[mchargue]

Henning wrote:
Pat, I think you’re describing something like this:
https://focusfusion.org/index.php/gallery/image_med/28/
https://focusfusion.org/index.php/gallery/image_med/84/

Yep.

This thread, and earlier ones, got me to thinking about this switch system. It was a fun thought experiment, and a good opportunity to contribute. It turns out that there is A LOT of prior art on this issue of high-voltage, high-current, spark-gap switches. Everything from materials to use, to triggering strategies.

I’ve become convinced that there’s likely a drop-in design available somewhere at, or – especially given the low repetition firing rate of the experimental FF system.

I hope that the team can use what ideas were presented here, and from other resources, to improve their design for the experiments they’re running, and for more robust demonstration systems.

Pat

[mchargue]

Help

I can’t seem to post my last submission to the thread, “Volunteer for LPPX Sparkplug Simulation Team”
https://focusfusion.org/index.php/forums/viewthread/722/

Can someone else do this, please, or grant me access?

TIA;
Pat

[mchargue]

Now this sounds quite close tho what you’re trying to do… It’s behind a pay-wall, but maybe you’re already a part of this?

Pat

—–
Design and testing of a multi-triggered spark gap switch for 2-15 kJ plasma focus device
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1228882

yun-Jong Woo; Hyun-Jong You; Yong-Sup Choi; Kyu-Sun Chung;
Dept. of Nucl. Eng., Hanyang Univ., Seoul, South Korea

This paper appears in: Plasma Science, 2003. ICOPS 2003. IEEE Conference Record – Abstracts. The 30th International Conference on
Issue Date: 2-5 June 2003
On page(s): 308
ISSN: 0730-9244
Print ISBN: 0-7803-7911-X
INSPEC Accession Number: 7797019
Digital Object Identifier: 10.1109/PLASMA.2003.1228882
Date of Current Version: 29 September 2003
ABSTRACT

A multi-channel spark gap switch has been widely used in high current and low inductance pulse forming network to obtain switching of fast rise time and high current pulses. Inductance and resistance of a spark gap switch rapidly decrease with increase of channels and electrode erosion is reduced by lower current density. In this work, an electrically multi-triggered spark gap switch was developed to be used as a multi-channel spark gap switch in a 2-15 kJ plasma focus device with a capacitor of Maxwell No 32169 (capacitance, 32 uF, inductance 65 nH). The geometry of the multi-triggered spark gap switch is similar to an annular-type rail gap switch. The large bodies of dielectric that surround the electrodes were designed to prevent arcing along the exterior of the gap. The dielectric material is translucent polycarbonate which has high Izod impact strength. By using translucent polycarbonate, the breakdown in the gap switch could be visually observed. The main electrodes and trigger are made of stainless steel. The minimum gap spacing in this switch is 7 mm and the trigger is located between two main electrodes. The trigger is similar to the main electrode of the trigatron switch. In parenthesis, five different trigger-pins are located in the main trigger-plate and these are isolated with dielectric material. Therefore, six different trigger signals can be generated with a time difference of a few micro-seconds.
—–

[mchargue]

Folks;

Just asking, but has anyone contacted Sandia Labs, and asked after the switching system they use for their Z-machine? The Z-machine also shunts huge currents – sending them through a fine mesh of wires to produce a ‘z-pinch’ – to produce X-rays. I think they’re using oil-bathed switches, but I may be wrong.

Pat

[mchargue]

Continuing with what I talked about yesterday, here are few ideas that may help…

Assuming that we’re using a plate to screw the rods into, (the electrodes) in order to make the whole thing easier to make, lets make that plate metallic. (conductive) This metallic plate would be used to host all of the rods (the electrodes) in bores that are threaded to accept the threaded rods. This plate is now the ‘common’ side of the switch, and connects the the FF device. (maybe directly, as in ‘bolted on’)

Across from the metallic plate that holds the threaded rods is a non-conductive plate that holds the opposite electrodes that complement each of the rods. These electrodes would be fixed into their plate, protrude through it into the switch body, and would not be movable. Each of these fixed electrodes would be connected to each of the capacitors that comprise the bank. One side of the high-voltage initiator would also connect to this fixed electrode, with the second pole of the high-voltage initiator connected to the ‘common’ plate.

The volume between the two plates, one a metallic ‘common’ plate hosting the threaded rods, and the other a non-conductive plate bearing the fixed electrodes, would be walled in to isolate the switch gas from outside air. This means that all switches use the same switch gas, so differences based on differences in switch gas should be minimized. Additional supports between the two plates may be needed to minimize flexure-induced changes in the gap distance.

With this, you should be able to make the fixed electrodes a bit more robust, and make the threaded rods a bit easier to work with, as they have no electrical connections made to them. (all the current is borne by the metallic plate) The metallic plate could be bolted directly to whatever needs to use the current. Mechanically, this might be easier to wire as the large gauge wire from the capacitors connects to fixed terminals on the non-conductive plate, and the metallic ‘common’ plate needs only a single connection from it to the FF reactor.

Service should be easier with the simplified connections. The non-conductive plate could be removed to open the switch, or the threaded rods could be individually removed for service without having to remove any wires.

Materials selection remains open, but note that all the machined parts comprise a plate, and the rods that thread into the plate. The non-conductive plate may need holes in it, but machining should be a lot simpler. I’m thinking a hole with a carriage-bolt in it. The threaded side of the bolt would hold the connections to the capacitor bank & initiator, while the smooth side would be in the switch body.

Switch gas selection is still up for grabs, but the switch should allow you to try a bevy of different gases. I’d certainly try something in place of SF6. Another enhancement would be the use of a transparent window on the switch body wall to visualize the switch process. Perhaps a port that can be opened to permit a ‘feeler gauge’ to be fished through to each gap for calibration.

Well, I think that’s it. Given the design, I’m not sure how much simpler this could be made. I’d love to hear more about what’s actually happening, thouh, and ideas other folks have.

Pat

[mchargue]

Folks;

Initially, I had made the suggestion of a LASER-mediated spark-gap switch in order that the experiments could continue. From reports, it seemed like the current implementation was a gating item to experimental progress. That said, if you can’t use it now, I hope that it helps in any future design.

FWIW, I’ll start looking at alternatives that would dove-tail well with your current design.

A few questions, then…

As I understand the design, you’re using a spark-plug that has had it’s negative return arm removed, and only the center electrode is used. In addition to having one pole of the capacitor bank connected to the center electrode of the spark-plug, you have also wired an additional, higher-voltage, initiator line to the center terminal. (through a diode of some kind) The other pole of the capacitor bank, and the return for the higher-voltage initiator, are connected to the second terminal of the spark-gap (not the spark-plug) through the FF chamber.

When the capacitor bank has charged to an appropriate level, a higher-voltage pulse is passed across the spark-gap to initiate a conductive plasma that then allows the capacitor bank to discharge through the plasma, and FF.

The problems that you’re having include: insufficient heat dissipation, (melting terminals) insulator breakage, (from heat expansion) and terminal wear. (too high current density)

How about…

If you’re having difficulty with the ceramic/plastic/peek/whatever insulator breaking down, why not dispense with the insulator entirely? As you’re not attaching both poles of the capacitor bank to the same spark-plug, the the insulator serves no purpose.

If that can be done, you could replace the spark-plug with just a metal rod. (threaded, maybe?) The metal rod would be the (former) center electrode of the (erstwhile) spark-plug. All other connections would stay the same.

Using this, you’d be able to machine a rod from raw stock with a larger diameter that should have better characteristics: lower current density; less thermal expansion; better heat dissipation; (especially if the rods are threaded into a solid, non-conductive plate) and no more insulator to break down.

If you like, you could thread the rod – all along the rod, or only along a portion – to mate with the threads that already accept the current spark-plugs. Another nut over the top of the rod with a compliant gasket between the nut and the shoulder of the threaded bore may be enough to insure a gas-tight seal. (or plumber’s tape?)

Now you can generate electrodes from raw stock, and a lathe. The machining should be simple, and you can play with the electrode tip design to help insure it’s life-time, and shot-to-shot characteristics. Using the threads along the electrode, you can control the gap by winding the rod in & out, and better thermal characteristics will get you more life-time after you tune the gaps.

The only problem I can see with this is that I didn’t understand your original design.

How’s that sound?

Other ideas include using more switches, and rotating them between shots to minimize thermal breakdown. That means, 12 switches fire 100/sec, or two banks of 12 fire at 50/sec, or 4 banks at 25/sec… You only need to use different initiators to control which set of switches is used. Assuming that you can place all the positive electrodes on a single, non-conductive plate, you need only drill/thread more holes, and set more rods in them.

Other improvements might include a better shaped initiator pulse. I would think that something as close to a square-wave is what you need, It may be that running the transformer that provides the initiator pulse from it’s own spark-gap switch may provide a better current pulse to the transformer than an electronic version. (you may be doing this already)

Anyway, there’s some thoughts.

Pat

[mchargue]

zapkitty wrote:

And finally, there’s nothing wrong with playing with LASERs!

http://www.mathematicianspictures.com/Images/375w_MATH_PG3020_WLAB_9000X.jpg
😛

Oh, I see. (or at least I used to)

Pat

Actually, that’s what happened to me…:)

Of course the alternative was losing all remaining fragments of vision, so… diabetes takes no prisoners.

Sorry to hear that, Zap. Glad you’re here. Good moniker, BTW.
Pat

[mchargue]

Breakable wrote:

And finally, there’s nothing wrong with playing with LASERs!

http://www.mathematicianspictures.com/Images/375w_MATH_PG3020_WLAB_9000X.jpg
😛

Oh, I see. (or at least I used to)

Pat

[mchargue]

Breakable wrote: I wish there were some numbers.
Ok lets say vaporization is slow,
so maybe we can ionize gas some simpler way instead of using lasers?

I think work has been done using plasma jets to close switches, and also microwave-generated plasmas to close switches. Based on my understanding of these processes, I considered them inferior to a LASER-mediated plasma switch. (though microwaves might work out, but I am not sure about the excitation frequencies that would have to be used)

Based on my understanding, you’d like a working gas with a low atomic weight, (H, or He) used at a high-pressure with a large distance between switch poles. I’m not sure that you can excite He with a microwave signal. Also, can microwaves be focussed along a line int he working gas? If not, I think that an ‘area-effect’ is going to be less consistent switch-to-switch, and shot-to-shot, than a consistent plasma ‘line’ drawn between the switch poles.

Pat

[mchargue]

Henning wrote:

Before someone starts playing with lasers, could someone please explain to me what is the problem with Ignitrons?
http://en.wikipedia.org/wiki/Ignitron

I think the difference here is, that you need to vaporize the mercury first before the current can flow. So it’s much slower, or you cannot orchestrate all ignitrons to switch simultanously enough.

I agree with Henning on this.

I see a problem with speed, and the synchronization of multiple switches. The process used – a puff of vaporized mercury – is, I think, slower, and would be harder to coordinate among multiple switches. You can imagine this puff travelling inside the envelope of the switches to close the switch. It may be that travel time is hard to characterize switch-to-switch, and shot-to-shot

Ionizing gases, on the other hand, assumes no travel-time for the conductive media – just the time it takes to flash the gas into a conductive plasma along the LASER path. Something that I would think is easier to coordinate among multiple switches, and between shots.

Based on the recommendations that have been made, I have to think that this has been brought up in discussions within LPP. I’d appreciate it if someone from LPP could comment on suggestions that have been made. I’d like to hear the reasons for/against the approaches that have been suggested. (though I suspect it will come down to cost, simplicity, and unexpected complications with the switches)

And finally, there’s nothing wrong with playing with LASERs!

Pat

[mchargue]

Here’s a book on using LASERs to ionize gases. Something that would lend itself to designing a LASER-based spark-gap switch.

Theory of laser-induced gas ionization
http://www.springerlink.com/content/w85365618m238vlj/fulltext.pdf

Pat

[mchargue]

Brian H wrote: If a patent is for 17 yrs(?), that one’s expired!
Here’s the document: http://www.freepatentsonline.com/4771168.html

I can’t quite decipher if the amperage capacity is high enough, or the response is fast enough, though.

Brian;

I have the feeling that the originators of the patent didn’t push their switch very hard. They’re application was for pumping LASERs at 12-15Kv. I would expect that, with a different switch gas, and higher voltages, the current would climb up there too. Not sure if it would make the 250KA that is claimed as a limit for the spark-plug spark gap-switches, though. (I don’t know if the spark-plug spark-gap switches will do that either)

Based on other reading, control of the hold-off voltage is done through gas selection, gas pressure, and switch electrode distance. (with an eye toward: higher pressure is better)

What I think would be a winner for this type of switch is that it would be more immune from switch-to-switch manufacturing (spec) differences. So long as each switch can hold-off the potential difference between electrodes, and initiate an arc when the LASER fires, minor differences in hold-off voltage between switches would, I think, make little difference.

I think that this contrasts well with the spark-plug based spark-gap design. The spark-plug switch would, to my thinking, be very sensitive to hold-off voltage differences (controlled by electrode distance, gas pressure, and gas mixture differences between switches). This is because changes to the values would change the hold-off voltage. Differences in the hold-off voltage, coupled with the rise-time of the triggering voltage, would mean that each switch would fire at a different time. (a different time along the trigger voltage ramp)

Varying the distance is a way to tune things, but then somebody melts, or the gas mix degrades, and the switches have to be re-tuned.

I think that a LASER initiated gap would render these complications moot.

Pat

[mchargue]

Some products…
—–
http://www.perkinelmer.com/Category/Category/KeyName/IND_DEF_CAT_Spark Gaps_051
http://www.perkinelmer.com/Category/Category/KeyName/IND_DEF_CAT_Thyratrons_054
http://www.reb3.com/pdf/sg_specs.pdf
http://pulsesciences.com/html/AdvPSwitching.htm

Some discussion
http://home.earthlink.net/~jimlux/hv/hvtrigsg.htm
—–
Though made for a different market, designs may be applicable for experimental purposes.

And, yeah, I know. Not suited to shoe-string operations, either. Who knows, though. Maybe one of ’em wants a lock on design-in?

I’d certainly talk to the folks. They do this bread & butter, so they might be able to steer you in a good direction on this… And if they’re trying to sell you something, what boots it? Explain the problem, listen politely, and maybe they’ll connect you to an engineer who just breathes this stuff. HE may have an interest, and a bench to crank out some test articles.

These folks can smell a paper, too. A by-line on a ‘how I did it’ article appearing in an industry-specific rag wuld be good for everyone.

Pat

Pat

[mchargue]

And here’s the patent:
—–
Light initiated high power electronic switch
United States Patent 4771168
The invention disclosed herein includes a low pressure, light initiated, glow discharge switch for high power application. The switch is comprised of an insulating envelope formed into a cylindrical shape having conductive plates at each end. Contained within the envelope is a cathode cup and an anode cup. Each of the cups has a plate at one end which defines two central apertures. The central apertures are positioned opposite one another a short distance apart and centrally axially aligned. A quartz window at the lower end of the cathode cup defines a visual opening to allow unfocused high energy electromagnetic radiation (UV light) to be shined upon the back side of the cathode plate. When UV light is presented to the back of the cathode plate, a photoemissive mechanism is initiated which causes an avalanche effect in the gas-filled chamber of the switch which leads to the discharge of current across the gap between the anode and cathode allowing the switch to close. A system includes the electronic switch is also disclosed which may be used to trigger a high energy flash lamp or excimer laser, as well as other high power applications. A system for controlling the flow of gas into the chamber defined by the envelope of the electronic switch is also disclosed.
—–

—–
Inventors:
Gundersen, Martin (San Gabriel, CA)
Kirkman, George (Los Angeles, CA)

Application Number:
07/046405

Publication Date:
09/13/1988

Filing Date:
05/04/1987

Assignee:
The University of Southern California (Los Angeles, CA)
—–

Someboody get these guys on the phone!

Pat

[mchargue]

Brian H wrote:

The manufacturer of LPP’s switches R.E. Beverley (see here) also offers laser spark gap switches (see specs).

I think Eric and team have a good reason not to use them, be it costs or complicated laser beam setup. Another reason could be, that with every shot the optical system gets misaligned, because of the whole construct moving with the applied force.

On a quick skim thru the specs, I note that they have a max firing rate around 100 Hz, and an expected life from 5-20,000 shots. This would be a minute at most of FF operation.

Given life time is for maximum voltage/amperage. If you stay well beneath it, a million shots are possible. Not very much either as it’s maybe a day of full operation.!
Not even! Assuming 330 Hz is possible, then 1,000,000 shots is 3030 seconds, or less than 1 hr. !

I think that the ‘rated maximum number of shots’ using a switch being <1e6 means that you're going to have to use a noble gas as the switch gas. SF6, (the switch gas that's now being used that came from research on switches that not suited to the FF application) while it has advantages, has the significant disadvantage that it breaks down, and coats the interior of the switch with sulphur. That would tend to alter the way each switch works – differently for different switches – and would degrade the switches for use with the FF device.

That’s why I think that you’ll have to come up with a better long-term solution for the switches. Something that allows you to build encapsulated switches that can be swapped out, rebuilt, and re-used. A switch gas that doesn’t break down, and with electrodes that can be large enough to withstand the thermal cycling.

I have to think that, using the spark-gap setup described here, you’re already experiencing switch-to-switch drift due to electrode break-down, and sulphur deposits.

Experimenters: Based on the results thus far, do you see these kinds of switches working for the final design, or for repeatable experimentation?

[Author]

Posts