Dual-Coil Choke for Tube Preamp

[CLS]

Hi there,

I got a pair of Lundahl LL1673/15H which consist of 2 coils each. I'd like to use them in the power supply of my new line stage.

Conventionally, to make good use of the whole choke, using them in series would get a larger inductance, which is mostly desireable.

But since there're 2 coils in one choke, I got an idea to use them as 2 chokes in the same iron core. (sorry if this is confusing).


Please see the schematics below.

The 1st two are conentional, coils in series as one, serve in CLC or LC.

The 2nd group add one more cap in the middle of the 2 coils.

And the 3rd group reverse the current flow in the latter coil, and also add a cap in the middle.

So, which one would be the best in filtering? What will happen when I reverse the latter coil?

 

[Sch3mat1c]

The coils are linked by the core so are better thought of as a transformer. Any voltage you put on one will appear on the other, and vice-versa. Only series and parallel arrangements will work for a single filtering job; you cannot put them between caps as they will fight the caps and put out essentially the same signal that a single CLC arrangement (using just one winding) would. With the added trouble that, since the windings are fighting each other, current runs up (depending on input ripple voltage) and may cause heating of the choke.

You can use it for filtering two supplies of the same voltage and arrangement (i.e., CLC and CLC, or LC and LC, not mixed), but that is rather redundant. Plus the voltage coupling negates any supposed "crosstalk" issues that some more eccentric power supply "designers" would not consider "kosher".

Tim

 

[Eva]

Assuming both coils are wound on the same ferrite or iron core and neglecting leakage inductance:

Series connection will result in four times the inductance of a single coil, half the saturation current and double the winding resistance.

Paralell connection will result in the same inductance as a single coil and the same saturation current, but half the winding resistance.

The coil loses most of its inductance when the saturation current is reached, so series connection may be practical or not depending on load current.

That may appear weird at first, but that's how inductors work.

I think that using a coupled inductor for a single LCLC or for two independent LC filters is not a good idea.

 

[CLS]

Thanks a lot for your replies

The coils are linked by the core so are better thought of as a transformer. Any voltage you put on one will appear on the other, and vice-versa. ...

Yes! a transformer.

Let me explain what I have & what I want to get from these.

I'm building a line stage pre amp with built-in crossover. To maintain the driving capability of each signal path, it became 4 channels, i.e., 2 for mid-high & 2 for bass. Or you may treat it as 2 pre amps in one chassis....

The gain device is 5842 (one for each channel of course), with output transformers at their plates....

I group them 2 by 2 for the power circuit. 2 tubes in each group share the same rectifier (6X5) & LC filter. Then finally, each tube gets its own big RC.

....

So, here I "refined" the original idea a bit by combining the 2 power circuits into the same iron core. (please see the attached diagram)

I made 2* LCLC into 2* dual-coil choke. Since the final loads for each circuit are mostly identical, and the upstream rectifiers are also the same, so their current & ripple must also be identical. (well, I know our world is not that perfect, I can tolerate)

Making the 2* 1st L in the same core & out of phase, this whole choke looks just like the pri winding of the PP OPT

And the following 2nd L is arranged in the same way.

So, " Any voltage put on one will appear on the other " and they are in the same strength & out of phase, they cancel each other out! Also, like the pri winding of PP OPT, the net flux in the core is nearly 0, so the inductance would be high & can not be saturated.

How's that? Is it better now?

 

[Eva]

You are still missing the point, that circuit would allow ripple to pass with little attenuation. You should try harder to analyze it, until you understand what happens.

 

[CLS]

Aren't those ripple in common mode & can be highly suppressed by such arrangement?

 

[Sch3mat1c]

You have it differential mode! Any voltage (or since it cancels, I should say current) appearing on one winding directly opposes the other and nothing gets done.

Tim

 

[CLS]

Thank you guys still hanging on with me, and believe me, I'm trying not to be annoying

But I'm confused, more confused.

Please see the diagram below, blue arrows means current & greens mean flux.

Once again, would you please point out the mistakes?

Thanks.

 

[Eva]

A differential filter appears as a short in common mode due to coupling between both windings, but the ripple that you want to filter is differential mode.

When you try to apply equal voltages to two equal windings in opposite directions, both windings act as if they were short-circuited.

 

[CLS]

...both windings act as if they were short-circuited.

Short-circuited, I suppose you mean in AC?

Sorry I can't follow.

Please see the diagram below, I just thought it's almost the same as a pp output stage.

 

[CLS]

I figure it out now

Inductor works when flux changes.
Cancelling out the flux, then there's no reactance.

I'll lookin for another solution.
Thanks for messing with my confusion.

 

[Eva]

But you have two channels and two dual inductors, isn't that enough?

PD:
http://www.jacmusic.com/lundahl/html/power-chokes.htm
http://www.lundahl.se/pdfs/datash/1673.pdf

 

[Sch3mat1c]

I just thought it's almost the same as a pp output stage.

It would be if the tubes were driven inverted, as they are in a PP amp. Rectifier circuits can't even show an inverted signal (it always looks like a series of humps), let alone shift the phase to make it look right. THAT is why your choke would've been useless.

Instead you have to make them bounce in sync. Dotted end of the winding to rectifier, other to caps.

Tim

 

[CLS]

Understood & thanks a lot.

I just finished the project (well, sort of).
Eventually I used them conventionally in series in the LCRC filtering.

So, that's 15H-> 220uF (with 68k bleeder), followed by individual 2k-> 1000uF for each CK5842 (with 47k bleeder).

The main transformer gives 210VAC, instead of 230 in the specs. (I'm a little confused because the output of all the rest taps are very accurate. This Lundahl unit is also very well built.... )

Via one & a half hybrid bridge by 2* 6X5GT & 2* 1N4007, I got 189V on the LC with 0.7V ripple (measured by DMM).

Then the final RC brings the voltage down to 155V with 0.52V ripple. Well, not quite up to the target, but acceptable.

I'll share the rest of the project in the Tube forum.