Double filament choke wound on single core

Design and use of Chokes for PS, anode and filament use.
Post Reply
mseddon
Posts: 35
Joined: Thu May 05, 2005 12:04 pm
Location: perth australia

Double filament choke wound on single core

Post by mseddon »

Hi, what happens if one winds 2 x choke coils for 2 unrelated but similar DC filament circuits on a single core?

Coils could be fed either 1) current summing, or 2) opposing for DC cancellation.

There would be inductance in both circuits, and looked at as an output transformer, coupling between the windings might offer useful cancellations.

My applicaton is for 2 output tubes per side on monobloc chassis (biamp), so channel separation is not an issue.

Any merit in the idea?

rgds,

martin
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Double filament choke wound on single core

Post by dave slagle »

mseddon wrote:Hi, what happens if one winds 2 x choke coils for 2 unrelated but similar DC filament circuits on a single core?
the AC ripple will become the same for the two nodes of the circuit since the coils will couple very well to each other at the ripple frequency.
Coils could be fed either 1) current summing, or 2) opposing for DC cancellation.
I think the ripple always is "in phase" with the DC so if the DC cancels the inductance will cancel too.
There would be inductance in both circuits, and looked at as an output transformer, coupling between the windings might offer useful cancellations.
common mode chokes are probably a better way to look at it.

I think cigna has a few ideas on how to get the ripple to sum and the DC to offset... Dave???

The whole in phase out of phase thing gets a bit confusing, if the ripple is in phase both will see the inductance, if they are out of phase, they will only see the DCR and leakage inductance.

dave
Dave Cigna
Posts: 96
Joined: Sun Jan 23, 2005 10:58 pm

Re: Double filament choke wound on single core

Post by Dave Cigna »

I think cigna has a few ideas on how to get the ripple to sum and the DC to offset... Dave???
Huh? Me? Well, I have had a few ideas. Unfortunately none of them work. I have, however, always learned something. Actually, I keep learning the same thing; I'm a dolt. It might be fun to put together a booklet of 'bad circuit ideas.' Anyone that has read Horowitz and Hill knows what I mean.

Here's one of my favorites. The input choke has two separate windings, one for each diode. The DC part is in opposite directions in each winding, so that cancels much in the same way that the DC cancels in the two half-windings of the transformer secondary*. As for the AC (ripple), only one diode at a time is conducting, so there's nothing to add or cancel. Whichever diode is conducting should see inductance in the winding that it's feeding, right?

It doesn't work. A prize to anyone that can figure out why. It's not an easy one.

* If you try to make a center-tapped tranformer from two separate transformers (separate cores) both cores will saturate because both see DC current. That DC cancels in a CT winding on a common core.

-- Dave
Attachments
an idea that doesn't work
an idea that doesn't work
BadIdea1.gif (5.55 KiB) Viewed 9731 times
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Whichever diode is conducting should see inductance in the winding that it's feeding, right?
then essentially each winding is its own half wave rectified supply. Doesn't that put DC on the core? Since only one side is conducting, how can the other side offset the DC if it is off?

did the choke you tried have an airgap? could you possibly have been saturating the core?

just thinking out loud,
mseddon
Posts: 35
Joined: Thu May 05, 2005 12:04 pm
Location: perth australia

Post by mseddon »

Interesting tangent, but I posed this as two separate circuits, ie 2 supplies, 2 loads. The only link between the circuits is a common choke core, and I suppose a common mains primary winding.

In the simple case where the DC currents are in same direction there seems no particular effect except the effect of the doubled DC / AC components on the core. ie if the circuit conditions are identical and parallel they could be substituted by one circuit of double current.

In the case where the DC directions oppose, DC effects on the core cancel presumably. AC components which were in phase are now out of phase.

Seems like one should get ripple A, minus inductive emf B, minus coupling C.

Therefore if coupling is perfect one gets a net component B and the choke ceases to work as a choke.

Oh well - no free lunch as usual - but at least seems one can wind 2 chokes on a single core in same direction - purely as a convenience - wouldnt you agree?

martin
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Oh well - no free lunch as usual - but at least seems one can wind 2 chokes on a single core in same direction - purely as a convenience - wouldn't you agree?
yes. essentially the two chokes are in parallel (coupled by the core) which makes me wonder what gains there would be over simply using a single winding. (or a clip pair of clip leads electrically bridging them) I guess it would be a good experiment to try.

dave
Dave Cigna
Posts: 96
Joined: Sun Jan 23, 2005 10:58 pm

Post by Dave Cigna »

Sorry to have hijacked your thread, martin. As dave suggested, I have thought about this sort of thing in the past and have convinced myself that there is no way to get it to work as you originally hoped.

If you'll let me indulge my hijack just a little more, I've redrawn the circuit above with the dual winding choke shown as a CT choke. If you look carefully, that's how the dual winding unit is actually used. In my own experiment I used a pair of identical power transformers with the second performing the job of the choke. Neither core saturated. At the point marked B+ I got a DC voltage MUCH lower than expected.

-- Dave
Attachments
The same idea redraw for clarity.
The same idea redraw for clarity.
BadIdea1b.gif (4.95 KiB) Viewed 9705 times
mseddon
Posts: 35
Joined: Thu May 05, 2005 12:04 pm
Location: perth australia

Post by mseddon »

Not at all Dave, it's quite apposite. DaveS opened up my specific application pointing out one might as well share a common conductor too. However as my filament circuits will be at different DC potentials I will need two windings.

The circuit drawn appears to cancel the whole half wave ripple - ingenious!! So yikes - no B+ at all...

An inductor joke book full of these will be a Christmas best seller. I think. well I'd like one anyway.

best rgds, martin
JeffreyJ
Posts: 315
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

ok.. maybe a bit out of my element here, but...

Post by JeffreyJ »

hello dave, dave and martin... :)

if you were using two power transformers, then I am assuming no airgap... then the dc would saturate the "ct choke", correct? isn't the dc flowing always in one side only (alternating 120 times a second) and therefore not offsetting like we want?

part 2: I always wanted to use the same core with two windings (dc going opposite directions) for chokes two and three in a power supply... but then I am assuming that I will just be transposing the ripple from winding one onto winding two... basically building a 1:1... but would it work if I had two seperate supplies for left and right and the windings are L1right and L1left? perhaps a way to meet critical inductance in a smaller package? maybe we could then afford nickel?

Peace,
Me
Post Reply