How to glue ferrite core halves together for flyback SMPS transformer?

[treez]

Hello,
We are doing a 60W isolated offline flyback (120W peak).
We are using an ETD44 ferrite core with 1mm gap. The transformer winding company say that they cannot glue the ETD44 ferrite core halves together because it is not a robust method of adhesion. Is this true?
They tell us that they must instead use the core clips to hold the core halves together and then varnish the cores so that they stick together.
As such, do you agree that glue cannot be used to glue the core halves together?

The 1mm gap is in the centre leg (only) of the ETD44 core, and I believe that the core halves can be held together while the glue sets in the outer core arms by squirting “hot melt” glue into the air gap to hold the core pieces together whilst the outer ferrite arms are glued together. (obviously ‘hot melt’ glue cannot be used to glue all the surfaces of the core together because it goes soft at 85 degC.)

So do you agree that “hot melt” glue should be squirted into the core gap to hold the core halves together while the proper glue in the outer core “arms” drys off?

I presume you agree with me that the permeability of air and “hot melt” glue is similar enough to mean that the inductance of the coils will not be affected by having “Hot melt” glue in the gap instead of air?

ETD44 core datasheet:-
**broken link removed**

(Sorry to ask, though I felt many could benefit from this knowledge)

 

[BradtheRad]

Hot melt glue is terrific for a lot of jobs. If it were just me, I'd try what you suggest. I realize I would have to 'do it right' the first time. Because if an unwanted blob of glue gets between the parts, and hardens, the unit will be misaligned, or there will be a gap, etc. Then you will need to heat the entire unit, in order to soften every blob of glue which is in between the parts.

The glue is gooey plastic. Once it has messed up a job, you cannot hope to fill in easily with a different type of glue. You would have to clean the surfaces completely, and hot melt glue is terribly difficult to clean away, especially from a rough surface.

And if I were handling a hot glue gun several times a day, I think I would be likely to get burned once a day.

In these regards, hot melt glue is not very 'forgiving' of mistakes.

Your agency may even have previously tried hot glue for a job. Perhaps there was one messed-up unit, which required extra time to fix. So he said 'It's not worth it, let's stick with more forgiving methods.'

 

[aashitech]

Ferrite core based transformer manufactures here uses Araldite to join ferrite cores together and leaves these transformers clamped overnight .

 

[FvM]

I agree with aashitech. Epoxy resin (e,g, araldit) has been sugested by core manufacturers since ever. Standard epoxy resin is applicable in all designs that don't operate at elevated temperatures (e.g. > 100 °C). You may prefer the mechanical clamps provided with core sets though, e.g. for ease of use.

 

[ysba]

I never used a big core as ETD40... But for smaller ones like EF20 or NPQ32/20 we hold them together with cyanoacrylate glue and then varnish them. The glue is not put between the core halves because it would create an addition gap. The core is manually held and the glue is placed aside the junction of the halves. It doesn't have a great resistance, but works well until the varnishing proccess.

We used to glue the halves with Araldite, but this proccess is a bit messy and very slow since we had to varnish the transformers in anyways.

I made some experiments with hot melt glue, but personally I didn't like it. I think it is messy, quite dangerous for who does this all day long, and doesn't give a great hold because it softens as temperature rises. I would not trust that for use in a product.

When using Araldite we developed a clipping device to hold the cores together. The images are attached.