No charge pump MULTIPLIES! A better phrase would be STACKER... stacks one capacitor on top of another.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
I still haven't seen an example where this is actually done. In my application, I need a high voltage power supply in the kilovolt range that is portable (lightweight) and can run off DC batteries in the 30V range AND at the same time can put out hundreds of watts. The usual way of getting high voltage is to use a transformer and then perhaps use a CW doubler or tripler to ramp up the voltage. The problem is that transformer usually has a ferrite core and that makes it heavy. Ideally, I would want the power supply (exclusive of the battery) to weigh less than 1 gm for every 1 watt generated while producing kilovolts. I wanted to eliminate the transformer to save weight. So I thought I could use *many* stages in a CW alone to get the high voltage. However, the usual CW circuit is additive. So if I wanted 30,000V from 30V I would need 1,000 capacitors and associated diodes. Even using those flat disc capacitors, this would probably weigh more than just using a transformer. Therefore my question arose, is it possible instead to feed one CW multiplier into another *multiplying* the voltage with each repetition?
(1) Guy Macon couldn't locate his asshole with both hands and a flashlight. His 2-3 year job cycles prove that ;-)
(2) It is NOT possible to feed one CW into another.
(3) "hundreds of watts" forces you to use a transformer.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |
formatting link
| 1962 | I love to cook with wine. Sometimes I even put it in the food.
You are off by a factor of two. The nominal output voltage (no load) assuming perfect parts is *twice* the peak input voltage multiplied by the number of stages.
I have never seen the topology I described above either, but Winfield says that he uses it, so obviously my experience isn't broad enough to have ran into it in practice or read any literature on it. It does indeed multiply, so from a voltage standpoint you can use three copies of a Cockroft-Walton voltage multiplier with a DC-to-AC inverter on the front end, and enough stages to get a nice 10X increase from each copy. That's 30 to 40 capacitors and associated diodes, plus whatever parts you need to make the 3 DC-AC inverters.
That "hundreds of watts" specification concerns me. Cockroft-Walton voltage multipliers aren't all that good at putting out current. Here is how to figure your currentL
Let's look at a single Cockroft-Walton voltage multiplier that with a DC-to-AC inverter on the front end.
Start with the oversimplified model of the nominal output voltage being twice the peak input voltage multiplied by the number of stages,
Now reduce it by the number of diode drops you have.
Now figure the voltage drop;
Edrop = I1oad /(f*C)*(2/3*n^3+n^2/2-n/6)
Edrop = voltage drop Iload = load current f = frequency C = capacitance n = number of stages.
Now figure the ripple; Assume all capacitances are equal.
Eripple = Iload/(f*C)*n*(n+1)/2
You can reduce the ripple by making the capacitors smaller at the top and larger at the bottom: C1,C2=10*C C3,C4=9C, etc. This gives you;
Eripple = Iload/(f*C)
As you can see, you want to keep that frequency as high as you can before the switching losses start kicking in. I have had good results with starting at a low frequency and ramping the frequency up - sort of a poor man's soft start.
(Disclaimer: I am away from my reference books and I don't have the above equations memorized, so I snagged them off the web. Make sure they are right before actually building anything.)
--
Cheers!
Rich
------
"Look out for yourself -- or they'll pee on your grave."
-- Louis B. Mayer
"The reason so many people showed up at Louis B. Mayer's funeral was
because they wanted to make sure he was dead."
-- Samuel Goldwyn
I agree completely. Fortunately, with sufficient skill, you can run a small, light resonant transformer at a high frequency and do far better than any C-W circuit. I suggest you look into that approach.
For how long a time period do you need to continuously supply 100s of watts of these kVs, and with what duty cycle? More detail, please.
Marx generators usually use spark gaps, but they're still additive per stage. There's a neat megavolt Marx out at Livermore that uses laser-triggered spark gaps and water-filled clear plastic tubes as the charging resistors.
Your mention of the ripple brings to mind another question. The ripple means there would be a more or less sine wave pattern about an average DC voltage on the output. Is there any way to shift down this sine wave so that the voltage will be AC? I was thinking since we already have this varying voltage in the output it would be nice if we could use it to create the AC we need to feed into the next CW. I know we can just put another DC to AC inverter on the output, but I was wondering if there might be an easier way given the output is already varying.
Make DC into AC, or more accurately, into pulsed DC. But I have a feeling at least half the voltage gets lost doing this. And will spark gaps scale indefinitely in voltage? Probably not.
The purpose is make a self-contained power unit for a lifter:
Experiment 14 (21/03/03)- 100g payload lifter.
formatting link
The power units used so far are heavy and have to be left on the ground, the power being delivered to the craft through wires. The best ratio of weight lifted to power required so far has been about 1 gm lifted to 1 watt of power, as in that example by Blaze Labs. In this example the payload weight is 100 gm, with total weight about
200 gm. On the page you see the voltage required is 46 KV and the power required 202 watts. So you would need both the battery and power supply total to be the 100 gm payload weight while putting out 46 KV and 202 watts. Here is a link to some lithium battery specifications:
Thunder Power Li-Poly Batteries.
formatting link
When you calculate the watt-hours they can put out by multiplying the mAh given (milliamp-hours) by the voltage and compare that to the given weight, you see the weight in grams is about 6 times the watt-hours. Then assuming the best lifter takes about 1 watt to lift 1 gram, you see lifting the battery alone without considering the weight of the power supply or the lifter body, could only work for 1/6th of an hour,
10 minutes. Adding in the power supply and lifter weight makes this even less. Still finding a self-contained power unit for a lifter would be important as a proof-of-concept even if it lasts only for minutes.
1 gram lift per watt! The things PhD's do to amuse themselves!
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Ozone doesn't turn water into nitric acid, does it? You're probably thinking of oxides of nitrogen, which, admittedly, probably also get generated by the arc - lightning supposedly "fixes" nitrogen, which is good for the plants, or so I've been told.
But I would have no qualms whatever about bubbling plain ozone through water - it dissolves about as well as O2 does, AFAIK, and sterilizes better than chlorine, and doesn't hang around the way chlorine does.
Hmm the output maybe DC how ever, until it reaches it's final series diode and filter cap, you still have a some- what looking AC signal there which can be decoupled via a Cap.
--
Real Programmers Do things like this.
http://webpages.charter.net/jamie_5
don't know about scaling but i do know it makes nice plasma, the ozone is nice especially when coming in contact with moisture! generates some nice nitric acid that will make your skin do alittle etching .!
--
Real Programmers Do things like this.
http://webpages.charter.net/jamie_5
Suppose you had a light, high-voltage, high-current power supply. You connect one power supply terminal to the lifter body. Where do you connect the other?
One lead from the power supply is connected to wires at the top of the lifter. The other lead is connected to metal foil at the bottom. You have insulators separating the wires from the foil. Then the air between the wires and foil serves as a dielectric for a capacitor. The asymmetric geometry of the two sides of the capacitor, the wires compared to the flat metal foil, causes a flow of ions from one to the other.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.