How to wind a Car AUdio SMPS Transformer

[thefish]

I have some ST102-267 Cores that I want to use to make some simple low power Push-Pull SMPS transformers. The input is 12V the output will be connected to a diode bridge to give +/- 30. I have finished the control circuitry and the MOSFET's, everything is running @ 45Khz. Now my question what is the best method to wind them? I need 12 total turns in the primary.

Would I:

A) wind 6 turns (around half the core) center tap then wind the next 6 turns in the same directions back to the beginning of the core?

B) Wind 2 parallel wires 6 turns each all the way around the core then connect the output of one into the input of the other (at this point also connect the input power)?

As for the secondary I know I should wind the secondary in the opposite direction of the primary but should they be wound all the way around the core or only half CT then the other half also?

The transformer will be connect as seen below, disregard the rest of the schematic as the controll side is totally different.

[N-Channel]

Fish,

This circuit is from Rod Elliot's website. Many good reviews on it.
One question: Do you know the magnetic properties of this core? This is VERY important, as the wrong core can lead to catastrophic failure, like overheating of the main switching MOSFETs, or the core itself. For SMPS push-pull 12V operation, you should stick to high-permeability ferrite, such as Amidon's #77 material.

The cores you have appear to be composite material, possibly made up of a low-permeability powdered iron bonded to a high-permeability ferrite. Their function is as the inductor portion of the L-C Output filter, and not as the main power transformer. A good source for #77 ferrite toroids is at www.amidoncorp.com.

At a frequency of 45kHz, your primary turns will be in the 3T+3T to 4T+4T range. Your secondaries will be about 12T+12T for about +/-36V out at 12V in.

Will this supply be regulated or unregulated running wide open at 50% +50% duty cycle?

Hope these questions help. You might want to pick up a copy of Chryssis', Pressman's , or Brown's switching power supply design books. Any, or all three are absolutely valuable to properly understand every aspect of design and prototyping of all kinds of SPMSs: flyback, forward, buck, boost, invert, SEPIC, PWM, PFC, etc.

Cheers,

Steve

[thefish]

Just a little clarification...

A:
Wind half way around the core, center tap, then continue to wind in the same direction all the way back to the start. Connect +12V at the CT and at both ends use MOSFET¡|s to switch to ground (with 0 overlap time).

B:
Use parallel-wound (bifilar), windings evenly covering the entire surface area of the core. At the series connection add +12V and at both ends use MOSFET¡|s to switch to ground (with 0 overlap time).

I have been refering to these web pages to calculate turns and other stuff.

http://www.hammondmfg.com/5CHook.htm <-- Transformer Hookup

http://www.bcae1.com/trnsfrmr.htm <-- Turn/Guage Calculator

[thefish]

Originally posted by N-Channel
Fish,

This circuit is from Rod Elliot's website. Many good reviews on it.
One question: Do you know the magnetic properties of this core? This is VERY important, as the wrong core can lead to catastrophic failure, like overheating of the main switching MOSFETs, or the core itself. For SMPS push-pull 12V operation, you should stick to high-permeability ferrite, such as Amidon's #77 material.

The cores you have appear to be composite material, possibly made up of a low-permeability powdered iron bonded to a high-permeability ferrite. Their function is as the inductor portion of the L-C Output filter, and not as the main power transformer. A good source for #77 ferrite toroids is at www.amidoncorp.com.

At a frequency of 45kHz, your primary turns will be in the 3T+3T to 4T+4T range. Your secondaries will be about 12T+12T for about +/-36V out at 12V in.

Will this supply be regulated or unregulated running wide open at 50% +50% duty cycle?

Hope these questions help. You might want to pick up a copy of Chryssis', Pressman's , or Brown's switching power supply design books. Any, or all three are absolutely valuable to properly understand every aspect of design and prototyping of all kinds of SPMSs: flyback, forward, buck, boost, invert, SEPIC, PWM, PFC, etc.

Cheers,

Steve

I was looking at www.amidoncorp.com. and I see they have some Ferrite Toroid FT-140-77 cores for $3.75 would these work well? I see the 140 is for the OD size what is the 77?

[dangus]

Randolph Vikan's article on building a car amp in Audio Amateur (around 1988) went into detail on how to wind a transformer. That article is included in "Audio Amateur Power Amp Projects".
http://www.audioxpress.com/bksprods/products/bkaa40.htm

[N-Channel]

Guys,

Check out the #4 of 1994 issue of what was then Audio Amateur. My version of Mr. Vikan's Car amplifier was covered in the "Showcase" section of that issue.

Rgarding #77, this is the material. Here is the description from Amidon's website:

"MATERIAL 77 (µ = 2000) Has high saturation flux density at high temperature. Low core loss in the 1 KHZ to 1 MHz range. For low level power conversion and wide band transformers. Extensively used for frequency attenuation from 0.5 MHZ to 50 M Hz. Available in toroids, pot cores, E-cores, beads, broadband balun cores and sleeves. An upgrade of the former 72 material. The 72 material is still available in some sizes, but the 77 material should be used in all new design."

Be careful to mind the maximum power a particular core can handle at a given frequency. The FT-140-77 can handle about 200W at 20kHz, I believe, though I could be wrong. Look at the charts.

As for winding the primaries, you should wind them bifilar. this way, all the primary connections will be on the same side as the core, and all the connections will be on the opposite side of the core. Never wind only half-way around, stopping for the center-tap, then continuing on the the original side. Each half of each winding should go completely around the core to ensure good magnetic coupling between the primaries and secondaries, and also to minimize leakage inductance.

[thefish]

I found a great webpage guide about how to wind a transformer, I will post it here just in case anyone comes after me and has a similer question...

Building a SMPS

P.S.
I think I will also take a look into that book

[thefish]

Originally posted by N-Channel

Be careful to mind the maximum power a particular core can handle at a given frequency. The FT-140-77 can handle about 200W at 20kHz, I believe, though I could be wrong. Look at the charts.

As for winding the primaries, you should wind them bifilar. this way, all the primary connections will be on the same side as the core, and all the connections will be on the opposite side of the core. Never wind only half-way around, stopping for the center-tap, then continuing on the the original side. Each half of each winding should go completely around the core to ensure good magnetic coupling between the primaries and secondaries, and also to minimize leakage inductance.

What charts are you refering to?

The design is unregulated (50% duty cycle) controlled by a LM555 timer whose output is fed into a D-Latch TTL chip whose Q and Qbar outputs go to the MOSFET's (4 per side). This SMPS will be powering a 100W x2 amp I built based on a Tripath TA2022 chip (I have been told works very well on a unregulated supply).

[N-Channel]

Originally posted by thefish


What charts are you refering to?

There is a chart on the datasheets that Amidon mails out. It appears that it is not on their website. Basically, this is a graph on what power levels various cores can handle at given frequencies.

I would contact them and request a datasheet and a full-line catalogue be sent out. It is chock-full of good info about core materials, properties, etc.

If I can snap a pic, I will post it.

[thefish]

I found on some webpage that the FT-140-77 can handle 300W @ 50Khz so I ordered it (should be ok for my current amp)... but for future more powerfull reasons I also ordered 1 x FT-240-77 which should be good up to 1500W.