120VAC to +V/-V devices?

I need a transformerless PSU, that can output something like +20/-20V @

3A. I see all the computer peripherals use such things already at different voltages/current.

Any schematics?

JJ

What you need is called a 'power supply' or PSU for short.

It can be done using a conventional 50/60Hz transformer and rectifier or by a 'switchmode' psu.

What's your specific requirement ?

Graham

On Wed, 19 Oct 2005 13:55:23 -0400, JJ top-posted:

Well, notwithstanding top-posting is frowned upon (start your reply _after_ the message you're replying to, and delete the portions of the previous message that don't apply - we call this "snipping".), and this sounds suspiciously like homework, the usual answer to questions like this is:

Good Luck! Rich

I dont think it can be easily done without some sort of magnetics. But try

they seem to be familiar with power conversion. Loads of app notes last time I looked

martin

Prefer as you will, but the protocol that is preferred by the majority of the people who post here, is that comments will be placed following the quote of the person to whom your post is addressing.

Unless you have an impairment that is helped by top-posting, you should follow the preferred format for this board.

Don

I prefer seeing the reply at the top, not having to scroll down to find a line or 2 of actual text.

Is it such a trivial homework task? I doubt it,I'm trying to avoid costly iron transformers for my own use. My old PSU is also bulky and generate a lot of heat.

I searched Nastional Semi and Texas Instruments site, none have a simple solution.

TI thinks Vin max should be 6V while NS max is around 60V.

Does any major US manufacturer have a one IC solution?

JJ

Perhaps by you but I prefer top posting especially with idiots like you who don't prune. So you can stick it.

Ted

Rich Grise wrote:

Usually, people in your situation buy switching power supply modules. Getting three amps at +/- 20 Volts from mains power with a single IC (no inductors or capacitors) is probably not possible.

Try lambda or power-one. I'm sure there are others. These are just the ones I've heard of.

--Mac

--Mac

When you stop being an arrogant top-posting snipper I might give you the answer you're looking for !

What's it worth ?

Oh, and the answer is yes btw.

Graham

But these are not transformerless! You simply cannot achieve a galvanic isolation from the mains without a transformer. The trick used inside the small wall-warts is that the mains is rectified directly, then converted again to an AC voltage with some electronics, but with a much higher frequency like 50kHz. This is then fed into a transformer again. But due to the high frequency, the transformer can be very small. After the transformer is a rectifier to create DC again.

Meindert

I have no problems accepting a small ferrite transformer. In fact, thats exactly what I need but what IC works up to 400V to do the switching? I have no problems adding a hexfet or 2.

JJ

I'm not trying to be difficult. My newreader has split windows and only show maybe top 20 lines. But if this group insists on that format I will ablige.

JJ

...

;-) Rich

Would this be your first switching power supply project? It sounds like a non-trivial design task, albeit, in the 120 watt range, it's conceivable.

Unfortunately, I'm also no help in the HV IC department - what do other switcher manufacturers do? You could go down to the local Radio Schlock or equiv, and pick up a little switching wall wart for about $20.00, and open it up and see how they do it. I've seen them with a zener regulator for the chip, and just discretes for the higher voltage and current parts.

As usual, google is your friend. :-)

Good Luck! Rich

--- Not true. Consider:

+------+ MAINS>--[C1]---|~ +|---+---[REG]--+-->+OUT | | |+ | | | [C] [C] | | | | MAINS>--[C2]---|~ -|---+----------+-->GND +------+

Where the dielectric of C1 and C2 completely isolates the bridge and all the rest of the circuitry from the mains.

For a bipolar supply:

+------+ MAINS>--[C1]--+----|~ +|---+---[REG]--+--> +OUT | | | |+ | | | | | [C] | [C] | | | | | | MAINS>--[C2]--|-+--|~ -|---+-----+----+ | | +------+ | | | +--> GND | | +------+ | +-|--|~ +|---+-----+----+ | | | |+ | | | | | [C] | [C] | | | | | | +--|~ -|---+---[REG]--+--> -OUT +------+

-- John Fields Professional Circuit Designer

--
AAARRGHHHH!!!!!

NO! NO! NO!

Check out

One of these won't quite supply your 120 watts. If you could back off to 100 watts then you could get both voltages from one xfmr and chip. For 120 watts you'd need 2 xfmrs and chips.

IR offers self-oscillating half-bridge drivers that can be line-powered (up to 600 volts), drive a pair of MOSFETs that could drive a single small ferrite xfmr at the 120-watt level; you could then rectify and regulate your + and - voltages on the low-voltage side of the xfmr with any of a large variety of regulators.

In that case may I offer you my commiserations.

I strongly recommend that you go to

and download their extensive range of app notes. And read them of course too. ;-)

They make numerous 'single chip' switchers that contain all the active electronics ( bar the opto-coupler for feedback of course ) including the power FET in a single package.

Very good supporting design software too.

120W is towards the upper limit of their flyback parts.

Graham

You can learn a few thing indeed from that line of approach.

From having had to learn about switchers recently I know that the magnetics will likely prove the trickiest bit to understand. You can't learn much about them this way though. It's down to good old theory.

Graham