Buck regulator

[vwdevotee]

Hi all! First I want to say that I'm a mechanical engineer, so electronics are a little outside of my training, but hopefully this (and future ones) will be a simple question for people here.

Ok, so I think need a buck regulator but I want to make sure I've got the fundamentals down. I have 300V DC on the high side, and I need 15V DC on the low side. The 15V load will range from 0A to 30A. Will a buck be able to handle the variable current and maintain the voltage? With the PWM of the high side switch (FET, thyristor, IGBT, whatever), will the high voltage parts (and the inductor) need to also carry up to 30A?

Thanks!

 

[OutToLunch]

a non-isolated buck regulator would not be an ideal choice for this application. The voltage drop is too large (a 5% nominal duty cycle) as is the input voltage. If the control FET ever failed short you would have 300V on the output.

I would suggest either a forward converter or a flyback. You want the transformer to step down the voltage and isolate the output from the input.

 

[dougy83]

I have 300V DC on the high side, and I need 15V DC on the low side. The 15V load will range from 0A to 30A. Will a buck be able to handle the variable current and maintain the voltage?

A buck converter can handle these high currents so long as the components you use are sufficiently rated. Personally, I wouldn't use a buck topology for such a great in-out voltage differential; I would lean more towards a forward converter. The primary switch current will never need to be as large (less stress on switch), and the duty cycle won't be as low (so you can increase the switch freq). The output diode also doesn't need to block as much voltage, and can be a more efficient lower voltage schottkey. There's also the advantage of electrical isolation from the 300V rail, which is a big plus.

With the PWM of the high side switch (FET, thyristor, IGBT, whatever), will the high voltage parts (and the inductor) need to also carry up to 30A?

For 30A out, the inductor will carry an average current of 30A, the switch will have a peak current of > 30A (by low average current) & flyback diode will carry ~30A.

 

[Chippie]

You could modify an old pc power supply to get you some where near....

 

[vwdevotee]

Thanks for the feedback! Will forward and flyback converters be able to handle the changing output current? I'd hate to have to explain having shoved 30A into a 2A device. Thanks again!!

 

[Chippie]

I'd hate to have to explain having shoved 30A into a 2A device. Thanks again!!

It's not a case of shoving current anywhere, the current is drawn from the supply by the load be it 2 amps or 20.....

As I suggested earlier, a pc power supply suitably modified can supply the power you need without having to design something from scratch... but does require a modicom of ability....

 

[smanches]

a non-isolated buck regulator would not be an ideal choice for this application. The voltage drop is too large (a 5% nominal duty cycle) as is the input voltage. If the control FET ever failed short you would have 300V on the output.

I would suggest either a forward converter or a flyback. You want the transformer to step down the voltage and isolate the output from the input.

Trying to find a flyback transformer to handle 30A is going to be near impossible though, you'll probably have to wind one yourself. Although I agree that the input-output voltage difference should have isolation.

This is not going to be an easy buck regulator. That's a lot of current.

 

[vwdevotee]

I see. The way I was reading the specs on the converters made it sound like they were fixed current sources. As a MechE I took one class in electrical engineering, and it was very superficial.

 

[vwdevotee]

One more quick question about these converters. I'm looking at the schematics of the flyback and forward converters and not quite sure of something. Is the output voltage determined by the PWM of the input or by the ratio of turns in the transformer? Thanks for everyones help!!

 

[bountyhunter]

Hi all! First I want to say that I'm a mechanical engineer, so electronics are a little outside of my training, but hopefully this (and future ones) will be a simple question for people here.

Ok, so I think need a buck regulator but I want to make sure I've got the fundamentals down. I have 300V DC on the high side, and I need 15V DC on the low side. The 15V load will range from 0A to 30A. Will a buck be able to handle the variable current and maintain the voltage? With the PWM of the high side switch (FET, thyristor, IGBT, whatever), will the high voltage parts (and the inductor) need to also carry up to 30A?

Thanks!

You are building a 500W converter powered from 300V. The best choice is a half-bridge converter which gives inherent input-output isolation. A buck is a horrible choice because the switch will be on only a tiny % of the time.

 

[bountyhunter]

One more quick question about these converters. I'm looking at the schematics of the flyback and forward converters and not quite sure of something. Is the output voltage determined by the PWM of the input or by the ratio of turns in the transformer?

BOTH. You design the transformer so it has enough turns ratio to supply full power at lowest input voltage, where the ON time would be maximum.

 

[vwdevotee]

supply full power, or voltage? So it looks like I would need 20 high side turns per low side turn? Don't transformers require a certain amount of modulation to maintain the variation in voltage across the inductor? So then I couldn't actually let the PWM reach 100 percent?

 

[Chippie]

Power is applied as required....that is the job of the pwm...

Take the TL494 as an example, the max pwm 90% I think, this ensures that there is dead time between both transistors/fets being on..else there's a big bang....

Btw...a typical pc power supply has 40 turns on the high side

 

[bountyhunter]

supply full power, or voltage? So it looks like I would need 20 high side turns per low side turn? Don't transformers require a certain amount of modulation to maintain the variation in voltage across the inductor? So then I couldn't actually let the PWM reach 100 percent?

PWM never goes to 100%, it is limited to a value which depends on a lot of factors including the PWM driver IC being used. Typically 90% or that ballpark. You have to do a worst case analysis of voltage drops on all sides of the transformer, ie the drops on the FET switches and rectifiers on the output side, as well as Vdrop on the inductor. The FET's also have a positive TC so their resistance goes up with temp. Bottom line, the min voltage the primary will see has to be enough to crank out the full secondary voltage and current.

 

[bountyhunter]

Don't transformers require a certain amount of modulation to maintain the variation in voltage across the inductor? So then I couldn't actually let the PWM reach 100 percent?

The transformer is seeing voltage square waves at a duty cycle up to about 90% max, the peak voltage of those waves is dependent on input voltage and limiting factors. The secondary side rectifies and applies the positive square wave voltage pulses to the L-C output filter which smooths it to DC with some ripple voltage.

 

[bountyhunter]

supply full power, or voltage?

BOTH. Any power supply is specified for an output voltage at a max load current, so that specifies the maximum power it has to deliver to the load.

 

[Mr RB]

Just buy an off the shelf 240v AC to 13.8v DC switchmode supply, they sell big ones to power HAM radio equipment etc, i've seen 25A ones you can probably find a 30A one.

Then just connect it to 300v DC instead of 240v AC. If you really must have 15v instead of 13.8v out, you only need to change one resistor inside in the voltage regulator section. Most of these switchmode supplies use a TL494 control IC which is well documented, datasheet avaliable etc.