toroid for car amp smps

[subwo1]

my only problem is that the core is getting hot(about 62C celsius ) when the amp is running at low volumes and getting colder(55C celsius) when i increase the volume .

This type of problem is one reason I avoid push-pull even for 12v powered SMPS circuits. Unless active monitoring and compensation of the duty cycle is performed, the transformer always experiences at least some net flux creep. There are at least two reasons why. The first is that the current * time product applied to the windings is different in each direction due to different pulse widths at the MOSFET gates, and maybe worse, that the on-resistances of the MOSFETs are not the same. The other is that those drive currents are applied to the core slightly differently since both windings cannot occupy the same physical location.

When the load on the transformer secondary is increased the reset current through the primary windings increases when they fly low. The result is that the core flux resets faster, getting closer to zero magnetism before the dead time period expires.

Well, that way is how I understand things, anyway. If the problem is not critical in your application, the answer would be to make sure the core has enough ventilation. Otherwise, you could reduce the duty cycle at low power output which would increase the flyback current through the primaries, as well as providing more dead time for the core to demagnetize. But then the output choke becomes more important and a feedback circuit becomes necessary.

 

[jackinnj]

Well, that way is how I understand things, anyway. If the problem is not critical in your application, the answer would be to make sure the core has enough ventilation.

Agreed:

1) Use a bigger core
2) Use a pulse-by-pulse current mode controller

 

[Eva]

Even simpler :

Increase turn counts and/or add something like 10% dead time

Also check for symmetry the pulses generated by the controller IC. I've found some SG3525 units from SGS-Thomson to produce asymetric pulses

EMI feedback into the IC signal paths may also cause asymetric pulses

 

[jackinnj]

or do a better job of winding the turns -- you could use a small computer fan to cool the core --

i really think that this discussion gets very complex for a bulletin board -- sanjay at National Semi has done a really good job of explaining the route to optimization -- again search "magnetics" on their website . With a core you pull out of a hamfest grab-bag it's difficult, but not impossible, to get really good results -- but you have to characterize aspects of the core so that you can apply the design equations.

 

[soundbag]

conquer the ferrite

I think the best advice anybody could give on smps is to push aside one"s linear attitude or knowhow aside for the moment and master the ferrite and is principles.

Honestly, that is what I am doing for the main time and that is what Eva and the smps guru did.

BWG, pls conquer the ferrite and you will be surpriced.

soundbag.
_________________
"DONT TRUST ANYMAN,
MEASURE EVERYTHING"

 

[jackinnj]

SMPS "Reading List"

This seems to be new on National Semi's "Power Bench" website, a reading list for prospective SMPS designers:

http://www.national.com/appinfo/power/0,1768,319,00.html

btw, I have been adding links to resources I find helpful at:
www.tech-diy.com/smps.htm

jack

 

[robin2d]

How to build ETD39 transformer

Hi,
I red your many thread replly. really you are very knowledgeable person.
I want to build a SMPS from this thread
http://www.diyaudio.com/forums/showthread.php?s=&threadid=31614&perpage=10&pagenumber=1

But problem that i did not find any toroid in my city. i found a ETD 39 transformer . Can you tell me how to winding a ETD transformer?
i want to use for 34 vols output transformer for this SMPS.

Let me know how meny AWG wire i have to use?
How many turn i have to use for primary and secondary?

I did not send mail directly to u, sorry for write to u from this thread
regards
Robin

 

[robin2d]

Me again

I am sorry that my last reply (Post #247 ) was for 'Eva'. sorry again for my mistake. I was forgeted to mention her name.

But any one can help me pls.


Regards
Robin

 

[gearheaddruid]

ETD 39.

Robin, I have done a lot with the ETD series of cores. First make sure you have an UNGAPPED core! Hold the two halves together and look at a light source through the joint of the middle legs. You should not see any light. There are also different ferrites available in the same core shape. The H7C ferrite is the one you want for smps.

Eight turns center tapped for the primary and 12 or thirteen turns center tapped for the secondary. For the primary, I would use the equivalent of #10 wire and for the secondary, #12. There should be plenty of window space for the wire on the bobbin. I would recommend using a Litz type wire or at least several smaller wires in parallel for the windings.

Aside to Eva, A woman who speaks smps, ...... I Love you Best regards, Steve

 

[DT]

Hi to all,

i am building my first smps from project 89. I wound my transformer(just primary), and have a question.
My ferite core(permeability is 2000) is 29mm inner diameter, 45mm outer diameter and 8mm height and width. I wound my primary with double solid cooper wire (2mm diam. each). Maybe anyone can look at my work and coment if i do this right?

What power rating can i expect from this transformer?

primary winding

thanks

 

[jackinnj]

Re: ETD 39.

Robin, I have done a lot with the ETD series of cores. First make sure you have an UNGAPPED core!

Of course a gapped core won't saturate as quickly as an ungapped core, but it is hardly a free lunch. At any rate, Sanjaya Maniktala has all of the equations on the National Semiconductor website -- you can find them if you plug his last name into the search engine at www.national.com

 

[xyz9915]

So please, forget about my gender, I feel much more enjoyable talking about electronics only

Anyway, electrons doesn't have such a thing as gender

Eva,

Although I think that this is a very old thread, but since in my view something hurt you, its my request to you that do not take any comment much seriously. I am a newbie in SMPS but I am amazed you have such a skill in Electronics (especially in SMPS) because it takes years to learn.

So, I request you to please do not try to change you nickname

 

[AudioEngineer]

Hi Eva. If you are using multi thin wires for your primaries rather than just 1 thick wire due to higher switching frequencies to reduce the skin effect problem would it be better to twist the thin wires together before winding them on the core or is it better to leave them as straight wires and wind them round the core?

 

[Hadighorbani]

Hi
i am use sg3525 with 75khz for car dc to dc converter
which of this core is beter for that?
EE or ETD?
ETD49 can handel how meny watt?
and ETD59?
thanks

 

[kASD]

This may be true, but if you are going for 40 amps primary current, you probably don't want to be dealing with a push-pull transformer with a foil primary. A full bridge at least has a single primary section. With any high power circuit of this sort, one will be dealing with paralleled devices. For a dissipation of 20W in the primary switches, you want 12.5 milliohms total resistance. Ths dissipation will be spread out between all the legs of the bridge, so it's not a lot of dissipation per device. 10 milliohm devices in a TO-220 package are pretty routine now, and not horribly expensive. Two paralleled switches will get you less than 10mohm total resistance, and three paralled devices will help to spread out the current to reduce the losses in the leads. You end up with 12 devices for a full bridge, which is not too bad considering the primary current.

Well said,

Full Bridge is used in 12V input DC sinewave inverters also. Full Bridge is better when it comes to high current and efficient use.