Which Core Material for an SMPS.

[si2030]

Hi all,

I have been doing some background on SMPS core types and I am confused. Here are a couple of pages from TSC International's catalog.

View attachment 61432

View attachment 61433

From the materials guide which is best suited for an SMPS core and why? What should I be looking for in a good SMPS core? There are a number of types displayed and I have no idea what I should be looking for...

Hope some one might shed some light on this..

Kind Regards

Simon

 

[tvtech]

SMPS design is an art that is best left to professionals who deal with this on a daily basis....

You cannot do a SMPS design without the help of pro's...

I speak from experience. Burnt my fingers with the Viper series of SMPS controllers. Blew 5 of them up and realized that the other 95 that I had left would also go the same way.

My specific problem was getting the correct SMPS transformer. To be honest with you....I am none the wiser even now. And that was around ten years ago.

My advice is stay clear of trying to design SMPS. It's a dark art. Leave it to the PRO's.

In fact I am so sure of my facts that I guarantee that nobody on this great Forum will try and tell me otherwise. Or try and prove me wrong

Or heaven forbid...build a SMPS....that is both reliable and all. Challenge to all out there.

Cheers,
TV Tech

 

[simonbramble]

TSF5099 because you have got a low change in Permiability over Flux Density. Keep with Ferrite as well. This gives less core losses than iron powder, even though the saturation is a lot steeper.

 

[tvtech]

TSF5099 because you have got a low change in Permiability over Flux Density. Keep with Ferrite as well. This gives less core losses than iron powder, even though the saturation is a lot steeper.

@ what Frequency ???? My Vipers were @ 100 KHZ....... Well within spec.....

Bring it on

Cheers,
TV Tech

 

[MrAl]

Hi,

As TV Tech pointed out, there's quite a bit of thought that goes into designing a transformer for a switching converter. A pulse transformer is much easier for example, and a line frequency power transformer is the simplest.

Let me give you a quick idea what goes into a transformer like this and what you need to know...

1. The transformer equation.
2. Leakage inductance.
3. Wire size vs current in an enclosed area.
4. Skin effect.
5. Watts lost per pound vs frequency.

If you've never heard of any of these then perhaps you would do well to read up on some of this stuff. A site with a lot of info on magnetics is none other than Magnetics Inc. They have a lot of literature from years back on their site.

The watts lost per pound vs frequency is important because if the core heats up too much it melts the insulation and the transformer is destroyed. I've seen large transformers melt down, where all the layer insulation melted and dripped out of the transformer on an overnight burn in test.

If you want to experiment the main thing to watch out for is overheating.

 

[si2030]

Frequency between 40kHz and 60kHz. Use a tl494 as the controller. Use two mosfets. Off line frequency transformer to 30 volts. Anyway thats the idea... See this for the idea...

 

[MrAl]

Hi,

What kind of power you looking at here?
What input output voltages?

 

[ronsimpson]

typeo

So far most comments say it is a complicated art form.

From the first graph one I would choose TFS5090 or TFS50all because their lines are at the bottom of the graph. (low loss)

From graph 2 I would look for a line that is flat and extends to the right as far as possible. But that depends on what you are trying to do! If you want maximum inductance and will have low flux then TFS10K. For power supplies I would not us that one. If you are switching fast (above 1mhz) then the flux will be below 1000 and it does not matter what happens at 3000. If you switch slow at 25khz then you can run flux up at 2500.

From graph 3; don't use TFS10k above 70C. My transformers often run 100C so I won't use a core that falls apart at high temperatures. It is hard to know the "hot spot" temperature in a core. The outside of a core might be 70C while down deep inside is well above 100C.

For those that think 100C is too hot for a transformer. Well, copper and ferrite can handle high temperatures. My supplies are tested at 50C room temp for a week and some are tested at 75C for two weeks.

 

[MrAl]

Hi,

The temperature limit is not something that can be specified out of hand. It depends on the rating of the wire insulation and if any additional insulation has been added. 105C wire might make it, but 150C wire would be better of course. If the wire is rated less than 100C, then running at 100C would be asking for problems.
Im sure you know a core has to be able to handle the higher temperature too without changing too much. Also, in the application the relatively hot device has to have ventilation or heat buildup could burn it up or ruin something else in the same chassis.

The design is more complicated when we're looking for high efficiency while keeping the cost low.

 

[ronsimpson]

If high efficiency is the goal then heat must be low.
If cost is important then size of a transformer must be small.
I do not use 105C wire or 150C wire.

 

[MrAl]

If high efficiency is the goal then heat must be low.
If cost is important then size of a transformer must be small.
I do not use 105C wire or 150C wire.

Hi ron s,

Oh what kind of wire do you use?

I also forgot to bring up the problem of flux ratcheting.

 

[ronsimpson]

I have a good variety of wire.
Belden Poly-Thermaleze 200C
Belden POLYURETHANE 270C can solder strip at 750C


I see that some is rated at 105C, 155C, 180C, 200C and 240C for magnet wire.
I like triple insulated (some times called double insulated) Teflon wire. It is rated at 200C. I like it because I can build power line transformers with out using tape.

 

[MrAl]

Hi again,

Oh that's some pretty nice wire. Back when i worked in the industry we had to use nearly the cheapest stuff we could find, well maybe one grade up in temperature but that was it, and of course it was always enamel.

 

[RCinFLA]

Core selection starts with what type of SMPS you are designing. It falls into two basic categories.

First is tight coupled transformer where current flows in both primary and secondary winding at the same time. These type of cores have no air gaps and have high permeability. This type of transformer is designed for tight coupling of energy transfer between primary and secondary and do not store much energy in core. Example of this type is what you might find in an AC inverters that hooks to battery and provided AC (modified step sinewave or sinewave). These have a DC-DC smps that takes battery voltage to high voltage DC that is chopped to provide AC output.

Second type is flyback type transformer. These type of smps build up energy in the core then discharge the energy stored in core to the secondary windings. These types of cores have lower net permeability. These type may be a torroid core or a high permeablity transformer core with an air gap. Designing air gap transformer is even more complicated and very difficult to do for DIY'ers. Because the core stores energy they generally require larger cores for equivalent power then power transformer based smps design. Most computer power supplies and smps wall mount power converters are flyback smps designs.

The core size is based on power required. This boils down to the maximum magnetic force that is created by the energized coil (Amp-Turns).

**broken link removed**

 

[si2030]

MrAl,

OK, 240 AC rectified, frequency as described, output 30 volts/10 amps.

This is not going to be focused on a flyback topology but simply one with a tightly coupled transformer.

So based on this type of design what should I be focusing on... There are different values of Permeability and Gauss for each... am I looking for a high Permeability or low permeability.. and the same goes for Gauss... I note that gauss is as high as 5100 or 0.51T (flux density) and Permeability is as high as 10000 Uo. In this situation should I aim for high or low values and what are the significance and trade offs....

Simon

 

[ronsimpson]

Here are two application notes on switching power supplies. Power Integration usually shown all the formulas and graphs. They often show transformer construction and where to get the parts. Read these. I can help you modify one of the designs.

Power Integration is not used that much in the USA but I find the parts easy to get in China. Use then as a resource. The parts are simple.

 

[si2030]

Hi Ron,

Could you let me know which companies you deal with for cores in China if possible...

Simon

 

[ronsimpson]

TDK for raw parts.
BI has finished product.
If you want more information on transformers see
Coil Craft
Pulse Engineering
https://premmagnetics.com/
**broken link removed**

In USA See:
CWS, Coil Winding Specialist, Inc.
ICE Components, Inc.,
Santronics
Wurth Electroncis Midcom Inc.

Also:

Premier Magnetics
20381 Barents Sea Circle
Lake Forest, CA 92630
USA

Renco Electronics
595 International Place
Rockledge, FL 32955
USA

Have not used these but you asked for China: They all make transformers.

Elytone Electronic Co., Ltd.
218, Chung Cheng Road, Sec. 2,
San Hsia 237
Taipei Hsien, Taiwan

Tamura Electronics (M) Sdn. Bhd.
27 Jalan Firma 2/1
Kawasan Perindustrian Tebrau 1
81100 Johor Bahru, Malaysia

VOGT electronic AG
Sales Centre Shanghai RM1201, Block C
Shanghai Everbright Convention & Exhibition Centre No. 70, Caobao Road
Xuhui District, Shanghai 200235, China


Sales Centre Shanghai RM1201, Block C
Shanghai Everbright Convention & Exhibition Centre No. 70, Caobao Road
Xuhui District, Shanghai 200235, China

Sales Contact:
Mr. Hu Yuejun
+86-(0)21-64325615

Wurth Electronics Singapore Pte. Ltd.
160 Paya Lebar Road #07-07
Orion @ Paya Lebar
SG-409022 Singapore
Tel.: +65 (0) 67 42 15 67

Wurth Electronics Taiwan
4th Floor, No. 18-1, Section 6
Mincyuan East Road, Neihu District
Taipei 114, Taiwan
Tel.: +886 (0) 2 27 91 86 25
Email: inqeisos@we-online.com

Wuerth Electronic China
9 D 501, Xinmao, Rongyuan, LU STR. 15
Huayuan Hi-Tech Industry Park
CN-300384 Tianjin, PR China
Tel.: +86 (0) 22 2385 8666
Email: inqeisos@we-online.com

Wurth Electronics Hong Kong
Unit 02, 22/F, New Tech Plaza
No.34 Tai Yau Street
San Po Kong, Kowloon, Hong Kong
Tel.: +852 (0) 23 27 84 88
Email: inqeisos@we-online.com

Shulin Enterprise
762 Chung Chen Rd.
Chung Ho City
Taipei Hsien, Taiwan

 

[si2030]

Wow comprehensive... thanks. I will start searching..

Thanks Ron

 

[Ubergeek63]

Hi all,

I have been doing some background on SMPS core types and I am confused. Here are a couple of pages from TSC International's catalog.

View attachment 61432

View attachment 61433

From the materials guide which is best suited for an SMPS core and why? What should I be looking for in a good SMPS core? There are a number of types displayed and I have no idea what I should be looking for...

Hope some one might shed some light on this..

Kind Regards

Simon

there really is no good answer. depends on your frequency. you need one that can handle probably 10x the operating frequency so as not to dissipate the harmonics.

to know for sure if you do not have a good analysis package you would need to do a Fourier and figure out the dissipation for each harmonic.