Voltage Doublers

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Thanks NY Dave. I always have to pull out an old reference to remember how the general times n multiplier looks.

One thing they don't discuss is the case when you tack on multipliers to existing bridge-rectified supplies. I have a way that I used to do a full-wave version of this that I will post as soon as I get my schematic capture-export-post act together. I'm sure it's been done but I've never seen it in print. It has utility because, otherwise, the half-wave tack-on puts some d.c. through the transformer and has half the ripple freq of the full wave. It can be used to generate complementary rails above and below the initial bridge-rectified supply, each about equal in delta V to the initial bridge voltage.
 
> as soon as I get my schematic capture-export-post act together.

Get the plan on-screen.

On Windows, Ctrl-PrntScrn will copy the screen to the clipboard. Mac OSX has a similar function.

In a graphics program, paste the clipboard to a new image. MS Paint will work. Trim the excess. Save. GIF is best for line-draw.

Post the file to public web space. ISP accounts often have 10MB of free HTTP space. Decipher the ISP's obscure help-pages to find out how to access it. Figure out what URL it appears at. This may be the hardest part. In fact, on a hunch I looked through SBC/PacBell (very popular ISP) help pages and don't see a public webspace feature. Golly, dumb old CompuServe had that years ago, a whopping 2Megs for free.

Put that URL inside the IMG tags in the forum posting editor.

Full-screen capture is awkward, and MS Paint makes easy things hard. I like PaintShopPro, which has a snapshot feature to grab part of a screen, and a simple Crop tool.

My other technique is pencil, paper, and scanner. If I don't already have the idea open in a simulator, this can be far faster. I scan to PSP so I can rotate (I can't draw or scan square), bump contrast (pencil scans as light-grey: it is specular), and replace sloppy lettering with font-text.
 
I'm quite dense. It often takes more than one explination from more than one source (person) to get this stuff through my thick skull. So if we are looking at this basic doubler:
voldoub.gif


Hmm, now I realize that I am assuming that the wires coming off the transformer are not shorting together there before reaching the diodes.

I understand the top diode takes the top half of the wave (minus the .6 drop) and the bottom diode takes the bottom half of the wave.

Now the top diode will charge the top cap.

Does the bottom diode feeding the negative of the bottom cap raise the ground (floating ground), and then when the other leg of the transformer goes between those two caps, it is seeing it as doubled voltage because of the raised ground?

I am probably WAY off.

Sigh.

Joel
 
Consider the circuit with just the top diode and cap. Every positive half-cycle of the a.c. out of the transformer, the diode charges the cap to the peak positive value minus diode drop. On negative half-cycles, and indeed as soon as the transformer positive half-cycle falls below the cap voltage, reverse-biasing the diode, the diode blocks---only conducts leakage current. It continues to block during the entire remainder of the full cycle.

Consider the circuit with just the bottom diode and cap. It works just like the top circuit did, except that when the top lead of the transformer goes to the peak negative value, the diode charges the capacitor to that negative voltage, relative to the other transformer lead (the one you have marked "+" for some reason), minus the (negative) diode drop. When the top transformer lead gets a little less negative than the cap voltage the diode blocks, and continues to do so during the entire remainder of the full cycle.

Now put the whole circuit together. You stack an upper nearly-peak voltage on top of another nearly-peak voltage.

If you want to consider the junction of the two caps as "ground" you have a positive voltage at the top and a negative voltage at the bottom. But this assignment is arbitrary. Call the bottom of the lower cap ground and the top the + d.c., and you have about doubled what would otherwise be the voltage developed by a half-wave rectifier or a bridge rectifier feeding a single cap. Call the top "ground" and you have a negative d.c. output on the bottom with about twice the negative voltage you would have otherwise.
 
> raise the ground (floating ground)

What ground? I don't see no ground.

The transformer winding is fully "floating". Think of a flashlight battery in outer space, except AC. No ground is present or needed. If you have a ground and want/need to connect it, you can connect it anywhere in the circuit. TO top, middle, or bottom of the cap-string, or even to the diode end of the transformer (though I don't see any use for that).

Practical items can matter. Common transformers can "float" up to about 500V across the insulation. The primary side is normally grounded at the utility company and fusebox. So if you want to multiply-up a high voltage winding to a very high voltage DC, you can get in trouble. If the winding is 200V, each cap charges to 280V. If you ground the bottom of the cap string, the center of the cap-string is 280V, the top is 560V, and the top terminal of the transformer will peak at 560V above ground. One end of the primary is at ground, the other end peaks at 170V or 330V, and maybe in opposite phase. So you could end up with 560V+330V= 890V across the paper insulation between windings.

That is why there is a second version with one end of the AC source (optionally) grounded. It is less efficient, but less abusive of the transformer insulation. If you need 20,000VDC, you will probably start with about 1,000VAC (which would probably have >2,000V insulation) and run about 20 stages of rectifiers and capacitors (on tall glazed porcelain insulators!). (Actually, for small current, you would pull the specially insulated 20KV transformer out of an oil-burner or neon-sign.)

> assuming that the wires coming off the transformer are not shorting together there

That won't work too well. While all schematics SHOULD have dots and bump-overs or equivalent notation to differentiate crossings from connections, sometimes you have to guess and pick the most likely (or less unlikely) interpretation.
 
[quote author="bcarso"]Thanks NY Dave. I always have to pull out an old reference to remember how the general times n multiplier looks.

One thing they don't discuss is the case when you tack on multipliers to existing bridge-rectified supplies. I have a way that I used to do a full-wave version of this that I will post as soon as I get my schematic capture-export-post act together. I'm sure it's been done but I've never seen it in print. It has utility because, otherwise, the half-wave tack-on puts some d.c. through the transformer and has half the ripple freq of the full wave. It can be used to generate complementary rails above and below the initial bridge-rectified supply, each about equal in delta V to the initial bridge voltage.[/quote]

Any chance of seeing this multiplier, it seems like it is just what I am looking for.

adam
 
adamasd, thanks for bumping this thread. It is the one where PRR explains to me how to print screens into notepad and do gif conversions (which I tried and promptly forgot afterwards!).

Actually, the circuit that I hadn't seen showed up in two other threads almost as soon as I made that post! It just looks different than I had drawn it when I happened upon the same topology.

The way it appears on other threads is as two caps from each end of a transformer going to a bridge rectifier. The - side of the bridge goes to the center tap ("common") and the plus bridge output to another cap. That cap is referenced to either common or to the normal plus voltage output cap.

Meanwhile the regular supply is just the rectified voltage either side of center for + and - relative to common---the standard arrangement.

Maybe someone remembers one of the posts showing this and can bump it or paste the schematic into this thread.
 

Latest posts

Back
Top