Tape wound core self resonant

[bowman1710]

Hi Guys,

I guess this question is an extension of a previous post from before:

https://www.edaboard.com/threads/343789/#post1468131

Basically in summary I have a high power, high turns ration transformer that needs a SRF of higher then the 20KHz upper frequency band. Now I am looking at nanocrystalline (linked below).

**broken link removed**


This material has low core losses at the higher frequencies, my only concern is the permeability of the material. Does the high permeability of this material mean that the self resonant of the core will be dragged right down?? Will the high perm (60,000) also mean that is any DC is present then the core will drop straight into saturation also?

Any help welcome to answer my questions

 

[schmitt trigger]

High perm materials usually are not very tolerant of DC bias. Unfortunately the curves provided are not very clear on this.

But you can use cut "C" cores halves, and introduce an air gap. Of course this will also reduce the effective perm.

 

[bowman1710]

I was then thinking you can get cut toroid cores, this was my initial thought, so I guess my thinking is correct then, this will be an issue with having this material as it is.

 

[mtwieg]

If permeability is high, then you should be able to reduce the number of turns to decrease inductance.

Self resonant frequency is often more limited by winding geometry than by core material.

 

[Warpspeed]

Basically in summary I have a high power, high turns ration transformer that needs a SRF of higher then the 20KHz upper frequency band. Now I am looking at nanocrystalline (linked below).

The core material itself does not resonate, and any resonant effects will be due only to a combination of inductance and parasitic capacitance of the actual windings themselves.

There are some time proven techniques to winding inductors with high SRF, and the key to this is reducing the capacitance and voltage gradient between adjacent turns.

One method is a simple single spiral coil, which is probably the ultimate, but often just not practical.
A classic example is the ubiquitous EHT transformer winding in old CRT TV sets. These have thousands of secondary turns, and typically about 1 Henry of secondary inductance and also require a very high SRF to work during the short horizontal flyback period, as well as a very high voltage rating.

Another solution is honeycomb or wave wound coils, and these are often broken up into multiple sections along a common core. Old style vintage RF chokes were often wound this way, and had surprisingly high inductance with high SRF.

If you are into really high power, another solution might be to use multiple smaller transformers with the primaries in parallel, and the secondaries in series.
The advantage is the turns ratio of each transformer can be kept low, and if you use a large cross section core, few turns will be required keeping the SRF high.

For instance four 4:1 ratio transformers so connected would produce an overall ratio of 16:1 without needing to have sixteen times the turns on one secondary winding.
The SRF of each transformer should be very easy to keep high, and the overall SRF combination would be no lower.
It may also be easier to design and wind (for example) four 500W transformers than one 2Kw transformer.

 

[bowman1710]

If permeability is high, then you should be able to reduce the number of turns to decrease inductance.

Self resonant frequency is often more limited by winding geometry than by core material.

The turns have been limited by the acceptable core losses on the material so the inductance that is created by the turns are fixed, so to speak. Trying to pick a material/winding technique is where I am at this point in time. I think I have a similar situation that Warspeed mentioned about EHT transformers.

The core material itself does not resonate, and any resonant effects will be due only to a combination of inductance and parasitic capacitance of the actual windings themselves.

Yes, I understand that the material itself does not resonate, my concern was with a core with high permeability would cause the secondary inductance to be very high, meaning that the minimal of capacitance would cause a low SRF.

Another solution is honeycomb or wave wound coils, and these are often broken up into multiple sections along a common core. Old style vintage RF chokes were often wound this way, and had surprisingly high inductance with high SRF.

I've never heard of these winding techniques before, are such winding techniques possible with a Toroid core shape?

If you are into really high power, another solution might be to use multiple smaller transformers with the primaries in parallel, and the secondaries in series.

I did think about this, but I was concerned about what extra complications this would bring to the table such as current circulation, or other issues cause by any difference in winding differences between the transformers

 

[Warpspeed]

Here is a youtube of a honeycomb coil being made :
https://www.youtube.com/watch?v=TXP1-fdYKVU
This winding method makes a more bulky coil, but it will have a low self capacitance and much higher SRF than a conventional layer wound coil.

The reason being the turns cross each other at an angle, instead of laying parallel which creates much higher capacitance between layers.
Its a very old fashioned method of winding, but its exactly how EHT transformers are wound with very few turns per layer, and a very great number of built up layers.

Its still applicable for very high turns ratios, and high voltages or high frequencies.
Its perfect for extremely high voltage flyback supplies on an open frame UU core.

Connecting multiple transformer primaries in parallel, and the secondaries in series will not present any problems of balanced load, or circulating currents.
Its a well proven solution.

Its also good where you might for example need say twenty isolated low voltage secondaries. Four individual transformers, each with five secondaries would probably be a much more practical solution than a single transformer.

 

[mtwieg]

Simplest thing you can do to increase self resonant frequency is to not overlap turns, at all. Your parasitic capacitance will scale roughly with the number of layers squared. You may need to use a core with a longer path length to fit all the turns on.

RF inductors are always single layer solenoids, usually with spacing between turns of about one wire radius.

 

[FvM]

The turns have been limited by the acceptable core losses on the material so the inductance that is created by the turns are fixed, so to speak.

Not necessarily. The usable flux of a nanocrystalline core at 20 kHz will be set by core saturation limits rather than losses.

The resonance frequency discussed in your previous thread is a resonance with the leakage inductance, which is mostly independent of the core permeability.

Regarding DC bias, it should be possible to keep the transformer DC free, e.g. by using a capacitively coupled driver, or any other circuit that allows flux self balancing.

 

[bowman1710]

Not necessarily. The usable flux of a nanocrystalline core at 20 kHz will be set by core saturation limits rather than losses.

No it is not necessarily in all cases but in my situation to get the losses down to an acceptable level it was well within the saturation limitation of the core material

The resonance frequency discussed in your previous thread is a resonance with the leakage inductance

So the resonant is not a factor of the L and C secondary, it is a factor of Ls(Leakage) and C secondary? My understanding was that the poor coupling (i.e leakage) only contributed to this resonant affect or a Ls and Cs??

**broken link removed**

Table 1 illustrates how each capacitance reflected to a reference winding creates the Cd "total".
(see EQ.1) Note that in the last column of the table, the SRF is calculated from the reference
winding inductance and Cd. The transformer action causes the reflected capacitances to appear
in a ratio that makes the SRF the same with different turns. The fact that all windings will have
the same SRF is not intuitive. Transformers with poor coupling between windings or a strong
drive sensitivity will have differences in SRF.

 

[Warpspeed]

The SRF is usually always determined by the winding with the largest number of turns.
Adding one or more extra windings to the transformer that themselves have fewer turns than the problem (reference) winding will not change the SRF.

There is not a lot you can really do about the inductance, its pretty much fixed by the number of turns and core parameters required for transformer action, but how you go about arranging the winding can have a large effect in inter turn capacitance.

 

[bowman1710]

There is not a lot you can really do about the inductance

Would adding a gap to the core not change the permeability/AL value of the material, thus reducing the inductance present on the secondary, lowering the SRF? I am trying to use a toroid if possible are there any know methods of reducing the effective capacitance of the secondary, I cant see how you would get the honeycomb technique on one, are you pretty limited with a toroid core on what you can do?

 

[Warpspeed]

Use two toroids each with half the secondary turns.
Secondary inductance per toroid will thus be one quarter, and SRF should more than double, because the stray distributed capacitance will be less as well.