Can a Forward Converter Work with 385V Input for 12V Output?

[kahei]

Hi. I would like to do a dc-dc converter which have output voltage equals to 12V and current is between 1.1A and 100mA. But, the input voltage is around 385V which is the voltage output of the PFC part.

Since the voltage into the buck converter is so large, I would like to make a forward converter. But, I don't know which IC should be used and I have tried UC3844.

Does anyone would like to give a suggestions or the design steps to me? Thank you.

 

[berkeman]

Hi. I would like to do a dc-dc converter which have output voltage equals to 12V and current is between 1.1A and 100mA. But, the input voltage is around 385V which is the voltage output of the PFC part.

Since the voltage into the buck converter is so large, I would like to make a forward converter. But, I don't know which IC should be used and I have tried UC3844.

Does anyone would like to give a suggestions or the design steps to me? Thank you.

I googled forward converter ic, and got lots of hits, including the IC you mention. Here's the hit list in case it helps:

http://www.google.com/search?source...z=1T4GGLL_enUS301US302&q=forward+converter+ic

Heck, there's even a Simple Switcher forward converter (I didn't know they had one).

 

[Mike_In_Plano]

If money isn't so much of an issue (like it ever isn't!), you may wish to look into the UC3801. It doesn't pull as much current as the 384x and has a soft restart feature that can save you. What it doesn't have is the ability to source much current from the error amp, so you have to be careful how you load them.

As for the supply, it's not so much a question of the chip as it is of experience, time and equipment. I'd avoid a forward converter at such a low power level. They provide their own issues because the transient currents during comutation of the output rectifiers. This adds to the troublesome current spike that appears across the current sense resistor and makes it harder to design a good current sense network.

The easiest solution is probably to buy a ready made chip from Fairchild. These come complete with the MOSFET and controller in one.

Next easiest is to use the chip and FET, but make the topology a discontinous flyback. This will avoid a great deal of noise and current sensing issues at the cost of efficeincy and transformer size. BUT good engineering isn't about perfection. It's about making good use of your resources.

At one time, the majority of the offline power supplies were comprised of a UC384X, a cheap FET, a cheap opto isolator, and a TL431 shunt regulator. I'm not up to going into this special combination, but there are plenty of devices from monitors, to computer supplies, and even wall warts that used this circuit, so I suspect there are schematics floating about.

Best luck,

- Mike

 

[Antiphon]

Consider designing a flyback converter instead of a forward. They are simpler, cheaper and more flexible in that you can adjust the output voltage as well as the current using the feedback loop. With a forward, the voltage conversion is defined by the transformer turns ratio alone.

Look up the application note for the NCP1053. You will not find an easier way to make a buck stage.

 

[Mike_In_Plano]

The forward converter is mroe than a transformer followed by rectifiers, and the voltage from it is easily regulated through manipulation of the pulse width, or even better, through manipulation of the peak current during the switching cycle (as this applies constant volt seconds across the output inductor and rejects the input ripple).

A buck converter from 380V to 12 is pretty silly, unless you need a few milliamps for an appliance. Even then, it isn't isolated for safety.

 

[Antiphon]

Mike,

the forward converter is an isolated buck converter. The maximum output voltage is defined by the input voltage and the transformer turns ratio.

The flyback is an isolated buck-boost. It can raise or lower the output voltage to match a varying load like for example an unknown number of LEDs.

The forward converter design is more complex than the flyback due to need for a proper core reset. A flux walkaway will destroy the converter in a fraction of a second. A flyback doesn't need a core reset circuit even when driven in the continuous conduction mode. It is inherently simpler, safer, cheaper and more flexible than a forward. The forward excels at utilizing the core so if you need a small supply with a high output power (like an ATX compeer supply) than the forward is for you. But a 15 watt supply? You're way way better off to go with a flyback which is why it's used in 80% of all commercial power supplies.

Open up your laptop's AC adapter. It's a PFC followed by a flyback. Gauranteed.

 

[kahei]

Thanks for the suggestions. But, my product needs to do a buck (forward) converter to step dwon the 380V d.c. Would the UC3844 also act as a controller in the circuit?

Besides, I would like to have the simulation first before making it practically. Does anyone have the PSpice model of the IC UC3844?

Thank you.

 

[berkeman]

Thanks for the suggestions. But, my product needs to do a buck (forward) converter to step dwon the 380V d.c. Would the UC3844 also act as a controller in the circuit?

Besides, I would like to have the simulation first before making it practically. Does anyone have the PSpice model of the IC UC3844?

Thank you.

A buck and forward converter are different topologies.

 

[kahei]

Then, if I would like to use a forward converter, how to use UC3844 to do it, or if UC3844 do as a controller in a forward converter?

Does anyone has the simulation PSpice model in this IC?

 

[Mike_In_Plano]

Hello Kahei,

You can get a free copy of LTspice at linear.com to open the given model
I think this path will get you a model for UC3844A
http://tech.groups.yahoo.com/group/LTspice/files/adventures_with_analog/my model files/

While you're on that page, pick you up a copy of the TL431A. It's good, cheap, and multisourced and will provide you with regulation from the output side.

As for the notion of forward converters, i'd get it out your head for his small a supply. It will cost you more money, be more grief to design and debug, and generally won't offer any benefits at you power level.

Don't forget when modeling to get an idea of the leakage inductance of your transformer. It's totally, without a doubt, essential to having a good model. I'd talk with your magnetics designer, or even better, build a transformer and test.

- Mike