Blaze Labs Resonant CW multiplier

[Mirage_662]

http://www.blazelabs.com/e-exp19.asp

I have seen blaze labs mentioned a few times on this site for CW multiplier construction. Experiment 19 details the use of an inductor to cancel out the reactance of the circuit. Its called the Blazelab Resonant Multiplier. Kinda like a PFC capacitor I guess. Has anyone actually verified his work? At the bottom of the page he sums it up...

"The standard CW multiplier requires 42 stages using 1.22uF capacitors! Our BRM uses much smaller capacitors and just 30 stages, very close to the ideal loss-less multiplier which would require 28 stages."

This would be a very cost effective way to build a DIY power supply. If this is so much better, why isn't it used in the industry? Comments?

Elijah

[Richard Hull]

Blaze labs are a good working group.

Your question is rather easy to answer. Modern switching systems are much more efficient, require less parts and assemble in a smaller space. Most designs that actually get manufactured shy away from more than 3 stages. 5 stages is considered obsence and poor design. 12 stages a laughable effort. 28 stages..well.....'nuff said.

Only amateurs would fiddle with such efforts.

The power actually transferable through 28 stages would be extremely limited for CW operation at power line frequencies. A good fusor needs access to between 200 and 800 watts of high voltage.

This is normally accomplished with a high frequency inverter to give a base voltage of 10-15kv and then multiplied using special, custom designed, fast switching multipliers

The fact that no one has done this yet in our community speaks volumes.

I would like to know from any amateur who is really fusing and in the neutrons club if....

1. You have any system that uses multipliers designed and constructed by you?
2. If so, how many stages?

Richard Hull

[Tyler Christensen]

I built my own stack, I've tried a number of frequencies and configurations. My first stack was for 80kv with 3-stages, then I went to 4 stages at 65kv, and finally settled on 6 stages at 80kv. The voltages are mainly just based on the capacitors I had on hand, and doesn't necessarily reflect the voltage I actually ran the fusor at, since I'm not shielded to run at 80kv safely. I found it easiest to wind ferrite for <10kv which is why I moved up in stage count. 4 and 6 performed almost the same (as in the v-drop going up to 6 negated the benefits of having 6 stages almost perfectly), but it was still a tiny bit ahead so I left it alone.

Most caps were in the 2-5nF range.

As far as frequency, I started at 60khz, moved to 40, moved to 30, moved to 20, and I currently run just under 15khz (the high pitch squeal is just wonderful). Most people run higher than me in frequency, but I just played with parameters until it worked best, and it worked by far the best at the lowest frequency I was willing to switch at. What can I say, optimize for what works.

So my current optimised multiplier is 6 stages of 4nF at 15khz.

[Carl Willis]

The multipliers I use are commercially built. One has eight stages (Wallis Hivolt, 80 kV). One has 10 stages (Glassman, 100 kV).

A broader generalization of the Blaze Labs idea is mentioned without extensive circuit analysis in Baldinger E. "Kaskadengeneratoren", in Handbuch der Physik, S. Flugge (ed.) Vol. 44, Springer-Verlag, 1959, p. 56. The following is a Google-assisted attempt at translating German:

>The logical extension of this idea leads to structuring the stages of the cascade as bandpass filters [Figure 50 depicts a full-wave C-W with series-resonant LC in the charging legs, parallel-resonant LC across DC equipotential points], the frequency of the alternating drive voltage being approximately in the center of the filter chain's passband. The switching capacitance and the inductance of the wiring is incorporated into the filter and thus presents less interference.

At this point in the article there is a citation to another of the author's works, "Kompensations-Drosselspulen hoher Gute fur Kaskadengeneratoren [Compensation inductors of high quality for C-W multipliers]" Helv. Phys. Acta 30, 282-287 (1957).

To conclude, although amateur builders probably often neglect the effect described by Blaze Labs, similar or identical concepts seem to have been communicated in the professional literature for a long time and it's hard to imagine modern professional designers being ignorant.

-Carl

[John Futter]

Carl
Well said
I used this technique in 2002 to provide a 200 volt rail in a telco line powered ISDN modem that had to provide pots emergency service when local power was not available ie line powered.
The 200 volt rail was for a ring amplifier to ring 6 ren with a specified cadence.
every mw counts in a situation like this so resonant multiplication was employed.

as the web site mentioned has a spice link
use it!!

no rocket science here
any good electronic engineer is aware of this for the last 40 years

[Richard Hull]

We are back to where one amateur fusioneer here has built their own stack multipliers and 4 stages worked almost as good as 6. I have found that most pro designs like the lower frequencies10-30khz, though there is little beyond magnetics and switching component limitations that would hold back higher frequencies.

I have a Hi-Volt stack rated for 120kv @60ma that uses 5 large multiplier packs and a much older Spellman rated at 160kv but limited to 5 ma Both function under 25khz.

Fast HV rectifiers that are demanded to be reliable are usually a limiting factor in these stacks.

As noted before in these forums, power line frequencies require larger energy storage caps for even half way decent filtering and energy delivery. HF driven systems use caps with less stored energy which make the finished product, more compact, less dangerous after shutdown and at lower production cost.

Richard Hull

[John Futter]

Richard

documented here is a design of my own for a 200kV supply for an x-ray system @ work
ie +-100kV about earth, and yes I did the board layouts and built it, winding the transformers myself.
viewtopic.php?f=11&t=4770#p27608
http://www.coultersmithing.com/forums/v ... p?f=9&t=98
Each multiplier 18 stage fed in the middle so there has to be 50kV plus safety margin primary to secondary. So this means two multipliers of 9 stages each way to get the desired 100kV for each side added to this is the need for 100kV secondary to secondary isolation this is why its all in oil.
output current entirely dependant on stage cap value.
Note the the feed into centre of multiplier stack used twice the capacitance values.

info while i was designing came from LTspice, voltage multipliers web site, phillips soft ferrite databook.

and note I drove this with a mosfet audio amp with a sine wave @ 33kHz. This done to save strain / recovery time on the high voltage diodes --none destroyed so far with a few spectacular blowups of transformer secondaries.

I'm about to do another design for a 200kV 5 mA PSU for one of our ion implanters using a very similar technique

PS most high voltage SMPSUs output to the multiplier between 5- 10kV so a 20 kv might be a quadrupler to make up for losses. Its not unusual to find 15 -20 stages in an 80kV supply

[Richard Hull]

I have seen a DC tesla Coiler who also liked feeding his multipliers with the kinder, gentler sine wave. He used a Classic Crown DC 300 to bull-head about 15khz sine tone into his custom wound ferrite xfrmr. Amazingly, Nothing blew up and his coil put out about 4 feet of spark.

DC power supplied coilers are rare who feed classic rotary spark gaps.

Many coilers today use solid state IGBT drivers and avoid the classic spark gap altogether.

Many ways to skin the proverbial cat.

Richard Hull

[scinut]

Hello Richard-Is the Spellman PS you mention below a XRF160N160X3273 Xray unit, or something close? Is the factory output variable, or must you vary the input? If variable, could the output current be raised by lowering the voltage output, or are there limiting components to prevent that? (80kV @10mA and so on...?) If not, could the supply be modified (cheaply and safely), to increase the output current?

Richard Hull wrote:

> I have a Hi-Volt stack rated for 120kv @60ma that uses 5 large multiplier packs and a much older Spellman rated at 160kv but limited to 5 ma Both function under 25khz.

Thanks-Jerry

[Richard Hull]

The spellman is fully variable and the current is limited to 5 ma whether you are outputing 1kv or 160kv. NO modification possible.

Richard Hull