Lapped ferrite cores not useable with Two Transistor Forward?

[cupoftea]

Hi,
We are doing a 130W Offline Two Transistor forward after a rect bridge input.
(Fsw=120kHz, vout=24v)
Vin is 100-373VDC.

We have chosen to use PQ26/20 ferrite cores from ferroxcube (3F3).

But our contractor has offered us, free of charge, a load of "lapped" PQ26/20 cores which he says are 3F3 equivalent.
Do you think we should take them?

The problem is, ayk, that if the supply is overloaded, when at max vin of 373V, then the duty cycle will go to
maximum for a few seconds (before overload timeout kicks in). During those few seconds, the lapped cores
will go into brick wall saturation, and the magic smoke will come.

So do you agree, that lapped cores are a no-no for our two transistor forward?

 

[Easy peasy]

lapped usually means mirror finish on the core faces

0.1mm kapton should add any gap you think you require.

 

[cupoftea]

Thanks, and presumably you would not advise bothering with such kapton if non-lapped cores were used? (ungapped and non-lapped).
As you know, and as D.A. (Tony) Stewart (of this forum) once kindly detailed, a plain "ungapped" core is actually gapped down at the microscopic level, since the surface is actually very unsmooth. As such, no kapton needed. Ive never known problems with unlapped+ungapped......but you never know...maybe some cores get a bit of a lapping at the factory when they are not so busy?

 

[FvM]

Thanks, and presumably you would not advise bothering with such kapton if non-lapped cores were used? (ungapped and non-lapped).

Why particularly? Apart from the question if core halves with raw unprocessed interface are available at all, adding 100 um film will mostly smooth "microscopic" surface irregularities.

You'll define a target AL value for your transformer core and set the gap respectively.

 

[cupoftea]

You'll define a target AL value for your transformer core and set the gap respectively.

Thanks, ayk, we dont want a gap as its an extra manufacturing step, and generally more cost.
As you would know, we dont have a tight AL requirement (seeing that non gapped ferrites have AL tolerance of +/-25% anyway).
As you know, one of the great advantages of a forward is that ungapped cores are acceptable.......but the downside that we are touching on here is always
a "back of the mind" worry.

Many PSU's of course, are flogged off in the hope that if they get overloaded, then the ferrite wont brick wall saturate....or just flogged off in the hope that they dont get overloaded. -Nobody would bother looking to see the degree of lapping that had been done to the core surfaces.

 

[FvM]

Your latest post seems to contradict post #1. In the former you are worrying about possibly too high AL of lapped core resulting in saturation. Consequently you would define a maximal AL value respectively minimal gap and see how it can be best achieved.

 

[D.A.(Tony)Stewart]

It is best to store most of energy in the air gap. The mu is somewhat inversely related. Lapped cores will be <= 1um.

The 100 um Kaptop solution is useful.

 

[Easy peasy]

Gapping the core gives extra slope in the current sense if using pk curr mode control - due to extra mag current - this can be a very easy way of adding slope comp to stabilise a fwd converter - a lot of designs have become stable by gapping when the designer thought the gap was preventing saturation - in reality it was a lack of slope comp at 45% D and above.

p.s. it is of course extremely difficult to get saturation in a fwd converter if peak curr mode is used with some slope comp - as the peak current is limited at a finite level determined by the volt EA output - as long as the peak level from the volt EA to the CS comparator is clamped you have a safe system.

 

[cupoftea]

Thanks,
To be honest its not high AL that we are worried about, but rather "brick wall" saturation.
ie saturation that's so sudden that even a clamped current limit isn't enough to stop the current skyrocketing
to ruinous levels.
Also, we can appreciate that a gap is useful for slope comp reasons, but the gapped core is more expensive, and more prone to
being nil-stocked....and Kapton
gapping is an extra manufac step, so we are very keen to avoid a gap.

p.s. it is of course extremely difficult to get saturation in a fwd converter if peak curr mode is used with some slope comp

Thanks, we dont have slope comp, and would rather not add it due to the extra components.

I take it that we are not "wanting our cake and eat it" by having a "non-gapped but non-lapped core and no slope comp except the bit of i(mag)".
We have current mode control via LM5021-2.

 

[Easy peasy]

this seems to be mainly hyperbole:

"brick wall" saturation.
ie saturation that's so sudden that even a clamped current limit isn't enough to stop the current skyrocketing
to ruinous levels. "

in that it can't happen it current mode.