A tale of two inductors.

Design and use of Chokes for PS, anode and filament use.
Post Reply
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

A tale of two inductors.

Post by dave slagle »

I was asked to do some CT chokes to serve as plate loads for a pair of PP drivers in a brook amp. I had access to the original brook choke and decided it would be wise of me to measure it and see what was going on and I must say it is a really interesting example.

Attached below are some preliminary measurements of the original choke and what I would do for the same situation which should open the door for some interesting discussions.

dave
Attachments
Multiply the Y axis by 100 for ACV@100hz.  Since It was a rare device that was not mine, II limited the ac to 125V on the Brook.
Multiply the Y axis by 100 for ACV@100hz. Since It was a rare device that was not mine, II limited the ac to 125V on the Brook.
Screen shot 2010-03-06 at 10.49.57 PM.png (34.23 KiB) Viewed 13727 times
Get Your Fix
www.hifiheroin.com
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

another split choke

Post by Aron8 »

Hi,

My first post. I am the guy who wants these chokes (for Brook amp clones).

In order to play music while I am waiting for Dave's chokes, I pressed into service a split choke from another maker who sells it as an autoformer phase splitter. This has a much lower DC resistance than the Brook original and less closely matched halves than the Brook, which Dave infers must has a Bifilar winding. I used some 2W Allen Bradley resistors in series with the ends of the new choke to approximate the total resistance of the original.

I haven't broken the temporary replacement in yet, so all of this may be premature, but my Brook 12A now sounds more "Hi-Fi", and more like other push-pull tube amps.

The low-level detail may even be improved, but there is an edgy quality to the sound, and the amp is less lucid and less relaxed. The live, unstrained-yet exciting quality that the amp had driving a single Spendor BC1, is largely gone. Clipping during loud passages is hard and more salient, and cracks and pops on records -which used to seem too separated from the music to annoy- are now more salient and irritating. Finally while there seems to be a more dramatic separation of instruments (even though I am listening in mono), that separation seems discrete and a little artificial, in the manner of overdone stereo effects or Viewmaster 3-D slides. Because of these changes the amp sounds less organic and is less emotionally moving.

I am technically ignorant, but I wonder if the bad non-linear behavior of the original Brook choke might not contribute to my amps exceptional qualities, particularly to it's single-ended like ease and ability to communicate emotion. Possibly by minimizing current differences between the two halves of the phase splitter? Does that make sense?

I would be more easily convinced that the amp sounded the way it did in spite of the non-linear choke, rather than because of it, if the behavior that Dave documented could result from common construction techniques. However if it would be hard to get that behavior then it would seem to point to a useful, if forgotten, Lincoln Walsh design innovation.

Could one easily wind a choke to behave this way? Does the apparent sloppiness make sense historically, based on state of the art in the late 40's early 50"s

I'm eager to hear an informed opinion on the question.

Thanks,

Aron
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hello Aron,
A typical autoformer/phasesplitter is only to be used for AC. So with DC you always have some kind of unbalance between the tubes making your autoformer saturate. If it also is bifilar things will get even worse.

So the unsuitable thing you have put into service is most certain the part to blame.

With a properly built CT-choke it should not contribute to any colouring.

If I have understood Daves measurements on the Brooks, it indicate that that one is also with a small gap. This makes it mandantory to use closely matched tubes.

Hey Dave,
Have you measured reactance at higher frequencies for the original Brook?
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hey Guys.

I cannot see into the brook chokes but i am guessing they are on a split bobbin not bifilar. The DCR's are within 0.05% at around 1875 ohms. Given the gauge of wire I suspect they are wound with (#40) that is insane. Assuming #40, we get just about an ohm per foot so over the course of 1800 feet of wire there not much room for error. I don't think you need to put too much emphasis on the DCR. Getting matched DCR's is as easy as removing a turn or 10 and in my experiences I rarely see DCR's matched this closely for the same number of turns. Even the stretching of the wire comes into play.

The biggest problem I see with the chokes is the low level inductance. Here is the above picture plotted on a Log scale. Looking at the dotted red line we see the inductance drops rapidly as signal goes down.
Image

Since the 6C4 has an Rp of around 7K we can calculate out the "theoretical" -1 and -3 dB points for the above plots. The plots below assume the choke is linear and In the real world this choke will show increased inductance as frequency goes down assuming you keep the drive voltage constant. I still think the graph nicely presents one of the concepts at hand which is with this particular nonlinear inductor your low end response decreases as signal level goes down.

Image

Replicating this behavior is really easy, just use a really crappy core and stack it with as small a gap as possible.

Now for the High end response. I think the goal here is to have wide matched bandwidth to prevent phase differences between the halves within tha audible band. The plots below show the two sections of the brook compared to the 4 sections of the ones I did for aron. I find the notch in the brook very interesting, first since it is there and secondly because they are identical. This in addition to the matched DCR's says they were very consistent. Mine behave well out to around 150Khz and have the expected bessel roloff. The resonance and behavoir between 150 and 200Khz shows slight differences between them but that is far enough away from the audio band for me not to worry. These measurements were done using a single ended signal on one winding with the other winding open. I'll take a peek at balanced drive later if i get time.
Image
Get Your Fix
www.hifiheroin.com
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

(Blushing)

Post by Aron8 »

Thanks Lars,

Of course! I forgot all about the fact that the part I swapped in is an autoformer meant to be used in a parafeed configuration. I'm such a numbnut.

So I have to acknowledge that my observations have no bearing on the contribution of the original Brook choke to the amps great sound.

But the presence of an air gap doesn't explain the non-linearity of the original choke does it? And I would still like to know if it the behavior that Dave documented makes sense as an unintended consequence of standard (or even sloppy) construction methods.

Thanks,

Aron
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Aron,

The way you gave the choke to me, there are two orange wires that were twisted together and had a single lead.

I assume that connected to B+ and the other two leads went to the plate of the 6C4's

Was this a factory wired amp?

dave
Get Your Fix
www.hifiheroin.com
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

Thank you!

Post by Aron8 »

Wow!

I guess you posted this while I was still typing my last message. Thank you Dave for all the work involved in this characterization.

And yes, the tap with two twisted wires spliced to one is the center tap connection (to B+). The amp was factory wired, but my step dad, who inherited it from his father, had it rebuilt before he gave it to me, and I assume that the extension was spliced onto the center tap during the rebuild.

OH, turns out not. I looked at a photo I snatched off the web of the underside on an unmolested 12A, and lo and behold, the same splice is clearly visible, so they were all apparently wired this way.

I am going to keep my (poorly informed) counsel until I have had a chance to listen to the linear choke Dave wound, and to compare it with the original. I'll post some observations then.

Aron
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

The glitch in the HF frequency response reminded me of a pattern That I have seen with dual winding chokes when only measuring one winding and leaving the other open circuit. Since both halves will be driven by 7K I used a balanced source and remeasured.

Lo and behold things changed the dip went up nearly an octave on the brook and the rest stayed the same. My chokes didn't show a difference with half the winding being driven SE or the complete winding balanced.

The dip at 40K didn't bother me too much and this makes me think that the HF is not the issue.

I spoke briefly with Aron and from my knowledge of the Nickel PP splitter choke (technically it is an inverting interstage) I believe it is minimally gapped and one winding has 10% more turns on it than the other (Aron, can you measure the DCR's of each half?) If this is the case then in order to have complete current offset one tube needs to be running 10% more current and be in the appropriate socket. Since I suspect that isn't likely I'll bet that that could be part of the sound.

Lets start with what I have wound and see what things sound like and go from there.

dave
Attachments
New plot of the balanced drive added.  the wiggle in the null is an artifact of to few samples (oh how I wish I could automate up to 500Khz.)
New plot of the balanced drive added. the wiggle in the null is an artifact of to few samples (oh how I wish I could automate up to 500Khz.)
Screen shot 2010-03-11 at 2.38.34 PM.png (42.58 KiB) Viewed 13647 times
Get Your Fix
www.hifiheroin.com
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

Post by Aron8 »

Hi Dave,

The substitute (parafeed) choke measures 394 ohms on one side and 406 on the other.

I can't resist commenting that the Brook choke is putting out more total energy above the audible range, even though its frequency response is awfully ragged up there. I bet you that this energy contributes to the amps sound, at least with speakers with extended high frequency response (like my Spendor BC1).

There is functional brain imaging data showing that while it is not audible, this part of the spectrum causes dramatic brain activations (both cortical and subcortical), and improves the subjective qualities of reproduced music. There is a pretty high threshold for this effect, so that the high ultrasonic energy output of the Brook choke could arguably make a real difference (see attachment)

The phenomenon in question explains the weird effects of super-tweeters, which cannot be heard and are usually only noticed when they are disconnected. It may also explain our subjective preference for vinyl sources.

Does anyone buy this?

Aron
Attachments
UltrasonicSOUNDSemotionpdf.pdf
(497.97 KiB) Downloaded 816 times
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

Delayed Report

Post by Aron8 »

Hi Everybody.

Apologize for the long delay. When I got the new chokes and the original back from Dave, and swapped them into the amp, I suddenly had motorboating. For some reason it didn't occur with the inappropriate (parafeed) choke that I had substituted temporarily while I waited for Dave’s choke, which he wound very quickly. After two days of teeth gnashing and fixes that didn’t work I finally got rid of the motorboating by removing the 12A's negative feedback.

Of course the gain was higher with the feedback disconnected, and I had changed several other things during my thrashing efforts to defeat the motorboating. But Dave's new choke sounded much better than the original in this new and unfamiliar context. It sounded so much better that I couldn’t bring myself to swap back and forth between the two chokes much at first. Right away it was clear that Dave’s choke sounded more vivid, more live, with less noise and a blacker background. Generated more of the feeling of excitement and dread (the frisson) that I value above everything else.

I realized, immediately, that I was wrong and Dave was right. Clearly the Brook sounds good in-spite of the original choke, not because of it.

Nonetheless, I was also more aware of overload and congestion during loud passages with the new choke, and I feared that it made the amp too revealing… taking away a pleasing euphony. And I still think that it is doing that, although only in a specific context that I will return to in a minute.

I had a hunch that I might be overloading my single Spendor BC1 speaker… that it wasn’t moving enough air, even in my smallish room, that I was pushing the Brook too hard and hearing overload and IM distortion more clearly. I read Dave’s comments about a system with three original Quad speakers per side (on the heroin blog that he links to here)… about this triple stack’s good (un Quad-like) dynamics. This made me want to hear a triple stack of BC1’s. I only have one working BCI, but I recently picked up a cheap pair of related speakers, B&W DM-3’s re-badged Sony (UK). These have the same tweeter and super-tweeter, but use a larger oval EMI woofer in a bigger more conventional thick-wall box.

And sure enough, when I wired kluged a double stack one of the DM-3’s in series with the Spendor and attached the resulting chimera to the 16 ohm OPT tap, the benefits of Dave’s choke were even more apparent. The most notable change was not increased headroom during loud or complex passages, but better downward resolution: an unmasking of low-level detail that gives things more color and more presence. This makes sense to me as lower IM distortion, and the benefits of allowing the amp to spend more time in its linear range, but it wasn’t what I was (naively) expecting.

Swapping the Brook choke back in now, with this more efficient speaker arrangement, produced a radically different sound that was more opaque, more distant, less live, less exciting, less immersive… but also familiar and nice. A sophisticated and burnished that somehow makes me think of patina on an old object… less immediate, less exciting, over-there instead of here-now, but familiar, and organic, and somehow comforting. And it occurred to me that with the original choke the Amp now sounded like a really good AM car radio. This prompted me to listen to some recordings that were intentionally compressed for radio play. Early Birds hits, 70’s soul hits, great compressed recordings of this sort indeed sound much better with the Brook choke, which has a similar, and in some ways very beguiling presentation.

By contrast, with my favorite recordings, and really with all good to great recordings Dave’s choke is vastly better sounding. The spooky subliminal-chills feeling that I love so much (and am chasing hard in a variety of ways), occurs a lot more often, and is more intense, and I am really happy.

That is my first take on the two chokes. I am going to add the other B&W DM-3 to make an even more sensative triple stack and also set up a permanent modular arrangement that allows me to swap chokes instantly. It is amazing how different the amp sounds with the Brook choke and with Dave’s choke, and I like the idea of having what amounts to different effect plug-ins to change the amps character for different music or different recording styles. I expect to use one of Dave’s chokes for most listening, and to plug in the Brook in occasionally for music that is recorded with high compression, or to remind myself of that familiar, comforting, (great song on the radio) Gestalt.

I’ll make another posting in a week or two, after further listening and rumination.

Thanks,

Aron
Aron8
Posts: 6
Joined: Mon Mar 08, 2010 5:09 am

gaps and autonomic thrill-seeking

Post by Aron8 »

Hello Again,

The chokes Dave gave me came with an amber mylar spacer that gives an inductance of about 200H and they sounded very good. Vastly better than the original in fact. After playing around with the shims, I liked the red spacer, which lowered the inductance to about 150H better, and the green, which brought it down to 112H better yet. However going up to the next largest blue spacer (70H) sounded much worse.

It seems to me that the smaller spacers give a more live, -loose- sound, but also distort more readily with complex material. Perhaps for that reason the smaller spacers add a sort of grunge. As I made the gap bigger the grunginess went down, but without reducing the live, charged-atmosphere effect that I like so much . At least until I hit the largest spacer, which sounded overdamped and uninspiring.

There seemed to be a pretty obvious sweet spot. Everything came into better focus with the green mylar spacer, and that (best) gap size minimized distortion/grunge without impairing the excitement that I am so addicted to. You know, that feeling of excited anticipation that is in a room even before the music starts (at least when a great performer is playing live). Feels like there is a mild electric charge in the air. This feeling definitely has an autonomic component: its like an attenuated version of the feeling you get when your hair stands on end, though on the thrill end of the thrill-terror piloerection spectrum. For me, the music that most reliably evokes that feeling is stuff with drones, or drone-like elements (synthesizers, concertinas, electric organs etc.).

I wonder if you all have discovered any approaches that made your equipment better at evoking this response. I know that you all are very technically minded, and that the question may be off-putting. But I am a biologist, and know that this is the subjective experiential correlate of a very real physiological state, and a pretty interesting one at that. I am very sharply focused on building audio that is better at evoking this specific effect. I think I've made some useful progress in this direction, and I would like to share what I've found, and kick some ideas around...

I suspect, as I mentioned before, that ultrasonic responses are involved, but I did want to bring up something else that I did while I was messing around with the choke gaps that also changed magnetic fields/gaps/linearity and also boosted the chills effect dramatically.

What happened is that I swapped an unusual variant of the Celestion HF1300 tweeter into my Spendor BC1. This HF1400 model has exactly the same moving parts and exactly the same distinctive focusing lens on the front cover (the line was apparently based on a very old compression driver and still bears the traces of that lineage). The only thing that distinguishes the two tweeters is that this one has a substantially bigger magnet.

When I put that driver in, the "galvanizing" effect of the system was dramatically enhanced. I cannot rule out changes in frequency balance resulting from increased driver efficiency, but I wonder if greater linearity (without over-damping) isn't also at play. Larger permanent magnet having a field-coil like effect?

Aron :shock:
Post Reply