To clarify my last post, Ae is the effective core cross section in cm^2. Ae is a basic parameter for any core, and should be found in the core specification. The power handling capability of the core depends on the copper cross-sectional area and winding resistance. This will depend on frequency, as the number of turns varies inversely with frequency. The power handling capability goes up with frequency until core losses and copper losses (from skin effect) dominate. It is in ones's interest to minimize the number of turns on the core (within reason) to reduce the winding resistance. For 12V input, 50kHz operating frequency, an Ae of 1cm^2, Bp of 2000 Gauss, the number of primary turns is
10^8 * 12/2000*1*5*10^4 = 12 turns. Since a push pull converter is used, one winds two bifilar windings of 12 turns. The twowindings should be applied simultaneuosly and be spaced close together to maximize coupling and ensure that they are identical, hence the bifilar winding. The secondary windings are scaled from the primary.
10^8 * 12/2000*1*5*10^4 = 12 turns. Since a push pull converter is used, one winds two bifilar windings of 12 turns. The twowindings should be applied simultaneuosly and be spaced close together to maximize coupling and ensure that they are identical, hence the bifilar winding. The secondary windings are scaled from the primary.
tennisballg :
The first image shows the right way to do the windings since in the second image you are maximizing inter-winding capacitance and this causes huge amounts of ringing
In the other hand, this core is absolutly useless as a transformer since its color code [yellow body with white side] means 'iron powder material' [usable only as a power inductor]
You must get ferrite cores, independently of the format [toroidal, EE, EI, etc...]. If you only have access to PC PSU components then your *only* choice is the main EE or EI transformer, these usually work at 30Khz and use 7 primary turns for 12V so at 50Khz you can also use 4 turns per primary. Use moderate heat to carefully soften the glue and dissasemble the core. Don't apply excessive force or you will crack the core. Count the numbers of turns when disassembling and take it into account. To know the real operating frequency of the unit look at Rt and Ct from the TL494 IC that most units use [see the datasheet and divide by 2]
The first image shows the right way to do the windings since in the second image you are maximizing inter-winding capacitance and this causes huge amounts of ringing
In the other hand, this core is absolutly useless as a transformer since its color code [yellow body with white side] means 'iron powder material' [usable only as a power inductor]
You must get ferrite cores, independently of the format [toroidal, EE, EI, etc...]. If you only have access to PC PSU components then your *only* choice is the main EE or EI transformer, these usually work at 30Khz and use 7 primary turns for 12V so at 50Khz you can also use 4 turns per primary. Use moderate heat to carefully soften the glue and dissasemble the core. Don't apply excessive force or you will crack the core. Count the numbers of turns when disassembling and take it into account. To know the real operating frequency of the unit look at Rt and Ct from the TL494 IC that most units use [see the datasheet and divide by 2]
Actually the maximum power you can get from a transformer depends on the rise of the temperature on it.
There is a book called Designing Magnetic Components for High frequency DC-DC Converters, from McLyman, with lots of information and many design examples.
I made many transformers based on that book and all of them worked fine.
Regards,
There is a book called Designing Magnetic Components for High frequency DC-DC Converters, from McLyman, with lots of information and many design examples.
I made many transformers based on that book and all of them worked fine.
Regards,
Yes, both the toroid and wires are useless for a push pull SMPS, it's one of my extra toroids, I just grabbed one to wind as an example. See my earlier post for the stats on the one I want to use.
By the way, I think the temperature rise is only a byproduct of core saturation, which is really the point that we're looking for(I guess).
By the way, I think the temperature rise is only a byproduct of core saturation, which is really the point that we're looking for(I guess).
No, the temperature rise is due to copper loss and core loss. If you are saturating, you are in trouble, unless you designed the transformer that way. A car inverter transformer wilkl be copper loss limited due to the high primary currents. Transformer capacitance will be of little concern due to the low voltage. An off-line switcher transformer can be an entirely different affair. BTW, I don't recall seeing the power range desired.....
Blmn
Absolutely correct, and the temperature rise is directly proportional (if no linearly) to the flux density which is decided by the time-voltage-area. The lower the switching frequency and the higher the voltage, the higher flux density.
Therefore it is correct to say that the amount of power that can be withdrawn from a core is decided by the amount of copper that can be fit onto the core.
In principle you could take out 1000W out of a 10mm diameter core if you only could find those supra conductors laying around.
Absolutely correct, and the temperature rise is directly proportional (if no linearly) to the flux density which is decided by the time-voltage-area. The lower the switching frequency and the higher the voltage, the higher flux density.
Therefore it is correct to say that the amount of power that can be withdrawn from a core is decided by the amount of copper that can be fit onto the core.
In principle you could take out 1000W out of a 10mm diameter core if you only could find those supra conductors laying around.
We've used ETD ferrites at 500KHz here at work. You have to wind them with Litz wire at these frequencies though, otherwise copper losses will go way up. Also, the switching loss in your MOSFETs will rise dramatically.sss said:
For a car amplifier SMPS, I'd stay in the 50-75KHz range.
If you need a lot of output power, you can use multiple transformers... You could wind two transformers with the primary and secondary windings in series, or you could get creative and use two identical transformers with single primary windings.
Wrenchone, that would be 3 turns, not 12. you left out the 4 in the lower half of the equation you initially gave. Also parenthesis would help us see what is multiplied first, then divided. The example cleared that up. thanks for the equation, though, it's the most specific advice that's been given.
I can understand the skinning effect due to frequency and copper losses, that's in all applications and unaviodable. Just have to choose appropriate conductors.
I can see how heat being generated can limit a cores flux capabilities due to the curie effect, but the generation of heat is caused by other factors, like the B-H curve, correct? Basically when the toroid starts to give less flux for each amp supplied, it gives off the energy as heat instead, saturation is the extreme. Is the core's BH curve based on the core's curie temp?
example B-H
See, you guys are saying that the power limit is when the core gets hot, and I am asking," what causes it to get hot?" Losses, what attributes in core selection cause losses? There must be some parameters to look at, some toroids handle more power than others, what characteristic allows this?
I have the calculation for flux density at a given current and no. of turns, if I stay within the rated flux density of the core am I good? Sorry, I am just trying to do my best with what I am being given. I appreciate the advice really, but it's rather general and seems that nobody understands anything but the basic concepts.
I can understand the skinning effect due to frequency and copper losses, that's in all applications and unaviodable. Just have to choose appropriate conductors.
I can see how heat being generated can limit a cores flux capabilities due to the curie effect, but the generation of heat is caused by other factors, like the B-H curve, correct? Basically when the toroid starts to give less flux for each amp supplied, it gives off the energy as heat instead, saturation is the extreme. Is the core's BH curve based on the core's curie temp?
example B-H
See, you guys are saying that the power limit is when the core gets hot, and I am asking," what causes it to get hot?" Losses, what attributes in core selection cause losses? There must be some parameters to look at, some toroids handle more power than others, what characteristic allows this?
I have the calculation for flux density at a given current and no. of turns, if I stay within the rated flux density of the core am I good? Sorry, I am just trying to do my best with what I am being given. I appreciate the advice really, but it's rather general and seems that nobody understands anything but the basic concepts.
If you study the specifications of a core material you will find a graph that shows losses per qubic millimeter. This loss parameter is decided by the flux density which is decided by the time - voltage - area. You can probably also see the temperature rise for a given flux density and for added resistive losses on the surface of the core. Check that this temperature in combination with the highest obtainable surrounding temperature does not get close to the temperature where the permeability drops.
calculating core losses for smps magnetics
To compute core losses, you must determine the ac flux swing, zero-to-peak value, which is half the peak-to-peak value. The core loss curves in the core suppliers data sheets give losses usually in milliwatts per cubic centimeter, or watts per pound, etc. The ac flux swing is what determines core loss, not the peak flux density. If the inductor is operating in the continuous conduction mode (current never decays to zero), then the transformer core flux consists of an ac component sitting atop a dc pedestal. Although the peak flux density may approach saturation (as an example, 3300 gauss for some ferrites), it is the ac flux swing that determines core loss. If the flux swings between 2000 gauss minimum and 3000 gauss maximium, then half of the peak to peak (1000 gauss) is 500 gauss. This is the value you will use when observing the core loss graphs. If the inductor is operating in discontinuous mode, then zero is the minimum flux, so that the zero to peak ac swing is half of the peak flux density. I hope this helps.
To compute core losses, you must determine the ac flux swing, zero-to-peak value, which is half the peak-to-peak value. The core loss curves in the core suppliers data sheets give losses usually in milliwatts per cubic centimeter, or watts per pound, etc. The ac flux swing is what determines core loss, not the peak flux density. If the inductor is operating in the continuous conduction mode (current never decays to zero), then the transformer core flux consists of an ac component sitting atop a dc pedestal. Although the peak flux density may approach saturation (as an example, 3300 gauss for some ferrites), it is the ac flux swing that determines core loss. If the flux swings between 2000 gauss minimum and 3000 gauss maximium, then half of the peak to peak (1000 gauss) is 500 gauss. This is the value you will use when observing the core loss graphs. If the inductor is operating in discontinuous mode, then zero is the minimum flux, so that the zero to peak ac swing is half of the peak flux density. I hope this helps.
I have 2 question, hope somebody can answer it.
Is it true we can use subwoofer magnet for smps power transformer? They are more common to find than ferrite core. If it can, what frequency will it be best operating? Does the magnet properties that is inside the ferrite (actually it is a subwoofer magnet!!) will influence something if it is used for power transformer, like 50khz?
What is the calculation for wire size for primary? In car it is only 12VDC swinging. In the mag-inc handbook, the formula is for home smps, does for 12VDC smps the same size wire can handle more current than the textbook equations? This is important, because in car smps, with audio power amp up to 1000W, the DC current can be up to 200A, what is the minimal wire can handle this in the primary?
Is it true we can use subwoofer magnet for smps power transformer? They are more common to find than ferrite core. If it can, what frequency will it be best operating? Does the magnet properties that is inside the ferrite (actually it is a subwoofer magnet!!) will influence something if it is used for power transformer, like 50khz?
What is the calculation for wire size for primary? In car it is only 12VDC swinging. In the mag-inc handbook, the formula is for home smps, does for 12VDC smps the same size wire can handle more current than the textbook equations? This is important, because in car smps, with audio power amp up to 1000W, the DC current can be up to 200A, what is the minimal wire can handle this in the primary?
I don't think you can, because they're magnetized. While ferrites are capable of holding flux, they dont have much of a residual magnetic field when you turn off the power. My thought is that if you wound a speaker magnet and ran AC through it, you'd end up fighting the core's magnetism, but that's just a guess.
As for wire sizing for current, just do a search online for an AWG wire table, it should show the ampacity of common wire sizes.
As for wire sizing for current, just do a search online for an AWG wire table, it should show the ampacity of common wire sizes.
Magnet material is entirely the wrong stuff to make switching power supply transformers. It has a huge hysteresis loop by design - that's what makes it a permanant magnet. Proper materials for frequencies like 50kHz or so are Philips/Ferroxcube 3C85, Magnetics, Inc. F or P material, TDK PC30 or PC40, Fair-Rite 77 material, Nippon Creamics NC-2A, etc. These materials are designed for reasonably low losses at switching power supply frequencies, permanant magnet materials are not.
As far as the calculations were concerned, oops- it was 3 turns after all. This is a pretty small number of primary turns, so it would be hard to get good coupling to the secondaries. Solution - lower the switching frequency, use a smaller core, or lower the peak flux excursion. A core with 1 cm^2 cross section is pretty big, anyway.
You can find appropriate ferrites on Ebay or at online surplus outlets if you search carefully. Ask the vendor if they know the ferrite material. C&H Sales lists a large ferrite toroid. The material is Magnetics, Inc. type J. This is a high permeability material rather than a low loss material, but it could be pressed into service using a low value of B peak, say 500-1000 Hz. It's a big toroid, so this is not really too bad a restriction.
As far as the calculations were concerned, oops- it was 3 turns after all. This is a pretty small number of primary turns, so it would be hard to get good coupling to the secondaries. Solution - lower the switching frequency, use a smaller core, or lower the peak flux excursion. A core with 1 cm^2 cross section is pretty big, anyway.
You can find appropriate ferrites on Ebay or at online surplus outlets if you search carefully. Ask the vendor if they know the ferrite material. C&H Sales lists a large ferrite toroid. The material is Magnetics, Inc. type J. This is a high permeability material rather than a low loss material, but it could be pressed into service using a low value of B peak, say 500-1000 Hz. It's a big toroid, so this is not really too bad a restriction.
check www.cwsbytemark.com as well, they have a wide selection of toroids in many different materials, ferrite , iron powder and others.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.