 Ohms output impedance feeding my Audeze LCD 4 headphones. The SII-45 uses EML 45B tubes, while the Kaiju uses EML 300B. The SII-45 was used as a straight power amp by setting the Mourquiet attenuators to 0dB attenuation. Both amps have the DC filament upgrade, and the SII-45 also has the Shunt Reg upgrade (a version of which comes stock with the Kaiju). So as close to apples-to-apples as I can get. So time for a second impression.

First off - a surprising result: my SII-45 sounds just as loud as the Kaiju at the same BeePre attenuator settings (normally, for me, -18 dB on the course attentuator, -0 dB on the fine). This is surprising since the the SII-45 is capable of no more than 2W per channel versus 8 wpc for the Kaiju. But this finding is consistent with my previous impression that my SII-45 sounds noticeably louder than my SII-2A3.  I don't know what explains this - perhaps the SII-45 has more 2nd and/or 3rd order distortion and the human ear/brain translates higher 2nd and/or 3rd order distortion into the experience of a louder fundamental? Dunno - just wild speculation on my part.

The Kaiju has a plainly wider bandwidth, with most of the addition being, to my ears, in the treble and it is here that I think the Kaiju really shines. The Kaiju sound signature seems centered in the higher frequencies, with incredible detail and space.  By comparison, the SII-45 sounds lower mid-centered. The highs are there in the SII-45, but with much less detail and more as accents to the mid-sounds that seem to "lead" or "center" the music.  Bass-wise, the Kaiju again has incredible detail and is quick and hard punching -- but, to my ears, the Kaiju's bass performance is still dominated by the treble, which leads and centers the music. The bass of the SII-45 sounds woolly by comparison - powerful, but lacking the detail and precision of the Kaiju.  With the Kaiju, I can hear the timbre of a string on an upright or even electric bass, whereas with the SII-45 I just hear the note. At the same time, though, the bass of the SII-45 seems to work harmoniously with the mids, resulting in a sound signature that is, as previously stated, lower mid-centered. 

Overall, the Kaiju is clearly the technically superior amp. But for simple, non-analytical listening enjoyment, the SII-45 still holds it own. Having concluded that the Kaiju was clearly better, I was all set to power down the SII-45 and re-insert the Kaiju but couldn't bring myself to do so ... "just one more song". And that turned into a few hours of really enjoyable listening, with my head happily bouncing to the music. For me, the "magic" of the 45 is real and the SII-45 has lots of it (not just with the EML 45B, but also with my other pairs of 45 tubes - I just prefer the EMLs). If I had to pick between the amps for very long listening sessions or background music to be played all day, i would probably choose the SII-45. It is never fatiguing; whereas, to my ears, the Kaiju can be after more than a couple of hours - it really is that detailed and precise.

So my next endeavor (after rebuilding my sadly dismantled BeePre) will be to rebuild the SII-45 with new and bigger iron on a bigger chassis to see if I can add some of the additional treble, detail and punch of the Kaiju.  I am leaning towards getting custom 4K M6/Mumetal OTs from Sowter, to be paired with one of their big plate chokes (maybe even their 100H / 50 mA version if that offers enough DC current handling).

cheers, Derek
« Last Edit: October 10, 2019, 04:13:24 AM by Deke609 »



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #7 on: October 10, 2019, 05:39:29 AM
First off - a surprising result: my SII-45 sounds just as loud as the Kaiju at the same BeePre attenuator settings (normally, for me, -18 dB on the course attentuator, -0 dB on the fine).
The amps have about the same amount of gain.  Gain and power are independent, so no conclusions can be drawn about power based on which one is louder at a given attenautor setting.  You could put a pair of pentodes into the Stereomour to drive the 45s and make it unbearably loud with the BeePre turned all the way down, and similarly you could redesign the driver stage of the Kaiju to use 6CG7s and you'd have to turn the BeePre up a bit more to get to the same listening level.


So my next endeavor (after rebuilding my sadly dismantled BeePre) will be to rebuild the SII-45 with new and bigger iron on a bigger chassis to see if I can add some of the additional treble, detail and punch of the Kaiju.  I am leaning towards getting custom 4K M6/Mumetal OTs from Sowter, to be paired with one of their big plate chokes (maybe even their 100H / 50 mA version if that offers enough DC current handling).
Be sure to make the plate 1/4" thick if you make it much larger, especially if you order 100H chokes, otherwise you may find that the top doesn't stay very flat. 

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #8 on: October 10, 2019, 05:52:30 AM
The amps have about the same amount of gain.  Gain and power are independent, so no conclusions can be drawn about power based on which one is louder at a given attenautor setting.

Huh! Well, lots more for me to learn here. I just assumed that more power meant at least some additional voltage gain, not just more current.

Quote
Be sure to make the plate 1/4" thick if you make it much larger, especially if you order 100H chokes, otherwise you may find that the top doesn't stay very flat.

Many thanks. Good advice.  I was thinking of using 1/8" but reinforcing with some 90 deg angle aluminum the full width of the chassis.

cheers, Derek



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #9 on: October 10, 2019, 06:01:50 AM
Huh! Well, lots more for me to learn here. I just assumed that more power meant at least some additional voltage gain, not just more current.
More power means more available output before distortion becomes problematic.  Take the Pass F4, for instance, which is a 25WPC amp with 0dB of gain.

Amps that put out a lot of power will tend to have more gain so you don't have to apply a ridiculous amount of signal to the inputs to achieve full power. 

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #10 on: October 10, 2019, 06:23:33 AM
Well, "huh!" again. I will read up on this.

I should have mentioned that I also plan on using the 5670 with Kaiju reg bias scheme in the new SII-45. I'll have to build my own boards - not sure whether to try my hand at designing a pcb and getting it professionally fabricated for me, or just McGyver something with protoboard.  I figure the latter approach should work since there aren't many parts.

cheers, Derek



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #11 on: October 10, 2019, 06:48:26 AM
Well, that will likely give you the effect of the lower power amp having less gain.  The SMR has a few dB of loss in the level controls and some extra step-down in the output transformer, as well as some loss of gain from the unbypassed driver stage cathode resistor.  This is offset a bit by the mu of the 12AT7 being higher than the 5670.  In the Kaiju, the biasing arrangement of the 5670 acts like a well bypassed resistor and the output transformer has a bit less step-down. 

There's also a bit of difference in gain of the output stages as well.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #12 on: October 10, 2019, 08:02:58 AM
Many thanks PB.

The SMR has a few dB of loss in the level controls and some extra step-down in the output transformer, as well as some loss of gain from the unbypassed driver stage cathode resistor.

If by "level controls" you mean the Mourquiet, this won't be an issue b/c I'll won't be installing it in the rebuild b/c I only use the SII-45 as a power amp with attenuators set for -0 dB.

If the resulting dB loss of switching to 5670 drivers is less than 6dB, then all is good.  If only -3 dB or so, I won't miss it at all.

If the step-down from the OT is a function of the 4K impedance, then I am stuck with that b/c I don't want to play around with operating points since I believe that would have implications for how much voltage swing the driver stage needs to provide, which in turn I suspect would make the shunt reg upgrade no longer usable - for me, the amp wouldn't be worth the effort of rebuilding if that meant leaving out the shunt regulator upgrade.

But if the step down is the result of, say, transformer losses, then maybe new iron might decrease such losses?

cheers and thanks, Derek
« Last Edit: October 10, 2019, 08:34:52 AM by Deke609 »



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #13 on: October 10, 2019, 10:41:36 AM
If by "level controls" you mean the Mourquiet, this won't be an issue b/c I'll won't be installing it in the rebuild b/c I only use the SII-45 as a power amp with attenuators set for -0 dB.
With the attenuators cranked all the way up, they still have about 4dB of insertion loss. 


If the step-down from the OT is a function of the 4K impedance, then I am stuck with that b/c I don't want to play around with operating points since I believe that would have implications for how much voltage swing the driver stage needs to provide, which in turn I suspect would make the shunt reg upgrade no longer usable - for me, the amp wouldn't be worth the effort of rebuilding if that meant leaving out the shunt regulator upgrade.
The 45 is pretty easy to drive, so you don't have to be too worried about that.  The Sowter parallel feed transformers that are 5K with the nickel sandwich would be an OK choice for what you're looking to do.  The 4K transformer will give a little more power and the 5K transformer a little less distortion.  There are some other differences that can be more pronounced with a more substantial change in loading impedance.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #14 on: October 10, 2019, 12:16:23 PM
Many thanks again PB.  This is very helpful info.

So leaving out the attenuators gives me a 4dB bump  ... maybe that will offset the loss from using 5670s.

When you say the Sowter's would be an "OK" choice do you mean, "Meh ... they'd be ok , I guess" or "Yes, they're well-suited"?  B/c if they'd  only be "meh", I can go with Electra Print - maybe their silver plated copper windings (don't think I can afford all silver). Or some other winder, if someone can suggest one (that is reachable). I just figured that since the MQ iron everyone raves about uses nickel interleaving, the Sowter mumetal sandwiches would be the way to go.

cheers and thanks, Derek