Power transformer with interleaved winding

Thread Starter

silkyre6xtenz

Joined Jan 8, 2018
15
interleaving AAC question.GIF Hello guys, i know that windings interleaving allows to reduce flux leakage and increase efficiency of a transformer.
It is like a sandwich - 1st primary layer, 1st secondary layer, isolation, 2nd primary layer, 2nd secondary layer...
But how to wind it this way? I am developing power transformer for dc/dc converter (2-switch forward), and i want to try to add a Faraday shield between windings to reduce capacitance (should i connect this shield to system ground?), and i also want to try interleaving to reduce flux leakage.
So, i have ETD54 core and etd54 bobbin (it has round cross-section). My primary is 34 turns (i use 5 wires to route primary to reduce losses - so the primary is 5 wires 0,5mm diameter each which are turned over the bobbin 34 times. These 5 wires are going in line - i attached a photo to show it ). My secondary is litz wire with 0.071mm diameter (1075 strands).
If put all secondary together, it is an isolated wire with 3.5mm average diameter.
So - how to use winding interleaving with such boobins and windings - is it even possible?
If u guys have ever made transformer with interleaved windings (it doesn matter which kind of transformers) please tell me how to do it (at lease partial interleaving).
And so - should i ground Faraday shield between winding and is it helpfull to reduce current spikes on MOSFETs?
Thanks a lot guys, i wish you can help me.
 

DickCappels

Joined Aug 21, 2008
10,153
Just wind it as I think you described. Divide each winding into layers and interleave the layers, as you said. It is fine if one winding has one more layer than the other. Use only the number of layers of Permacel tape as necessary for the breakdown voltage so that intimate coupling is maintained.

The Faraday shield will reduce electrostatic coupling between the windings but it will increase the capacitiive load on the winding and might result in undesirable ringing and loss of risetime.

0.5 mm is pretty thick wire. That you are using Litz wire on the secondary suggests that you will be using either a very high switching frequency or a very large flux gradient within the bobbin. You might wish to go with multiple primary strands insulated from one-another in parallel to cut down on eddy current losses in the primary.
 

Thread Starter

silkyre6xtenz

Joined Jan 8, 2018
15
Just wind it as I think you described. Divide each winding into layers and interleave the layers, as you said. It is fine if one winding has one more layer than the other. Use only the number of layers of Permacel tape as necessary for the breakdown voltage so that intimate coupling is maintained.

The Faraday shield will reduce electrostatic coupling between the windings but it will increase the capacitiive load on the winding and might result in undesirable ringing and loss of risetime.

0.5 mm is pretty thick wire. That you are using Litz wire on the secondary suggests that you will be using either a very high switching frequency or a very large flux gradient within the bobbin. You might wish to go with multiple primary strands insulated from one-another in parallel to cut down on eddy current losses in the primary.
Sorry, i dont understand one thing.
You say that electrostatic coupling between windings will decrease if shield is applied, but this shield also increases capacitive load? How it looks like?
I thought that if i apply the shield between the layers, capacitance between layers decreases and it results in less losses, but you say opposite...
Feels like i'm losing something.
Thanks you for your reply, but i think i'm stuck at all...
Can you please tell me more about this electrostatic coupling between windings and its effects in overall converter operation?
So, the capacitance between layers won't be decreased if shield is applied and grounded?
P.S. You are right about 0.5mm in primary - converter now operates at 83kHz frequency, but we are looking to increase power by increasing frequency up to 100kHz. At 100kHz 0,5mm wire is not effective.

Well, if i get it right, the shield between primary and secondary works like a single turn of wire?
 
Last edited:

LesJones

Joined Jan 8, 2017
4,174
The capacitance to ground increases. (To both windings.) But the screen will decreace the capacitive coupling between the windings. Note also that the ends of each screen must not make contact with the other end of THAT screen or they will form a shorted turn.

Les.
 

Janis59

Joined Aug 21, 2017
1,834
The last one I wound was 100 kW 50 kV thus if I wind a HV without sections, the lower layer would give a spark to the next neighbour layer. Therefore I used one layer coming from left to right, then going externally by outside to next layer the same position and have the same direction. Thus, the voltage between layers are not the zero in the right and 2V at the left, but all the places are 100%V. Later epoxy vacuum-infusion made the rest. By the way, carpet linoleum of PVA type 3mm , thread-less, compact not puffy type, is superior good for HV. The extreme it may stand the 40kV as shown my previous experiments with adjustable 60kV source and infrared thermal camera.
The HV section with 1,2 mm wire was most deepest, the 600V "low voltage" section with 15 mm wire was over that. The core is glued by epoxy from 1x1x2 inch "bricks" made at India (2 USD piece, 0,39 Teslas, ferrite capability for F(max) around 2 MHz (I worked far lower).
 

DickCappels

Joined Aug 21, 2008
10,153
To emphasize the key point made by LesJones, be careful that your shield does not form a shorted turn. One way to avoid that is to put insulating tape on one side of the shield.
 

shortbus

Joined Sep 30, 2009
10,045
Therefore I used one layer coming from left to right, then going externally by outside to next layer the same position and have the same direction. Thus, the voltage between layers are not the zero in the right and 2V at the left, but all the places are 100%V.
I've read that winding that way also reduces "inter winding capacitance". Read it claimed in an old patent, from the 1950's or 60's if i remember correctly.
 
Top