high watt smps

There are many ways to go, many things to consider when designing a switching supply. For starters the fact that you have opened this thread here without many detailes could suggest that you lack the proper experiance, so it would be that much harder to get it right, but still i would have some questions designed to get things started:

-Where will the PSU get it's power from, a car battery ( or something like that ) or will it be an offline SMPS ( 220/110Vac )?
-What is the prefered configuration? ( be aware that each has it's own advantages as well as dissadvantages ), will it be Forward ( Half-Bridge; Full-Bridge; Push-Pull, etc. ) or Flyback?
-What is the prefered control method, Voltage or Current?
-Does it need to be a regulated voltage PSU or not? ]
-100V output means 2 identical rails ( possitive and negative with respect to a ground ) +/-50Vdc or just one output of 100V?
-What is the prefered action of the current limmiting section, block the unit at some current threshold with reset needed, or does it need to be a Constant Current Source?

It is important to pay enough attention to those detailes, they will set the way to go and make getting help much easyer.
At 1000w the flyback SMPS would be out, they are only good upto about 300w.

I have just finished a project using the IRS27952 and am getting 70 volts at 5 amps so that's 350 watts. There is no reason why you could not scale it up to 1000 watts.

If you ask lorylaci nicely he might let you have an excel spread-sheet that helps with the design a lot.
At 1000w the flyback SMPS would be out, they are only good upto about 300w.

That depends on the available input voltage, off line high wattage flyback smps's are commonly used as capacitor charging supplies, there are two 5kW ones at work. Currently I am building a 100W flyback capacitor charger which is a 12V input version of a 1000W 110V input design. Admittedly they are fairly specialised and do not make best use of the the magnetics or switching device.


2013-03-05 11:01 am
sorry for delay
-its input voltage is about 220v and 50hz
-not matter about its configuration...(fly back or forward)
-it must have variable voltage (0-100v) output with feedback for voltage drop
-its regulation must be about 1%
-and feedback for protection of overload

i have a project like this but it has AVR programming for fire angel of tristors and i wanna remove that! i dont want to use tristor
Ok now we are getting somewhere...

So it will be an offline SMPS, ok now at 1000W of power i think everyone agrees with me that it should be a forward one, ( flyback would be insanely complicated at this power ). Now let us choose first the precise configuration ( forward PSU are comprise of some subcategoryes ), you may choose from theese 3:

1.- Half-Bridge, it has the advantage that relatively low voltage power devices can be used, 400V or so would be ok, this comes with one main dissadvantage, the voltage on the primary of the transformer is rather low ( somwhere around 140V ) so the bigger the power needed the more current must flow trugh the primary ( thus the power switching devices and everything else in the HV stage ), you must also take into account the lower vallue of the line voltage witch sometimes can be maibe under 200V, so the primary voltage drops even lower, i have found that a minimum primary voltage of about 100 is rather fine for ensurring that the output will not suffer with line drop, so it would be theoretically 10A trugh the primary but practically it's more than that, the efficiency will not be 100% so you have to take that into account when calculating the primary side current rating, about 80% keeps you on the safe side, so that would mean about 1250W in the primary side and so something arround 12A. So this is the dissadvantage of the half-bridge, a rather high current in the primary side, i usually chose the power devices as souch that the current trugh each is about 1/3 of it's max rating, i think 2 power transistors ( be they MOSFET or BJT that's your choice ) of minimum 400V and 15A each rail keeps you on the safe side with this config.

2.- Full-Bridge is somewhat a more complex version of the half bridge, it has the advantage of high voltage on the primary ( ~Vin ) so a much smaller current needed ( about half that of the HB ), the power devices can still be rated at about 400V but the dissadvantage is increased complexity for the drive section.

3.- Push-Pull like FB allows a smaller current in the primary side cus of the high voltage involved, but also a bonus witch is a more simplified control/drive section, there is no need here for High side/Low side driver, just some totem poll ( if mosfet is used ) to supply the needed current for high speed switching ( there are specialised drivers, simple to use an cheap ). But all of this comes at a price, this time the power devices has to be rated a a voltage of at least 2*Vin, ( usually it is used 3*Vin ) so for offline SMPS you would need power devices with at least 800V rating, and they are not so cheap.

Now knowing all of the above you can choose the config you think it suits you, and also what tipe of power devices you prefere, BJT or MOSFET ( i would suggest the latter ), after theese things are well stated and chosen we can go forwarth with the design.

Personally I'd use the following architecture:

- PFC stage, stolen from a Fairchild app note, outputting 380-400V.
- Full bridge architecture, built with 560-600V MOSFETs. Put a big film capacitor in series with the transformer primary to prevent flux walking, and leave plenty of room for snubbers.
- A pair of IR2110 or similar gate drive ICs to drive the FETs
- Voltage mode control, but use a chip with OCP.
- Synchronous rectified output, to keep the converter in CCM over its entire current/voltage load range, making compensation easier. Drive the rectifier gates off extra windings on the transformer.

0-100V is difficult to pull off with a switcher, because of minimum duty cycle limitations - going for something like 10-100V is more practical. If you really need 0V, I'd put a linear post-regulator on the output of the supply.
Personally I'd use the following architecture:..
All of that sounds good ( maybe even hi-end ) as an config but you have to take into account the thread author's experiance in developing souch units ( switching PSU ), what you propose needs a good knowledge and understanding of the switching technology, it is not something for the begginers, and although i cannot possibly know if @hamedyas is a begginer ( i appologise if it seams like an offence, i give you my word it isn't ), but i think he is not verry experianced designer otherwise he would have came up with a much more elaborated first posting, at least some basic design witch only would need some improoving, but giving only so little and asking for so much has to say something. What is my point? well, i think it should be kept simple, as simple as the requirements allows, loose the PFC ( witch adds to complexity and cost ), loose the syncronus rectifyer witch also adds to complexity and cost, half bridge to keep the drive section simple and easy to get good parts ( power switchers mainly ), voltage mode control for good transient response and load/line regulation ( granted compensation might get tough but it can be done easily enough ), and if the load powered by it does not protest a simple overload/short protection witch will block the unit at the desired threshold.

PS: I agree with the voltage range at the output, a minimum of at least let say 5V would be required, it all depends on the chosen controller's possibilityes. Anyway, for the life of me, i cannot understand what use do people find for 0V output?