Off-the-shelf shims to gap ETD ferrite cores are available?

Hello,
Does anyone know if 0.1mm gapping pieces are available to put between the core halves of an ETD44 core so as to give a 0.2mm gap?
This is needed because the cost of gapping by grinding the cores is ridiculously expensive

Ok that it would be not a professional solution, as well as unreliable because should decrease its thickness under compression, or in presence of moisture, but the standard thickness of an A4 sheet paper is exactly 0,1mm.

Thanks, but as you kindly discuss, we need more reliable gap size

Use polyester film cuts. 100 µm is pretty standard.

Thanks do you agree that by adding a 0.1mm polyester cut between each outer leg, we will then get an inductance that has less tolerance variation than using two ungapped cores together?

ETD44 3C90 has an AL value of 3800+/-25% when ungapped.
the "+/-25%" is far too wide for us.

ETD44 datasheet
https://www.farnell.com/datasheets/1719195.pdf

we calculate that with a 0.2mm total gap, we end up with an AL value of 841...and we believe that the tolerance on this will be some +/-5%....you agree?

Relative tolerance will be lower. But those applications that work with ungapped core rarely need exact inductance values. I don't understand your motivation for adding 0.2 mm air gap.

its for a 600W full bridge transformer.
If you don't gap, you need too many turns to get below Bsat.
If you use that many turns, then it just increases the leakage term in the transformer.
This is not acceptable.
The leakage delays the rise of I(pri) and robs us of duty cycle, -we need to know as accurately as possible what leakage we are getting, and we just cant do that if we don't gap.

There are other reasons too why we want the gap.
Such as the problem of primary magnetising current staircasing up to Bsat during transients where the error amplifier is railed high, and because of the leading edge blanking time of the control chip, the Imag can staircase high as discussed....we cannot have this happening with ungapped cores as the I(mag) can fly away into saturation levels....hence we need the gap.

- - - Updated - - -

The following two threads have already discussed gapping of full bridge smps transformers….

https://www.edaboard.com/threads/341980/

https://www.edaboard.com/threads/340530/

..These threads don’t go into enough depth to really realise the problem of ungapped bridge transformers

If you actually go about the process of making out the transformer manufacturing document in relation to the design, it becomes obvious that having a small gap is best.

Our 600w full bridge smps is as follows
pout=600w
vout=380v
vin=390v
fsw=100khz
core = etd44 3c90 with 0.2mm shimmed gap
NP = 36 NS = 58

...if you go about this design, you see that you cannot do it as well without a "small" gap.....unless you use a much bigger core.


This thread also becomes a discussion on gapping, after about 15 posts
https://www.edaboard.com/threads/348624/

.....the subject of staircasing of primary current was raised here (due to the leading edge blanking time), but glossed over.
The subject of how core remenance flux is helped by gapping was also explained, but again glossed over.

The +/- confidence range (I think it is 95% confidence limit but I can be wrong) if expressed as a percentage, will not change much (you need to compute at the two limits and at the middle; you may even end up with an asymmetric limit)- that is my guess. After the gap, the reluctance will increase and the AL value will decrease but the confidence interval will not be changed much in the percentage terms. But it will be a bad design if it depends critically on one elusive parameter (robust designs are insensitive to small changes in material parameter).

sorry but I cannot accept that the small gapped core has the same tolerance on AL value as the ungapped core.

If you are talking "lapped ungapped" cores, then I accept.

But normal "ungapped" cores are actually gapped anyway....due to the microscopic roughness of the core surfaces, and this causes the wide tolerance range.....when you put the gap in there, you prevent that microscopic roughness from having such an effect, and the AL value of the overall gapped core is more determing by the gap size.

Thus adding a little gap vastly reduces tolerance on AL value.......and this is even more so when you consider the variation over temperature.

Kapton film is widely available in a variety of thicknesses,
can take plenty of heat and is pretty resistant to compression.

I think this will be about the best answer you can have.

I paged through the Ferroxcube catalog until i found examples of gapped cores and there Al value. Depending on the gap the % was +-5% to +-25%.

Ferroxcube catalog 2013 page 206 E13/6/3.

I could not find any gap values for ETD cores. So with the correct gap you can have +-5%, but what about variation with temperature, long term stability?

I think you are really asking for trouble depending on +-5%.

I am wrong in post #8. Once you introduce an air gap, the reluctance increases and the effective permeability decreases in a geometry dependent manner. The inductance factor also decreases, somewhat like the effective permeability, in a geometry dependent manner and the new error limits need to be computed point by point as the relationship is no more linear. The best way is the take the minimum value of the parameter and compute fresh...

Once you introduce an air gap, the reluctance increases and the effective permeability decreases in a geometry dependent manner. The inductance factor also decreases, somewhat like the effective permeability, in a geometry dependent manner and the new error limits need to be computed point by point as the relationship is no more linear. The best way is the take the minimum value of the parameter and compute fresh...

Click to expand...

Yes, but a simple first order estimation should be sufficient.

Presuming a temperature independent gap width, the original ungapped µr TC will be reduced by the same factor as µr. Ending up with about +/- 5% for the example in post #5, as assumed by treez.

I could not find any gap values for ETD cores. So with the correct gap you can have +-5%, but what about variation with temperature, long term stability?

Click to expand...

Thankyou for your excellent findings
Regarding temperature variation of AL value, ungapped cores have far more AL variation of AL value over temperature, than gapped cores, as the attached shows.

All the “gap doubters” have never explained how non-gapped cores handle the situation of primary current staircasing which can occur during the “leading edge blanking” time of the controller, .. when the error amplifier is railed high due to some transient.

When I worked for a company which manufactured hundreds of types of transformers and inductors, we routinely used plastic films or paper sheets as gaping material.

Once that the magnetic structure is properly clamped and vacuum impregnated, they would be extremely stable, impervious to temperature and humidity (related to the air gap, of course. The core's permeability and the copper resistivity do have some temperature dependency).

schmitt trigger said:

When I worked for a company which manufactured hundreds of types of transformers and inductors, we routinely used plastic films or paper sheets as gaping material.

Once that the magnetic structure is properly clamped and vacuum impregnated, they would be extremely stable, impervious to temperature and humidity (related to the air gap, of course. The core's permeability and the copper resistivity do have some temperature dependency).

Click to expand...

I plotted a graph using the data given by FlapJack (#11)-

Once the gap becomes larger than some value, it is no more the critical factor. The graph becomes almost horizontal...