High temperature, high frequency magnetic circuit

[Fabio88]

Hi all,

I'm trying to build a magnetic circuit (custom geometry, see draw) and having a lot of trouble with the material choice..

Biggest problem: temperature (probably up to 350-400C) if necessary a cooling system can be realized but temperature would not be less than 200C for sure.

Now I describe to you the purpose of the device:

We are trying to build a magnetic circuit which can be described as a C magnetic core with a coil on it driven in AC (with a DC bias); the goal is to have a good changing in the magnetic field in the air gap of the 'C'.

We want a fast system that shows fast change in the magnetic field in the air gap accordingly whit the modulation of the voltage applied to the coil; so we would like to have a small time constant and small phase lag in the reluctance of the material.

Also good behavior at high frequency is required (over the DC bias we are going to have a sinusoidal signal up to 30 kHz but we are probably also going to use other time of signal which may have some higher frequency component (step signal)).

Probably no laminated metal is going to withstand the temperature and also the custom geometry could be a problem so I think the only solution is powder material, but you are the expert

Thank you really a lot for your help.

Please find attached a very rough draw of the system.

 

[MagBen]

although the curie points of core material for powder core are higher (>400C), but the max rated temperature (for continuous operation) normally <200C. you may try some types of ferrite core with lower Curie temepratures, but the oprating point can be higher than 200C. ferrites lose their magentics when above the Cuire point, but not damaged. so ferrites dono thave a practical max rating for temperature. technically, the absolute max temperature rating is its sintering temperautre, 700-900C. but keep in mind ferrites loss magnetics at above Tc (<300C), and there is shift (over temp) of losses, permeability and Bs.

 

[MagMike]

What magnetic field strength are you looking for? Sometimes it is more effective to build the electromagnet around the oven/heating zone.

 

[EdStainless]

The DC bias should be through your 5 coils, there should be coil on either side of the air gap for the AC component.
Don't try to push the high freq AC through a return path.
You could use 2V Permendure for the structure, high saturation and high Curie temp.
Coil insulation will also be a trick, perhaps PEEK or one of its cousins.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[MagBen]

I am not sure if there are any FeCo-V powder cores commericailly available.
Metal Fe-Co, bulk or lams, would generate large heat under 30kHz, so this form seems to be not appropriate. Actually for powder cores, the tempearature constrain is hte epoxy coating, the core alloy itself is rarely the limiting factor, the high Cuire temp (maybe the highest) of Fe-Co can not contribute much for high tempeature applicaitons. .

 

[Fabio88]

Thank you for your help!

I’ll try to give you some more info:

The most strict requirement is the B value in the air gap: it should be more or less 300 Gauss in a radial direction between inner and out circular plate and so I’m worried about saturation in the ferrite since the plate near the air gap has to be quite thin (5mm maximum) and can’t figure out how intense the B field should be in it in order to have the desire value in the air gap.
The external coils can have a bigger core so maybe saturation there is not an issue, but central coil and front plate (inner one in particular) can’t be bigger..

I also found some ferrite plate that more or less match the dimensions of mine, but I’m not sure it is feasable to cut the big hole in it..

What about some other powdered material?

I like also the idea of EdStainless, but my concern was exactly how to close the AC Path... Could you please exlplain it to me a little more?
(I thought that Permendure has to be laminated to work good at that frequency, is it not necessary?)

 

[Fabio88]

With regards to the wire, I’m probably going to use Litz Wire with fiberglass or Kapton insulation and maybe run a spiral copper pipe as a first row on the core..

I probably need an equivalent 20AWG Litz Wire so that I can push a relatively high current and use less turn in order to have a lower inductance a so a faster system..

 

[Fabio88]

I can confirm you that we'll develop a cooling system to prevent temperature going too up..since it will be a very simple cooling loop we cannot be sure that temperature will be perfectly homogenous but let's say that there are not going to be parts with a temperature over 200C; this should simplify the material cliche (wire included).

Some guys proposed me to used 80% nickel iron laminated at 0.014" for the plates (I still have to understand how to calculate the minimum thickness that can provide the 300G in the air gap without exceeding saturation in the material itself)

For the coil core I found some stock parts in catalogs of ferrite vendors but my concern is that they could easily saturate.

Do you have any advices?

Ps I still have not completely clear the problem connected to the return path of the AC field.

Thanks as usual!

 

[MagBen]

not sure if i read your drawing correctly. I am guessing you use DC bias to build 300G at the air gap, AC component for field switch. i am trying to figure out how your flux goes, as long as your center pole has opposite polarity with your 4 surround poles, there will be not flux in oyur air gap.
it seems your gap is about 1'' wide, with a max of 5mm thick plate, i estimated roughly that you will need a filed of more than 5 T (50,000Gs) to gererate a 300G file din hte air gap!
Also, with that huge air gap, if you magnetic path must go throught the gap, forget about permalloy, the effective permeability will be decreased drastically.

 

[Fabio88]

Hi Magben and thanks for your opinion! At the moment the magnetic circuit is already built, the core of the coil are probably made of steel (we don't know which kind) and the plates are pure iron.

It works more or less as you thought: we use a dc biase to generate the radial B field (roughly speaking 200G) in the air gap (and now the upper plate is only 3.5mm).

The goal is add a sinusoidal modulation over the DC bias (30kHz, +\- 100 Gauss more or less)..

Do you think is impossible? With iron it doesn't work because at that frequency we have huge losses, my hope is that with another material the situation will be different..

Thanks again!

 

[IRstuff]

Seems to me that the physical design is suspect for this application. I would have thought that making it look like a motor stator would have made more sense. All these right angle turns is not conducive to making strong fields.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[EdStainless]

IR, interesting thought.
Use a stack of motor lams, wind coils around each ligament, all with the same polarity.
Then wind a small coil around each also. Place this into a cup with the center post and pole piece.
Use the main coils for the DC, the smaller secondary ones for the AC. That way they could each be sized.

You could put a couple of Cu lams in the stack, they will hurt the max field but they could have a cooling coil soldered to the outside of them to cool the assembly.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[Fabio88]

EdStainless could you please explain to me again the issue about the return path of AC? You wrote to me 'Don't try to push the high freq AC through a return path' but I can't exactly understand what you mean..

With regards to the shape of the circuit, I'm not sure that I weel understood the motor 'look' you are suggesting (I'm not familiar with electric motor), probably this would be a good suggestion but we don't have much time to fix the problem...in DC this stuff is working fine (probably can be optimazed but this is secondary), the problem now is the superposition of the AC modulation over the DC bias that creates the 300 Gauss in the air gap..so basically I have to better understand how to modify the material in order to do this!

Thanks for helping, I really appreciate.

 

[EdStainless]

Let's forget a different geometry and go back to yours.
If you put small thin coils on each side of the air gap for the AC yes the steel will act as the return path, but you will not be relying on the steel to deliver the flux to the air gap. These coils would work with air as the return path.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[Fabio88]

So do you think that the steel of the inner central rod can conduct the AC field even if the material is not good for AC (but it will be a poor result)? And that's why you are suggesting me to use air as a ruturn path?

I'm going on asking this because the problem is that the inner round plate is quite small and for some other reasons it can't accomodate anything else so I can put thin coils around the gap (externally) but not internally.

One feasaeble option is to convert the external part of the circuit in 'AC proof' material, but the inner part is really difficult to be changed: I am wondering what happens if I modify the outer coils (or add small thin ones dedicated for AC), and the two big plates, leaving however the inner coil and the inner central round plate made of plain iron and then drive the outer coils in AC but inner one in DC..I can't imagine what happens :-(

 

[IRstuff]

Your drawing says that it's 60.33mm, which is larger than some pancake motors.

However, the issue with the AC, I presume, has to do with the ability of the magnetic domains in the material to follow the frequency of the signal, for which 30 kHz is challenging. Air does not have this limitation.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[EdStainless]

There will not be a 30kHz signal in any of the steel, the losses will be so high that it will effectively be dampened or filtered out. Your only option is to push the AC right at the air gap. I doubt that a single coil will give you enough field. Try cutting a shallow groove into the edge of your center pole piece and laying a few turns of wire in. That may be enough along with and OD coil.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[Fabio88]

IRstuff: I can't get what do you want to suggest me with the comment on the dimension of the plate..
With regard to the 30kHz I think it's only a matter of finding the right material and proper geometry, there are a lot of transformer working at that frequency and much higher!!

EdStainless: What about using a laminated Permalloy plate (and rods)?

 

[MagBen]

Your desgin maybe genius, or I missed something, but it seems confusing and in-effective. I still boubt even your DC worked fine: 3.5mm (.138'') plate create a 300Gs field in a 1'' air gap. it is almost impossible even you could make a very strong field at the edge of plate using a superconductive system.

With that huge air gap, your permeability at aip gap will be very small, close to 10, regardless of materials. For example, if oyu use permeally with a permeability of 300,000, the permeability at air gap will be 300,000/30,001=10; or if you use powder core with a permeability of 300, the permeability at ais gap will be 300/31=9.7 (in both cases, assuming the path length of 10'').

 

[Fabio88]

Hey MagBen, actually the DC operation of this circuit has been already extensively tested: I'm working on an existing satellite thruster called Hall Effect Thruster (there's a lot of material on google just to have an idea of how this stuff works) and whit a front plate (pure iron) of 5"x5" and 3.3mm (more or less 0.13")of thickness we daily achieve 200-300 Gauss in the 1" air gap without saturation in the core.

The problem now is that we won't to add a sinusoidal signet at 30kHz on that DC Bias in order to produce let's say a +/-40% in the gap..

And since I'm totally new in AC (and somehow DC too) magnetic circuit I'm worried that I'm doing some major mistake without realizing it...

Thanks A LOT for your time