Finding the allowable pk-to-pk flux density in our Full Bridge SMPS transformer core?

[treez]

Hello,

We are trying to work out the peak-to-peak B (flux density) allowed in our 40KHz Full Bridge SMPS transformer. The maximum ambient for the transformer will be 45 degrees C.

We are using a Epcos TDK PM62/49 core, centre leg gapped to 0.2mm. We have it in N87 ferrite material.

PM62/49 Ferrite core datasheet:
https://www.farnell.com/datasheets/1734383.pdf

Firstly, we need to know the maximum allowable operating temperature of the ferrite. This information does not appear anywhere in Epcos TDK literature. Do you know where we can find this?

Also, pages 4-8 and 4-9 of the following (slup126 document) state that we need to assign our maximum allowable temperature rise of the core, and then using the “core thermal resistance”, we need to assess the maximum power dissipatable by the transformer. Then, the slup126 document says we must assume that core losses will form 50% of this power level…

Transformer design document.
https://www.ti.com/lit/ml/slup126/slup126.pdf

We must then find the volume of the core, and then calculate the Watts/meter cubed power dissipation due to core loss. We then have to find the data which tells if this power density is OK. –Except we cannot find that data anywhere in Epcos website.
When we have the allowable maximum power density , we must find the graph of power density vs delta_B. –This shows different curves that accord to different frequencies. However, these graphs do not exist for the Epcos PM62/49 core.


The only actual information that we can find on Epcos cores is the following ferrite databook…

Epcos TDK ferrites databook:
https://en.tdk.eu/blob/519704/download/2/ferrites-and-accessories-data-book-130501.pdf

On page 49, we can see a graph stating that at 40khz, the N87 ferrite material has a “PF value” of 12000. This is in units of kHz.mT.
Anyway 12000 hertz.tesla’s , at 40000 Hz , means a B value of 0.3 Tesla’s.

…But what does this B refer to? Does it refer to the Peak B allowable?, or the peak-to-peak B allowable? –and if it is pk-to-pk B, then is that with sinusoidal or square wave excitation?

Does anyone know of the data which we need to find our maximum allowable peak-to-peak B?

 

[Easy peasy]

PM62 N87 40kHz peak B = 180mT, Bp-p = 360mT else core will overheat due to its size, unless you have conduction heatsinking or forced air...

plus you have to allow for heating contribution of Cu, for PM62 PCu+core = 7W for convection cooling in Tamb up to 45 deg C, assuming peak temp of 100 C in the middle of the Tx.

 

[treez]

As you know, N87 ferrite can run up to well above 100degC?...

(as attached, from the Magenetic Design Tool download of Epcos TDK)

 

[Easy peasy]

yes but the R of Cu has already gone up by 1.3 from 25C to 100C, why make the Cu losses even higher by pushing up the core temp and hence the temp of the whole...?

 

[treez]

I think many managers would say "so what", the copper isn't going to melt, and then they'd say "keep that transformer size down".
If you tell them "no its already at 100degc", they will then say, "what is the maximum allowable temperature of the ferrite?", and then of course you have to confess that the ferrite can work up to 180 degC.

 

[Easy peasy]

Have you heard of thermal runaway in transformers.... its what keeps most designs to the 100 - 110C range for peak internal temp...

- - - Updated - - -

Also the lifetime of the insulation is affected by temperature...

 

[treez]

thanks, is that temp of the thermocouple buried down at the core, or is that the temperature of the deeply buried ferrite, well inside the volume of the core? As you know, the internal ferrite temperature is greater than the thermocouple reads on its surface.

 

[Easy peasy]

as stated - peak internal temp, where the wire meets the core near the centre leg of the core...