If you put a capacitor and an inductance in series with your transformer, you can have a LC tank that resonate (I hope the English word is correct) at the exact frequency of your oscillator.
In this condition the output transformer is not fed with a square wave, but with a sinusoidal one; the IGBT switch on and off smoothly, cause the current rise and fall following the sine wave, and thus when you switch the IGBT on and off the current is almost zero.
You must calculate the impedance of your transformer, then tune the circuit (tank) in manner to set the center resonating frequency on your oscillator frequency (115Khz?).
600V IGBTs are fine if you're using, as told, resonant circuit: in hard switching mode you have hard spikes and overvoltage that could destroy your mosfets.
In my application usually I do prefer to use 1200V IGBTs, just to be sure...
In this condition the output transformer is not fed with a square wave, but with a sinusoidal one; the IGBT switch on and off smoothly, cause the current rise and fall following the sine wave, and thus when you switch the IGBT on and off the current is almost zero.
You must calculate the impedance of your transformer, then tune the circuit (tank) in manner to set the center resonating frequency on your oscillator frequency (115Khz?).
600V IGBTs are fine if you're using, as told, resonant circuit: in hard switching mode you have hard spikes and overvoltage that could destroy your mosfets.
In my application usually I do prefer to use 1200V IGBTs, just to be sure...
Well, I usually ignore the transformer leakage inductance, at least with frequency up 100Khz.
To calculate the exact frequency, give a look HERE.
Remember to place a trimmer in the oscillator circuit of your SMPS controller, in manner you can tune the frequency to match the components tolerance.
My circuit use a 84uH inductor in serie with a 0,12uF capacitor, giving an oscillating frequency of 50154 Hz; Actually, my oscillator is now running at 60Khz (I can regulate it from 35Khz up to 70Khz).
You can see when the frequency is correct when you'll have a really good output sine wave shape on your scope.
To calculate the exact frequency, give a look HERE.
Remember to place a trimmer in the oscillator circuit of your SMPS controller, in manner you can tune the frequency to match the components tolerance.
My circuit use a 84uH inductor in serie with a 0,12uF capacitor, giving an oscillating frequency of 50154 Hz; Actually, my oscillator is now running at 60Khz (I can regulate it from 35Khz up to 70Khz).
You can see when the frequency is correct when you'll have a really good output sine wave shape on your scope.
Well, if I were you I would substitute the 1 uF capacitor with a smaller value (0,047uF).
I would also reduce the frequency.
Remember: if your SMPS oscillator runs at 115Khz, on the bridge output you have only half of such a frequency (57,5Khz), because you use two cycles to drive the two IGBTs.
If this is the case you would redo your calculation.
The transformer itself has a really big inductance; mine has a calculated inductance of 2,3 milliHenry (2300 uH), so it does not interfere at all with the resonant tank.
For the same reason the secondary coils/inductance do not interfere too much.
Ciao,
Giovanni
I would also reduce the frequency.
Remember: if your SMPS oscillator runs at 115Khz, on the bridge output you have only half of such a frequency (57,5Khz), because you use two cycles to drive the two IGBTs.
If this is the case you would redo your calculation.
The transformer itself has a really big inductance; mine has a calculated inductance of 2,3 milliHenry (2300 uH), so it does not interfere at all with the resonant tank.
For the same reason the secondary coils/inductance do not interfere too much.
Ciao,
Giovanni
I have been looking at a couple of high power amplifiers with a half bridge smps, and they all have a primary series capacitor of about 1uF. If I use a 47nF cap it will have an impedance of 30Ohms at 115kHz. I think this is a but too much isn't it? So I think 1uF will be ok.
Oscillator frequency in my design is 230kHz, so the calculation is correct.
But, I think the primary series inductor is in series with the transformer leakage induction. So if you put a coil of 84uH in series with 2.3mH I don't think it will make much difference?
Is the 2.3mH measured/calculated with the secondary windings shorted? Otherwise I think this is not only the leakage inductance?
Oscillator frequency in my design is 230kHz, so the calculation is correct.
But, I think the primary series inductor is in series with the transformer leakage induction. So if you put a coil of 84uH in series with 2.3mH I don't think it will make much difference?
Is the 2.3mH measured/calculated with the secondary windings shorted? Otherwise I think this is not only the leakage inductance?
0.47mfd cap is also advantageous when the power supply is overloaded, because it will also tends to limit the current going through it in series. Secondly it will also help you in sustaining the resonance with lighter or offloads....If you increase the value of cap, then i think its difficult to sustain resonance during no load conditions.
What is the leakage inductance of your primary with secondary shorted......??
Can you put some waveforms?
Also see the qsc patent
http://www.pat2pdf.org/patents/pat5767744.pdf
What is the leakage inductance of your primary with secondary shorted......??
Can you put some waveforms?
Also see the qsc patent
http://www.pat2pdf.org/patents/pat5767744.pdf
I think the patent describes the first series of powerlights. I know that in the PL4.0 the resonant capacitors are 2x 470nF in parallel (approx. 1uF). There is no inductor, so it uses the leakage inducance of the trafo and secondary coils. (as the patent describes).
My LCR meter has a resolution of 1uH, so the measurement is not very accurate for smal inductances, but it shows a reading of 1uH. I think this means that it is in between 0.5 and 1.5uH.
What waveforms are you interested in seeing?
My LCR meter has a resolution of 1uH, so the measurement is not very accurate for smal inductances, but it shows a reading of 1uH. I think this means that it is in between 0.5 and 1.5uH.
What waveforms are you interested in seeing?
croccodillo said:
Use a higher value turn-on resistor, and a lower value turn-off resistor with a series schottky in parallel to achieve slower turn-on and "stiff" turn-off. I have never been in the need to use negative bias, at least with the modern IGBTs that I've been using and 70ns turn-on and 30ns turn-off @ 600V/30A. Note that IGBT turn-on threshold is usually around 5V.
Also, note that part of the spike could be actually EMI being picked up by the oscilloscope probe and also that most of the spike seen is actually produced at the leakage inductance of the emitter terminal rather than at the own G-E capacitance, so there is probably no problem to solve. Remember that you should put a common mode filter in the probe and measure directly at IGBT leads without any probe wire loops suitable to pick EMI.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.