Big EMF Pulse.
Phillip R. Cline
Greetings,
I need to find a way to snub the pulse that is present after I dump
a bank of caps totaling 1000uF at 4KV. into a coil. As the field
collapses a very large pulse is generated which charges the caps in
reverse polarity. The pulse is short duration (less than 2mS).
Any ideas?
Phil Cline.
Mike
Yikes! Hope you got rubber underwear :)
It seems the coil and caps are making a resonant circuit. The energy in
the magnetic field is collapsing, causing the voltage swing negative.
The "Q" is not too good, else you would see a decaying transient that
swings positive and negative with reducing amplitude.
If you want to kill the negative swing, maybe you could try putting a
very hefty diode across the coil. During the initial pulse, it would be
back-biased, but when the voltage tries to swing to the opposite
polarity, it would forward-bias and keep the voltage to one diode drop.
You could try to model it in SPICE, but you need to make some assumptions
about series resistance in the caps and coil. The results would not be
very trustworthy.
I'd get a diode with a reverse breakdown rating substantially higher than
the operating voltage, and a peak forward current rating as high as
possible. Then measure the forward current in actual operation and see
how close it is to the diode rating.
Also, I'd put a metal shield around the diode. If it ever decided to fail
with that much energy in the caps, it would explode and spray shrapnel
everywhere.
If you try this, please let us know what results you get!
Mike
Bill Sloman
In article <3392FD...@iquest.net>, "Phillip R. Cline" <pcl...@iquest.net> says:
>
>Greetings,
> I need to find a way to snub the pulse that is present after I dump
>a bank of caps totaling 1000uF at 4KV. into a coil. As the field
>collapses a very large pulse is generated which charges the caps in
>reverse polarity. The pulse is short duration (less than 2mS).
> Any ideas?
> Phil Cline.
You've got an LC circuit, and it is ringing. The "less than 2ms" suggests
that the inductance of the coil is of the order of 1mH. You could make
the LC circuit critically damped by pushing the series resistance up
to of the order of an ohm. Most of the several kilojoules of energy
initially stored in the capacitors would end up in the resistance, so I
wouldn't use a single resistor for the job.
An alternative approach is to use clamping diodes to let the ringing
current bypass the capacitors. The diodes need to be rated for 4kV
reverse voltage and several kiloamps of forward current, and they have to
soak up several kilojoules of energy, so again we aren't talking about a
single device. Of course, the instantaneous power dissipation in the
diodes will be much lower than in a snubbing resistance; the former will
take of the order of several seconds to dissipate the excess energy,
while the latter will kill it in of the order of 1msec.
Bill Sloman (slo...@sci.kun.nl) | Precision analog design
TZ/Electronics, Science Faculty, | Fast analog design and layout
Nijmegen University, The Netherlands | Very fast digital design/layout
| e-mail for rates and conditions.
Winfield Hill
Phillip R. Cline, <pcl...@iquest.net> said...
>
> I need to find a way to snub the pulse that is present after I dump
> a bank of caps totaling 1000uF at 4KV. into a coil. As the field
> collapses a very large pulse is generated which charges the caps in
Hey Phil, What's a ham like you doing trying to make 8kJ EMP pulses?
I assume you're using some math to work this all out. For example,
I_peak = V sqrt (C/L) etc. For a current peaking after 2ms, I get
your L as about 1600uH and a peak current of about 3100A (assuming no
resistance losses - HA HA HA!). That must be some coil!
Last year I made a 10kJ electrolytic capacitor bank to create a 3500A
pulse. I divided the bank into two parts - the current was supplied
from the charged half and delivered to the uncharged half. I used a
Powerex hockey-puk style SCR to start the pulse, and a coil made from
2/0 cable to stretch it. The half-sine current pulse stopped when the
SCR opened as the current dropped to zero (when 2nd bank had all the
energy from the first, minus some lost in resistance).
Actually, I further refined this with a 3rd charged cap bank and two
more SCRs to quickly stop the current when at its maximum (creating a
second fast 7000A pulse in the process), and added two more SCRs to
reverse the current, but that's another story.
It's true the SCRs required some snubbing to handle reverse-recovery
currents, but these snubbers worked in the 200 to 300A arena with
greatly reduced energy levels. The funny thing was, to get everything
under control, I had to snub the 300A snubber with a 10A snubber!
Now, if as an alternate you use a single cap bank and try shunting it
with a diode, well you'll really stretch out your pulse (maybe that's
good?) plus that'll be some diode - 4kV and maybe 3000A!
How are you turning on your pulse?
Personally, I'd consider say 1kV with 16,000uF (the same amount of
physical capacitance) and wire the coil with 1/4 as many turns (but
the same amount of copper, paralleled). This would keep all the time
constants, energy and physical sizes unchanged, but allow one to use
standard 1200V power components.
Let us know what you're doing, Phil. Also, if we don't hear from you
for a while, should we send out a search party?
--
Winfield Hill hi...@rowland.org _/_/_/ _/_/_/_/
The Rowland Institute for Science _/ _/ _/_/ _/
Cambridge, MA USA 02142-1297 _/_/_/_/ _/ _/ _/_/_/
_/ _/ _/ _/ _/
http://www.artofelectronics.com/ _/ _/ _/_/ _/_/_/_/
Dave VanHorn
> Hey Phil, What's a ham like you doing trying to make 8kJ EMP pulses?
> I assume you're using some math to work this all out. For example,
> I_peak = V sqrt (C/L) etc. For a current peaking after 2ms, I get
> your L as about 1600uH and a peak current of about 3100A (assuming no
> resistance losses - HA HA HA!). That must be some coil!
Must be something in the air.. (KC6ETE here)
I've been looking at similar things, and I got one big question:
> How are you turning on your pulse?
At dayton, I ran into this guy who had some interesting devices for sale.
It's a GE GL-5822-A ignitron. So far, I haven't been able to find any
ratings
or data on the care and feeding. It's got a 1" x 5/8" busbar as cathode
terminal, and about a 1" diameter wire as anode. The water jacket is about
5" dia and maybe a foot long.
Does anyone have a pointer or two on these beasties?
John Woodgate
>Greetings,
> I need to find a way to snub the pulse that is present after I dump
>a bank of caps totaling 1000uF at 4KV. into a coil. As the field
>collapses a very large pulse is generated which charges the caps in
> Any ideas?
> Phil Cline.
overswing without slowing the discharge time too much by adding a small
amout of resistance in series with the L and C. R = 2*sqrt(L/C) will
give just no overshoot with the fastest possible discharge.
--
Regards, John Woodgate Tel. +44 (0)1268 747839
Fax +44 (0)1268 777124. OOO - Own Opinions Only
Alternative e-mail address: jm...@thenet.co.uk
That means I get double spam with everything (;-(
Mike
If it's the kind I'm thinking of, mount it vertical so the mercury
collects in the bottom. There is an electrode mounted to contact the
mercury pool - dump a 1 - 10 uf capacitor charged to 200 volts into this
electrode and sit back and watch the fireworks.
Not much you can do to hurt these things, unless you happen to have a
direct connection to a nuclear power plant.
Mike
Dave VanHorn
> > Does anyone have a pointer or two on these beasties?
>
> If it's the kind I'm thinking of, mount it vertical so the mercury
> collects in the bottom. There is an electrode mounted to contact the
> mercury pool - dump a 1 - 10 uf capacitor charged to 200 volts into this
> electrode and sit back and watch the fireworks.
>
> Not much you can do to hurt these things, unless you happen to have a
> direct connection to a nuclear power plant.
>
> Mike
>
That's what I was hoping, it certainly gives one the impression of not
being
easily damaged, except maybe with a hammer.
I thought the guy just had the one tube so I bought it, ($20) and later,
when I had unfortunately spent my limit, I discovered he had more.
The best part is it was factory-sealed, absolutely brand new!
Phillip R. Cline
Dave VanHorn wrote:
>
> > Hey Phil, What's a ham like you doing trying to make 8kJ EMP pulses?
> > I assume you're using some math to work this all out. For example,
> > I_peak = V sqrt (C/L) etc. For a current peaking after 2ms, I get
> > your L as about 1600uH and a peak current of about 3100A (assuming no
> > resistance losses - HA HA HA!). That must be some coil!
>
> Must be something in the air.. (KC6ETE here)
> I've been looking at similar things, and I got one big question:
>
> > How are you turning on your pulse?
>
> At dayton, I ran into this guy who had some interesting devices for sale.
> It's a GE GL-5822-A ignitron. So far, I haven't been able to find any
> ratings
> or data on the care and feeding. It's got a 1" x 5/8" busbar as cathode
> terminal, and about a 1" diameter wire as anode. The water jacket is about
> 5" dia and maybe a foot long.
>
This is the device I chose to use to switch the cap bank. The device
must be in a perfectly verticle position with the cable on top. To fire
it you can charge a small cap say around 1 or 2 uF to around 100 volts.
This is probably overkill but prevents not triggering. Discharge the cap
from the cathode to the trigger terminal. This vaporizes a little of the
mercury inside the tube and it then conducts. These things are good for
a bunch of current in pulse mode and shouldn't require the cooling with
water.
Phil Cline.
Mike
Dave VanHorn wrote:
>
> > > Does anyone have a pointer or two on these beasties?
> >
Well, you certainly got a bargain, but there's really not much you can do
with it.
When it is conducting, there's about a 18 volt drop across it, and it
won't stop conducting until the current falls to zero. About the only
thing I can think of to use it for is a variable-conduction power supply,
which of course is what it is designed for. Aluminum refiners may still
use them - they need lots of amps.
You could always mount it on a base and put in on your coffee table as a
conversation piece.
Mike
Winfield Hill
Dave VanHorn at dvan...@127.0.0.1 says...
>
>
>> > Does anyone have a pointer or two on these beasties?
>>
>> direct connection to a nuclear power plant.
Not at all! So maybe someone can tell us this source, or repost if
missed it somehow...
> I thought the guy just had the one tube so I bought it, ($20) and later,
> when I had unfortunately spent my limit, I discovered he had more.
>
> The best part is it was factory-sealed, absolutely brand new!
What we need is names and phone numbers (or email addresses!).
--
Winfield Hill hi...@rowland.org
Rowland Institute for Science
Cambridge, MA 02142
Dave VanHorn
> Not at all! So maybe someone can tell us this source, or repost if
> missed it somehow...
Hi Win! Me source, you expert.. ug :)
It's a GE GL-5822-A, I havent' seen too many similar devices, so when
I saw a nice big one in the factory-sealed card. I grabbed it. For $20,
I figured it's a deal. The seller was just some ham at the Dayton
Hamfest, I probably should have gotten his number, but didn't think
of it at the time.
Pat Kelley
Phillip R. Cline wrote:
>
> Greetings,
> I need to find a way to snub the pulse that is present after I dump
> a bank of caps totaling 1000uF at 4KV. into a coil. As the field
> collapses a very large pulse is generated which charges the caps in
> reverse polarity. The pulse is short duration (less than 2mS).
> Any ideas?
> Phil Cline.
Rather than wasting all that energy in snubbers,
why not do a little switching and recover some of it for
the next shot? Assuming that your cap bank isn't polarized,
of course.
--
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
"I think, therefore you are. Don't make me regret it."
http://home.swbell.net/manaret/
//////////////////////////////////////////////////////
Winfield Hill
> a bank of caps totaling 1000uF at 4KV. into a coil. As the field
> collapses a very large pulse is generated which charges the caps in
> reverse polarity. The pulse is short duration (less than 2mS).
Now that you've told us the purpose is to create a high field to
charge small magnets, it's interesting to play with the various
paramters to momentarily maximize this field. I used these
relationships:
magnetic field: B ~ NI/d
peak coil current: I = V sqrt C/L
coil inductance: L ~ d N^2
at 1/4-cycle time: t ~ sqrt LC
heating energy: E ~ I^2 t
The sole idea is to maximize B. No matter how the equations are
manipulated, one thing becomes immediately clear: minimze d, the coil
diameter! This means wrapping the magnetizing solenoid coil right
around the magnet.
Since the field is proportional to NI, one could use more turns N at
less current for the same effect, but this would increase the
inductance by N-squared, which looks bad. However, using small N and
high I increaes the heating by I squared, which also looks bad.
One thought is to maximize the field, subject only to the constraint
of keeping the copper temperature below white hot. As the current
increase, or maximizes and i^2 t heating is taking its toll, a series
fusing link of appropriate size, external to the coil, can open and
stop the current before the coil is damaged. A sufficiently large
capacitor bank and high enough voltage are used to achieve the
maximim field / current just before impending coil failure.
Rather than guess at an optimum tradeoff, I mashed all the equations
with a grinder and found the following proportionalities:
B ~ d^-2.5 Emax^0.5 C^-0.25 N^0.5
We see making the diameter small is very very important. Natch.
Then the maximum allowed energy should be as high as practical, short
of (or even including) coil destruction! This means making the coil
"large" around the magnet - more copper volume. The small capacitor
size (at 1/C^1/4, not a strong effect) merely reflects the value of a
shortned pulse time, 1/t^1/2.
Surprisingly, the wire-size vs. turns tradeoff appears to favor more
turns of smaller wire! Of course this also means higher voltages at
lower currents. Hmmm. Maybe winding-insulation volume should also
be a tradeoff factor.
Be careful, Phil!
Ian Stirling
Winfield Hill (hi...@rowland.org) wrote:
: Phillip R. Cline, <pcl...@iquest.net> said...
: >
: > I need to find a way to snub the pulse that is present after I dump
: > a bank of caps totaling 1000uF at 4KV. into a coil. As the field
: > collapses a very large pulse is generated which charges the caps in
: > reverse polarity. The pulse is short duration (less than 2mS).
: We see making the diameter small is very very important. Natch.
: Then the maximum allowed energy should be as high as practical, short
What about mechanical strength of the coil, re: tearing it'self apart?
: of (or even including) coil destruction! This means making the coil
: "large" around the magnet - more copper volume. The small capacitor
: size (at 1/C^1/4, not a strong effect) merely reflects the value of a
: shortned pulse time, 1/t^1/2.
--
Notice: Anyone mailing to this account should check the notices in the header.
Ian Stirling. Currently designing a new PDA, see homepage.
Homepage: http://www.mauve.demon.co.uk/
Money is a powerful aphrodisiac, but flowers work almost as well.
Robert A Heinlein.
John Woodgate
In article <5n8umo$k...@fridge-nf0.shore.net>, Winfield Hill
<hi...@rowland.org> writes
>Phillip R. Cline, <pcl...@iquest.net> said...
>>
>> I need to find a way to snub the pulse that is present after I dump
>> a bank of caps totaling 1000uF at 4KV. into a coil. As the field
>> collapses a very large pulse is generated which charges the caps in
>> reverse polarity. The pulse is short duration (less than 2mS).
>
>charge small magnets, it's interesting to play with the various
>paramters to momentarily maximize this field.
Where did he say that? By e-mail or telepathy? For an impulse magnetizer
a *slight* overshoot is not undesirable becuase it 'conditions' the
magnet by moving it to a more stable part of the BH loop. With no
overshoot, the inherent demagnetizing effect causes the strength to fall
during the first few hours afetr magnetizing unless a keeper is put on
immediatley.
Phillip R. Cline
Winfield Hill wrote:
>
> John Woodgate, <j...@jmwa.demon.co.uk> said...
> >
> > Winfield Hill <hi...@rowland.org> writes
> >>Phillip R. Cline, <pcl...@iquest.net> said...
> >>>
> >>> I need to find a way to snub the pulse that is present after I dump
> >>
> >> Now that you've told us the purpose is to create a high field to
> >> charge small magnets, it's interesting to play with the various
> >> paramters to momentarily maximize this field.
> >
> > Where did he say that? By e-mail or telepathy?
>
>
> -- Win
Greetings,
Thanks for your help Mr. Hill. I really wished that I had not asked
the question. Now I am seeing everyone telling me that I can't need this
much field strength and the materials never need that much field density
to saturate and I am getting a bunch of people telling me that the pulse
is because of the inductance of the coil. No shit! These people must
think that I am totally ignorant of the entire subject and all of it's
facets.
Hell, all I wanted to know was how to snub a pulse. If it were as easy
as buying a big fast recovery diode that was rated at 2000 amps and
5KV(Yes, that was a suggestion I recieved!) then I would. I guess that I
will play an inductor in series with a large rectifier and see what
happens. All I can do is destroy a few rectifiers.
Anybody that reads this reply can forget replying. I will figure it
out on my own. I don't need all the smartasses telling me how much that
I don't know. You can fill a warehouse with stuff that I don't know so
don't think that I think I know everything. I do however know enough to
realize that this NG is full of people that rather than answer a
question would like to tell you how little you know.(You are the
exception Mr. Hill!)
Thank you,
Phil Cline.
Peter Gutmann
hi...@rowland.org (Winfield Hill) writes:
>Phillip R. Cline, <pcl...@iquest.net> said...
>>
>> I need to find a way to snub the pulse that is present after I dump
>> a bank of caps totaling 1000uF at 4KV. into a coil. As the field
>> collapses a very large pulse is generated which charges the caps in
>> reverse polarity. The pulse is short duration (less than 2mS).
>Now that you've told us the purpose is to create a high field to
>charge small magnets, it's interesting to play with the various
>paramters to momentarily maximize this field.
If you *really* wanted to maximise the field, you could try turning the
existing setup into a flux compression generator, although it's not really
something which you can do in a lab environment where there are things which
could be damaged by EMP (and since the FCG uses explosives, it's rather a
one-shot affair). If people are interested in the details of these beasties
I'll try and dig up some (unclassified) references.
Peter.
Winfield Hill
Peter Gutmann, <pgu...@cs.auckland.ac.nz> said...
I'd say we're certainly interested. Phil Cline said he was trying to
charge rare earth magnets such as such as Samarium Cobalt and Neodymium
Boron Iron. He said he hoped to be able to overcome the coercivity of
magnets, and to achieve a maximum field density of around 65000 Gauss.
[Ha! Superconductor city!]
Even much lower fields saturate most materials, eliminating the simple
flux magnification techniques I know about. So Peter, educate us here!
Winfield Hill
John Woodgate, <j...@jmwa.demon.co.uk> said...
>
> Winfield Hill <hi...@rowland.org> writes
>>Phillip R. Cline, <pcl...@iquest.net> said...
>>>
>>> I need to find a way to snub the pulse that is present after I dump
>>
>> Now that you've told us the purpose is to create a high field to
>> charge small magnets, it's interesting to play with the various
>> paramters to momentarily maximize this field.
>
Samuel M. Goldwasser
Winfield Hill wrote:
<snip>
> I'd say we're certainly interested. Phil Cline said he was trying to
> charge rare earth magnets such as such as Samarium Cobalt and Neodymium
> Boron Iron. He said he hoped to be able to overcome the coercivity of
> magnets, and to achieve a maximum field density of around 65000 Gauss.
> [Ha! Superconductor city!]
>
> Even much lower fields saturate most materials, eliminating the simple
> flux magnification techniques I know about. So Peter, educate us here!
Maybe he really meant 6,500 Gauss which is a much more reasonable value?
Those little rare earth magnets can be made strong but not that strong!
Now if I only had a big crowbar, I could pry all those disk drive
positioner magnets off of my filing cabinet. Oops, then they just
stick to the crowbar. Oh well. :-).
BTW, I really like all those linear motors they are manufacturing
these days using rare earth magnets - substitute for ball screw or
air cylinder drives in many applications.
--- sam : Sci.Electronics.Repair FAQ:
http://www.paranoia.com/~filipg/REPAIR/
Coming soon to a computer screen near you:
http://www.repairfaq.org/
Sam Goldwasser
You might consider that yes there are indeed many smartasses as well as
many knowledgeable people who monitor this group. Without enough information
in the original posting, there is no way to which category you fall into.
Often people ask the *wrong* questions so just replying to the original
question may result in erroneous recommendations.
This was your original posting:
"I need to find a way to snub the pulse that is present after I dump
a bank of caps totaling 1000uF at 4KV. into a coil. As the field
collapses a very large pulse is generated which charges the caps in
reverse polarity. The pulse is short duration (less than 2mS)."
Please don't take this the wrong way but based on your posting,
for all anyone knew, you could be a pimply-faced teenager trying to build
a rail gun (if they didn't kill themselves first). More information never
hurt. How could anyone know that you knew this was due to the inductance?
I am not necessarily defending the actions of anyone acting like a smartass
but you have basically asked the entire world for an answer to a partially
specified problem.
The Usenet newsgroups represent an excellent resource but one does need to know
how to separate the solid information from the B.S.
--- sam : Sci.Electronics.Repair FAQ: http://www.paranoia.com/~filipg/REPAIR/
Coming soon to a computer screen near you: http://www.repairfaq.org/
Winfield Hill
>> John Woodgate, <j...@jmwa.demon.co.uk> said...
>>> Winfield Hill <hi...@rowland.org> writes
>>>> Phillip R. Cline, <pcl...@iquest.net> said...
>
> Greetings,
> ... I really wished that I had not asked the question. ...
> Hell, all I wanted to know was how to snub a pulse. If it were as easy
> as buying a big fast recovery diode that was rated at 2000 amps and 5KV
> (Yes, that was a suggestion I recieved!) then I would. I guess that
> I will play an inductor in series with a large rectifier and see what
> happens. All I can do is destroy a few rectifiers.
> Anybody that reads this reply can forget replying. I will figure it
> out on my own. I don't need all the smartasses telling me how much
> that I don't know. You can fill a warehouse with stuff that I don't
> know so don't think that I think I know everything. I do however know
> enough to realize that this NG is full of people that rather than
Hey Phil, at least we're having a good time playing experts! Come on. get
with the program! That's really the main purpose of these news groups,
isn't it?!
Actually. as a novice in these matters (magnet charging), I do have one
suggestion, involving an technique used in an interesting old physical-
chemistry instrument, a "temperature-jump" apparatus. This instrument
uses a hummungus charged 10kV capacitor bank quickly discharged into a
small (1cm^3) cell of solution, for the purpose of raising its temperature
say about 5 degrees in a few microseconds. A few simple calculations will
convince one that electronically speaking this requires a non-trival
amount of energy, and that the power and current levels are awesome.
In this apparatus, the method used to turn on the current is _mechanical_
and consists of something like a high-pressure-operated high-velocity
wedge shorting two contacts. Anyway, no snubber is needed there I
suspect!
John Woodgate
In article <dRYkOuA+...@jmwa.demon.co.uk>, John Woodgate
<j...@jmwa.demon.co.uk> writes
>In article <5n8umo$k...@fridge-nf0.shore.net>, Winfield Hill
><hi...@rowland.org> writes
>>Phillip R. Cline, <pcl...@iquest.net> said...
>>>
>>> I need to find a way to snub the pulse that is present after I dump
>>> collapses a very large pulse is generated which charges the caps in
>>> reverse polarity. The pulse is short duration (less than 2mS).
>>
>>charge small magnets, it's interesting to play with the various
>>paramters to momentarily maximize this field.
>
>a *slight* overshoot is not undesirable becuase it 'conditions' the
>magnet by moving it to a more stable part of the BH loop. With no
>overshoot, the inherent demagnetizing effect causes the strength to fall
>during the first few hours afetr magnetizing unless a keeper is put on
>immediatley.
>
The group may be interested to know that Mr. Cline told me impolitely by
e-mail that he knew the above already and didn't need me to tell him.
Well, for a person who asks about snubbing, when it's clear that
snubbing is impracticable because of the high current and voltage, and
then objects because people point this out, he is doing well as a
snubber himself.
I think it is unlikely that many people who have not studied impulse
magnetizers in detail do know about the advantage of a little overshoot,
unless the magnet has a keeper. Apparently, Mr. Cline's magnets do have
keepers. I wonder if their owner does.
What did Shakespeare say about it? Blow, blow, thou winter wind/Thou art
no so unkind/As man's.... I forget (;-).
Winfield Hill
> John Woodgate <j...@jmwa.demon.co.uk> writes
>>
>> For an impulse magnetizer a *slight* overshoot is not undesirable
>> becuase it 'conditions' the magnet by moving it to a more stable part
>> of the BH loop. With no overshoot, the inherent demagnetizing effect
>> magnetizing unless a keeper is put on immediately.
John, what do you mean by overshoot? When I first read your post, it
didn't seem significant to me because I imagined a condition where you
want as much field as possible, and the maximum reached could be
considered an "overshoot" of some arbitrary lower field. Did you mean
"undershoot" or field reversal? What's the effect of that?
John Woodgate
In article <5ne2ps$1...@fridge-nf0.shore.net>, Winfield Hill
<hi...@rowland.org> writes
>> John Woodgate <j...@jmwa.demon.co.uk> writes
>>>
>>> For an impulse magnetizer a *slight* overshoot is not undesirable
>>> becuase it 'conditions' the magnet by moving it to a more stable part
>>> of the BH loop. With no overshoot, the inherent demagnetizing effect
>>> causes the strength to fall during the first few hours after
>>> magnetizing unless a keeper is put on immediately.
>
>John, what do you mean by overshoot? When I first read your post, it
>didn't seem significant to me because I imagined a condition where you
>want as much field as possible, and the maximum reached could be
>considered an "overshoot" of some arbitrary lower field. Did you mean
>"undershoot" or field reversal? What's the effect of that?
>
Daniel Haude
John Woodgate (j...@jmwa.demon.co.uk) wrote:
[...]
>Replied by e-mail in case I post something else to upset Mr. Cline.
Oh, I found this thread very interesting (although this subject is quite
new to me). Since Mr. Cline announced that he wouldn't participate in this
discussion any more, you can't possible upset him, could you?
--
Daniel Haude /--------------------------------------
Institut fuer Angewandte Physik /
Universitaet Hamburg / ha...@physnet.uni-hamburg.de
---------------------------------/
Tony Williams
In article <5ngdaq$pak$1...@rzsun02.rrz.uni-hamburg.de>, Daniel Haude
<URL:mailto:HA...@alpha5.physnet.uni-hamburg.de> wrote:
>
> John Woodgate (j...@jmwa.demon.co.uk) wrote:
>
> [...]
>
> >Replied by e-mail in case I post something else to upset Mr. Cline.
>
> Oh, I found this thread very interesting (although this subject is quite
> new to me). Since Mr. Cline announced that he wouldn't participate in this
> discussion any more, you can't possible upset him, could you?
>
Seconded, the tech details were just getting interesting....
--
[Tony Williams, Ledbury, Herefordshire, UK.---Pagewidth=64-----]
Mike & Peggy Cline
There's one thing I was curious about. After the ignitron fires and
current pulse falls to 0, Why doesn't the ignitron stop conducting and
totally block the reverse current? We've tried it on a smaller scale
at only 400v with both an SCR and an ignitron and both device continue
to conduct the the reverse current and charge the caps in reverse :( .
Mike
--
mailto:mcl...@iquest.net
Walter Gray
You don't say what your triggering circuit is. Could be the reverse
pulse triggers via the cathode?
If you have ignitrons available, why not use one as your snubber?
Just trigger it when the reverse pulse starts building up.
Anyway, how many Clines do we have on this job?
--
walter
------
Phillip R. Cline
Daniel Haude wrote:
>
> John Woodgate (j...@jmwa.demon.co.uk) wrote:
>
> [...]
>
> >Replied by e-mail in case I post something else to upset Mr. Cline.
>
> Oh, I found this thread very interesting (although this subject is quite
> new to me). Since Mr. Cline announced that he wouldn't participate in this
> discussion any more, you can't possible upset him, could you?
>
> Daniel Haude /--------------------------------------
> Institut fuer Angewandte Physik /
> Universitaet Hamburg / ha...@physnet.uni-hamburg.de
> ---------------------------------/
I have gotten mostly replies to this straightforward question that do
everything but answer my question. There were exceptions and they
probably know who they are. This became very frustrating! Although this
is a very easy problem to take care of with cubic dollars, I was hoping
for a less expensive answer. Sure, if there was a rectifier that had a
PIV rating of 5KV at 1000 amps for $100 I would not need to ask this
question. Unfortunately they don't exist.(If they do, please give me a
phone number where I can order one.)
I got lessons in basic arithmetic to advanced magnet theory but almost
no practical or in most cases no answer.
I suppose that I should have given a bit more info when I asked the
question. Things like what the heck are you doing with that much energy
dumped into a coil.(Charging rare earth magnets in 1/24 scale slotcar
motors.) Also I should have explained the primary reason for the pulse
to be eliminated.(To prevent the capacitor bank from being charged in
reverse polarity to 1000 Volts.) But I didn't. That was my fault.
I want to again apologize to everyone concerned!
I am still open to answers as long as they don't involve reinventing
the wheel or things that will cost a bunch of bucks.(Commercial magnet
chargers are only $9000 so try to keep that in mind if you will.)
Thanks,
Phil Cline.
ps. This NG is very well mannered as a whole. Much more civil than if
you piss off someone on alt.radio.pirate.
Stefan Wimmer
In article <339C15...@iquest.net>, mcl...@iquest.net wrote:
>Tony Williams wrote:
>
>There's one thing I was curious about. After the ignitron fires and
>current pulse falls to 0, Why doesn't the ignitron stop conducting and
>totally block the reverse current? We've tried it on a smaller scale
>at only 400v with both an SCR and an ignitron and both device continue
>to conduct the the reverse current and charge the caps in reverse :( .
Thats because Ignitrons and SCRs stop conducting after the _current_ falls
under a certain thereshold and not the voltage. The collapsing field in the
inductor works as your energy source then until your LC oscillator did a half
cycle.
--
__________________________________________________________________
| Stefan Wimmer __ |
| hardware designer _|__| |
| |_|__ |
| Email s...@cellware.de |__| |
| Tel (+49 30) 670 08-135 Cellware Broadband |
| Fax (+49 30) 670 08-108 Rudower Chaussee 5 |
| WWW http://www.cellware.de/ 12489 Berlin, Germany |
|__________________________________________________________________|
Do 't tou h the f op ydi ks su f ce! (Anonymous)
It's the government. It doesn't have to make sense. (B. Kaplow)
John Woodgate
>John Woodgate (j...@jmwa.demon.co.uk) wrote:
>
>[...]
>
>>Replied by e-mail in case I post something else to upset Mr. Cline.
>
>Oh, I found this thread very interesting (although this subject is quite
>new to me). Since Mr. Cline announced that he wouldn't participate in this
>discussion any more, you can't possible upset him, could you?
>
disapproval of a shower of rather naive responses rather wider than he
intended. So, basically what I mailed to Win was that if in an impulse
magnetizer the circuit Q is arranged to be slightly greater than 0.707,
the current reverses after the first pulse, but is not nearly as large.
This slightly weakens the magnet but moves its magnetization to a more
stable point on its hysteresis loop, so that it doesn't lose strength so
much if not keepered. This is called 'conditioning' or 'stabilizing' the
magnet.
John Woodgate
In article <5nha9d$imd$1...@trog.dra.hmg.gb>, Walter Gray
<wag...@dra.hmg.gb> writes
>Anyway, how many Clines do we have on this job?
>
Well, we have the originator, Philip R Cline, and then there's Win
Hill, who must have a syncline and an anticline, so I'm inclined to say
it's three.
Sorry, it was completely irresistible!
John Woodgate
In article <339C15...@iquest.net>, Mike & Peggy Cline
<mcl...@iquest.net> writes
>> >
>> > John Woodgate (j...@jmwa.demon.co.uk) wrote:
>> >
>> > [...]
>> >
>> > >Replied by e-mail in case I post something else to upset Mr. Cline.
>> >
>> > Oh, I found this thread very interesting (although this subject is quite
>> > new to me). Since Mr. Cline announced that he wouldn't participate in this
>> > discussion any more, you can't possible upset him, could you?
>> >
>>
>> --
>> [Tony Williams, Ledbury, Herefordshire, UK.---Pagewidth=64-----]
>
>current pulse falls to 0, Why doesn't the ignitron stop conducting and
>totally block the reverse current? We've tried it on a smaller scale
>at only 400v with both an SCR and an ignitron and both device continue
>to conduct the the reverse current and charge the caps in reverse :( .
I expect the current reverses too fast for the charge carriers to have
left the conduction region/die (dpending on whether it's an ignitron or
a thyristor)before the voltage is again high enough to maintain
conduction.
Dave VanHorn
> I am still open to answers as long as they don't involve reinventing
> the wheel or things that will cost a bunch of bucks.(Commercial magnet
> chargers are only $9000 so try to keep that in mind if you will.)
>
> Thanks,
> Phil Cline.
>
> ps. This NG is very well mannered as a whole. Much more civil than if
> you piss off someone on alt.radio.pirate.
>
Please keep it on the NG. I deal with small ND magnets in my work that
are sometimes apparently not well charged, so this is interesting to me.
At one point we decided to test the neo to see if it was the material or
the
charger, and ran smack into the price of a charger.
An interesting thing I learned, is that the field on the rectangular face
is
bathtub shaped, if you map intensity over the surface. The center is NOT
the peak, neither are the edges. It's quite pronounced, and easy to see
with a gaussmeter and small probe.
As to your snubbing problem.. Would a shorted transmission line help
here? The idea is that the initial pulse reflects off the short and arrives
back at the input after a delay that is dependent on the length of the
cable (and it's velocity factor). It MIGHT be possible to cancel your
problem..
You are referring to the induced current as the field in the coil
collapses?
Daniel Haude
Phillip R. Cline (pcl...@iquest.net) wrote:
-> I suppose that I should have given a bit more info when I asked the
->question. Things like what the heck are you doing with that much energy
->dumped into a coil.(Charging rare earth magnets in 1/24 scale slotcar
->motors.) Also I should have explained the primary reason for the pulse
->to be eliminated.(To prevent the capacitor bank from being charged in
->reverse polarity to 1000 Volts.) But I didn't. That was my fault.
-> I want to again apologize to everyone concerned!
-> I am still open to answers as long as they don't involve reinventing
->the wheel or things that will cost a bunch of bucks.(Commercial magnet
->chargers are only $9000 so try to keep that in mind if you will.)
Oh, so we DO have quite some financial headroom for our suggestions ;-)
One question: How do you turn your current ON? With an SCR? Probably not;
if you had an SCR rated that high, you could also afford an appropriate
diode...
Roy McCammon
Mike & Peggy Cline wrote:
> There's one thing I was curious about. After the ignitron fires and
> current pulse falls to 0, Why doesn't the ignitron stop conducting and
> totally block the reverse current? We've tried it on a smaller scale
> at only 400v with both an SCR and an ignitron and both device continue
> to conduct the the reverse current and charge the caps in reverse :( .
Assuming the capacitors are grounded and given a positive charge,
the inductor is between the caps and the scr which grounds
the inductor, and that there is no stray capacitance with which
the inductor rings, the answer is: the current does not reverse.
When the scr first fires, a current flows from the caps thru
the inductor and thru the scr, which discharges the caps. When
the cap voltage equals the scr voltage (inductor voltage = 0),
the current will still be the same direction and will continue
to tdischarge the caps. It can continue even to the point of
discharging the caps right thru zero volts and thus reverse
charging them. Its not a reverse current that reverse charges the
caps, its a forward current that lasts too long.
Opinions expressed herein are my own and may not represent those of my employer.
Phillip R. Cline
There is a reverse current. It is visible on a scope.The voltage
clearly reverses.The voltage changed polarity across a very low
resistance that was used to measure peak current. The reverse current is
due to the magnetic field collapsing in the coil.This is true of all
inductors to some extent.
Help Mr. Hill. Am I crazy or just totally confused.
Phillip R. Cline.
Pat Kelley
Mike & Peggy Cline wrote:
>
> There's one thing I was curious about. After the ignitron fires and
> current pulse falls to 0, Why doesn't the ignitron stop conducting and
> totally block the reverse current? We've tried it on a smaller scale
> at only 400v with both an SCR and an ignitron and both device continue
> to conduct the the reverse current and charge the caps in reverse :( .
> --
> mailto:mcl...@iquest.net
As it happens, I have access to a commercial magnet
treater, which apparently uses an ignitron. I'll see if I
can find a schematic on it.
--
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
"I think, therefore you are. Don't make me regret it."
http://home.swbell.net/manaret/
//////////////////////////////////////////////////////
Sam Goldwasser
> There is a reverse current. It is visible on a scope.The voltage
> clearly reverses.The voltage changed polarity across a very low
> resistance that was used to measure peak current. The reverse current is
> due to the magnetic field collapsing in the coil.This is true of all
> inductors to some extent.
> Help Mr. Hill. Am I crazy or just totally confused.
But when and for how long? Have you simultaneously monitored capacitor
voltage and coil current? After the cap charges up in the reverse direction,
it will then attempt to discharge in reverse. Is this what you are ssing?
Dave VanHorn
>
> There is a reverse current. It is visible on a scope.The voltage
> clearly reverses.The voltage changed polarity across a very low
> resistance that was used to measure peak current. The reverse current is
> due to the magnetic field collapsing in the coil.This is true of all
> inductors to some extent.
> Help Mr. Hill. Am I crazy or just totally confused.
>
> Phillip R. Cline.
>
I think this is a metering artifact. The current can keep going forward
through the coil, but the voltage on the input side of the coil goes below
ground at that time.. Same phenomena in a boost switcher, where the
Q turns on the L for a while, then turns off. The I through the L continues
to flow, and the output side, hanging in the breeze, goes WAY positive,
(at least if there's no output circuit to catch it.)
If you are gating the input as I suspect, then the current continuing to
flow
would drag the input side of the coil way negative, but the flux isn't
reversing
direction, it's just collapsing, and inducing current in the coil as it
does so.
CAPS-------IGNITRON--------COIL--------GND. Is that basically right?
Winfield Hill
Pat Kelley, <man...@swbell.net> said...
>
> As it happens, I have access to a commercial magnet
>treater, which apparently uses an ignitron. I'll see if I
>can find a schematic on it.
That will raise more than a few eyebrows!
Winfield Hill
> Help, Mr. Hill. Am I crazy or just totally confused.
Phil, I'm sorry, you probably know much more than I about Ignitrons! At
this point, my knowledge is restricted to the dozen or so pages devoted to
ignitrons in some of my older reference books. They do mention avoiding any
reversal of the ignitron trigger voltages. I supposed a goal is to keep the
trigger electrode intact and clean. Hmmm. Their concern may be for a long
life with 60 firings a second, while yours would be only for a few dozen at
most.
But I've never even seen an ignitron, nor an ignitron data sheet. In fact,
even tho I frequent FLEA at MIT (usually spending about $600 each time on old
goodies), I don't remember any ignitrons available, new or used, so if you see
more of them I'd love to hear about it. I'll give you $75 for a new one.
OK. I have noticed some good comments in this thread. For example, surely
the plasma is responsible for a considerable amount of current flow after the
voltage sign reverses, analogous to the reverse-recovery time of a rectifier.
This would mean the inductance has a pathway to continue driving the current
derived from its stored energy. Note that -V = L di/dt must be a very large
voltage to get your inductor's current stopped, since dI is large and dt is
short. Only when the discharge pathway has disappeared, with an insulating
ability of many many thousands of volts, will everything quiet down.
One solution is to take this current elsewhere with another device able to
handle this environment. Having just finished a 250kW pulser, I'm reminded
that there are few real-world devices available for use. Even the Powerex SCRs
I used for my "3500A pulse" could only handle 7 to 10,000A peak and 1400V max.
And they had a considerable reverse-recovery time. So, although ignorant about
Ignitrons, I like the suggestion that you turn on another Ignitron to take the
inductor's current. Did you buy two?
You could carefully experiment and find the right setup time to turn on a
second ignitron so as to discharge the coil's inductance as fast as possible
while protecting your magnitizing-current switch. Some circuit manipulation
will reveal to you a way to quickly extinguish the current in the magnet-
charging coil, with a reverse-charged capacitor bank. This is what I did for
my own high magnetic-field pulse design. I'd like to spell this out for you in
more detail, but I really must run, because my wife is insisting I pack for our
short vacation starting in a few hours. And before that I must download some
stuff from my lab computer into my laptop. See you next week!
Keep in touch!
Mike
> Winfield Hill hi...@rowland.org _/_/_/ _/_/_/_/
> The Rowland Institute for Science _/ _/ _/_/ _/
> Cambridge, MA USA 02142-1297 _/_/_/_/ _/ _/ _/_/_/
> _/ _/ _/ _/ _/
> http://www.artofelectronics.com/ _/ _/ _/_/ _/_/_/_/
This is getting much too complicated. It seems a much simpler
answer has already been posted - but you have to filter out the
ones that point to the solution.
Check John Woodgate's reply from June 3:
"You have an underdamped resonant circuit, and you can eliminate
the overswing without slowing the discharge time too much by
adding a small amount of resistance in series with the L and C. R
= 2*sqrt(L/C) will give just no overshoot with the fastest
possible discharge."
Then, part of John's posting from June 6:
"For an impulse magnetizer a *slight* overshoot is not
undesirable because it 'conditions' the magnet by moving it to a
more stable part of the BH loop. With no overshoot, the inherent
demagnetizing effect causes the strength to fall during the first
few hours after magnetizing unless a keeper is put on
immediately."
It seems you could add a small amount of resistance (very heavy-
duty to handle the peak currents!), then tune the damping to give
enough overshoot to properly condition the magnet.
To see how the overshoot changes with damping resistance, check
my posting on June 11. This supplies a SPICE model and a URL for
MicroCap V, as well as a program to calculate the inductance of
the coil.
To run the simulation, load BIGEMF2.CIR into MicroCap, and press
Alt-1, then F2. A little bit of study in the Help will show you
how to change the circuit, but if you just want to change the
damping resistor, click on the value and enter a new one.
And, for God's sake, be careful. You cannot blow up anything in
SPICE, but don't let that carry over to the real world!
If anyone is experienced in this area, please let us know how to
charge a magnet - I lie awake at nights thinking of poor Phillip
tinkering with lethal circuits.
Mike
Terry Montgomery
Winfield Hill is being rather inconsistent in his postings.
Just yesterday he was concerned about his overinflated reputation.
Today he mentions wife, packing & vacation.
Win, the solution is obvious. Once your wife starts posting to this
group, I am sure that your reputation will be quite bounded :-).
--
Terry Montgomery
mo...@pitot.dfrc.nasa.gov
My opinion in my opinion.
Dave VanHorn
> >CAPS-------IGNITRON--------COIL--------GND. Is that basically right?
> >
> >
> Well, you need to start from ground as well, otherwise there is no
> circuit, but I suppose you meant that.
Thanks, yes.. :)
> The set up that PC has is an underdamped resonant circuit. The current
> must reverse: you cannot get reverse charge on a cap without putting
> reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
>
> The descriptions which claim to show that the current does not reverse
> appear to be dealing only with the epoch when the current is decreasing.
> What may be confusing the issue is that everything happens too fast for
> teh ignitron to de-ionize, so that it conducts in reverse. This is not
> the same as in a SMPS, where the switch must turn off and the frequency
> had better be low enough to let it
> --
Ok, that's a feature of ignitrons I suspected, but had no data on.
Would it be possible, you think, to fire some charge into a shorted
transmission line, and use the reflected pulse (opposite polarity)
to help shut off the ignitron?
The transmission line might be interesting to construct, but it
shouldn't take a LOT of energy.. What we want to do is cause
the ignitron current to go to zero for a while,.... hmm.
Thinking this through, the magnetic field <MUST> collapse,
and we don't get any choice about it. Can it, in a practical
sense, collapse without inducing current flow somewhere??
While in theory, one can imagine a perfect switch, I suspect
this is impossible in practice. The voltage across the inductor
ought to rise until current flows..
Hmm. what happens if you build up a huge field around a coil,
then use explosives to cause the coil to stop existing?
Dave VanHorn
As long as it's relevant to the thread:
Does anyone have data, specs, etc on a GE GL-5822-A
ignitron? I'm going down the same path in the near future,
and I would like to find out all I can about the care and feeding
of these tubes before I start making expensive mistakes!
Thanks
John Woodgate
In article <01bc7611$46592460$35f1...@hammer.stonewest.com>, Dave
VanHorn <dvan...@127.0.0.1> writes
>
>
Well, you need to start from ground as well, otherwise there is no
circuit, but I suppose you meant that.
The set up that PC has is an underdamped resonant circuit. The current
must reverse: you cannot get reverse charge on a cap without putting
reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
The descriptions which claim to show that the current does not reverse
appear to be dealing only with the epoch when the current is decreasing.
What may be confusing the issue is that everything happens too fast for
teh ignitron to de-ionize, so that it conducts in reverse. This is not
the same as in a SMPS, where the switch must turn off and the frequency
had better be low enough to let it
--
Roy McCammon
John Woodgate wrote:
>
> In article <01bc7611$46592460$35f1...@hammer.stonewest.com>, Dave
> VanHorn <dvan...@127.0.0.1> writes
> >CAPS-------IGNITRON--------COIL--------GND. Is that basically right?
> The set up that PC has is an underdamped resonant circuit. The current
> must reverse: you cannot get reverse charge on a cap without putting
> reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
> The descriptions which claim to show that the current does not reverse
> appear to be dealing only with the epoch when the current is decreasing.
I guess its time to define terms.
Lets call current going in the direction from the caps
thru the ignitron, thru the coil to ground as positve
(or forward) and assume the caps are initially
charged to a positive voltage.
The current that initially flows when the ignitron is fired is
positive and also discharges the caps. There is no need for this
current to reverse in order to discharge the caps to a negative
voltage. In fact, if this current reversed, it would be a
charging current, tending to take the caps positive,
rather than reverse charging them.
Phillip R. Cline
John Woodgate wrote:
>
> Well, you need to start from ground as well, otherwise there is no
> circuit, but I suppose you meant that.
>
> must reverse: you cannot get reverse charge on a cap without putting
> reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
>
> The descriptions which claim to show that the current does not reverse
> appear to be dealing only with the epoch when the current is decreasing.
> teh ignitron to de-ionize, so that it conducts in reverse. This is not
> the same as in a SMPS, where the switch must turn off and the frequency
> had better be low enough to let it
> --
> Regards, John Woodgate Tel. +44 (0)1268 747839
> Fax +44 (0)1268 777124. OOO - Own Opinions Only
> Alternative e-mail address: jm...@thenet.co.uk
> That means I get double spam with everything (;-(
I am glad that you are in agreement. The caps can't be charged in
reverse without the current changing direction. I read one gentlemans
explanation(not Mr. Van Horn's) and came away scratching my head.
I then asked Mr. Hill if I was missing something. I was beginning to
think that everything I thought that I knew was wrong.
Phil Cline.
Adrian Jansen
From what little I know about SmCo and similar magnets, the 65000
gauss figure is more likely correct - the coercive force is on the
order of 12000 oersted, and a rough rule of thumb is that you need
5 * coercive force to achieve saturation.
And no, you dont need superconductors to get this sort of field,
but it is pretty tricky.
Mr Hill, as usual, has managed to define the problem in a nutshell.
In article <33998A...@stdavids.picker.com>,
"Samuel M. Goldwasser" (s...@stdavids.picker.com) writes: >Winfield
Hill wrote: >
><snip>
>
>> I'd say we're certainly interested. Phil Cline said he was trying to
>> charge rare earth magnets such as such as Samarium Cobalt and Neodymium
>> Boron Iron. He said he hoped to be able to overcome the coercivity of
>> magnets, and to achieve a maximum field density of around 65000 Gauss.
>> [Ha! Superconductor city!]
>>
>> Even much lower fields saturate most materials, eliminating the simple
>> flux magnification techniques I know about. So Peter, educate us here!
>
>Maybe he really meant 6,500 Gauss which is a much more reasonable value?
>
>Those little rare earth magnets can be made strong but not that strong!
>
>Now if I only had a big crowbar, I could pry all those disk drive
>positioner magnets off of my filing cabinet. Oops, then they just
>stick to the crowbar. Oh well. :-).
>
>BTW, I really like all those linear motors they are manufacturing
>these days using rare earth magnets - substitute for ball screw or
>air cylinder drives in many applications.
>
>--- sam : Sci.Electronics.Repair FAQ:
>http://www.paranoia.com/~filipg/REPAIR/
> Coming soon to a computer screen near you:
>http://www.repairfaq.org/
>
--
Regards,
Adrian Jansen aja...@brisbane.DIALix.oz.au
Design Engineer J & K Micro Systems
Ph (076) 394 676
Microcomputer solutions for industrial control
John Woodgate
In article <01bc76ae$bbe89420$29f1...@hammer.stonewest.com>, Dave
VanHorn <dvan...@127.0.0.1> writes
>
>> >CAPS-------IGNITRON--------COIL--------GND. Is that basically right?
>> >
>> >
>> circuit, but I suppose you meant that.
>
>> The set up that PC has is an underdamped resonant circuit. The current
>> must reverse: you cannot get reverse charge on a cap without putting
>> reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
>>
>> The descriptions which claim to show that the current does not reverse
>> appear to be dealing only with the epoch when the current is decreasing.
>> What may be confusing the issue is that everything happens too fast for
>> teh ignitron to de-ionize, so that it conducts in reverse. This is not
>> the same as in a SMPS, where the switch must turn off and the frequency
>> had better be low enough to let it
>> --
>
>Would it be possible, you think, to fire some charge into a shorted
>transmission line, and use the reflected pulse (opposite polarity)
>to help shut off the ignitron?
>
>The transmission line might be interesting to construct, but it
>shouldn't take a LOT of energy.. What we want to do is cause
>the ignitron current to go to zero for a while,.... hmm.
>
>Thinking this through, the magnetic field <MUST> collapse,
>and we don't get any choice about it. Can it, in a practical
>sense, collapse without inducing current flow somewhere??
>
>While in theory, one can imagine a perfect switch, I suspect
>this is impossible in practice. The voltage across the inductor
>ought to rise until current flows..
I do not believe that there is any need for this. All that is needed is
to damp the LC circuit so that the current does not reverse.
>
>Hmm. what happens if you build up a huge field around a coil,
>then use explosives to cause the coil to stop existing?
That is the principle of the flux concentrator already mentioned in this
thread. I believe you have a coil, surrounded by a copper cylinder. You
set up a large current in teh coil and then use explosives to implode
the copper cylinder. The flux cannot escape through the highly
conducting cylinder, so it gets sorta squeezed in the middle.
John Woodgate
In article <339F2B...@mmm.com>, Roy McCammon <rbmcc...@mmm.com>
writes
>John Woodgate wrote:
>>
>> In article <01bc7611$46592460$35f1...@hammer.stonewest.com>, Dave
>> VanHorn <dvan...@127.0.0.1> writes
>
>> >CAPS-------IGNITRON--------COIL--------GND. Is that basically right?
>
>
>> must reverse: you cannot get reverse charge on a cap without putting
>> reverse current into it. V = Q/C, after all, and Q = Integral (i)dt.
Whoops! Cancel the bit about reverse charge. I should take more water
with it.
>I guess its time to define terms.
>Lets call current going in the direction from the caps
>thru the ignitron, thru the coil to ground as positve
>(or forward) and assume the caps are initially
>charged to a positive voltage.
That would be startlingly conventional, yes.
>
>The current that initially flows when the ignitron is fired is
>positive and also discharges the caps. There is no need for this
>current to reverse in order to discharge the caps to a negative
>voltage. In fact, if this current reversed, it would be a
>charging current, tending to take the caps positive,
>rather than reverse charging them.
Well, apart from not liknig the phrase 'taking the caps positive', I
agree, BUT that doesn't invalidate anything I put into this thread apart
from the bit farther up the screen, which I now withdraw.
What mechanism do you envisage that makes the voltage go negative? Do
you agree that it is simply the damped oscillation of an LC circuit?
Mark Zenier
In article <01bc76ae$bbe89420$29f1...@hammer.stonewest.com>,
Dave VanHorn <dvan...@127.0.0.1> wrote:
>
>Thinking this through, the magnetic field <MUST> collapse,
>and we don't get any choice about it. Can it, in a practical
>sense, collapse without inducing current flow somewhere??
The problem doesn't seem to me to be the collapse, but the recharging
of the cap to the opposite polarity, followed by this discharging back
through the coil. One fix I can see is to have a second ignitron (with
a large wattage resistor) across the capacitor bank, with a delayed
trigger to go off half a cycle after the first just when the capacitor
voltage is starting to reverse.
Or a bunch of microwave oven diodes in reverse across the capacitor bank.
A look in the NTE book shows that they are rated at between 25 and 250
amps surge, so that's not as wild as it sounds. That would only be a
couple of hundred of them, and at that quantity he should be able to get
a better price than the $3 each the service parts supply places charge.
Mark Zenier mze...@eskimo.com mze...@netcom.com
Peter Gutmann
I wrote:
>hi...@rowland.org (Winfield Hill) writes:
>
>>Phillip R. Cline, <pcl...@iquest.net> said...
>>>
>>>I need to find a way to snub the pulse that is present after I dump
>>>a bank of caps totaling 1000uF at 4KV. into a coil. As the field
>>>collapses a very large pulse is generated which charges the caps in
>>>reverse polarity. The pulse is short duration (less than 2mS).
>>Now that you've told us the purpose is to create a high field to
>>charge small magnets, it's interesting to play with the various
>>paramters to momentarily maximize this field.
>If you *really* wanted to maximise the field, you could try turning
>the existing setup into a flux compression generator, although it's
>not really something which you can do in a lab environment where
>there are things which could be damaged by EMP (and since the FCG
>uses explosives, it's rather a one-shot affair). If people are
>interested in the details of these beasties I'll try and dig up some
>(unclassified) references.
Here's a quick overview. Note that this almost certainly falls into the
impractical/silly category for most people, but it does show what you can do
with the right resources and government funding.
The flux compression generator was first developed in the late 1950's, and is
now the subject of extensive research in several countries, with current
models being capable of a peak power level in the order of terawatts to tens
of terawatts (did I mention that it was unsuitable for use in a lab
environment?). An FCG works by using an explosive charge to compress a
magnetic field, transferring the energy from the explosive into the magnetic
field. A typical FCG looks something like this:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Non-magnetic
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% structural jacket
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ Stator winding
-------------------------------------------- Copper armature tube
/ ############################################ Explosive core
\ ############################################
--------------------------------------------
Explosive
lens
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The initial magnetic field is produced by discharging a battery of HV
capacitors through the device. When the current peaks, the explosive lens is
used to produce a uniform planar explosion wave through the device which
compresses the armature against the stator. This forms a short-circuit
between the ends of the stator coil, shorting the start current source and
trapping the current within the FCG. As the explosive wave travels down the
device, it compresses the magnetic field, which slowly ramps up to a maximum
just before the device disintegrates. Peak currents are in the tens of
mega-amps with peak energies of tens of megajoules. Some references on FCG's
are:
"Production of Very High Magnetic Fields by Implosion", C.Fowler, W.Garn,
R.Caird, Journal of Applied Physics, Vol.31, No.3 (March 1960), p.588.
"Tests of an Explosive Driven Coaxial Generator", R.Caird, Proceedings of the
5th IEEE Pulsed Power Conference, 1985, p.220.
"The Mark IX Generator", Proceedings of the 5th IEEE Pulsed Power Conference,
1989, p.475.
(other pulsed power conference proceedings often have material related to this
sort of thing as well). It's interesting to note that the first few papers
were published in the early 1960's before the effects of nuclear EMP had been
observed much, after this there is an almost complete silence lasting around
25 years before newer papers began to appear in the late 1980's. With the
winding down of the cold war it is apparently acceptable to talk about this
subject again.
Peter.
Daniel Haude
Phillip R. Cline (pcl...@iquest.net) wrote:
> I am glad that you are in agreement. The caps can't be charged in
>reverse without the current changing direction. I read one gentlemans
>explanation(not Mr. Van Horn's) and came away scratching my head.
>I then asked Mr. Hill if I was missing something. I was beginning to
>think that everything I thought that I knew was wrong.
+ ------------->----------------
| Current |
| )
--------- )
--------- Capacitor ) Coil
| )
| |
- ------------------------------
Take a look at this humble ASCII drawing. It shows the situation just after
the start of the pulse: The capacitor is charged (upper terminal positive)
and discharges into the coil (note the current arrow at the top). The
ignitron is left out because it's like a closed switch at this time.
As the charge in the capacitor decreases, the current steadily rises,
reaching its MAXIMUM when the voltage across the capacitor is ZERO (this may
sound inplausible, but keep in mind that voltage and current are 90 degrees
out of phase in this circuit).
The inductivity now forces the current to keep flowing (so to speak). What
happens to the voltage across the cap? Will the upper terminal become
positive again, although the current is still flowing OUT OF it? No! It
will become negative, the cap will charge "in reverse", but the current
still flows IN THE SAME DIRECTION as in the beginning!
The rising "reverse voltage" (upper terminal negative) across the capacitor
eventually causes the current to stop. What's the situation now? Just like
in the beginning, capacitor fully charged and current zero, buit with
opposite voltage. It is only at this point that the current will reverse,
yes, you DO have a reverse current after all, but this is only AFTER the
VOLTAGE across the cap has reversed -- i.e. it's too late to do anything by
this time.
Check it out with an oscilloscope; maybe with a less dangerous circuit.
This is basic LC circuit behavior. Now of course the resulting oscillating
process can't go on forever, the resistive parts of the circuit will
eventually drown both current and voltage. They may even be so dominant
that the oscillation doesn't even make half a period; in that case, there
will be no reverse voltage indeed.
Roy McCammon
John Woodgate wrote:
> What mechanism do you envisage that makes the voltage go negative? Do
> you agree that it is simply the damped oscillation of an LC circuit?
Yes
There is an added complication that the inductance
changes from a high to a low value and stays low as
the magnet saturates.
There is the posibility that the circuit may be
initially underdamped and then become overdamped
as saturation sets in. Thus we may see initial
oscilatory behavior that then becomes
non-oscilatory. That might lead to observing
a single negative pulse on the caps.
It may possible to arraigned the time constants
so that the switch over occurs before the caps
become negatively charged. Of course that means
the entire setup is tuned to some specific magnet.
Roy McCammon
Appologies for posible duplicate posting.
News server is cranky.
Phillip R. Cline
Ted Park wrote:
> >
> >Surprisingly, the wire-size vs. turns tradeoff appears to favor more
> >turns of smaller wire! Of course this also means higher voltages at
> >lower currents. Hmmm. Maybe winding-insulation volume should also
> >be a tradeoff factor.
> >
> >Be careful, Phil!
>
> I'm curious now about what kind of wire you would use. I've only made
> small pulsed electromagnets but they haven't been nearly strong enough
> to magnetize anything with greater coercivity than a screwdriver.
> How long does the pulse have to last to magnetize materials such as
> NdFeB? I've seen figures like 12500 gauss for some materials, but
> that kind of seems like a lot of current going through a lot of turns.
>
> I guess that a few turns of wire wrapped around a pop tin won't cut it.
>
> --Ted
>
> --
> ---------------------------------------------
> Ted Park tp...@world.std.com
> also <A HREF="mailto:tp...@canuck.com"> tp...@canuck.com </A>
> On the web as <A HREF="http://www.beer.org/~tpark/"> Ted's Home Page </A>
Greetings,
My brother wrote a program that will give you the field density from
the coil size, wire size and number of turns, length and no. of layers
and the peak current. We have plugged in several sets of numbers . The
original plan was to use 450 volts. With this the wire size of choice
was 16 ga. The coil was to be six windings in parallel with each
winding consisting of 150 turns per winding with each winding
consisting of three layers.(50 turns down then up then back down.)The
coil form was to be 2 inches O.D. This yielded approx. .06 ohms as I
remember. This yielded
over 65000 gauss at 5000 amps. as I remember.(Please you guys, don't
calculate this and tell me I am wrong as this data is from memory)
This sounds like a lot of current but the pulse is a very short
duration. The reason for such seemingly small wire is to get the most
ampere turns/inch. Amp. turns by themselves are not important.
With the higher voltage I will go with much larger wire as the current
will be increased a bunch and I will shorten the overall length of the
coil but still employ the multi layer multi winding approach. I will
probably use 14 or 12 ga.
I am not sure how long the pulse has to be there but I suspect that
100's of microseconds is sufficient. I know it is with ceramic material.
I hope this helps,
Phil Cline.
Ted Park
In article <5n8umo$k...@fridge-nf0.shore.net>,
Winfield Hill <hi...@rowland.org> wrote:
>
>One thought is to maximize the field, subject only to the constraint
>of keeping the copper temperature below white hot. As the current
>increase, or maximizes and i^2 t heating is taking its toll, a series
>fusing link of appropriate size, external to the coil, can open and
>stop the current before the coil is damaged. A sufficiently large
>capacitor bank and high enough voltage are used to achieve the
>maximim field / current just before impending coil failure.
>
>Rather than guess at an optimum tradeoff, I mashed all the equations
>with a grinder and found the following proportionalities:
>
> B ~ d^-2.5 Emax^0.5 C^-0.25 N^0.5
>
>We see making the diameter small is very very important. Natch.
>Then the maximum allowed energy should be as high as practical, short
>of (or even including) coil destruction! This means making the coil
>"large" around the magnet - more copper volume. The small capacitor
>size (at 1/C^1/4, not a strong effect) merely reflects the value of a
>shortned pulse time, 1/t^1/2.
Ian Stirling
Mark Zenier <mze...@netcom.com> wrote:
: In article <01bc76ae$bbe89420$29f1...@hammer.stonewest.com>,
: Dave VanHorn <dvan...@127.0.0.1> wrote:
: >
: >Thinking this through, the magnetic field <MUST> collapse,
: >and we don't get any choice about it. Can it, in a practical
: >sense, collapse without inducing current flow somewhere??
: The problem doesn't seem to me to be the collapse, but the recharging
: of the cap to the opposite polarity, followed by this discharging back
: through the coil. One fix I can see is to have a second ignitron (with
: a large wattage resistor) across the capacitor bank, with a delayed
: trigger to go off half a cycle after the first just when the capacitor
: voltage is starting to reverse.
: Or a bunch of microwave oven diodes in reverse across the capacitor bank.
: A look in the NTE book shows that they are rated at between 25 and 250
Oldish catalogue (95) has a GE diode that lists 6000A (10ms) at 1200V,
So you'd need 4 of them, at $50 each or so.
(RS compnets, not radio shack (http://www.rscomponents.com/))
: amps surge, so that's not as wild as it sounds. That would only be a
: couple of hundred of them, and at that quantity he should be able to get
: a better price than the $3 each the service parts supply places charge.
Of course 1N4007's are only $.05 each, assuming they will take 12A
surge, you would only need 2000 of them, for $100
Course, you would be all week soldering them :)
: Mark Zenier mze...@eskimo.com mze...@netcom.com
--
Ian Stirling. Designing a linux PDA, see http://www.mauve.demon.co.uk/
-----******* If replying by email, check notices in header *******-------
Among a man's many good possessions, A good command of speech has no equal.
Ian Stirling
Phillip R. Cline <pcl...@iquest.net> wrote:
: Ted Park wrote:
: > >
: > >Surprisingly, the wire-size vs. turns tradeoff appears to favor more
: > >turns of smaller wire! Of course this also means higher voltages at
<snip>
: Greetings,
: My brother wrote a program that will give you the field density from
: the coil size, wire size and number of turns, length and no. of layers
: and the peak current. We have plugged in several sets of numbers . The
: original plan was to use 450 volts. With this the wire size of choice
: was 16 ga. The coil was to be six windings in parallel with each
What diameter (mm, inches) is this? Alternatively, what is the
weight of the copper?
Copper has a specific heat of 900J/C/Kg, this means that if the
coil weighed 100gramms, and was allowed to heat by 100C, it
could absorb 900*.1*100 = 9000J
..06ohms*5000 implies 300V, peak power = 1.5MW, so, more than 6ms, and
your coil is toast.
: winding consisting of 150 turns per winding with each winding
: consisting of three layers.(50 turns down then up then back down.)The
: coil form was to be 2 inches O.D. This yielded approx. .06 ohms as I
: remember. This yielded
: over 65000 gauss at 5000 amps. as I remember.(Please you guys, don't
: calculate this and tell me I am wrong as this data is from memory)
Impressive.
: This sounds like a lot of current but the pulse is a very short
: duration. The reason for such seemingly small wire is to get the most
: ampere turns/inch. Amp. turns by themselves are not important.
: With the higher voltage I will go with much larger wire as the current
: will be increased a bunch and I will shorten the overall length of the
: coil but still employ the multi layer multi winding approach. I will
: probably use 14 or 12 ga.
Can anyone point me to a site listing wire gauge diameters, I could
work out what 16ga was from the above details, way too much calculus tho :)
: I am not sure how long the pulse has to be there but I suspect that
: 100's of microseconds is sufficient. I know it is with ceramic material.
: I hope this helps,
: Phil Cline.
--
Ian Stirling. Designing a linux PDA, see http://www.mauve.demon.co.uk/
-----******* If replying by email, check notices in header *******-------
"I am the Emperor, and I want dumplings." Austrian Emperor, Ferdinand I.
Phillip R. Cline
John Woodgate wrote:
>
> In article <5nha9d$imd$1...@trog.dra.hmg.gb>, Walter Gray
> <wag...@dra.hmg.gb> writes
>
> >Anyway, how many Clines do we have on this job?
> >
> Well, we have the originator, Philip R Cline, and then there's Win
> Hill, who must have a syncline and an anticline, so I'm inclined to say
> it's three.
>
> Sorry, it was completely irresistible!
> --
> Regards, John Woodgate Tel. +44 (0)1268 747839
> Fax +44 (0)1268 777124. OOO - Own Opinions Only
> Alternative e-mail address: jm...@thenet.co.uk
> That means I get double spam with everything (;-(
We have two. Myself (Phillip) and my brother Mike.
Thanks,
Phillip Cline.
Pin 2 Hot
Ian Stirling wrote:
> Of course 1N4007's are only $.05 each, assuming they will take 12A
> surge, you would only need 2000 of them, for $100
> Course, you would be all week soldering them :)
Solder??? Naah.
Just get about 40 feet (About 2000) of the 4007's on a spool meant
for auto-insertion. You know the stuff, with tape on the edges to
hold them all together. Now, cut about 82 pieces of 1/4" X 1" X 12"
copper stock, and drill 8) 9/32" holes in each one, evenly spaced.
get about 20' of 1/4-20 threaded rod, and cut into 16) 16" long pieces.
Then, cut the "roll-o-diodes" into 12" ladders.
To assemble, start by putting a 1/4-20 nut on each length of threaded
rod on one end. Put all the rods through the holes in two copper
bars. Lay the pair of copper strips on the bench, parallel, about 1"
apart with the rods sticking up. Put the first diode ladder on top of
them, and bend any leads out of the way of the rods. Next, two more
copper strips, and then another diode ladder (watch the polarity).
Keep going until you're about a foot high, and you're out of diodes
and bars. Put some more nuts on the ends to hold this sandwich
together, and torque down evenly and securely.
Now you have a 12" X 12" X 3" diode block that will handle a whole
bunch of amps, has a lot of thermal mass, and is fairly easy to
dissasemble. And you never even plugged in your soldering iron.
Of course, copper is pricey, and this beast is going to be heavy.
Posted "tongue-in-cheek",
--
Gene
When replying by e-mail, please remove the asterisks from my e-mail
address, they are there in an attempt to thwart the automailers.
Winfield Hill
> the coil size, wire size and number of turns, length and no. of layers
> original plan was to use 450 volts. With this the wire size of choice
> was 16 ga. The coil was to be six windings in parallel with each
> consisting of three layers.(50 turns down then up then back down.)The
> coil form was to be 2 inches O.D. This yielded approx. .06 ohms as I
> remember. This yielded over 65000 gauss at 5000 amps. as I remember.
> (Please you guys, don't calculate this and tell me I am wrong as this
Phil, can you tell us about the capacitor(s) you've used to date, and
anything you know about their esr, etc. Also, are you measuring the
5000A current?
-- Win
\
FRID...@psfc.mit.edu
->Mike & Peggy Cline wrote:
->>
->> There's one thing I was curious about. After the ignitron fires and
->> current pulse falls to 0, Why doesn't the ignitron stop conducting and
->> totally block the reverse current? We've tried it on a smaller scale
->> at only 400v with both an SCR and an ignitron and both device continue
->> to conduct the the reverse current and charge the caps in reverse :( .
->> Mike
Ignitron will conduct current in any direction, as long as it's conductive
conditions met. Ignitron conducts by igniting plasma in mercury vapor. The
plasma is ignited by firing high current pulse through ignitor. It takes time
(few milliseconds) for that plasma to dissipate when current drops to zero.
If current crosses zero point and get up high enough to keep plasma on,
ignitron will conduct in reverse. It will also conduct in reverse if you reaply
(or did not remove) ignitor pulse. It will also conduct in reverse if you
using holding anode. I word of caution, though - not all ignitron are designed
for reversed current. If you reverse current on ignitron not designed for it,
you'll significantly reduce it's lifetime, due to rapid electrode pitting.
Mike.
Adrian Jansen
In article <33A16D...@iquest.net>, "Phillip R. Cline" (pcl...@iquest.net) writes:
... snip ...
I guess the spin-flips in the electron state happen in pico seconds
or less to change the magnetisation.
The pulse length is probably only inportant to overcome the eddy
currents set up in a conductive material by the applied field.
With ceramics, the conductivity is essentially zero - 10E10 ohm cm
or so, so there are no eddy currents. I dont know what the
conductivity of your rare-earth magnets is, but it may be
significant for microsecond pulses.
> I am not sure how long the pulse has to be there but I suspect that
>100's of microseconds is sufficient. I know it is with ceramic material.
>
> I hope this helps,
> Phil Cline.
>
--
Fred E. Davis
I was having a fascinating conversation with a retired gentleman just the
other day (who builds MRI systems in his garage!), when the conversation
turned to this thread, charging magnets and the difficultly of suppressing
the reverse charge to the caps. It turns out that he used to charge his own
magnets using a similar, capacitive discharge technique, ignitrons and all.
The method of suppression he used consisted of a fuse in series between the
capacitors and coil, and a lamp (large, no doubt) in parallel with the coil
as a damping resistor. By using the fuse manufacturer's data for current
overload vs trip time, he was able to select a fuse that would sustain the
peak current, then open just before the reverse charging would occur. It
would obviously require fuse replacement with each charge, but the cost
shouldn't be prohibitively expensive. BTW, he said this method was
documented in an article back in the 1920's. I guess that one slipped by
me! (:D
Ian Stirling
Ted Park <tp...@world.std.com> wrote:
: In article <5n8umo$k...@fridge-nf0.shore.net>,
: Winfield Hill <hi...@rowland.org> wrote:
: >
: >One thought is to maximize the field, subject only to the constraint
: >of keeping the copper temperature below white hot. As the current
: I'm curious now about what kind of wire you would use. I've only made
: small pulsed electromagnets but they haven't been nearly strong enough
: to magnetize anything with greater coercivity than a screwdriver.
: How long does the pulse have to last to magnetize materials such as
: NdFeB? I've seen figures like 12500 gauss for some materials, but
: that kind of seems like a lot of current going through a lot of turns.
: I guess that a few turns of wire wrapped around a pop tin won't cut it.
On the subject of wire wrapped round a pop tin, check out
http://hibp.ecse.rpi.edu/~chowm/can_crusher.html
For some Big fields.
: --Ted
: --
: ---------------------------------------------
: Ted Park tp...@world.std.com
: also <A HREF="mailto:tp...@canuck.com"> tp...@canuck.com </A>
: On the web as <A HREF="http://www.beer.org/~tpark/"> Ted's Home Page </A>
--
Ian Stirling. Designing a linux PDA, see http://www.mauve.demon.co.uk/
-----******* If replying by email, check notices in header *******-------
The fight between good and evil, an epic battle. Darth vader and Luke,
suddenly in the middle of the fight, Darth pulls Luke to him, and whispers
"I know what you'r getting for christmas!" Luke exclaims "But how ??!?"
"It's true Luke, I know what you'r getting for christmas" Luke tries to ignore
this, but wrenches himself free, yelling "How could you know this?",
Vader replies "I felt your presents"
Ian Stirling
Fred E. Davis <FED...@worldnet.att.net> wrote:
: I was having a fascinating conversation with a retired gentleman just the
: other day (who builds MRI systems in his garage!), when the conversation
: turned to this thread, charging magnets and the difficultly of suppressing
: the reverse charge to the caps. It turns out that he used to charge his own
I was wondering about the possibility of using vapour phase silicon
switches for this.
(to92 SCR, turned on at the correct time.)
If I was wanting to build one of these at home, what do I use to complete
the field paths?
some sort of magnetically soft iron?
What form? Would a simple can round the solenoid, with two pole pieces
in the axis, leaving a hole in the middle for the magnet.
: shouldn't be prohibitively expensive. BTW, he said this method was
: documented in an article back in the 1920's. I guess that one slipped by
: me! (:D
--
Ian Stirling. Designing a linux PDA, see http://www.mauve.demon.co.uk/
-----******* If replying by email, check notices in header *******-------
Get off a shot FAST, this upsets him long enough to let you make your
second shot perfect. Robert A Heinlein.
FRID...@psfc.mit.edu
In a previous article, Ian Stirling <000033A9B...@mauve.demon.co.uk> wrote:
->Fred E. Davis <FED...@worldnet.att.net> wrote:
->: I was having a fascinating conversation with a retired gentleman just the
->: other day (who builds MRI systems in his garage!), when the conversation
->: turned to this thread, charging magnets and the difficultly of suppressing
->: the reverse charge to the caps. It turns out that he used to charge his own
->
->I was wondering about the possibility of using vapour phase silicon
->switches for this.
->(to92 SCR, turned on at the correct time.)
The way I see it, there 3 ways you can handle it:
1. By breaking the current with mechanical switch or circuit breaker and,
preferably inserting dumping resistor in series to avoid very high voltage
spikes
2. By injecting current of opposite polarity into the circuit
(also known as forsed inversion). Can be done with cap bank and ignitron/SCRs
3. By shorting out the coil ( also known as crowbar). Also can be done with
ignitron or SCRs.
No.3 is the easiest way to do it, in my opinion. And ignitrons are cheaper
for HV then SCRs
Mike.
N1FBW
If the caps discharge through the SCR via a saturable reactor (switch) the
SCR will shut off if a slight mismatch is designed in to the discharge
path cicuitry. This will drive the discharge voltage slightly negative
(inverse voltage) and help to turn off the SCR. Otherwise, latchup of the
SCR will occur and you'll hang up your cap bank charging supply. Look into
publications of semiconductor/magnetic switching modulators from M.I.T.
and other institutions for more info.