News: Trouble logging in? Log in from the login page.
February 25, 2020, 04:23:43 PM

Author Topic: New OTs - ok "drop in" primary values: 2K6 for Kaiju 3K and 4K5 for SII 4K?  (Read 207 times)

0 Members and 1 Guest are viewing this topic.

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
My parafeed amorphous core versions of the Lundahl 1679 OTs just arrived. They along with some 70H Lundahl plate chokes are intended for a rebuild of my SII-45 in the hopes of coming closer to the performance of the Kaiju, but with the EML 45B sound signature (the working theory being is that a good part of why the Kaiju sounds noticeably better than my SII's is the the better Kaiju iron).

The primaries of the 1679 can be configured for 2K6, 4K5 and 9K7 ohms. My eventual plans are to experiment with some higher plate-k voltages and different grid biases with the EML 45Bs using the 4K5 and 9K7 options. Jac at EML provides some example operating points to try out and I'll start with them first.[[Edit 2 - put first edit in wrong spot] Edit 1: and of course that will mean new power supplies with new PT and I'm guessing the need for a new shunt reg too - I've got some of that lined up, but not all, so it may be a while before I actually take a crack at SII-45B rebuild]


BUT ... these new toys have arrived and I want to try them out NOW, "just to see".

@PB or @PJ: just to try them, can I drop them in configured for 2K6 in place of the Kaiju 3K OT, and for 4K5 in place of the SII 4K OT?  I figure the primary impedances are close to enough to stock to permit this, but I just want to make sure I'm not courting destruction of my beloved amps.

[Edit 3: and I realize that changing the primary impedance of the OT and/or the choke will affect the ideal parafeed cap value - but that's easy enough to figure out using PJ's 2L/R-sqr cap sizing formula. My concern is whether changing the primary impedance of the the OT will cause some kind of destructive mismatch between the output tube operating point and other parts of the amp - for example, the C4S]

MTIA, Derek
« Last Edit: February 05, 2020, 11:49:21 AM by Deke609 »
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 4974
No problem with either change.
Paul Joppa

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
Awesome!  Thanks PJ.
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
I decided to try out my new Lundahl 2765 headphone OTs instead of the fancy amorphous core 1679s b/c I was feeling lazy and the 2765s fit on the stock SII chassis - they're about the same size as the Kaiju OTs.  And their mounting holes are spaced identically to the mounting holes of the new 70H 60mA chokes. So 2765s and new chokes went into the SII-45.  The 2765 has 3 turns ratio options 12:1, 6:1 and 3:1.  I went with 12:1 going into my 200 ohm headphones - which translates into an insanely large primary impedance of 25K+ ohms. I've ordered some 50W non-inductive resistors in a variety of values from 18R to 300R, and will play around with using them to hit different primary impedances.

I've only listened to the amp for a couple of hours - and I've read that new iron needs 50-100 hrs to break-in -- so no detailed impressions to share yet - but I like what I hear. It's still far short of the Kaiju, but noticeably closer. It has the punch, but not the clarity, soundstage or shimmer. Maybe it it will get better. I'll post again in about 4 weeks with impressions.

 But one thing is for certain: zero noise/hum! The 12:1 setup is notionally a 4K6/32 ohms setup, whereas I was previously using the OT-2 as notionally 4K/16 ohms (which suggests an approx. turns ratio of 16:1).  I say "notionally" b/c, as I just recently learned, the stated impedances are only obtained if you actually use the specified secondary load - e.g., 32 ohms or 16 ohms.  But what I take from this is that the DCF circuit is really quite amazing. I thought I might get a bit of hum in going from 16:1 to 12:1 - but nothing, not even a hint with the *Quiet attenuators full open on both the Beepre and SII with no signal playing.  I'll try the 6:1 ratio sometime soon -- who knows, maybe even that will be quiet.

cheers, Derek
« Last Edit: February 09, 2020, 06:20:29 PM by Deke609 »
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
Rewired the LL2765s for 6:1 today just to see. Paired with my 200 ohms planar headphones, that comes out to about 7K primary impedance. I was worried about two things: (1) noise/hum and (2) too much power/volume. Neither turned out to be an issue.  I hear just a slight amount of what I'm guessing is 120Hz hum (it's so faint that I can't be sure) with no music in playing - I have to listen for it, and wouldn't notice it otherwise. Cannot hear it when music is playing, even in silent passages. So no problem there.  So hat's off to PJ for designing such a quiet circuit.

As for power/volume - I'm now thinking that the high power recommendations for these phones aren't off. I only had to turn down the BeePre coarse attenuator by a single click in going from 12:1 to 6:1, so about 9dB (with MourQuiet attenuators on the SII-45 fully open).  And wow does the system sound good.  It's like the phones sucked up the extra power and used it to make the bass and low-mids a lot fuller and tighter. Sounds amazing.  I am crazy pleased.   :)

I have no idea how big the planar headphones market is, but wow does the SII make a helluva planar headamp when fitted with lower turns ratio OTs. Might make an easy addition to the BH headamp line up. Not telling anyone how to run their business - just passing along some data/experience based on a setup that few if any other Bottleheads may have tried.

cheers, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 12155
I have no idea how big the planar headphones market is, but wow does the SII make a helluva planar headamp when fitted with lower turns ratio OTs.
I think you might also be hearing the constraints of an output transformer being used at too high or too low of a ratio.  There's no free lunch, and with a design like this, it's not unreasonable to assume that the middle of the three configurations will yield the best overall compromise. 

Here are some measurements of a very different Lundahl transformer, but they illustrate the same idea:
http://www.shine7.com/audio/LL1545a.htm

The Hammond 125 single ended transformers are another decent datapoint for this:
https://www.hammfg.com/files/parts/pdf/125D.pdf
The 125 is a little hobbled because the secondary is just one long winding with taps, so these effects can get really exaggerated for high step-down ratios using just a tiny bit of the secondary winding.
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
Thanks for this PB. Very interesting. I can't even pretend to understand why/how different loadings on the same windings ratio has such a big impact on high frequency roll off (and to a lesser extent, low freq. roll off), but the charts you linked to are pretty clear.

So this may indeed explain some or even all of what I'm hearing. I'll do my own freq. response tests sometime soon - maybe this wknd - and report what I find.

cheers, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Doc B.

  • Administrator
  • Hero Member
  • *****
  • Posts: 7951
    • Bottlehead
Don't get too wrapped up with matching the output impedance of the amp with the nominal impedance of the headphone. The manufacturer is most likely using an amp with a more standard (relatively low) output impedance when they design and voice the headphones, even high impedance ones. If you go to the trouble to match the amp output impedance with high impedance cans it may end up sounding fat. There's no real guarantee that theory matches the actual result with this stuff.
Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
Sage advice Doc. Thanks.

I'm just playing around with this stuff to see how different configurations turn out. I really wasn't expecting the 6:1 turns ratio to work. I only bought the LL2765 headphone OT's as an experiment, and really only for the 12:1 turns ratio (or rather, for the 4K6:32 ohms configuration, b/c at the time I had no clue about how the turns ratio and the load on the secondary determines the primary impedance). I was already getting some 70H 60mA chokes and amorphous core parafeed OTs with notional 2K6/4K5/9K7 primaries and 4/8/16 ohm secondaries. Adding the headphone OT's didn't cost me much if anything extra in shipping. So I got them "just to see", with my real focus being the more "traditional" OTs. Haven't tried the latter yet.  I have a whole bunch of listening experiments and frequency response tests to do first with different combos of parafeed cap values, turns ratios and secondary loads. All of which I will enjoy immensely.   :)

cheers, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 12155
I'd suggest getting together the equipment to do frequency response sweeps in the near future.  My original setup cost me about $50, so it's not a huge investment, but it will come in handy for what you're doing.
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Sr. Member
  • ****
  • Posts: 901
Re: New OTs - ok "drop in" primary values: 2K6 for Kaiju 3K and 4K5 for SII 4K?
« Reply #10 on: February 12, 2020, 03:08:09 PM »
Thanks PB. Already got the equipment. Cost a bit more than $50, but I think it was a sound investment (no pun intended). I really enjoy learning about this stuff, and I'm finally starting to get a bit of traction with some of the basic concepts. I think I'm in this hobby for the long haul. 

cheers, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4