Half Bridge transformer help?

[ABOELLEEL]

I've built a half bridge DC-DC converter but have a problem that transformer is overheated. It reached about 120 degree with following specifications:

Vin = 220V AC leads to 310V DC after uncontrolled rectifier
Vout = 50V DC
Io = 1A
freq = 70kHz
transformer etd34
primary 130 turns
secondary 65 turns

 

[Prototyp_V1.0]

Hi.

The datasheet I found on net doesn't mention resistance in windings. You may meassure this manually.

 

[ABOELLEEL]

primary is 0.5 mm diameter
secondary is 0.7 mm diameter
can it help?

 

[ABOELLEEL]

I readed about proximity effect and i will try to solve.

 

[Prototyp_V1.0]

My guess is that the transformer simply isn't rated to deliver that much effect.

 

[smxx]

tell me about transformer core material. for 70kHz you can't use Iron type.

 

[DeepOne]

half bridge 220V AC leads to 310V, 1A
freq = 70kHz
transformer etd34
primary 130 turns may be 31 turns
secondary 65 turns may be 2x13 turns

 

[ABOELLEEL]

---------- Post added at 14:16 ---------- Previous post was at 14:09 ----------

[/COLOR]

half bridge 220V AC leads to 310V, 1A
freq = 70kHz
transformer etd34
primary 130 turns may be 31 turns
secondary 65 turns may be 2x13 turns

thnx
but at turns 31 we will reach saturation( B = 0.3T)

i tried
primary 70
secondary 35
and reduced proximity
but transformer reach 70 degree i see that may be acceptable?

 

[DeepOne]

at turns 31 we will reach saturation( B = 0.3T)

on my mind 31 whorls in primary winding may give not more than 0.2T. Certainly if this is a half-bridge.

transformer reach 70 degree

I think 40 turns is optimal value and transformer must be no more than mildly warm.

 

[ABOELLEEL]

on my mind 31 whorls in primary winding may give not more than 0.2T. Certainly if this is a half-bridge.

I think 40 turns is optimal value and transformer must be no more than mildly warm.

Eqn delta B = Vi * T / (2 * N1 * Amin)

Vi = I/P voltage at worst case 190V AC leads to 270V
T = period 7 uS
N1 = primary turns 31
Amin = min cross section 91.6 mm2
delta B = 0.67T so B = 0.34T we must not exceed 0.2T

@ 40 turns i 'll try it

---------- Post added at 16:32 ---------- Previous post was at 16:30 ----------

tell me about transformer core material. for 70kHz you can't use Iron type.

Ferrite N87 material and i tried N27

 

[DeepOne]

Vi for half-bridge *0,5

 

[ABOELLEEL]

I know and i tried it but in that sit
**broken link removed**
numbers leads to that Vi is not the half its all the Dc pass
**broken link removed**
for calculation
and i saw reference have another Eqn and say that all the DC pass
tell me if i wrong plz

---------- Post added at 17:02 ---------- Previous post was at 16:59 ----------

my problem that i wonna to reduce temperature of transformer

 

[DeepOne]

possible I mistaken in this, but this easy check practically.
Quite well else see form of the signal on oscilloscope.

 

[ABOELLEEL]

Quite well else see form of the signal on oscilloscope.

its my problem how can i see that by oscilloscope.
i make calculations because i can not see that by oscilloscope.

 

[DeepOne]

I have calculated on inexact formula
w = V*0,5 / (4 * f * B * S)
150/(4*70000*0.15*0.00009)=39
Or possible use program "Lite-CalcIT", in russian, but understanding not difficult.
However similars that deal not in amount of whorl.
Possible connect keys of the inverter through light bulb and be orientated on current and temperature without load on secondary coil.

 

[akerue]

Hi Guys,
Based on the design information provided, and assuming correct circuit behaviour, the only two reasons I can think of the transformer overheating are
1) You have a gapped core
2) You have too many layers of copper windings in your bobbin window.

Peak flux density should be 37mT (which is peanuts at this frequency in terms of loss); but the skin depth in copper wire (before the effect of layering your windings) is 0.25mm.

My supplementary questions are does your transformer get hot on no/light load or only at/near full load? This would point to the copper wire type being the problem; but if you get the problems at no load then you have a problem with the core being gapped (which you don't want) or there being some asymmetry in your half bridge.

If you must use enamel wire (rather than litz), you might try twisting together several strands of smaller diameter wire (eg 0.2mm) to give a similar x-sectional area to the 0.5mm and 0.7mm wires. You should also consider interleaving the primary and secondary windings (with appropriate inter-winding tape insulation), as this will maximise coupling and reduce the H field build up across the primary winding.

Good luck.

 

[ABOELLEEL]

First thank you for your help

You have a gapped core

No i have not gapped core

You have too many layers of copper windings in your bobbin window.

yes i have 2 layers primary and 2 layers secondary but i put inter layer insulation (6 layers)
i 'll test at light load but a i ask you about 70 degree of temperature of transformer if it normal or not?

 

[DeepOne]

may be temporary insert transformer from old PC power supply unit (with half-bridge topology) instead of present transformer, аnd then compare?
Or screw on ventilator and downtrodden on this deal ).

 

[ABOELLEEL]

Hi Guys,

Peak flux density should be 37mT (which is peanuts at this frequency in terms of loss); but the skin depth in copper wire (before the effect of layering your windings) is 0.25mm.

so conductor diameter is 0.5mm is the best

 

[ABOELLEEL]

I make a test at light load and there is no temperature rise