Inductor and transformer design methodology!

[goldsmith]

Dear Friends!
Hello
These are methods that i designing my magnetic elements with them and these are very good and simple methods!
Enjoy from them!




Best Regards
Goldsmith

 

[srizbf]

appreciation for sharing your design methodolgy.

if all the pages can be made as pdf , then it can be easily viewd.

 

[goldsmith]

Dear Friend!
Hi
sure i will change them as PDF file. thanks for your remark.
Respect
Goldsmith

 

[BradtheRad]

Your calculations appear to take most everything into account.

I have not done such calculations myself. Making a transformer is work. And when I'm finished it may not perform according to my calculations.

I know it is important to find the proper balance between many parameters:

* current
* number of windings
* turns ratio
* metal mass
* frequency.

It requires experience to discover what obstacles pop up. And it takes experience to find the secrets that lead to constructing a transformer so it will perform according to specifications.

My other thoughts...

You wish to obtain 5.5 amps from your secondary. Therefore it should be 14 AWG or thicker.

Your turns ration is 9. It's wise to increase this by 15 percent, in order to make up for losses and inefficiencies.

You will need to supply 50 amps to your primary. This will require about 5 gauge wire (4.62 mm diam.), or thicker.

Such thick wire may be hard to wind around the toroid. It may be easy to break the toroid (if it is made from ferrite material).

Your primary loop must not have more than 1 ohm total impedance ( 50 V / 50 A). A portion will be ohmic resistance (including the power source). The remainder will be reactance.

 

[goldsmith]

Dear BradtheRad
Hi
Thank you very much for your remark . i will try to improve them as you told me . ( so what is the tolerance of that way , in practice , in your idea ?( tolerance between calculations and actual event ).

Best Regards
Goldsmith

---------- Post added at 02:22 ---------- Previous post was at 02:21 ----------

By the way , i think i should , consider the effect of loss power , and i think if i consider it about 15 or 20 percent it will be ok and nearest to the actual . isn't it ?

 

[Insider2009]

AOA
Dear Goldsmith

I have seen the pages and then the discussion about wire size....
do you think it would be better to use Litz wire instead of thick wires as for high frequencies they seem to be more efficient.

btw I dont get Litz wire here in the market so i have developed a small gadget to twist my own wires and make Litz wire myself for my own use.

P.S Recently I have found a source for chinese Litz wire... but have not tried that yet.

Fahim Baig
fahimbaig2@gmail.com

 

[goldsmith]

Dear Insider 2009
Hi
Yes the LITZ wires are good points for high frequencies . and if you can't find litz wire , you can use parallel winding .
Consider , that we want a 5 mm wire . we can't use 2 wire with 2 mm size , and a wire with 1 mm sizw in parallel . do you know why ?
because of this :
you can calculate the space of 5 mm wire . and then you can calculate the place of those wires ! you'll see those are not equal together .
Good luck
Goldsmith

 

[Insider2009]

Thank you Goldsmith brother for a quick reply, I was not expecting that people read posts daily...

you are right about the sizing, but over the years I have found out that Designs are not based on 100% calculations, the calculations serve only for a guide line.
first thing you look around is what is the approximate size of core you would need depending on the total load in Watts. thats a good starting point in my opinion secondly I would find out the window area of the core which I intend to use. Then I find out what number of turns i would be needing I use Autocad and Excel spread sheet to work out what are the possibilites, and luckily mostly I am right about the size and load. I have found out from experience that using 80% of the window space gives the optimum results.
Also layering is important and sometimes I have to modify the layering scheme to get better balanced results from Full Wave converters.

thank you again for the support u are providing to ignorant people like me who are still in a learning stage.

Fahim Baig
fahimbaig2@gmail.com

 

[goldsmith]

well . i forgot to talk about what core we can select for .
Ok tomorrow , i'll attach a table of power instead of core .
( it's table is common with the other methods , such as methods that you can see at pressman's book or the other books .)
Good luck
Goldsmith

 

[Insider2009]

Dear Goldsmithl
can you suggest a ferrite transformer design for the following specs.
Buck converter
Input 220 VAC ( 312 VDC at the transformer Pri.)
Output 56.6 Volt DC 20 Amps with Current and voltage control.

Fahim Baig
fahimbaig2@gmail.com

 

[goldsmith]

Again Hi
You can use PQ series with gap space ( don't forget to create gap space .)
Or you can use some cores such as FM36 in parallel together ( cores can be used in parallel ) to obtain higher permeability and higher currents . and how much is your frequency ?
BTW : don't forget to use copper screen on your gap .
Best Regards
Goldsmith

 

[sky_123]

Btw, depending on your frequency, Litz is very good for a certain range of frequencies, but beyond about 1MHz (at least for smaller transformers), then Litz causes more losses, and a single wire becomes more preferable. The reason is that the skin effect starts to have an impact at those frequencies even on the litz wire, and becomes overall poorer performing than a single wire. There are also square cross-section wires, but I'm not very familiar with those.

 

[goldsmith]

but beyond about 1MHz

Dear sky_123
Hi
I'm disagree with you . about this , because , the skin effect , will has affect on the wires , at frequencies , more than 100 KHZ ( we have this effect , even at 50 HZ , but it is negligible at that frequency) . and at frequencies , above 100KHZ , we should pay more attention on it .
Regards
Goldsmith

 

[sky_123]

This is true, but what I meant was for the same diameter (i.e. compare 1mm of Litz to 1mm of single wire), then at the lowest frequencies (e.g. 50Hz)
then the single core can be a better choice, because less space is wasted in insulation compared to Litz, and therefore the wire can have less resistance compared to
the same-sized Litz). Then, as you say, Litz becomes the better performer (100kHz etc), but at the much higher frequencies, then the skin effect per strand also gets compounded by each strand affecting the other strands, and causing currents to be induced in the other strands, which makes it worse than a single wire. I'm not sure on the overall detail (I'm only familiar with low-power RF, not power transformers) but certainly for the smaller RF transformers, a point is reached (about 1MHz) where the single wire becomes more preferable again, if we're comparing an equivalent-sized litz wire.

 

[WimRFP]

This is true, but what I meant was for the same diameter (i.e. compare 1mm of Litz to 1mm of single wire), then at the lowest frequencies (e.g. 50Hz)
then the single core can be a better choice, because less space is wasted in insulation compared to Litz, and therefore the wire can have less resistance compared to
the same-sized Litz). Then, as you say, Litz becomes the better performer (100kHz etc), but at the much higher frequencies, then the skin effect per strand also gets compounded by each strand affecting the other strands, and causing currents to be induced in the other strands, which makes it worse than a single wire. I'm not sure on the overall detail (I'm only familiar with low-power RF, not power transformers) but certainly for the smaller RF transformers, a point is reached (about 1MHz) where the single wire becomes more preferable again, if we're comparing an equivalent-sized litz wire.

I think you are correct.

When going from single wire to stranded, there is a change that the loss first increase and when further decreasing the stranded wire diameter, the loss will reduce. One may look for "dowell curves" with "proximity loss".

 

[goldsmith]

Again Hi
If you are talking about RF , all things will change . at RF , we will pay attention at temperature / material of wire / distance between two wire / the length of wire / power / voltage / Impedance of wires and ...... etc .
Regards
Goldsmith

 

[sky_123]

Agree, unlikely to experience >1MHz for a power transformer! : )

 

[goldsmith]

to experience >1MHz for a power transformer! :

But i'm disagree with this , because i saw many power transformers for frequencies , more than 50MHZ and the powers about 12KW .
And i designed some transmitters with 1.6MHZ frequency , and the powers up to 5KW , with HF transformers simply . at higher frequencies , we have to use copper pipes as wires that has water in the center ( very high powers will need water )
Regards
Goldsmith

 

[sky_123]

I see; I've never experienced this, not really done much power stuff. That's very interesting about the copper pipes.

 

[goldsmith]

Yes it is interesting , my friend !
The reason of that , is , at high frequencies , any current can't go from the center of wire . and for economical reasons , and decreasing the weight , and ability of cooling the wire we can use from copper pipes , simply .
Respectfully
Goldsmith