Push pull converter

[milli1001]

Hello, I have designed a push pull converter to step up 12V DC to 200VDC to be used as the DC link of an inverter. The issue I am facing is when I connect the transformer I get basically a shirt circuit and the voltage at the circuit input drops down to 4V. I am using a 14V input from a 3amp transformer (rectified and filtered). The High frequency transformer is rated about 200W. It's 8+8 turns primary and 140 turns secondary. I use the EE42/21/15 ferrite core. Any idea why this is occuring?

 

[BradtheRad]

Your transformer admits entire power to the primary but fails to convey equal power to the secondary.

High frequency transformer

Then your switching frequency must be high. It's a mistake to apply a slower frequency than a transformer is designed for. Current ramps up to a greater level in the primary because the longer cycle allows more time for current to ramp up. However the flux field may reach saturation early and stop building.

Take gradual steps. Start out applying low voltage AC at the input. (It may even be a good idea to use sine waves.) Limit current into the primary windings.
Notice what waveform emerges from the secondary to your load. Is it true bipolar AC? Is step-up ratio what you expect?

Increase applied voltage. Increase current into the primary. Continue to check waveforms, amplitude, temperature rise in components.
When load voltage starts dropping then it's a sign the system is no longer working efficiently. Then start diagnosing for problems.

 

[milli1001]

Thanks for your response, note that the frequency I'm switching at is 50khz (from a tl494 in push pull mode) however this is divided to 25kHz since the mosfets are off half the time. Is 25Khz considered too low? Also at the output I do get a voltage (20-30V) measure after rectifying with 4 fast recovery diodes. I will try your suggestion of using a lower voltage and then gradually increasing. Do you suspect at all that it could be the supply since it's rated at 3 amps I'm wondering if the HF transformer needs more but based on the #of turns, under no load it should not even be drawing as much as it is now (2.5-2.8) amps.

 

[Easy peasy]

check the phasing on the primary windings

--- Updated Mar 29, 2023 ---

On an EE42 - take the turns up to 18 +18

--- Updated Mar 29, 2023 ---

Bifilar ... for lower turn off losses

 

[BradtheRad]

I once tried to make a power inverter from a center-tap transformer (operated in reverse direction). My attempt burned up several mosfets. I started at a low power level, then as I increased supply voltage, each shut-off event generated a severe spike. It's a characteristic of inductors switching high current.

I gave up because all I could get was anemic voltage from the secondary. Of course I barely knew what I was doing then.

when I connect the transformer I get basically a shirt circuit

The large step-up ratio magnifies everything going through the primary windings. A light load on the secondary allows higher output voltage. With no load attached the voltage can get alarmingly high. Then a severe spike in the primary might generate sufficiently high voltage that arcing occurs. In other words a shorted winding. Examine your secondary to make sure no turns are welded.

To build a successful transformer manufacturers evidently apply special knowledge to make it efficient at all power levels, and to draw minimal current when no load is present.

 

[milli1001]

check the phasing on the primary windings

--- Updated Mar 29, 2023 ---

On an EE42 - take the turns up to 18 +18

--- Updated Mar 29, 2023 ---

Bifilar ... for lower turn off losses

What exactly do you refer to as "phasing on primary windings"
Also since posting I have increased primary turns up to 12+12 now and still seeing the same results so I'm not confident 18+18 will fix the problem much but I will try it. As for bifilar for lower turn off losses, I have used 5*0.5 mm wire for each turn on the primary side. Is this ok?

 

[dick_freebird]

Every core+winding duo has a volt-seconds (dt=Isat/L) limit and
you have to look at both open and shorted secondary as your
"goalposts". It would be good to instrument up the primary
current (or both horns, using a small B+ resistor?). You can
look for driver strength and core saturation issues in the
waveform - weak driver will "roll over" or flatline, core sat will
"take off like a rocket" (until backstopped by winding and
driver and power supply R).

 

[Easy peasy]

Phasing refers to the DOT NOTATION on the push pull windings - you cannot connect the primaries in any fashion

( usually if you have 2 wires - bifilar - these are called the start(s), and the ends of both are called the finish(es) - you connect one start to one finish - not the same wire - to provide the centre tap for vin - if this is done the phasing will be correct - i.e. the dot of one wdg goes to a drain, the dot of the other goes to Vin )

it must be so that when one wdg is energised it raises the volts on the other mosfet drain wdg - usually to 2x the Vin ( + any spikes )

else you will ramp up the magnetising current and have large turn off volt spikes

9 + 9 turns - bifilar - gives approx 66.5mT each way in your core @ 14V drive - @25kHz switching - which is a good starting point for an unknown core material - and allows for a bit of asymmetry in the driving waveforms - unless you are using peak current mode control any small asymmetry will cause flux walking and Tx saturation and draw excessive current from the power source

time to research the push-pull more thoroughly

--- Updated Mar 30, 2023 ---

https://www.ti.com/lit/an/slyt813/slyt813.pdf?ts=1680080337165&ref_url=https%253A%252F%252Fwww.google.com%252F

--- Updated Mar 30, 2023 ---

Also if you provide enough dead tine between power pulses, say 45%, 5%, 45%, 5% max then there is an element of self reset provided by the energy in the core

 

[milli1001]

Thank you for your response. I have actually referred to that doc link you provide numerous times. It seems I overlooked this design principle. Upon doing some research on the concept I'm still not fully sure how this is implemented, is it done by the directions in which the wires are wound onto the transformer? In my transformer, I have kept all the turns going in the same direction. How do I go about ensuring I get the correct orientation based on the Dot notation?

 

[Easy peasy]

look up bifilar - and then re-read the above