Re: Marx generators, was RE: door knob caps

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Jim Lux by way of Terry Fritz <teslalist-at-qwest-dot-net>" <jimlux-at-earthlink-dot-net>

Actually, there is a fair amount of info out there, it's just buried..
Lyonel Baum has a good page summarizing most of the salient details. My own
Marx page doesn't have a lot of design rationale (out of sheer laziness on
my part).

The other thing is that once you've researched it and built a couple, you
tend to figure things are obvious.  Dr. Resonance runs 60 kV in his stages
because he's got plenty of experience at that voltage level, and probably
instinctively knows by now how to build stuff that doesn't leak everywhere.
The other problem with things like Marx generators (at least on the amateur,
one-off level) is that they are big and complex enough (mechanically) that
you tend to overdesign, because it's such a pain to rebuild.  They also have
very long lifetimes.  There are lots of 50 year old Marx systems around
still chugging along happily.  Unlike a TC, where the caps get stressed
highly, in a Marx, a commercial unit might do a few shots a day on the
average (actually, probably more like a couple dozen shots in one day, then
2 weeks of setup for the next test...).  Even if your caps only had a life
of 10,000 shots, that's a lot of years of use.

High rep rate Marx's (as in radar modulators, flash X-ray, etc.) are a
different animal, and are expensive enough that the mfrs can do some
research and cut and try.

The hard part of a Marx is the spark gaps, by the way, not the caps.  They
really need to fire together to get expected performance. That means
consistent spacing, consistent voltage, and good alignment.  In many hobby
built Marxes, the charging resistors are quite large (100s of K, to several
Meg), resulting in a fairly long time constant, which means that the top
stages don't get full voltage.  Commercial Marxes, which can have fairly
high power charging supplies, tend to use fairly small charging resistors
(5-10K).  That gets the arc current (5 Amps is a good design number) in the
gaps up, which helps with consistency.

----- Original Message -----
From: "Tesla list" <tesla-at-pupman-dot-com>
To: <tesla-at-pupman-dot-com>
Sent: Wednesday, August 20, 2003 3:13 PM
Subject: RE: Marx generators, was RE: door knob caps


 > Original poster: "Timothy Myers by way of Terry Fritz
<teslalist-at-qwest-dot-net>" <tjmyers-at-nortelnetworks-dot-com>
 >
 >
 > I've been searching the Web for details on Marx generator construction
with
 > little luck.  You just pretty much summed up the answers to all my
questions.
 >
 > Many times I find the design reasoning lacking when analyzing a project.
I
 > don't know if it's just assumed everyone is on the same page or not enough
 > reasoning is put into the design.
 >
 > Your details provide enough information to design a custom Marx generator
 > of you own choosing.  Really good information.  Thanks,
 >
 > Tim
 >
 > -----Original Message-----
 > From: Tesla list [<mailto:tesla-at-pupman-dot-com>mailto:tesla-at-pupman-dot-com]
 > Sent: Wednesday, August 20, 2003 12:54 PM
 > To: tesla-at-pupman-dot-com
 > Subject: Marx generators, was RE: door knob caps
 >
 > Original poster: "Jim Lux by way of Terry Fritz <teslalist-at-qwest-dot-net>"
 > <jimlux-at-earthlink-dot-net>
 >
 > At 08:16 AM 8/20/2003 -0600, you wrote:
 >  >Original poster: "Dave Halliday by way of Terry Fritz
 >  ><teslalist-at-qwest-dot-net>" <dh-at-synthstuff-dot-com>
 >  >
 >  >Besides, looking at their website, they only sell Ceramic Disc
 >  >capacitors with a 2KV range.  These simply will not work.  If you buy
 >  >these for this application, you are taking dollar bills and burning
 >  >them.  No, not even that because you would get some satisfaction from
 >  >the flame.  You are wasting your money.
 >  >
 >  >The reason is that the cap may be rated at 2KV but this is only one part
 >  >of the problem.  The capacitance-at-that-voltage is minimal.  It is the
 >  >capacitance that stores your energy and if you don't have any energy
 >  >storage, you are going to get a less-than-zero performance.
 >  >
 >  >Check this site for a discussion on MMC - they are using them for Tesla
 >  >Coils but they are very applicable to Marx generators too:
 >
 ><http://www.thegeekgroup-dot-org/projects/MMC/>http://www.thegeekgroup-dot-org/proj
ects/MMC/
 >
 >  >
 >  >You _will_ be spending about $100 and this _is_ a good chunk of change
 >  >but, any money you spend at radio shack will be dollars pissed down the
 >  >drain so don't do that, start saving.  What you will be buying will work
 >  >perfectly and deliver good solid fat performance.
 >
 > Well, I don't know that I'd be so hard on Radio Shack.  Their component
 > quality isn't marked worse than any other vendor selling in qty 1 in nice
 > blister packed packages. Sure, if you buy 100 pcs, you can easily do
better
 > other places (Digikey, Mouser, Newark, etc., we all have our
 > favorites).  Time is money too... you could spend 10 hours tracking down
 > the very best deal on a $100 purchase for a 10% savings.  Hmmm.. $10
saving
 > for 10 hours work.
 >
 > Ceramic caps, while not ideal, will work quite nicely in a Marx (although
I
 > wouldn't advocate stacking up zillions of little disk bypass types, for
 > structural and labor reasons).  Doorknob caps ARE ceramic, and I daresay
 > many perfectly functional Marx generators have been made with them, by
such
 > illustrious workers as Craggs & Meek, among others.  Not to say that they
 > are optimum, though.
 >
 > Here's some practical stuff on building a Marx:
 >
 > The circuitry is trivial. The mechanical and assembly aspects are not. Pay
 > careful attention when designing to such things as:
 > a) how long will it take to make thousands of solder joints (if you're
 > going that route)
 > b) what about corona from sharp edges/small wires (ceramic disk caps are
 > horrid for this, by the way)
 > c) How will you insulate the thing, what will support it.
 >
 > Look at some tradeoffs.. Ceramic doorknobs are about $25-30 each (list
 > price from C&H or RF parts), much cheaper other sources of varying
 > provenance. Say you find some 2200 pF 25 kV units at $30. That's 2.2 nF
 > (about 0.687 Joule)
 >
 > For the same cash, one could gang up 10 geek caps (150 nF -at- 2kV each) to
 > make a 15nF, 20kV equivalent.  (that's 3 Joules...)  One could be a bit
 > bold, and run the cap at 2.5kV each to give you 4.6875 Joule. (hmmm
voltage
 > squared...)
 >
 > In terms of "bang for your buck), the geek MMC is a MUCH better deal. But
 > wait, there's more to consider..
 >
 > You've got to solder all those geek caps together (as opposed to screwing
 > in the connection on one doorknob), and the inductance of the MMC might be
 > a bit more, slowing down the rise time of your Marx.  And, of course, the
 > physical size of the  MMC is bigger, making a bit of a packaging problem.
 >
 > So, you've got a bit of analysis ahead of you.  I will say, though, that
 > for a Marx bank, stored energy is where it's at, from a visual, audio, and
 > visceral display standpoint.  Nobody watching is going to care whether
your
 > rise time is 0.5 microseconds or 50 microseconds.  Unless you need a
 > particular waveform for a particular reason, optimize for stored energy.
 >
 > Say you want 500kV..You're going to need 200-250 geek caps, and it's going
 > to store a kilojoule. (and it's going to cost $1000 to build the thing, by
 > the time you're done)
 >
 > For lowest buck designs, the 0.001 uF 10kV ceramic caps from All
 > Electronics might work. I think some folks on the list have tried them.
 >
 > Your next big decision is going to be what the stage voltage should
 > be.  Fewer stages is better from a number of sparkgaps and complexity
 > standpoint.  However, high stage voltages are a pain to charge and have
 > huge corona problems.  If you're going with MMC technology, you've got
 > pretty fine grain adjustability (i.e. you can go in 2-3 kV steps for your
 > design). If you're using doorknobs, then that will pretty much determine
 > your stage voltage (1x or 2x doorknob voltage) I find that stage voltages
 > above about 25-30 kV start to get real troublesome from a corona
 > standpoint, and 50-60 kV is just plain evil.  If you're contemplating
 > 40kV+, I suggest you do a bit of fooling with a single stage and charging
 > it.  You're going to spend some serious rebuilding time learning how to
 > fabricate things that have large radii of curvature, no sharp points,
 > etc.  Either that, or immerse the whole thing in oil or pressurized gas.
 >
 > You need to think about your charging supply.  If your bank stores a
 > kilojoule, and you've only got a 50 watt supply, then you're going to
spend
 > 20 seconds charging it (minimum).. actually, you'll probably never charge
 > it with 50 watts, because corona losses might be in that ballpark.   A NST
 > makes a fine charging supply for lower stage voltages.  The inductive
 > current limiting makes it kind of a constant current source, which reduces
 > losses in the charging resistors. (What you really don't want is a
constant
 > voltage supply).  An NST feeding a voltage multiplier might make a good
 > charging supply for a bit higher stage voltages, but you're going to have
 > to invest in some more MMCS for the multiplier caps.  Again, the poor
 > regulation on a CW multiplier doesn't hurt you here.
 >
 >
 >
 >
 >
 >  > > -----Original Message-----
 >  > > From: Tesla list [<mailto:tesla-at-pupman-dot-com>mailto:tesla-at-pupman-dot-com]
 >  > > Sent: Monday, August 18, 2003 6:56 AM
 >  > > To: tesla-at-pupman-dot-com
 >  > > Subject: Re: door knob caps
 >  > >
 >  > >
 >  > > Original poster: "John by way of Terry Fritz
 >  > > <teslalist-at-qwest-dot-net>" <fireba8104-at-yahoo-dot-com>
 >  > >
 >  > > Thanks Dr. Resonance,
 >  > > I was considering using MMC(tons of radio shack 2kv caps) for
 >  > > caps to if I
 >  > > couldn't find doorknobs. Unfortunately I live in New Jersey(no jokes
 >  > > please) and Chessehold is to far away. Also I hope you win
 >  > > that 250 kv
 >  > > x-ray transformer. I say at least a 4 stage Marx.
 >  > > Thanks,
 >  > > John
 >  > >
 >  > >
 >  > > Tesla list <tesla-at-pupman-dot-com> wrote:
 >  > > Original poster: "Dr. Resonance by way of Terry Fritz "
 >  > >
 >  > >
 >  > > I have a very nice design for a Marx using MMC's and they
 >  > > only cost $3 per cap. If you make it to our Cheesehead
 >  > > Teslathon I can share the data with you.
 >  > >
 >  > > We charge at 60 kV/stage using either a flyback tripler or an
 >  > > x-ray xmfr (faster charging) and run 60 MMC caps on each
 >  > > stage. This gives up .01 MFD total cap per stage at 60 kV and
 >  > > makes a very bright output spark at 600 kV with 10 stages.
 >  > >
 >  > > Less expensive than doorknobs.
 >  > >
 >  > > Dr. Resonance
 >  > >
 >
 >
 >
 >
 >