Hi Guys,

I need help on designing a Floating Power Supply. The voltage of this power supply can be set from 0V - 10V and has a maximum continuous output current of 100A.

 

Most conventional transformer types fit that description. Most wall warts too.

Would 1.25V as a minimum work? There is a really cheap and easy to use voltage regulator chip (the LM317) that is available, but it can only go down to 1.25V.

Have you read this?

Basic Bench Top Power Supplies

 

Hi Guys,

I need help on designing a Floating Power Supply. The voltage of this power supply can be set from 0V - 10V and has a maximum continuous output current of 100A.

100A that is a lot of current. What are you building for?

 

I'm not sure with the 1.25V minimum voltage. Our application requires a minimum voltage of 1V. If 1.25V will be the minimum voltage, I think producing a 1V won't be a problem. I guess using a resistor network can get me a 1V output. Am I correct?

We are building a power supply that will be use for Mosfet's. We have different type of Mosfets and one of it has a spec of 1V, 100A.

If I will be using the LM317, can I have build a floating power supply because that is what I need, a floating power supply?

 

I'll get back with a schematic when I can, if someone doesn't beat me to it. I'd go with a simple transistor emitter follower, probably Darlington or Sziklia pair.

 

100A that is a lot of current. What are you building for?

 

Guess I need to get my eyes checked, I saw 100ma. What is your source for the voltage and current?

There are two ways to go about this, one is an analog supply. It is simpler to build and the output is a lot quieter, but it is going to be a space heater.

The other way is a switching power supply. It will run a lot cooler but will be a lot noisier.

Is noise important in this application?

It might be possible to create a hybrid, a SMP feeding an analog regulator, to get the best of both worlds.

 

Even if the mosfet can withstand 100A that doesn't means that you need 100A to flow in the circuit. What are you specifically going to do with the mosfets?

If you really need 100A, your only chance is to use a switching supply, because a linear regulator regulating 1v 100A from, say, 12v, would dissipate 1100 watts of heat, which is basically an electric stove.

 

As of the present, our application is only up to 10V,4A and we are using an external power supply for this. Our management is trying to bring in another device but with a 1V, 100A specs. We have a board from our ATE that can supply this current since it is only a 1us pulse. We want to eliminate external components so we decided to incorporate the power supply into our hardware. So if we will make our power supply it must have a specs of at least 10V, 100A to meet the 2 specs requirement.

 

You want a 1 us 100 A pulse?

I think you'll need to look at multiphase buck converters like they have on computer motherboards.

Edit:
Found this whitepaper.
http://www.irf.com/technical-info/whitepaper/multiphasebuckapec03.pdf

 

You may not need a 100A continuous supply.

Even if you have a 1V to 10V at 100A supply you will have problems to get that supply to give up a 10V, 100A pulse accurately for only 1 uS. The electronics just won't react fast enough. You will need to store the energy in capacitors first and then dump them into the pulse.



The attached, 100mA continuous, 100A pulse, 7% duty-cycle, supply circuit has 3% error at 10V and 30% error at 1V. Is this good enough?

• The lower the ESR of the caps the better.
• Keep leads short.
• Use 12 gage wire.
• The tantalums go next to the supply.
• The ceramics go next to the pulse.

Good Luck,
Ifixit

 

It is ok if the components cannot react fast. We have a fast analog switch on the output of the supply where we can just cut off the connection of the supply to the circuit. With this switch we can disconnect the supply after 1us. I will try this circuit.