Discussion in 'Power Electronics' started by Robert Clark, Aug 11, 2016.

1. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
I wondered if staging the CW voltage multiplier can be arranged so that adding stages increases the voltage by factors of 2. As they are used now, the voltage is only increased additively by adding additional stages. That is, adding another stage increases the voltage to 3 times, another to 4 times, another to 5 times, etc. I was seeking an arrangement that would instead multiply the voltage times two with each stage. The response I got was the CW takes AC then sends it to DC. So I would need to then convert the output to AC to apply another CW under this arrangement. But then this would bring it back to the same AC range at the beginning.

https://en.m.wikipedia.org/wiki/Voltage_doubler#Greinacher_circuit

I'm looking at the quadrupler discussed there not the doubler. Then even if when adding a DC to AC converter the peak AC voltage got halved, the peak AC out would still be double that of the initial AC in.

So repeating this circuit consisting of a Greinacher quadrupler followed by a DC to AC converter, I would get a voltage doubling at each stage.

Correct?

Bob Clark

Last edited: Aug 24, 2016
2. ### BR-549 Well-Known Member

Sep 22, 2013
2,169
416
What does CW mean?

3. ### Aleph(0) Member

Mar 14, 2015
356
339
But you can make the cascade as long as needed!
If your idea is to get by with fewer stages you'll be better off just using self or mutual inductor (which is boost coil or multiple wdg mag. transformer) for most of _step up_

If you want more current from cascade just do a full wave cascade but there are practical limits you can't get around.

Last edited: Aug 12, 2016
4. ### Aleph(0) Member

Mar 14, 2015
356
339
BR-549 I think OP means Cockcroft Walton cascade

Last edited: Aug 12, 2016
5. ### atferrari AAC Fanatic!

Jan 6, 2004
2,663
783
Cockcroft Walton

6. ### Aleph(0) Member

Mar 14, 2015
356
339
Guess that makes it unanimous!

Aug 6, 2016
35
16
8. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Thanks for responding. I'm investigating the ionocraft, commonly called "lifters":

https://en.m.wikipedia.org/wiki/Ionocraft

They require thousands of volts. The high voltage is usually provided by transformers. But transformers are heavy. The lifters have never been able to raise their own power supplies. I need a lightweight means to increase the voltage to that extent.

If I just used CW multipliers I would need thousands of stages to get the voltage I need starting from battery voltage. But if I get a circuit that could double the voltage at each stage I would only need ten.

Bob Clark

Last edited: Aug 12, 2016
9. ### AnalogKid Distinguished Member

Aug 1, 2013
4,685
1,297
Something to manage in a CW circuit is the ratio of the diode voltage drop to the voltage boost per stage. For a low starting voltage like a 12 V battery, it doesn't take very many stages for the boost per stage to be less than the diode voltage drops. If you start with a higher voltage source, such as 48 Vdc from a telecom power supply, the power losses are a much smaller percentage. For a home experimenter, the safest way to greater efficiency and lower weight is to start with a small offline isolation transformer, two diodes, and two 400V filter caps. The result is an isolated, energy-limited, 350 V DC source. A CW string with 1 kV diodes should add up to a multi-kV output very quickly.

ak

Robert Clark likes this.
10. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Thanks. But the best power to weight I've seen for a transformer is 1 watts per 3 grams weight. But the lifters typically manage only about 1 gram weight of thrust produced per watt of power used. So you can't use a transformer. That is why I'm investigating lightweight means of producing the high voltage.

Do a google image search on "lifters", "power supplies", "high voltage". You'll find the lifters are flimsy little devices that weigh no more than a few grams but the power supplies are large and heavy weighing in the range of kilograms.

Bob Clark

11. ### nsaspook AAC Fanatic!

Aug 27, 2009
3,009
2,352
Aleph(0) likes this.
12. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Thanks for that link. 2 hp is about 1,500 watts. So 2 hp per 1 kg of lift is about 1.5 watts per gram of lift force, in the range achieved by experimenters.

Bob Clark

13. ### Aleph(0) Member

Mar 14, 2015
356
339
Robert Clark ferrite xfmrs plus drivers operating in 40kHz area are light and can be more than 85% efficient to 50kV+ even b4 cascades! So I say just streamline an AC or for your uses DST (cuz of integral rectification) LOPT and your PS will be lighter than with other methods I say biggest problem you will face is weight of battery. Now in theory you can get even better mass to pwr ratio with air core xfmr (like Tesla coil) but lightweight driver can be tricky

Last edited: Aug 12, 2016
14. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Take a look at the transformers here:

http://www.ultravolt.com/uv_docs/HPCDS.pdf

The best I could find in these small transformers in power to weight ratio is 1 watt per 3 grams, or 1/3rd watt per gram. I need better than 1 watt per gram.

Bob Clark

15. ### Hypatia's Protege Distinguished Member

Mar 1, 2015
2,844
1,309
@Robert Clark

The transformer shown below is typical of a 'class' readily capable of 600W continuous and up to 800W intermittent -- Following removal of the steel mounting hardware and 'changing out' of the 'factory' primary (lower leg), the device weighs ≈ 280g -- (hence a power/mass ratio of >2W/Gm for continuous operation) -- The device pictured immediately below may safely be operated to 30KVP AC output 'as is' or up to 50KVP AC output following 're building' (please see second image below) -- Note that, in addition to increased performance, reliability and versatility, the device will have 'shed' even more weight following rebuilding (inasmuch as said procedure removes a significant amount of unused copper) --- For best reliability I suggest you operate the rebuilt device (an example of which is shown in the second image below) at 40kV into a full-wave doubler for an output of > 70kV at >8mA

Annotated image of 'stock' AC LOPT

Image of optimized LOPT -- Note that your application would 'omit' the steel mounting frame/hardware -- Note also that the mass of the added insulation and required primary windings is trivial by comparison to that of the auxiliary windings removed during the rebuild process.

Best regards
HP

Robert Clark likes this.
16. ### Hypatia's Protege Distinguished Member

Mar 1, 2015
2,844
1,309
@Robert Clark

The following image may grant you some intuitive insight as regards the (collateral) 'mass sparing' consequences of the LOPT optimization process Note that all of the loose Cu and bunting/adhesive debris is the remains of the (unneeded) coaxial Aux winding following removal of same from the EHT winding assembly

Best regards
HP

Robert Clark likes this.
17. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Thanks for that. The primary winding is necessary for the transformer operation. So by changing it out, are you referring to decreasing the number of windings on the primary to increase the multiplication factor of the voltage?

Bob Clark

18. ### Hypatia's Protege Distinguished Member

Mar 1, 2015
2,844
1,309

Correct -- In practice the original primary/Aux windings (on the lower form leg) are removed and replaced with (typically) 10 to 20 turns each side of a 'center tap point'.

For your information/edification here's an example of a typical 'high power' driver circuit -- NOTE: post continued below image...

A few points:

1) -- Please don't be discouraged by the 45V input requirement of the above modeled circuit -- such is readily 'worked around'
2) -- Output may be either polarity with respect to ground or 'floating'.
3) -- R5 represents the load and was chosen (in the above) to illustrate 600W operation -- much greater power is available with minuscule 'EMF load-down' - Howbeit such operation would rapidly prove thermally abusive to the LOPT
4) Although the electrical quantities stated in the safety warning (at the bottom of the schematic) apply to a 'dual/complementary implementation' of same -- please be advised that this circuit is very, very hazardous!
5) Please be advised that the circuit leaves much room for improvement/tailoring to specific applications

In case you're interested, HERE'S A LINK to Aleph(0) and my treatment of this and related circuits/techniques -- in context...

Please don't hesitate to make any further inquires!

Have fun and take care!

Best regards
HP

Last edited: Aug 16, 2016
19. ### Robert Clark Thread Starter Member

Aug 6, 2016
35
16
Thanks. I replied to the transformer suggestion in the lightweight transformer thread.

Some more questions about the voltage multiplier idea. It occurred to me it doesn't have to be the Greinacher voltage quadrupler. Any voltage quadrupler would do.

So using the Cockcroft-Walton circuit, but at two stages then followed by the DC to AC converter would have the same effect. This is the circuit that would be iterated, doubling the AC voltage at each step.

See the two-stage CW here:

A two-stage Cockcroft–Walton multiplier.
https://en.m.wikipedia.org/wiki/Cockcroft–Walton_generator#Design

As described there, the two-stage CW multiplies the peak input AC voltage times 4 for the DC output voltage. So even when you attach a DC to AC converter, which simply shifts the voltage range down so it's half negative, half-positive, the AC output voltage peak is still twice what it was initially.

But a key difference here is that unlike the usual case of adding simply a single CW stage at each step, the voltage drop that each added stage would be subjected to in this new arrangement would increase. That is, in the usual arrangement where you add repeatedly the same CW stage at each step, they each see the same voltage drop so the power ratings of the components stay the same. This is no longer so with the new arrangement. Now, with each added circuit the voltage drop that circuit sees would double. So you would need to make a consideration of the weight of these circuits using the higher power ratings of the components.

Also, in regards to the DC-AC converter part of the iterated circuit, you would have to take into account which of the various means of doing this would have the least weight penalty when you take into account the power ratings have to be increased as the stages are added. For instance, which of the Mazilli, Royer, or 555 oscillator DC-AC conversion methods, require the least weight increase with increased power ratings?

BTW, in that link you provided on the earlier discussion on producing a high-voltage generator I saw you were able to simulate your circuits in a computer program. I'm curious what would the results of this repeated voltage doubling actually look like. Perhaps you could simulate this.

For instance since there are losses it would not look like a sine wave. The losses wouldn't be a big issue at a single stage but might be important at iterated stages. Trying to save weight be eliminating a transformer would not be of much value if all the iterations destroyed the power level.

Bob Clark

20. ### Hypatia's Protege Distinguished Member

Mar 1, 2015
2,844
1,309
The relative complexities attending implementation of additional power oscillators or 'choppers' (i.e. "DC to AC converters") --prominent among them being greatly increased component count, 'conversion losses' and, hence, overall system mass -- will negate any ostensible 'advantage' attainable thereby -- Moreover, cascades are egregiously lossy under the best conditions -- Ergo my recommendation that you directly produce the requisite p-p EMF (via a power oscillator) then rectify same via a Delon arrangement (as shown in the schematic) such that the output EMF = 2X the transformer's peak output.

Please note that the maximum working EMF of the suggested transformer's EHT winding = 50kV -- Hence (equipped with appropriate rectifier diodes/Caps), the suggested circuit may safely be operated to 100kV @600W continuous (i.e. 100kV @ 6mA) - Said output/performance being far superior to a full wave quadrupler --- Don't get me wrong! Cascades have their place! -Medium and high power EHT systems, however, is not it!

Very best regards
HP

PS
Certainly -- Please post, link to a schematic or minutely describe the circuit you wish to simulate -- I feel that's a great idea inasmuch as it will likely 'clear up' any 'communication issues' we may be experiencing

Last edited: Aug 18, 2016