SMPS 30 V 20 A

[mikai]

I am working on a 30 V 20 A SMPS. Following is the specification-
I/P - 180 VAC to 250 VAC.
Output - 30 V/ 20 A with current limit.
Configuration - Full bridge.
Frequency - 85 kHz.
Control method - Voltage mode using SG3525A.
DC blocking capacitor -1.5 mfd.
Transformer - EE42/15 core (EPCOS) with 27: (6+6) turns ratio.
O/p inductor - 20 micro henry.
O/p Capacitors - 4x2200 mfd 35 V general purpose 105 deg. C.
I/p capacitors - 2x330 mfd 450 V general purpose 105 deg. C.
What should be the type of the DC blocking capacitor? I am using VISHAY MKP1839HQ type. Will this do?
The problem I am facing is as follows-
Upto around 16 Amps of current the system works OK. Above that current the circuit starts generating noise after operating for around 10 seconds. If I keep the circuit ON after that, the MOSFETs (12N60) blow. I hv not included the current limit circuit yet.
Pls note that the stability network is type II, but the stability components are selected based on trial and error method, as the calculated components are not solving the problem of instability.
The feedback and control network is based on the secondary side and the MOSFETs are driven through optocoupler 6N139 and IR2110.

Can anybody pls help in solving the problem?

regds,

 

[crutschow]

Suggest you post a circuit diagram otherwise it's a guess as to the problem.

 

[mikai]

enclosed pls find the diagram. sorry it is hand drawn.

 

[mikai]

enclosed pls find the diagram. sorry it is hand drawn.

 

[ronsimpson]

It is almost imposable to know what is happening. Your circuit diagram is hard to read. Pictures of the board will help. Do you have a oscilloscope?
Is anything getting hot?
How did you design the transformer? I think you have 11 volts/turn. That seems high.

 

[mikai]

I have an oscilloscope.
Enclosed pls find the equations I hv used to calculate the primary number of turns. This is according to the Pressman book. I admit the drawing is not very clear. I will try to upload a better drawing. I will post the pictures of PCB shortly.
No component is getting very hot upto 16 Amperes. At full load, they do not have enough time to get hot as I have to switch off the system pretty fast due to generation of noise, or the MOSFETs are blowing.
regds,

 

[mikai]

Enclosed pls find the pictures of the PCB.

 

[ronsimpson]

The length of wire from the MOSFET driver to the MOSFETs is longer than I like.

Please show me how you got 27 turns for the transformer.

Do you have a current probe so you can see the current in the MOSFETs or current in the primary of the transformer.

Current in MOSFET with good transformer/inductor.
AND Current with core saturation.

Core saturation is not good!

 

[KeepItSimpleStupid]

Ron:

Do you think the OP would benefit by having some ferrite beads on the drivers?

Bypass caps in general?

 

[ronsimpson]

If the IR2110 does not have good capacitors around it then there will be all kinds of strangeness.

 

[mikai]

I don't hv a current probe. So seeing the current is difficult.
I am trying to reduce the length of the driver wires. What should be the maximim length? Currently it is between 3 to 4 inches and the wires are twisted inside the protective sleeve.
About getting the 27 turns, do you want to see the calculation? or the physical winding?
Calculations I hv written in the attachment equations.doc

 

[mikai]

The ir2110's are having 10 mfd tantalum at lo side and 2x1 mfd box in parallel at hi side.

Anyway as per your suggestion i am increasing x'mer primary no. of turns to 37. As per my calculation it should hv flux excursion of 1600 gauss at lo line. but current density will be very high.

But i am having reservation about the dc blocking capacitors. Is the type (vishay mkp1839hq) capable of carrying approx. 5 A of current at 85 khz? if not what type of capacitors should i use?

 

[ronsimpson]

Do you have a current transformer? Maybe one from another power supply. You can make a current probe.

The Metallized Polypropylene caps look good. They must be de-rated when used at high frequencies. Do you know what the AC voltage is across the cap? Is it getting hot?