Talk:Voltage multiplier

Thanks
Thanks Atlant....


 * You're welcome. As the Wiki FAQ says, be bold, because there's always someone who can help out!


 * By the way, you can easily "sign" your talk posts by ending them with four tildes ( ~ ). When you press (Save page), these will be replaced by your username (or IP address) in a handy Wikilinked format. This will also contain a timestamp for your posting.


 * Atlant 12:58, 26 August 2005 (UTC)

It works, Thanks, You are the man Scott 16:27:59, 2005-08-26 (UTC)

An error?
Correct me if I'm wrong, but I thought that the term voltage multiplier was primarly used to describe an AC-to-DC converter (constructed with capacitors and rectifiers). You can obviously transform it into DC-to-DC converter by means of an oscillator converting input DC to AC. Am I wrong? --filu 11:21, 28 October 2005 (UTC)


 * Yes you are absolutely right. Does need changing Scott 11:23, 28 October 2005 (UTC)

I also suggest not to merge as this is AC to DC converte when the charge pump is normally the DC to DC converter. Audriusa 19:16, 1 March 2006 (UTC)


 * See my comment here. -- Femmina 21:42, 2 April 2006 (UTC)

Other circuit topologies
I have done my best to interpret the very unclear text in the "Alternative diagrams" section (now renamed as above) and convert it into understandable circuit diagrams. Please let me know if I have anything wrong here and I will amend the drawings. I know the first one is actually used because it appears in the Cockcroft-Walton generator article, but the others I have not come across so I could have made some errors here.  Sp in ni ng  Spark  14:52, 2 January 2009 (UTC)

Capacitor size
How does one calculate the best size of capacitor for a particular circuit? Am I right that the capacitor size should be halved at each stage because the voltage is doubled? Biscuittin (talk) 19:38, 25 July 2009 (UTC)


 * No, the capacitor values should remain constant throughout the circuit. In series with many stages, capacitors lower on the cascade may have slightly higher values to reduce reactance. —Preceding unsigned comment added by 71.101.131.135 (talk) 23:54, 22 May 2010 (UTC)

Another Error?
I believe this circuit suggests the voltage is doubled by each stage but I believe this circuit simply adds the input voltage to each stage so Vout = Vin x no of stages. Not Vin x 2^(no of stages). —Preceding unsigned comment added by 69.117.53.189 (talk) 19:51, 30 June 2010 (UTC)


 * There is no such error in the article as far as I can see. If you can see one then fix it yourself.  Sp in ni  ng  Spark  07:42, 2 July 2010 (UTC)

Suggest merge
Suggest merging the contents of Voltage doubler and Voltage tripler to provode context and reduce redundancy. --Wtshymanski (talk) 17:30, 15 November 2011 (UTC)

Not really in favour of merging voltage doubler. Several circuits are specifically voltage doublers and cannot be extended to a general multiplier. There is also some history information that is specifically related to voltage doubler. A further objection is that voltage doubler is GA status: the target article should be brought up to the same standard before even considering a merge. Not such a strong case for voltage tripler though.  Spinning Spark  18:53, 15 November 2011 (UTC)


 * Oppose in part. I'd leave doubler alone; it has plenty of material. Tripler says little (do we care about dampness?) and could be merged into here. Glrx (talk) 17:16, 16 November 2011 (UTC)

Caption of Figure "Other circuit topologies"
In Fig. “Other circuit topologies”, what does this mean, “and upon any collapse from arcing capacitive energy can cancel”? That’s not a complete concept. Should it say something like this instead, replacing “and” with a period, “. If a capacitor breaks down, all the energy will be suddenly released in an electrical arc.”Michael McGinnis (talk) 07:00, 7 June 2012 (UTC)

Dickson charge pump - "break before make"
The driving waveforms for the Dickson pump are described as "...non-overlapping so that the switching behaviour is break-before-make" (edit by Spinningspark, 25 September 2011‎). This analogy with a switch is misleading. While there is no advantage in overlap, there is no need for "break before make". No larger current will flow during any overlap: the circuit will simply become less effective as the overlap increases. Indeed, the waveforms shown do not comply with "break before make" if you regard "make" as "negative drive". Maximum output is achieved when the driving signals are 50% square waves in exact antiphase. I can find no reference to "break before make" in connection with Dickson multipliers (except in the numerous copies of this article). I have removed the reference to "break before make".Mike Shepherd (talk) 09:28, 7 June 2012 (UTC)

Questions and comments
1. Why should the first image say "Villard cascade voltage multiplier" if the circuit was actually invented by Greinacher/Cockcroft–Walton? Shouldn't the caption use "Greinacher/Cockcroft–Walton cascade voltage multiplier" instead? The section on the Greinacher circuit of the voltage doubler clearly states "Villard" is a misnomer (Voltage_doubler).

2. I have a few questions about the images for the topologies. First image: where is the multiplication if the output is +V? Second image: why does the lower branch has only +V. Third image: text says "Stacking the two cascades provides an output of twice the voltage" so why does the image use +V and -V instead of +2V and -2V?

3. The use of VT could actually be misleading because VT is also the thermal voltage. I think VD is safer and avoids confusion.

4. It's not clear to me the purpose of having the image with caption "TV cascade (green) and flyback transformer (blue)."

ICE77 (talk) 19:03, 6 August 2015 (UTC)


 * Q2. The voltages are merely relative to the voltage from one stack of unspecified multiplication. For instance, the third image has a potential difference between the two stacks of +V--V = 2V, ie, twice the voltage of one stack.
 * Q3. VD would be even more misleading, as it is commonly used to mean the drain voltage. VT is clearly defined, so there should be no confusion, and is actually used by some textbook authors (eg ) whereas I have never seen anyone use VD.
 * Q4. "TV cascade" is a voltage multiplier and driving a multiplier from the flyback transformer was a common way of producing the HT supply in old CRT television sets. SpinningSpark 23:03, 19 August 2015 (UTC)

Spinningspark, thanks for the feedback.

1. Should the image continue to use "Villard" which appears to be wrong?

2. I see the difference in potential for the third circuit. I do not see the other two circuits.

3. All the links you posted point to VT as used in fact typically for a MOSFET for threshold voltage and not for a diode for forward voltage. The use of VD for diode voltage is not at all uncommon and in my opinion it's not misleading. Wikipedia itself uses it: https://en.wikipedia.org/wiki/Diode#Shockley_diode_equation. The book entitled Microelectronic circuits by Adel Sedra in section 3.3.1 clearly shows VD in equations 3.6, 3.7 and Figure 3.10. Since VT is used twice before the image of the MOSFET implementation of the Dickson charge pump I assume it still refers to diodes.

4. If "TV cascade" is a type of voltage multiplier it would help to clearly state it in the article. I am myself not familiar with that naming.

ICE77 (talk) 01:52, 20 August 2015 (UTC)


 * 1. It's not wrong if that is what it's commonly called.


 * 2. The output of the first one is between the end of the cascade and the reference at the centre tap; so the potential difference is +V-0V = +V. The output of the second one is between the end of the top cascade and the reference at the bottom of the first cascade; +V++V=+2V and the potential difference is +2V-0V=+2V.


 * 3. Yes, but the switching device is not necessarily a diode, making VD inappropriate. In fact, the switching device in IC circuits is very commonly a MOSFET so VT is appropriate both for that specific case and for the general case.


 * 4. The article already repeatedly uses the term cascade. A TV cascade merely means a cascade used in a TV set and is not a special type.


 * SpinningSpark 16:00, 20 August 2015 (UTC)

1. Just because a food is commonly and erroneously called lasagna it does not mean that we should legitimate its spelling when the proper name is lasagne. That's the way I see it and the fact the section on the Greinacher circuit of the voltage doubler clearly states "Villard" is a misnomer is the indication that even if Villard is popular, it's not proper and it should not be called like that (Voltage_doubler). Maybe the article should have a note regarding this.

ICE77 (talk) 06:18, 27 August 2015 (UTC)


 * So would you rename the Morse code and Wheatstone bridge articles on that basis? I don't think you would get very far unless you first got the WP:COMMONNAME policy changed. It might be unfair that the original inventor is not honoured with an eponymous name, but Wikipedia is not the place for Righting Great Wrongs and in any case, the name is a long way from actually making the claim that Villard invented the circuit. SpinningSpark 15:19, 27 August 2015 (UTC)


 * I second Spinningspark. Glrx (talk) 03:24, 31 August 2015 (UTC)

A much clearer and more complete possible presentation
In voltage multipliers, the diodes are always set in series, but it is possible to arrange differently the capacitors on both side (the ones crossed by AC currents and the ones charged with DC voltages),see for instance: http://www.voltagemultipliers.com/pdf/Multiplier%20Design%20Guideline.pdf The capacitors can be set in series or parallel (see Half-wave Parallel Multiplier and Full-wave Serial Multiplier in the preceding document), leading to four implementations of the same idea. Unfortunately in the present page only the serial-serial configuration is presented. The modified Dickson charge pump is a possible introduction of the parallel-parallel configuration (although presented in a different manner that prevents to see the point discussed here), but the two serial-parallel and parallel-serial implementations are totally missing. Unfortunately I have not enough time to do the cleaning job now but the idea is very easy to grasp and to implement and will lead to a well constructed page instead of a patchwork of old ideas, any volunteer ?.--Henri BONDAR (talk) 13:22, 24 March 2016 (UTC)
 * "diodes are always set in series" may be an oversimplification, and I would like to see a better source than a manufacturers data sheet giving this classification. In any case, I don't think that source actually says diodes are always in series and all multipliers can be classified as series or parallel capacitors, it only lists the ones VMI make without making any such global statements.  How, for instance, would you classify a Delon voltage doubler under this scheme? SpinningSpark 14:32, 24 March 2016 (UTC)
 * Diodes in series is an empirical fact linked to the stacked voltage idea. I think that this point could be demonstrated (probably done somewhere). The Delon doubler (formely Latour doubler in France), is a single step full wave doubler. As such it implies a single capacitor for each alternation and then can be seen as a serial or a parallel case. The Greinacher doubler (formely Schenckel doubler in France) is also a degenerated case as it involves a single capacitor crossed by alternative current and a single capacitor charged with a constant voltage. So it will be classified either as serial-serial and parallel-parallel half-brige doubler. More generaly this classification does not apply to doublers (degenerated cases) but to higher order multipliers. In practice the serial-serial arrangements, if well tuned, imply to use capacitors able to sustain the same voltage but with capacitance values that decrease upwards, whereas the parallel-parallel arrangements, if well tuned, imply to use capacitors with equal capacitance but able to sustain higher voltages upwards. However it must be noticed that for large voltage ratios, mixed configurations are possible, as some capacitors can be grouped in series and other in paralell, albeight without clear practical interest.
 * I agree that other references are necessary to sustain this general classification. The background idea is to start the page with a modern description without removing the historical side that will remains in subsections. For the moment this page looks like a century old one to me. --Henri BONDAR (talk) 07:33, 25 March 2016 (UTC)

Breakdown voltage error
Presently this section says, "the individual components do not need to be rated to withstand the entire voltage range. Each component only needs to be concerned with the relative voltage differences directly across its own terminals and of the components immediately adjacent to it." This is not correct: Once the circuit is fully charged, each capacitor has a potential equal to the voltage at its stage. E.g., the last capacitor in a cascade generating xV needs to be rated to xV. — Preceding unsigned comment added by Dbooksta (talk • contribs) 13:38, 25 May 2016 (UTC)
 * Since the capacitors are essentially in series in a ladder circuit, it cannot possibly be true that they all carry the full voltage. However, what you say may be true of a charge pump design. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 14:59, 7 June 2016 (UTC)

Dickson charge pump using diode-wired MOSFETs


I was looking at the circuit on the right. The drain of the FETs is on the left and the source is on the right. I tried to simulate the circuit and when I do it like that I get a negative output voltage which doesn't make sense. I flipped the FETs so that source is on the left and drain is on the right. I get a positive voltage. What's the right polarity?

ICE77 (talk) 04:59, 15 March 2019 (UTC)


 * No, that doesn't make any sense – you don't have a source of negative voltage. Of the sources in the article, Liu gives the polarity shown (iirc, that's what the diagram is based on), and Yuan gives a diagram but does not show polarity.  This book also agrees with Liu and our article. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 14:18, 15 March 2019 (UTC)


 * The accompanying text specifically says "MOSFETs wired to behave as diodes". I simulated that aspect in isolation in the falstad circuit simulator (unfortunately, a direct link to the simulation is blocked by wikimedia, but it's easy enough to set up). Clearly, the N-channel MOSFET behaves as a diode when connecting G and S, not G and D. 2.244.168.131 (talk) 09:32, 4 November 2023 (UTC)
 * The picture is nominally correct. As wired, Vgs=Vds,  A positive Vds would turn the mosfet on.  However, this is a terrible way to do it, due the value of the Vgs threshold voltage.  The threshold voltage is high and variable compared to a Vbe of an NPN transistor with B wired to C.  I think this picture originally had NPN transistors that were switched out for Mosfets without actually analyzing the resulting design.  Iside an IC using mosfets, the gates of the arious transistors would be actively driven by a dedicated circuit, that is not shown.
 * Here is a more complete equivalent circuit of a mosfet: MISFET mit Inversdiode (Schaltzeichen).svg
 * When you wire G to D, you essentially turn-off the mosfet, leaving only the body diode shown here as a Zener diode. That is what is generating voltage in your simulation. <b style="color: #4400bb;">Constant314</b> (talk) 14:17, 4 November 2023 (UTC)

Spinningspark, thanks for the comments. I see the links you provided (Liu and Ma/Bondade) are consistent with the image in this thread. I went back to simulate and the results do not make sense. If source must be on the left, if I have Vin=2V, what should be the amplitude of the phases and the polarity of the output?

ICE77 (talk) 07:40, 30 March 2019 (UTC)


 * I don't know how much more I can help, I really don't know any more than is in the sources. Yuan says you should get about 4V from a 4-stage circuit with 1.5V input.  That would be a positive polarity.  The amplitude of the clock phases depends, of course, on how they are generated, which isn't shown, but if I had to guess I would say they are swinging between $$V_\text {DD} - V_\text {T}$$ and $$V_\text {T}$$. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 08:26, 30 March 2019 (UTC)
 * There is indeed a weakness in the description of the Dickson cascade: "D1 is then turned off and D2 turned on and C2 begins to charge to 2Vin. " Well, not quite.
 * If C1 and C2 are the same size (implicit assumption) then that makes a capacitive voltage divider and the voltage goes up only to 1.5Vin.
 * This is what a simulation and also what real life shows. Cost me a redesign ... 31.193.165.228 (talk) 15:28, 25 October 2022 (UTC)

Incorrect caption of Cockcroft-Walton multiplier
I think the caption of the picture of the Cockcroft-Walton multiplier is wrong. The caption says it's a doubler, but the circuit show in a a 4x multiplier. (the picture itself is from the separate page https://en.wikipedia.org/wiki/Cockcroft%E2%80%93Walton_generator where the caption simply calls it a multiplier, and the accompanying text clear describes the 4x operation. — Preceding unsigned comment added by WestfW (talk • contribs) 23:09, 6 September 2019 (UTC)
 * I agree. The caption also (correctly) states that this is twice the pk-pk voltage of the supply.  But peak voltage is the conventional reference, not pk-pk. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 11:52, 7 September 2019 (UTC)
 * I corrected the caption. Addition: It's totally the SAME circuit as the first image (Villard multiplier). Just drawn slightly differently. So do we need the same circuit twice? Also, in the first image "Us" is not widely understood I think - we usually indicate RMS voltage next to AC sources, while here the output is 4x the peak (if unloaded). It could also be mentioned that not only the output ripple but output peak depends strongly on Rload and C values. Hoemaco (talk) 08:46, 22 April 2021 (UTC)
 * Okay so I don't understand Spinningspark - you said peak is the conventional reference, not pk-to-pk, yet you reversed my edit in the image caption, which said 4x peak (which is of course same as 2x peak to peak). Also, you reverted to the misleading name of voltage doubler - it doubles the peak to peak, but as mentioned, most people use peak only and as such it is a quadrupler circuit. And there's still the problem I said that the first circuit in the article is totally the same, but with a different name and with an output formula written as 4 Us. Which in itself contradicts this other image. Hoemaco (talk) 20:36, 26 July 2021 (UTC)
 * My mistake, I thought I was reverting something else. I've self reverted. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 10:32, 8 August 2021 (UTC)

Okay so the first image (Villard) and third image (Cockroft-Walton) are still the same circuit. Which one should be deleted? Which name is more appropriate? Hoemaco (talk) 19:35, 22 June 2022 (UTC)
 * Why do you want to delete either of them? The lead needs an image, and it doesn't really matter if that is the same/similar circuit shown later in the article.  <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 14:01, 23 June 2022 (UTC)


 * By the way, I am not at all sure that it is correct to label this circuit a Villard circuit. Villard's orginal paper gives only a single diode device (see Voltage doubler) without a smoothing stage.  This book shows two different topologies for multiple stage Cockcroft-Walton and Villard multipliers, although I am pretty sure the diagram annotation has "c" and "d" the wrong way round since "d" (as marked) is definitely the Cockroft-Walton. <b style="background:#FAFAD2;color:#C08000">Spinning</b><b style="color:#4840A0">Spark</b> 14:14, 24 June 2022 (UTC)

India Education Program course assignment
This article was the subject of an educational assignment supported by Wikipedia Ambassadors through the India Education Program.

The above message was substituted from by PrimeBOT (talk) on 19:59, 1 February 2023 (UTC)