Messed up my chassis cuts -- need advice about layout

Guest · 1241

0 Members and 1 Guest are viewing this topic.

Deke609

  • Guest
As I've mentioned in other threads, I am rebuilding my Kaiju.  I made a mistake in laying out and cutting for the PT.  The attached pic, "Layout 1" is what I intended: axis of the PT coils pointing along the back of the amp where there will be a big CLCLC rectifier/filter.  But when it came time to layout the PT, I went into auto-mode and unthinkingly followed the stock PT layout -- so now the PT is pointing directly at the plate choke. I can potentially deal with that by rotating the PC so that it's coil axis is at 90 deg with the the PT coil axis.  But then the PC points at the 5670. See attached pic, "Layout 2". Merlin Blencowe Morgan Jones only mentions beam tetrodes as needing to be sheltered from magnetic fields, but I wonder whether the 5670 driver tube should be too.

I am inclined to modify the PT cutout and mounting holes in line with Layout 1. It's doable, and both the PT and the PC will have steel transformer covers, so the metal work errors won't be visible. And more importantly, the steel enclosures will attenuate the magnetic fields a bit (and I will likely add a layer of amorphous-type metal film shielding to the inside of the PC enclosure that's reported to reduce magnetically induced noise by 30 dB).

But before I start cutting again, I need a sanity check. What are the BH experts' best guess as to the "safest" approach: layout 1 (original intention) or layout 2?  Also: in case it's not clear from the labeling on the pics: the coils of the OPTs point straight up (vertically). 

I will drill for both layouts in any event so that I can, if necessary, switch orientations if I run into a hum problem -- but I'd like to avoid having to do that by doing the initial rebuild using the layout that is least likely to cause me problems.

(And, yes, I know I should have built a prototype ...but it all looked good on a 1:1 scale drawing which failed to capture the mis-orientation of the PT).

MTIA, Derek

[edited to change "Merlin Blencowe" to "Morgan Jones"]
« Last Edit: August 18, 2020, 09:55:36 AM by Deke609 »



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19319
Reply #1 on: August 17, 2020, 04:37:25 PM
I would make the corrections and cover it all with cans.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #2 on: August 17, 2020, 04:42:28 PM
Many thanks PB. That's my thinking too, but I just wanted to check.

cheers, Derek



Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5751
Reply #3 on: August 17, 2020, 06:54:24 PM
It's more complicated. The simple rule of the coil axis applies to EI laminations (and to the far field). Your plate choke is a shell core, so a large portion of the core is exposed running 90 degrees to the coil core.

In Layout 1, the portion of the plate choke core that is adjacent to the output transformer has vertical flux lines. This is counterbalanced by the opposite side, but the distances are very different, so the adjacent side will predominate at the output transformer. The output transformer is not visible, and I can't see what kind of core it has, or how it is oriented, but it might well interact with the plate choke. As you can tell from the stock arrangement, we have not found that to be a problem, since both plate choke and output transformer carry the same signal - but the potential is there and I have not done enough experimental work to make definitive statements.

The plate choke is well placed with respect to the power transformer in Layout 1, though there is a small possibility of interaction if they are not vertically centered on each other.

The same vertical flux lines are present in Layout 2, but the core is more distant from the output transformer and the vertical flux lines of the plate choke go in opposite directions. Again, the output transformer is invisible so I can't see how susceptible it may be.

The plate choke/power transformer orientations in Layout 2 look good.

It's difficult enough to visualize these flux lines in two dimensions, three is much worse!

Paul Joppa


Deke609

  • Guest
Reply #4 on: August 18, 2020, 03:14:29 AM
Many thanks PJ.

Quote

The output transformer is not visible, and I can't see what kind of core it has, or how it is oriented

The OT is dual c-core finemet. The manufacturer advises that the coil center axis runs perpendicular to the chassis/OT-bottom-plate - so that if you were looking down at the OT (as mounted) directly from above, the coil center would hit you in the eye. I'm unable to figure out more b/c the OTs are potted and cannot be opened and checked. (My original description of the OT's coil orientation as "vertical" was confusing and unclear -- I meant perpendicular to the chassis, with both the PT and the PC oriented horizontally, either north-south or east-west, as viewed when looking down at the chassis as if at a map).

Quote
In Layout 1, ... The output transformer ... might well interact with the plate choke. As you can tell from the stock arrangement, we have not found that to be a problem, since both plate choke and output transformer carry the same signal ...

That was my thinking. I am (whether justifiably or not) more concerned with the PT - PC interaction than with the PC - OT interaction.

Quote
The plate choke is well placed with respect to the power transformer in Layout 1, though there is a small possibility of interaction if they are not vertically centered on each other.

[...]

The plate choke/power transformer orientations in Layout 2 look good.

The iron will be centered, but I can't guarantee precision. I'm expecting to be within +/- 2mm of dead center.

So, both layouts look good in terms of PT - PC relative orientations.  In that case, I think I will go with layout 2 b/c on further study I've realized that making layout 1 work is more involved than I first thought -- I'd have to install the PT on a small plate that mounts over top of the chassis (doable, but time consuming), and  the rectifier diodes and voltage doubling caps would need to be squeezed in between the PT and the first PSU choke - assuming it would be doable, it would be very tight and a real PITA to wire and solder.

One final question (hoping I''m not trying your patience): would you expect any adverse consequences of having the PC coil axis point directly at the center of the 5670 tube?

Postscript: And to make clear for anyone who might later read this, the rebuild is NOT motivated by any sense that the Kaiju needs improvement. Quite the opposite: it is my favorite amp and b/c it sounds so good I want to tweak it so that it works as a dedicated headphone amp for my LCD4 headphones. Previous experiments with different iron revealed that I like having more power delivered to the headphones - or perhaps what I like is the sound of the Beepre when the Beequiet is closer to center rather than closer to fully open (less harmonic distortion?). 32R secondaries sounded better to me than stock 16R, but I felt that I could have done with a bit more. But 150R secondaries was too much -- using Doc B's phrasing, it sounded "fat" -- every note had too much of a halo around it. So I ordered custom parafeed output transformers with 16/32/64R secondaries.  But the new OTs are huge -- too big to fit on the stock chassis. Hence the rebuild.

All that said, the larger chassis also gives me an opportunity to experiment with regulating the B+ feeding the 300Bs. I was blown away by the driver shunt reg upgrade in the Stereomour II - so I want to hear what if any difference voltage regulation makes when applied to the output stage. I recognize that the output stage has much less voltage gain than the driver stage, so the benefits of regulation are likely much greater at the driver stage.  But maybe there will be an audible difference even at the output stage.

cheers, Derek



Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5751
Reply #5 on: August 18, 2020, 05:19:54 AM
It is possible that some tubes will be affected by magnetic fields, but I think significant effects are very unlikely. I have never seen any analysis of the effect except for cathode-ray tubes used in old televisions....

Paul Joppa


Deke609

  • Guest
Reply #6 on: August 18, 2020, 06:24:39 AM
Excellent. Many thanks again, PJ.

cheers, Derek



Deke609

  • Guest
Reply #7 on: September 08, 2020, 04:37:39 PM
Done!  After fixing a dumb wiring mistake, the rebuilt Kaiju is cranking out awesome headphone goodness.  Here are a couple of pics. the last shows the rebuilt Kaiju and BeePre together.

The big black boxes toward the front are the OPTs with rotary switches below for switching between 16, 32 and 64 ohm secondaries. The OPT primaries have a 3K3 impedance, so very close to stock.  The rows of little toggle switches toggle parallel load resistors -- 18R, 39R and 100R -- that when parallel with my 200R LCD-4 headphones result in an effective load that's pretty close to the spec'ed 16, 32, and 64 output impedances.

The PSU has a CLCLC filter: 440 uF electrolytic voltage doubler caps (220's in parallel), 10H choke, 200 uF film cap (tall centered cylinder that extends another 2 inches below the top), 10H choke, and 1500 uF film cap (fat silver guy in the corner).

The additional HV solid state regulator didn't work out -- not sure why. I must have messed up stuffing the board somehow. Whereas the stock shunt regs tightly peg the driver anode voltage, the 450VDC coming out of the additional reg bounced around with line voltage fluctuations. And not from a lack of slack - I was feeding it 470 VDC, and it's only supposed to need a couple volts over and above the target output voltage. I'll play around with it and see if I can get it working. I'd still like to hear what if any difference it might make to regulate the B+ feeding the 300B's.  In the meantime, I removed it and added a dropping resistor between the first choke and first film cap to get 450V B+.

The new iron only has a few hours of use, but my initial impressions are very positive. I think the Hitachi marketing that Finemet combines the punch/power of silicon steel with the micro-detail finesse of amorphous steel is true (to my ears). 

I'm a happy camper -- not the least b/c I am finally done.  I don't even want to know all the hours I put into rebuilding the amp. Most of the time was spent on the metal work and agonizing over layout.  My appreciation of BH kits has only increased as a result.  It's so much more relaxing to put together something that you know will work and sound great as long as you follow the steps carefully.  Although I voted for a mega power tube amp in Doc's survey - I'm now really hoping BH comes out with a new guitar amp (or just re-releases the Tode).  That would be awesome.

Edit: and zero audible hum. Thank heavens! 

cheers, Derek
« Last Edit: September 08, 2020, 04:45:32 PM by Deke609 »



Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5751
Reply #8 on: September 08, 2020, 05:03:40 PM
Congratulations - that was an epic journey!

Paul Joppa


Deke609

  • Guest
Reply #9 on: September 08, 2020, 05:32:18 PM
Many thanks PJ -- I really appreciate all the input from you and PB about this amp and all the unrelated crazy notions I get worked up about. I will undoubtedly have more of the latter when I turn back to the output stage voltage regulation issue  ;D  And I think I will in fact need to add thermistors to soften the inrush current spike - i blew a bunch of fuses this morning.  No problem with a 1.5A fuse if I ramped up voltage slowly with a variac. But a hard power on vaporizes them instantly (even slo-blo's). Had to resort to a 3A fuse, but that's not ideal from a circuit protection perspective.

And now that the Kaiju is rebuilt, i have the stock Kaiju chassis just waiting for a new build. I think it would work nicely for a a rebuild of one of my Stereomour II's.  Probably the 45 (which now has the Kaiju iron).  But that's for next year. I fully scratched my major-building-project itch for the year.

cheers and thanks, Derek



Offline 2wo

  • Hero Member
  • *****
    • Posts: 1240
  • Test
Reply #10 on: September 08, 2020, 05:36:42 PM
Congratulations it looks great!

" I don't even want to know all the hours I put into rebuilding the amp. Most of the time was spent on the metal work and agonizing over layout.

But isn't that the best part? 😋...John

John Scanlon


Deke609

  • Guest
Reply #11 on: September 08, 2020, 05:58:18 PM
Thanks John.

Quote
But isn't that the best part? 😋

Hah! Yeah, I don't know. These projects are kinda like my yearly canoe camping trips. In March I'm pouring over maps looking for those less traveled lakes in the interior of Algonquin Park. "Less traveled" means lots of long portages.  "Bah", I think, "It'll be awesome". Fast forward to mid-June. Peak black fly season. Me with pack on back and canoe on shoulders, a literal cloud of black flies under the canoe with me just 2 km into a 5 km portage. "This is the stupidest thing I have ever done. This is brutal. Never again. Next time: easy out and back."  Following March I'm planning something even more insane... 

cheers, Derek



Deke609

  • Guest
Reply #12 on: September 25, 2020, 03:46:34 AM
Just posting a couple pics of the internals -- I always enjoy looking at pics of layout and wiring more than external aesthetics.  And sharing some listening impressions and speculations.

The big open spot in the middle with exactly 169 hand-drilled holes(!) was for a high voltage regulator that I've failed to make work with the amp.  So now there's just a big vent in the middle of the amp -- not a bad thing, I guess.

The second pic shows the parallel load resistors that, in parallel with my 200 ohm headphones, let me match whichever OPT nominal output impedance I switch to: 16, 32 or 64 ohms.  There's a toggle switch for each load resistor, so I can listen both with and without.  Without the parallel load resistor, the reflected load that the output tube sees is much higher, particularly when the OPT is set for 16 ohms and loaded with 200 ohms.  PJ has advised in other threads that the higher reflected load reduces distortion a bit, but at the potential cost of some bandwidth/frequency response.  More recently, on Audio Asylum, PJ commented that a going theory (which he did not necessarily endorse, but simply reported) is that 2nd harmonic distortion contributes to the experience of "air"/"space" in the music, with more 2nd = more air/space, and less 2nd = less.  [Correction] "dynamics" versus "air/space/clarity" - where more 2nd harmonic = more dynamics and less clarity, and vice versa.

My listening impressions when toggling the parallel load resistors in/out are may be consistent with the above theory -- although perhaps there is more at play than just increasing/decreasing 2nd harmonic distortion when I do this. 

Without the parallel load resistors, music is tighter and more precise, but the soundstage loses depth -- it's as if all the instruments are in line across the front of a stage, or occupy a two dimensional plane. Whereas with the load resistors, the music occupies a 3 dimensional space, with different instruments/notes located in different left/right, fore/aft, up/down positions.  Pretty cool.  In the stock build, I tended to prefer the tightness and precision of using the amp without parallel load resistors (particularly the added precision of bass notes). But the new Finement OPTs are very detailed and punchy, and I tend to prefer the spaciousness I get from the parallel load resistors -- depending on what I'm listening to.

cheers, Derek

« Last Edit: September 25, 2020, 06:13:18 AM by Deke609 »



Offline Thermioniclife

  • Sr. Member
  • ****
    • Posts: 752
Reply #13 on: September 25, 2020, 04:48:40 AM
Congratulations Derek, It's amazing what some people can whip up in no time at all, LOL. Lotsa sweet parts and transformers in
there. where did you get those opt covers, they look nice. And is there any chance that 2 Lundahl's could be squeezed in one?
Enjoy the fruit of your labor. What's next?

Lee R.


Deke609

  • Guest
Reply #14 on: September 25, 2020, 05:27:55 AM
Thanks Lee.

It's amazing what some people can whip up in no time at all, LOL.

Yeah, those 100 or so hours just flew by   ;D

Quote
  where did you get those opt covers, they look nice. And is there any chance that 2 Lundahl's could be squeezed in one?

The Monolith Magnetics OPTs in front  came potted in covers - I really like the finish on them - matte with sandy texture.  I got the covers for the plate chokes (behind the OPTs) here: www.aliexpress.com/item/32851122312.html?spm=a2g0s.9042311.0.0.57444c4dvw8dMT You could probably squeeze two lundahls into one of the two bigger ones. They're powder-coated steel. I am pleased with the quality. No nicks or scratches or visible seams, and the steel seems to be a good 1 mm or so thick.  Amazingly, since the beginning of the summer shipping from China to Toronto has been faster than USPS from the States. The covers took less than 2 weeks to get to my door - I paid a few extra $ for Ali's express shipping.

Quote
What's next?

I'd like to figure out what's going on with the series HV regulator - but I don't want to keep desoldering/resoldering the Kaiju rebuild - some of the wiring is already getting a bit kinked/messy, and it's just a matter of time before I absent-mindedly melt one of the big film caps with the barrel of the soldering iron.  So I think I'll breadboard a simple parafeed output stage and see if I can get the reg to work with it.

And I should really get around to building the mains power buck/boost box that I have all the parts for but just never get around to building -- just not that exciting. 

I have dreams of a parafeed amp with both driver stage and output stage regulated and CCS-loaded (no plate chokes). But that will require some high voltage for the output stage -- e.g., circa 900VDC regulated if I go with the 300B, and that's after whatever the regulator drops. I think I'll go with shunt regulation, and I've only seen tube shunts. I suspect they drop a fair amount of voltage. So 1kV raw B+?  A bit scary. And I think I'll need to build the CCS myself -- and I don't even know enough to have a mistaken sense that this is something I'm ready to tackle.  Much more learning required.  But I already have some IXYS reg ICs that can handle 900V -- I just need to figure out how they work  ;D

[edited to correct some obvious typos]

cheers, Derek
« Last Edit: September 25, 2020, 05:36:48 AM by Deke609 »