dc-dc core design problem

[xaccto]

34063 mosfet

what I got is a MC34063 controlled switcher, switching a mosfet
the coil is 8T 1mm wire, output Voltage is 13.75V

That circuit works pretty good, nice efficient too, no trouble supplying 1A load.
its switching freq at that load ~73khz

Only trouble is if you short the output, the diode is smoked!

So since there is plenty of room to wind a secondary coil, I put on 26T of 0.5mm wire,
changed the diode to the secondary, and moved the f/b resistor to there.

The best now I get with 1A load is 5.8V, the frequency of the switcher is now at 33khz. efficiency is very poor, maybe 50% if that.

I am having trouble understanding this? anyone got some comments?
why is the controller switching freq so low ? it doesn't seem to be attempting to regulate the output....... but the no load output is 13.5V...

I'm guessing winding on the secondary coil effects the inductance of the primary coil, but but, anyway is is supposed to be this hard to design switching transformers?

Thanks

 

[VVV]

34063 design

Designing a switching transformer is not a trivial task.
Can you post your schematic, so I can comment?

 

[FvM]

34063 circuit

If I understand right, you have a good and a bad design, why not using the good one?

Regarding shorted output: This won't work without a carefully designed current limit, 34063 has on option for it, but you didn't mention if you used it at all.

 

[xaccto]

34063 circuit irfz44

sorry,reread my post, it doesn't explain it too well.
standard step-up boost circuit for this 34063

the switch element being a mosfet.
See, if you short the output, the current flows thru the Rsc the coil and the diode, which then proceeds to release smoke.
Because the switch device is not in series, the current limit feature of the controller has no effect.
The circuit otherwise runs well and efficient.

So I proposed to change the coil to a transformer, so its a feed-forward design.

and this i'm having trouble with.

As a reference I'm looking at a design in "Motorola Power Mosfet Transistor Data" book,figure 8-9 showing a automotive dc-dc 9-24V=>54V, only it uses a MC34060. claims 78% efficient.

I've been doing some more research, whats a reset winding? I think i know now, but not sure how to wind one, or number of turns etc.
but that reference design I just mentioned doesn't use one and its 78% efficient??

I did notice when in one configuration of wiring up the 2 secondaries I wound, and I was monitoring the drain on the switching mosfet, and there was like 60V peaks occurring.......60V thats the breakdown voltage of the mosfet IRFZ44 !!
and it was getting HOT.

I don't think! its damaged....its all running from a current limited PSU atm. But when it runs from a battery, potential of smoke release is very very much greater.

here is a pic of the secondary windings,
not too neat is it? should it make much difference ?

How important is it in how carefully it is wound?? You'll notice the core is glued to the PCB, so winding it was a bit of a pain.
[/url]

 

[VVV]

34063

So you changed the design from a simple boost to an isolated flyback.
Did you pay attention to the winding start? If you reverse the secondary you will not get the output you are looking for.

Yes, neat winding makes a difference, as does sandwiching the secndary between two halves of the primary. It seems you did exaclty the opposite. But I don't think that is the reason the converter performs poorly. I think the secondary is reversed.

Also, what kind of gap do you have on that core? In a flyback the core needs an airgap to be able to store energy. Without that, the effect may be twhat you are describing, the inability of providing the required load.

For an example of flyback calculations, see my post and Excel file here:

 

[xaccto]

boost converter 34063

Hmm, thanks vvv, you gave me much to think about,
I looked at that zip file,re calc of other thread...may take me a while to figure out values,so might have some questions.

Though what is the bootstrap winding?

Ah the core, yes there is no airgap! the original primary is simply wrapped around the core - no core former. So it needs an airgap?

Can't say I understand what the magnetic difference is between simple boost, and isolated flyback, both use the coil to store energy, no ?

secondary reversed, hmm,ok don't remember if i had already tried that, will do again.

your comment re

" as does sandwiching the secndary between two halves of the primary."

i'm trying to understand this, is it like you have one of those coil formers that have a spacing "split" in the middle to separate windings ?

The image in the calc.zip, if that is a section view of the windings, the indicated primary has the expected mirror image, yet the secondard and bootstrap don't ??


I've found a book by Pressman "Switching Power supply design" on Google books

https://books.google.com.au/books?i...a=X&oi=book_result&resnum=6&ct=result#PPA2,M1

Its hard to read through an online book, alot of it is there though I think, however i'm envious of engineers that may have the real book for perusal, geeze $US395!!

 

[VVV]

34063 design circuit

No, the split coil former is a poor choice. Normally half the primary is wound firs, then all secondaries on top of that and finally the second half of the primary.

Pressman's book is the best book on power supplies, in my opinion. You should be able to purchase it from McGraw Hill for less than $100, though.

 

[FvM]

dc/dc isolated vs non-isolated

The most simple method to protect from output shorts would be an additional PTC fuse, to my opinion. A non-isolated boost converter is simpler and can work with a smaller core in most cases, at least with a moderate boost ratio.

When designing an isolated converter, you should understand the difference between forward and fly-back and the respective requirements for the transformer dimensioning. You mentioned a reset-winding. This implies a forward converter, normally. But I don't know the said application and can't compare the design.

 

[xaccto]

coil for 34063

I used one of the secondarys I wound as a reset winding. Have some interesting results.


without


I don't like that oscillation on the waveform, any suggestions as to cause of it?

I couldn't measure much difference in current consumption in the above 2 configurations, I would expect with reset winding to improve efficiency.


To answer some previous questions.

The application is to charge 2x 13A/H AGM batteries in a car. Max current up to 3A.
Hence why its a boost design originally, and it is a moderate boost ratio.

Fvm, thanks mention of PTC, i'm starting to be persuaded thats the way to go,
but do you think they trip fast enough to protect the diode? Specs say 1sec.

 

[FvM]

34063 pcb

To my opinion, your results are showing impressively, why an isolated forward converter isn't suitable for the application, at least as it is designed now.

 

[xaccto]

to conclude this topic, i'll add that for this application I did revert back to the original boost circuit, with a beefer diode and a polyswitch in series with Vin.
So its been working for some time now with no problem.
However this episode has sparked my interest in dc-dc converters.......
thanks for the replies.