EER43 Windings

SMLinear

New member
Hello Everyone,

Ever since I registered in this forum, I have been collecting information about SMPS and hardware. Although I had some luck in getting some hardware (ferrite cores, Electrolytic Capacitors, IGBTs and Chips, I am still not fit in calculating the windings and the best strands for the power transformer.

Here, I mainly have two transformers for the project of a power supply 12V/>50A

Dimensions are: length A:42.5mm, height:44mm, depth:19.6mm, inner diameter 18mm, window area 8mm. I'll upload the image, please check for details and precision.

I have four pieces which make two transformers. I need to make an SMPS power supply 12V/ more than 50Amps, I aspire 100 amps for an induction heater.
How many windings do I need for the primary and the secondary, I won't have and auxiliary as I'll use other current sensing methods.

I have SG3525, IR2153, and TL494, I personally prefer the TL494 old timer as I experimented with it a lot and its comparators, but never mind the chip.

Could someone please help me do the calculations for a transformer? I'll make two identical to use in parallel. The winding programs got me nowhere as the one (All in one) is in Russian,
by the other versions I only get space numbers and signs! TransEER43.png
 

Silvio

Well-known member
Hello Everyone,

Ever since I registered in this forum, I have been collecting information about SMPS and hardware. Although I had some luck in getting some hardware (ferrite cores, Electrolytic Capacitors, IGBTs and Chips, I am still not fit in calculating the windings and the best strands for the power transformer.

Here, I mainly have two transformers for the project of a power supply 12V/>50A

Dimensions are: length A:42.5mm, height:44mm, depth:19.6mm, inner diameter 18mm, window area 8mm. I'll upload the image, please check for details and precision.

I have four pieces which make two transformers. I need to make an SMPS power supply 12V/ more than 50Amps, I aspire 100 amps for an induction heater.
How many windings do I need for the primary and the secondary, I won't have and auxiliary as I'll use other current sensing methods.

I have SG3525, IR2153, and TL494, I personally prefer the TL494 old timer as I experimented with it a lot and its comparators, but never mind the chip.

Could someone please help me do the calculations for a transformer? I'll make two identical to use in parallel. The winding programs got me nowhere as the one (All in one) is in Russian,
by the other versions I only get space numbers and signs! View attachment 7177

Hello SMLinear,

I am not sure from your post if you have ever built a SMPS before, I can see you are in for quite a task especially with the current involved in your PS. Having 12v@ 100A at the output I believe you also need your psu to be regulated. We are talking here of 1200w. I do not know the duty cycle needed for the smps either, this parameter is needed because the winding will need to carry the current.

It is also the case that it will be more difficult to parallel 2 smps together. These do not work like the linear ones due to the output sense has to be individual and there could be a conflict regarding the output regulation.

Regulation can be achieved either by using a boost power factor circuit in the input which I think will work better for you and make life easier. Having output regulation will bring the need of an output inductor which will have to carry the current. You will also need synchronous rectification to minimise losses and heat (100A is no joke)

Regarding windings for the secondary you will definitely need a copper sheet for it. An estimate for calculating the current handling at full duty you will be needing 3.2A per mm² so taking this into account 100A/ 3.2A will get an estimate of 32mm² of copper sheet cross sectional area. For the primary you will need for 1200w depending on the topology used if its half or full bridge or other. The input will also need to cater for the efficiency loss and usually we take an smps having 80% efficiency so another 20% have to be added considering the power factor at the input is 1. This is not the case if you do not use power factor correction. Usually the power factor will vary with load.

Another factor to consider is the input capacitance this we usually take a minimum of 1uf per watt of power. You will be needing around 1500uf across the 320v rail considering some more input power due to loss in the smps.

Lastly I think EER43 trafo is not so popular and I did not find any data for it. I guess you will be better off using a larger core like ETD49 or ETD54. This will make it more practical to fit your windings in it.

The oscillator chip is another option to consider I guess a current mode control chip will work better at the current involved.

Pcb layout also need to be good and traces need also to carry the 100A needed. Your setup need to be well planed.

Showing you all this do you still like to try? It is not going to be easy especially keeping the output stabilised at no load and full power not to mention stray inductances and capacitances which will disturb the sensitive areas on your pcb.

Regards Silvio
 
Last edited:

SMLinear

New member
Thanks a lot Silvio for your immediate reply and your effort to clarify things out.

In fact I have never built an SMPS before. The Chinese SMPSs and Welders are far cheaper for me to buy but I’d like to experiment and learn by doing using my salvaged parts.
I started to prepare things for a welder a few months ago (2 *25A Rectifier bridges, 2* 650u/450V Capacitors, 4*BUX23 etc, see images please), but I gave it up.
You are right it’s a challenging task. I took current regulation into consideration, and I have been studying current regulation methods and experimenting with current transformers.
I also took precautions and safety measures into account (PCB traces and transformer windings). I learned from your and others' videos how to wind ferrite transformers. I used to wind 50Hz EI in the past!
Surely I am not going to build the SMPS and immediately feed it with live mains before a lot of knowledge and precautions! “Popcorn” inevitability is excluded.
As for the duty cycle I left it for suggestions as people know better what is suitable for the parts I have.
I agree with you in all but unfortunately it's a bit difficult for me to get the up-to-the-task Cores. Here I uploaded some of the parts I have.
Unfortunately most of the cores are suitable for 350W or so. The powerful ones are (glued!!!) = ><worthless!
I also have a lot of gapped ones!
I was hoping I could combine toroids together as I have 3 big ones (the black ones), but I am not sure.
I have some 560u/450V Capacitors, IRFP460s, High Amp Schottkys, G20N60B3D, but I wanted these for later projects.
For the welder I wanted to get rid of BUX23 as I have 4 on the heat sink and 2 spare ones.
As for parts and regulation are not a problem except for the a pair of ETD59 for instance. Copper strips or sheets are available.

I wonder if I can use toroids as the guy does here in this video: https://www.youtube.com/watch?v=isZ9tApXwc4
How about continuing with the welder scenario into an SMPS power supply 12 Volts 35 Amps each transformer and getting the BUX32 to drive both transformers interchangeably or simultaneously??
I thought of switching both transformers simultaneously or connecting them in series where each gets half cycle.
What about the two (broken and glued U*2) I salvaged from a new TV line transformer could I get 12V/ 50Amps from it???


Glued.jpgPower_PCB_Solder.jpgPowerTransistors.jpgFerriteCores0.jpgToroids.jpgUngapped.jpg
 

Silvio

Well-known member
@SMLinear.

Well I gave things another taught and you might use two of your EER43 cores. These cores can readily supply more than 600w each at 20Khz. It is important that they be of the same material.

My taught is here. You will opt for a full bridge topology and you can use a TL494. The 494 will also give you a chance of current regulation. The windings of the primary will be in series using half on each transformer. The secondaries will have the full voltage needed say 15v-16v each. You will have a single secondary winding on each trafo. The windings on the secondaries will also be in series. You will then fit a single diode on each winding and joining the positives together. The other ends will be tied together and for the centre tap.

This way the transformers will be handling around 50A each. Looking at the setup you will be having a single transformer electrically but mechanically it will divided in two. Symmetry is very important when making the windings. You will still need copper foil for the secondary.

I opted for 20Khz as a switching frequency as this allows you to use 0.8mm copper wire for the primary. Your primary current with a full bridge circuit will be around 5A @ 320v this gives a power of around 1500w. The cross sectional area (CSA) for 0.8mm copper wire is 0.5mm². IF using 3 wires by 0.8mm then we have a CSA of 1.5mm². We multiply the CSA by the current density which is 3.2A/mm². This gives us a rating of 4.8A which is quite close to 5A. The CSA needed for 50A secondary is 50/3.2 = 16mm². The copper sheet needs to be at least 0.4mm thick estimating that the centre core length is around 35mm.

My current density estimates are for full duty cycle of 100A however if for example you will be using the supply at say 60% of the current then you can change the current density from 3.2 to around 4.5A/mm².

It will be very important that you will have a suitable ring core (green type) which has high permeability for the pulse transformer. The inductance of each winding on the pulse transformer should be around 2.5mH for a 15vdc input. IRFP460 mosfets will be more than enough to drive the primary. As for the output diodes well they should be fast switching types with around 40-50v rating and capable of 60-80A each. You could parallel these to aid with current handling. Ex having diodes of 30A rating and putting 2 in parallel.

For the output inductor well you need a large core to handle the current without saturating, The winding in this if its a ring one or maybe 2 rings stacked and glued could get you there, should carry the full current with minimal heat, using litz wire or multiple strands of wire to wind the core. 4 to 6uH will be adequate I guess. A minimum load will always be needed as the smps will be working in discontinuous mode which will result in a ticking sound in the transformers

A good clean waveform is necessary arriving at the gates of the mosfets. Any large spikes will create problems. Your primary design on the pulse transformer and its output waveform needs to supply at least 1 amp of power and providing at least 10v at load. This is necessary to switch the FETs quickly.

You can kindly draw and send a picture of the schematic so that we can discuss your ideas

Regards Silvio and Merry Christmas
 

Silvio

Well-known member
15772654284786640839631531903360.jpg

Here is my idea for you. I just noticed I made a small mistake the output voltage should read 16v and not 50v I was considering 50 Amperes not volts when I wrote 50v.
 
Last edited:

SMLinear

New member
Lots of thanks and Merry Christmas, God bless you and grant you power and health Silvio.
I dropped by to wish you and everybody all the best for Christmas and the new year.

Your suggestion is exactly what I have been looking for. I'll study things and come up with practical details for discussion and benefit for newbies too.

till later
All the best for you.
 
Top