Help with a DC-DC stepup power supply or other solutions

Thank you for your fast and professional answer.

I would like to mention that this source will be used for a ignition system. The interface i build will gave a life time of approximately 2.000 hours. Out of these 2.000 hours i may need maximum current only in 5% of the cases and when this happens the time frame will never exceed 400ms. For the rest of functioning time the current will be made of 2-3A spikes. Considering this situation i would go  for the inductor solution and I expect no issues. Am I right?

I am looking forward to hearing from you.

Laurian

I designed and built this circuit and did not work at more than 500mA(voltage drop 5-6V) with an inductor(4,7uH at 10A), diode DPG60IM400QB(400V, 60A) and 450kHz switching frequency. At higher currents(>500mA) the mosfet(no heatsink) is smoking in aprox 1 second.  If I use lower inductor 0.24uH(result from formula) the output does not reach 42V as it should and it stops at 36V. What do I do wrong? What can I do to improve the circuit and obtain more current?

Thank you

Laurian

I designed and built this circuit and did not work at more than 500mA(voltage drop 5-6V) with an inductor(4,7uH at 10A), diode DPG60IM400QB(400V, 60A) and 450kHz switching frequency.

At higher currents(>500mA) the mosfet(no heatsink) is smoking in aprox 1 second.

I checked the datasheet for the IRLU2905PbF MOSFET.  At 44V, this FET has a switching charge (Qgs + Qgd) of 33nC.  Even with 2A of gate drive, this will take 16ns to switch.  At 450kHz with an average switch current of 1.5A (and I suspect this design has higher switching currents than that) the switching losses estimate at over 900mW, so I suspect that your problem with the FET smoking is switching losses causing the FET to get too hot.  Based on that, we'd look at lowering the switching frequency.

If I use lower inductor 0.24uH(result from formula) the output does not reach 42V as it should and it stops at 36V.

You have a 10mOhm current sense resistor.  This implements a peak current limit of 12-18A.  with a 0.24uH inductor @ 12V, you'll see 50A/us inductor current slew-rate.  A current limit of 15A will trigger after 300ns which provides a maximum duty cycle of 13.5% @ 450kHz switching frequency.  The 36V point that you're seeing is the ouptut voltage that the repeated OCP faults are allowing the ouptut voltage to charge up to.

What do I do wrong? What can I do to improve the circuit and obtain more current?

First, your switching frequency is likely much too high for the power level you're driving.  I would try to work at about 100kHz switching frequency.

12V to 43V and a 0.5V diode gives 72% switch duty cycle and 35A average inductor current when sourcing 10A to the load.  Using 12V, 72% duty cycle, 10A and 450kHz I get an inductor value of 1.92uH.  Reducing this to 100kHz will drive a 10uH inductor for 10A peak to peak inductor ripple current.  Since your design will typically be running a much lower currents, you might want to consider an even higher inductor to reduce the ripple current further.  A 22 or even 47uH inductor.

In addition, to support 35A switch current with 10A ripple current, the peak switch current would be 40A, so you'll need to have a current sense resistance (Source of the MOSFET to GND) of less than 2mOhms.

All that said, I don't think the FET you've selected can support 35A @ 72% duty cycle.  This is 29.7Arms, which would dissipate 23.8W of power across a 27mOhm FET and a MOSFET with a lower Rdson will have higher switching losses, unless a more powerful driver is used.  For this much power, you may even have to consider several MOSFETs in parallel to disspate the energy.

Thanks a lot

Now it's more clear but I still need a little help with some advices. Because it;s high current on I found only 3 inductors on the market  that match what I need.  It's a 10uH with 30A  SC(Saturation Current), 5.6uH at 40A SC and 4.7uH at 55A. I think I use it 5.6uH at 40A saturation current. I considered 90% efficiency and at that high amperage in the battery voltage will drop to 9V. Output was calculated at 6A, 42V and peak-to-peak ripple is chosen to be 40%(Output current now can have 40% peaks too?) of the maximum input current. I want to exchange mosfet with this one: PSMN3R0-60PS who have more power dissipation and more current and less RDS (on).

• MOSFET, N CH, 60V, 100A, TO-220
• Transistor Polarity:N-Channel
• Current Id Max:100A
• Drain Source Voltage Vds:60V
• On Resistance Rds(on):2.4mohm
• Rds(on) Test Voltage Vgs:10V
• Voltage Vgs Max:20V
• Power Dissipation:306W

You can tell me if i make a good choice with this one? Or what features to look for when choosing a mosfet for my application?

Laurian

I redesigned the scheme and the PCB considering your advice but I came across some problems.
1. Measured switching frequency is much smaller than it ought to be (from the resistance and capacitor used). In order to check I put the resistance and capacitor in the example given in the data sheet. More specifically with a capacitor of 100pF and a resistance of 220K I got a frequency of 475kHz instead of 600kHz, just as it should be(data sheet).
2. I have problems starting the source if I use the resistance Rsense and SoftStart. I managed to start the power supply without Rsense and SS and no problems occurred even if I put a consumer of 2 - 3A (more it could not work because I had no inductor available with these features). I used a frequency of 90kHz and a 22uH inductor with a saturation current of 22A. In 8A, which is the power I need, I will use a 5.6uH inductor with a saturation current of 40A and for that a frequency of 250kHz resulted from the formula.

Observations. I have seen on the oscilloscope that their is noise on the ground path. I think there's a problem with the test PCB I have made because the ground path is quite long.

Thank you

Laurian