Help on buck converter

Matheus Lopes

Prodigy 70 points

So people, i am having some problems on how can i saturate a simple mosfet to make a simple buck converter.

My ideia is to get 12V on the input and have 9V on the Output with 2A.

I started my simulations but it is not working.

My circuit:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/buck_converter.jpg

The problem is that i need to put a signal of 18V to put the mosfet to conduct.

If I put only 5V, i can not get the same result.

http://img.photobucket.com/albums/v222/ahhh/Eletronica/buck_converter_2.jpg

So, how can i saturate the mosfet with a 5V signal to make my design to worK?

For now, i do not want to use any dedicated ci.

Thanks.

over 14 years ago

0 JohnTucker over 14 years ago

TI__Guru**** 191445 points

So which device are you trying to simulate? In general you have to drive the gate above Vin, so 18 V will work, but 5 V will not. Our devices have an internal BOOT circuit to derive the high side gate drive voltage.

0 Brigitte over 14 years ago

TI__Guru 65280 points

Hello Matheus,

A NMOS transistor needs higher voltage on the gate to turn it on. You will not be able to turn the NMOS on with 5V. How about replacing the NMOS with a PMOS? The PMOS is turned on when the gate is lower than the source and therefore this should work with your available 5V. You just have to add a resistor to VIN to turn the PMOS off.

In your circuit you also need an input capacitor to smooth the input current.

Best regards,
Brigitte

0 Matheus Lopes over 14 years ago in reply to Brigitte

Prodigy 70 points

Thanks Brigitte for your help.

I took your adivice and make some research here and found one project that can help me.

Lets slowly....So, I made this circui:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/circuit_buck-1.png

It is working but not has I want.

Now I can thrigger the FET with 5V, but let me show something:

This is the input of 5V and the output with 1Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out.jpg

Hum.. It is fine, hun? Ok, let me show now with a 10Khz 5V input:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_2.jpg

Still working...

But now with 50Khz, something strange starting to occur:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_3.jpg

Here you can see better:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_4.jpg

And finally with 400 or 300Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_5.jpg

The problem is that the mosfet is still conducting even when I put 0V on the gate. This is taking some time to "discharge" the energy on the gate.

Why? This is causing my problems now.

The project that I using is this one:

http://www.bineco.dk/ellert/smps.htm

The project is not in english, but I can see the circuti and interpret it...

Hope some help.

Thanks again.

0 Matheus Lopes over 14 years ago in reply to Matheus Lopes

Prodigy 70 points

I ploted some cicles of the input to see the voltage on the gate:

5V on the lower transistor and 300Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/vg.jpg

bye.

0 Brigitte over 14 years ago in reply to Matheus Lopes

TI__Guru 65280 points

Hello Matheus,

Please zoom into the drive signals with low frequency that you can see the details and you will observe that the signal looks very close to the signals with increased frequency. The problem you face here is that there is a lot of energy necessary to drive a MOSFET. Whenever you want to change the voltage on the gate of a MOSFET you have to discharge and charge the parasitic capacitances of the MOSFET. The most important capacitance here is the Miller capacitance.

When you now increase the switching frequency you also have to increase the speed in turning on and off the MOSFET, because the time to switch is getting lower and lower the higher you go in switching frequency. This is what you observe. And if you do not change the turn-on and turn-off times of the MOSFET, at high switching speeds, the MOSFET is no longer able to follow the switching. This is the problem you face and this is the reason why there are such a high amount of DC/DC controllers on the market.

In DC/DC converters or controllers, the integrated MOSFET drivers are using a lot of current (in the Amps range) to turn on and off the MOSFETs fast enough to support the high switching frequencies. Please check for example the following application note: http://www.ti.com/litv/pdf/slvae05 In addition you will find a lot information about buck converters when you search the web.

Best regards,
Brigitte

0 Matheus Lopes over 14 years ago in reply to Brigitte

Prodigy 70 points

Thanks for your help.

I took a driver mosfet called Totem Pole!? And simplified it to just to test. The ideia is the same:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/buck-1.jpg

So, output with 10% duty and 15Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_ok_10_duty.jpg

So, output with 50% duty and 15Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_ok_50_duty.jpg

So, output with 80% duty and 15Khz:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_ok_80_duty.jpg

Let see the ripple:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_ok_50_duty_10m_30m.jpg

High? No, it is not:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/out_ok_50_duty_29m_30m_ripple.jpg

Only 16mV.

So now, lets connect the PIC and try to read the output voltage with a resistor divider and adjust the duty to obtain the Vref of 9V:

Sweet:

http://img.photobucket.com/albums/v222/ahhh/Eletronica/vout_50khz_com_pic.jpg

This last test was with a 50Khz pwm into the lower transistor.

Well, it looks like that is working! Hun?

Coments?

0 JohnTucker over 14 years ago in reply to Matheus Lopes

TI__Guru**** 191445 points

So I still did not find out which TI dc/dc converter you are trying to model. What is your application?

Please let us know.

Power management

Power management forum

Help on buck converter