designing an intelligent and efficient power supply

hi,

i have to design such a power supply which is provided with only one dc source while it has to generate:

2 positive polarity signals , one greater than while the other one is smaller than the input

2 negative polarity signals with above mentioned conditions

3 phase signal with some hundreds of frequency

a carrier signal of kHz frequency.

the design needs to be efficient and highly stable as well as compact and cost effective, as it is my final year project.

can any one suggest a design please.
it would be very well if the design would be transformer_less.

 

Since this appears to be a homework assignment, it should've been posted in the Homework Help section.

While we can help you if you are "stuck", we cannot design it for you. That would defeat the entire purpose of the exercise. You will need to do a lot of research.

But first, you need to define the requirements as completely as possible.

 

You definitely need to state the output power and voltage required for all of those to get somewhere.

For the first two points I would use some SMPS similar to an ATX computer power supply, if the voltages and currents are similar.

 

Are we talking millivolts or megawatts?

 

no im not asking the exact and accurate deign, thats why i have not mentioned the exact ratings. i just want suggestions about the design. i mean the best of the possible ways and reason for that. it may contain new things as well as already searched things then i will select
or if there is a site that can help designers then indicate it to me.

 

The problem is that suggestions about the design are useless until you tell us the power, voltage and frequency range of the power supply and generator.

You simply cannot use the same design for voltage reference and for 5KW source.
Is that understood?

 

The design process begins by defining the required parameters. Without defined parameters, there can be no design.

 

So this is my patented hi-tech design that covers all your needs. Is it enough?

 

ok .
the input is 28v dc with 112w power approx
dc signals generate power of 75, 33 and 15 watt
per phase power of ac signal is 7watt aprox

 

112 is less than 75+33+15+7+7+7
You can´t have more power on output than on input.

I don´t think the number of powers matches the number of outputs either. You wanted 4 DC, 3 AC and one clock, is that right?

 

ok let me tell you the exact ratings

its 28v dc input with 4a of dc current.
it has to generate

+5v with 3A
-5V with 0.1A

+15V with 1.5A
-15V WITH 1.5A

+50V WITH 1.5A
-50V WITH 1.5A

I WISH THE DESIGN TO BE AS MUCH PRECISE AS IT COULD BE AND EFFECIENT AS WELL.
i have prepared a block diagram and right now serching for the particular components but its getting troublesome.

leave the portion of three phase generation. i am about to accomplish that much portion.

 

OK, that's 112 Watts.

15W.

0.5W.

22.5W.

22.5W.

75W. We're now 23.5W in the hole.

75W. We're now 98.5W in the hole.
If you take into account reasonable efficiencies for power conversion, you might get somewhere between 85% and 93%. So there's another (approximate) 25W in the hole.

Oh, no 3-phase AC yet.

There is absolutely no way that you can build this supply without increasing your 28VDC supply current considerably. I'll venture that you'll need at least 250W, or 9A from your 28V supply just for the above voltages @ current. You'll need more for your 3-phase AC.

But what you're asking for as of right now is simply not possible - unless those are maximum currents, not typical currents.

 

o yes.
after posting here, i also did consult to my project supervisor for this issue of low input and high output power and he has suggested to increase the supply current to 10A(or as high as i would require to generate these ac dc signals).

now plz suggest me how can i built a compact design. i mean i wish to use buck and boost converters and pic microcontroller for three phase signal.

plz indicate me of your suggestions and of any possible issue in this regard

 

I think you should use half-bridge topology like this for the +/-50V and +/-15V part, because they are symetrical and have the same current.
Then two buck converters for the 5V, and don't know about the 3phase.