ISSTC Discussions
Edited/Updated: October 23, 2004
This page needs editing - John C
1) 5-27-4. I think it should be OK to consider the ISSTC as a "Quasi-CW" system. The burst length seems long enough for it to reach the steady state (where power out of inverter=power into streamers). Steve C.
2)
5-25-4. Original poster: "Steve Conner" OK I messed up... I
have been corresponding with Richie and I am forced to admit the "magnetizing
current" model is a good one and the "leakage inductance" one is a red herring.
The model is described on his website. We are now working on extending the
model to DR/ISSTC's, and trying to factor in a quantitative way of dealing with
corona loading. We have already figured out that the tuned primary works its
magic by acting as an impedance matching device and can be modelled as an
L-match.
Now here is a question for the math geniuses... Can we take the corona load
impedance, as seen at the top of the resonator, and transform it to an
equivalent impedance at the base, using the well-known quarter wave stub
formula: Zbase=(Zo^2)/Ztop
Where Zo is the characteristic impedance of the "transmission line", which I
suppose would be sqrt(L/C), L being the secondary inductance and C=(Ctoroid+Cself+Ccorona)?
>The odd experiment I've done in terminating a resonator with a fixed resistance
suggests the formula holds, >at least loosely. That was some time ago when it
was suggested as an explanation for the behaviour of a >large CW coil in terms
of blowing fuses and MOSFETs vs being able to feed a useful amount of power in.
>Some time around '96 as I recall.
If so it all becomes rather nice and we can model the ISSTC after breakout as an
"antenna" with "radiation resistance", being base-fed from a "transmitter"
through a L-match "antenna tuner" :))). There is the complication that the
corona load varies with power as the streamers get longer and hotter, but we'll
burn that bridge when we come to it :)))).
>That is the real problem and gets much worse when the streamers connect with a
target. Malcolm.
3)
5-26-4.
OTOH, the new DR and ISSTC's are capable of going "nuclear" when
shorted out by an arc to ground. The low impedance at the resonator top
transforms to a near infinite impedance at the base, and the tuned primary sees
no load, allowing the primary current to ring up without limit.
>
> We get round this one by running them in pulsed mode, i.e. the inverter operates
for say a 200uS burst every 10ms. Even if the secondary was totally
What if we used a microcontrol with something like the following algorithm:
1) [power up] - At ludicrously low power, seek to the resonance of the
system.
2) Run a short burst of full power and using lead/lag, correct Fo for the
streamer loading.
3) Repeat step 2 a couple of times.
4) Let her rip at full power, tuned very close, but not attempting to track
dynamically.
5) Drop into a safe-area limiting control loop- drop the duty cycle as
needed while monitoring current.
I'm kinda thinking about using this approach as a $2.50 PIC can really
baby-sit $50 MOS/IGBT/FET's very easily. But I haven't built an SSTC yet...
Any merit in this approach? Thoughts? Comments? Anyone?
5-27-4. It's definitely
the right kind of way to go. But I think it could be even simpler than what you
describe. After all the ISSTC is self resonant so you don't need to bother with
tuning. The DRSSTC is not, but the tuning seems broad enough to set and forget.
So what I'm thinking is a current limiter hooked up to the interrupter. It
measures the peak primary current, and terminates the burst if it gets above a
safe...ish :) value. If the burst is terminated by the current limit, rather
than the interrupter timer, it should also light some kind of warning light to
make you aware of the fact.
There would have to be some gating to make sure that the burst was only
terminated at a natural switching instant, IOW, the enable signal to the gate
drivers should go through a D-type flip-flop that is clocked by the
gate drive signal. Otherwise it can try to turn off at a peak of primary
current, which is not good.
The current limit would be easily done with a Rogowski coil and a window
detector made from a couple of comparator chips. I try to avoid using a
microcontroller in any "mission critical" part of a SSTC where it could blow out
your pricey IGBT's if it glitched or crashed. Steve C.
4)
6-1-4. >Not this power vs. energy crap again . . . ;).
It's true, from what I have seen, the new ISSTC type coils
can produce a longer arc for a given power input than a SGTC. I think there are
a few reasons for this-
1. The long burst length of the ISSTC (much longer than a
SGTC) is conducive to growing long, straight streamers that don't fork and
wiggle too much. They seem to like that long steady burst of RF better
than the sharp "kick in the butt" they get from a SGTC or OLTC. I believe this
because I have seen the performance of an OLTC is somewhat less than a ISSTC
with the same BPS and bang energy.
2. The ISSTC's work at a lower BPS than a SGTC (Dan and Steve
run their coils at 100 BPS) and it is well known that lower BPS gives longer
sparks for a given power.
3. The solid-state switches have lower losses than a spark
gap (I estimate they're 90% efficient)
***
BTW - I think John Couture just muddies the waters with his
talk of power vs. energy, and controlled sparks. TC spark length is indeed a
random phenomenon, but that doesn't make it any harder to measure as long as you
are careful to be consistent with runtimes etc.
The spark length plotted against frequency of occurrence probably follows a
Poisson distribution or such like, and if you were really pedantic, you could
take measurements of how often a target at various distances is hit,
then fit a Poisson distribution to it, and quote your TC's spark length as
"There is a 95% chance of striking a grounded target 66 inches away during a one
minute run". Steve C.
5)
6-1-4. Thanks for the info regarding your Tesla coil tests. I believe you
have done a great job in advancing the design of Tesla coils especially the
SSTC. You not only have advanced the TC technology you have also the ability to
build well crafted devices.
The best way to make comparisons is to test real coils. Simulations can account
for only a few of the many variables involved. The three most important
parameters in the tests are the TC Power, TC energy, and the TC Overall
Efficiency. Note that in an AC circuit the power is normally called the average
power. The RMS power has no significance. Refer to an EE book or Physics book.
The RMS voltage and current are based on energy conditions.
Power has to do with instantaneous conditions. Because of the complex circuitry
and unique operation of Tesla coils it is difficult to properly test for
instantaneous conditions. The proper method of comparing Tesla
coils is to compare the overall energy efficiency of the coils. John
Couture.
6)
6-4-4. Driving a half bridge TC system with POWERMOSFETs I would
like to make the necessary dead times as short as possible.
We have found that with MOSFETs no dead time at all is needed. If you drive both
your MOSFETs from a single gate drive transformer with bipolar drive (i.e.
swinging from +12V to -12V) it works fine. The reason is that MOSFET's turn off
instantly when the gate voltage goes below the threshold, they don't hang around
like BJTs or IGBT's.
If you are using a more complex gate drive circuit then you may need to add
delays here and there. Steve C.
7)
6-4-4. ISSTC dead time: A good way to test the dead time is to run
the bridge with a resistive load, and put a resistor in series with the supply
to the bridge. Use a relatively low voltage (~12V) supply for these tests.
Use a scope and measure the voltage across the supply resistor. If there are
spaces where the voltage drops, but the drive signal is still high (or low,
whichever side of the half bridge you're looking at) the dead time is too long
(current flow stops).
If, on the other hand, the voltage across the resistor spikes, there is shoot
through, and the dead time needs to be increased.
This is a pretty good test, but because of the exponential decay behavior of the
turn on/off times, it will be hard to get perfect waveforms. As well, the
waveforms will change with applied input voltage, as the effective gate
capacitance will change, and the gate waveform will change. Sean Taylor
8) 6-4-4. I think the way to longer sparks is to increase the step-up ratio of your primary circuit. This is done by reducing the primary L, and increasing the C to keep it in tune. Increasing the coupling also helps in theory, but I think it's a dead end, as it encourages flashovers. In the long run you may need to LOOSEN the coupling, to get the primary out the way and stop it getting incinerated. More voltage helps too- going to 240V should double your spark length. And using heavier wire for your primary will be a must, and a bank of electrolytic caps on your H-bridge DC supply. You'll probably need to shorten the on-time too as you get beyond the IGBT ratings :)))) Steve C.
9) You need to tune the ISSTC just as you would a spark gap coil.
10) 6-5-4. I've been doing a lot of work on the theory of operation of ISSTC's... if not the practice :'( I now think that the ISSTC can be simply modelled, if we assume it perfectly tuned, as two quarter-wave matching transformers in series. (an untuned primary SSTC is just a single transformer.) I'm pretty sure that the resonator is a quarter-wave transformer, with characteristic impedance Zo equal to sqrt(L/C) or alternatively 2*pi*fres*L. And along similar lines I can argue that the primary circuit is another transformer of the same kind, with Zo calculated in the same way, although I don't think it does impedance inversion as the resonator does. Radio guys use two quarter wave transformers of different Zo in series to match two widely different impedances, and maybe this can explain why the ISSTC works so well- the two stage approach does a better job of matching a high streamer impedance to the <1 ohm output impedance of a high power inverter. Anyway there are equations for designing these transformers in the ham radio handbooks, and I'm going to try applying them to an ISSTC design. Steve C.
Very interesting. I myself have a lot of experience with designing microstrip versions of 1/4 wave transformers and similar. Most of it involving matching 50 ohm impedance lines with large numbers of planar antennas which are combined in parallel combinations which may result in impedances of 10 ohms, etc... so the quarter wave transformer is a must! I'll have to look at the some more. Dan.
6-7-4. Sorry, I think this needs some
more explanation. Here is my theory...
In a ISSTC, the secondary is a 1/4 wave transformer and the primary is an
L-match. The fact that they are inductively, rather than directly coupled
complicates things, but not much. (The inductive coupling can be modelled as an
ideal step-up transformer, as Antonio showed in his diagram, and Richie Burnett
explains on his site)
Anyhow, a 1/4 wave transformer is equivalent to a L-match, at its resonant
frequency. So if we are modelling the ISSTC at resonance (and since it's a
feedback system, it will ALWAYS run at resonance) then we can picture it as two
L-matches, OR two 1/4 wave transformers, with the inverter at one end (stepped
up by that ideal transformer) and the streamer load at the other.
If you use PSpice you will end up modelling it as L-matches with lumped
components, since PSpice didn't have a transmission line resonator model, last
time I looked. Steve C.
I disagree with a lot of this discussion, at least insofar as it calls a top-loaded secondary a "quarter-wave transformer". No way! The typical top-loaded secondary has an almost constant current distribution and therefore must be considered as an inductor, not a transmission line. The only transmission line which has a constant current (or voltage) distribution is one terminated in its characteristic impedance and is thus equivalent to a transformer with a ratio of 1. Ed.
11)
If it's assumed that the system is feeding a constant resistance, representing
the streamer loading, and that the other parts are lossless, it seems to be
possible to model the coil as a lumped band pass filter. I am imagining the
structure:
k12
. o---C1---+ +----+----+--o output
. | | | |
. Vin L1 L2 C2 Rload
. | | | |
. o--------+ +----+----+--o ground
C2 is the distributed capacitance of the resonator, as in a spark-gap coil.
This structure can be designed exactly to have any bandwidth, and any voltage
gain or any input resistance, by using the same procedures used in the classical
design of passive filters. It's also possible to make approximate designs
by considering the network as a series of two L-match impedance matching
networks, one C-L and the other L-C. The transformer can be treated by a
simple transformation, that converts it into an ideal transformer leaving two
inductors in the right positions for the matching network, or the filter.
The circuit below would be designed first, and then converted to the form above.
1:n
. o--+ +---C1'--+---L2'--+----+--o output
. ) ( | | |
. Vin ) ( L1' C2 Rload
. ) ( | | |
. o--+ +--------+--------+----+--o ground
The calculations are too complicated to explain here, but I can try to make a
design and put it in my web pages.
Antonio Carlos M. de Queiroz
12) 6-7-4. Original
poster: "Antonio Carlos M. de Queiroz"
> Original poster: "Steve Conner"
>
> >This structure can be designed exactly to have any bandwidth, and any voltage gain or any input resistance, by using the same procedures used in the classical design of passive filters.
>
> OK... cool... I agree totally. But what I am trying to find out is the magic values of bandwidth, voltage gain, and input resistance, that will give biggest streamer output from a given set of IGBTs, with reasonable tolerance to detuning by streamer load.
The input resistance is determined by the amount of power that you have
available, and your switching devices. The voltage gain, is as mysterious as in
a conventional coil. I guess that anything that produces more
than about 80 kV will work. The bandwidth affects mainly the coupling
coefficient between the coils, and the sensitivity to mistuning. The big
problem is to guarantee good matching for any load, what seems impossible.
Another problem is that the load is not purely resistive, but has a small
capacitance in series with it. No big problem, as this can be converted to a
parallel equivalent, with the capacitance absorbed by the terminal+secondary
coil capacitance.
> >It's also possible to make approximate designs by considering the network as a series of two L-match impedance matching networks, one C-L and the other L-C.
>
> I like this approach. I think that a lot of the design choices we _could_ make in the full bandpass filter network approach are constrained by other things. For instance the coupling will be limited due to clearances for primary-secondary flashover, and the voltage gain achievable in the primary will be limited (I imagine) by such things as flashovers between primary turns.
Surely. Most of the constraints are determined by proper insulation.
> With the L-match approach, the problem is reduced to matching the real part of the streamer load (several hundred kOhm) to the inverter output impedance (around 1 ohm) The inverter doesn't have an output impedance as such, but you can express the maximum current you want to draw as an impedance.
Yes.
> (fundamental of inverter output voltage /(peak IGBT current you want/sqrt(2))
Maybe: R=(4/Pi)*peak inverter voltage/peak current. This takes the fundamental
of a square wave voltage waveform and the current as peak values.
> I say fundamental because the inverter output voltage is a square wave but our analysis will assume a sine wave. The current is a sine wave anyway.
What I assume above.
> So my plan is to design the secondary according to good HV practice, choose the highest coupling I can without any risk of flashovers, then choose the primary L and C by treating it as an L-match. What do you think?
I was deriving a set of exact design equations. Not so complicated at the
end, using the filter approach.
Start with the specifications:
R = input resistance, Ohms
w0 = 2*pi*operating frequency in Hz, rad/s
B = 3 dB Bandwidth in rads/s (2*pi*bandwidth in Hz)
n = Voltage gain
Design a bandpass Butterworth (maximally flat) filter:
. o--C1---L1--+----+----+---o
. | | |
. +-> L2 C2 R2
. | | | |
. R1 o-----------+----+----+---o
R1=R2=R
C1=B/(w0^2*sqrt(2)*R)
L1=sqrt(2)*R/B
L2=B*R/(w0^2*sqrt(2))
C2=sqrt(2)/(B*R)
Convert it to the final structure:
kab
. o----Ca---+ +----+----+---o
. | | | |
. +-> La Lb Cb Rb
. | | | | |
. Ra o---------+ + ---+----+---o
Ca=C1
La=L1+L2
Lb=n^2*L2
Cb=C2/n^2
Rb=R*n^2
kab=sqrt(L2/(L1+L2))
Note that the relation La*Ca=Lb*Cb holds, as in a capacitor discharge coil.
Ex:
With 180 V of input voltage (a full bridge from 127 V rectified and filtered)
and peak current of 100 A:
R=2.29 Ohms
With n=500, operating frequency of 300 kHz, and a bandwidth of 50 kHz, the final
elements result as:
Ra: 2.29 Ohms
Ca: 27.3 nF
La: 10.5 uH
Lb: 35.8 mH
kab: 0.117
Cb: 7.86 pF
Rb: 573 kOhms
Not very different from the usual for a capacitor discharge coil. This coil
presents a constant resistance to the driver of 2.29 Ohms at 300 kHz, with
little deviation for errors up to 25 kHz to each side. It remains to be seen
what is the effect of different loads, and what happens before breakout, and if
something can be done to improve the characteristics in these cases.
Antonio Carlos M. de Queiroz.
13) 6-8-4.
Original poster: "Steve C.
Hi Antonio, thanks very much for your input.
>The big problem is to guarantee good matching for any load, what seems
impossible.
At first sight it does, but I think it's not really that bad. If the streamer
load is not heavy enough, the output voltage (and primary current) will just
ring up, lengthening and heating the streamers until they present the right
load. (as streamers get longer/hotter their impedance decreases, but this
impedance is inverted by the "matching network" so the impedance seen by the
inverter INCREASES)
So, if you design to a streamer length that is appropriate to the power/bang
energy you are running, I think the ringup will level off at your target current
and it will all work out.
You might argue that the system could get "trapped" in a state where the current
rings up but the streamers for some reason don't grow to limit it. But I
think this will only happen if you try to produce a streamer longer than your
inverter can handle- your IGBTs will explode before the streamer reaches target
length.
Anyway this sounds like serious non-linear math, so time for me to throw away
the calculator and get soldering ;)
>No big problem, as this can be converted to a parallel equivalent
Agreed, the only problem is measuring the streamer impedance in the first place
:( Again probably time for some experiments now...
>I was deriving a set of exact design equations. Not so complicated at the
end, using the filter approach.
excellent! I think I understand, in your equation, the voltage gain (n) is the
term we use to get the desired impedance match? And you just chose the bandwidth
arbitrarily? If you choose a much wider or much narrower bandwidth, what does
this do to the resulting component values? I suppose it would affect the
coupling and the tunings of both "matching networks".
>Not very different from the usual for a capacitor discharge coil. Spooky :))))
In fact I have such a coil (the Tesla-2) with similar parameters to the ones you
posted- but roughly twice the primary inductance and half the tank capacitance.
However I run my stuff on 377V DC here (240v AC mains) so if I understood you
right, the extra voltage would compensate for the higher primary impedance, and
the peak primary current would still be 100A.
(100/sqrt(2))A * (4/pi)*(377/2)V= 16.9kW flowing for 200uS= 3.3 Joules bang
energy, which is more than it had as a spark-gap coil 8-o
I'll try to run this coil as an ISSTC and get some measurements. Thanks to your
equation, I can easily calculate what the streamer load impedance actually is,
by measuring the primary current. Steve C.
14) 6-7-4. Yes I believe tune is the key point (at least one of many!). Last night I did some testing and now I get reliable get 22" sparks to an ungrounded Ring Stand on a shelf. But I wondered just how far anyone has pushed a Half-Bridge? So far I am vary happy with my results, I mean I am getting 22" sparks at ~0.8KVA! Try that with a vacuum tube!? Another note I would like to say; while I am using a proto-board and not providing any RF shielding to it (only 4-5" away from primary on a table, same plane) it must be earth grounded. I tried "floating" the ground and everything was great until I upgraded to my 6" secondary (due to greater RF interference), then it became very unstable. Once I grounded the negative, things are vary stable now. But I must note on just how hot the primary can get. When the sparks are flying, I must watch the primary tank, as it can get very hot. So far I have kept run times to less than one minute without breaks. BUT, the IGBT's stay rock cold with a fan running over the heat sink! I think that I could upgrade my primary to copper tubing and increase the current capacity of the tank cap, and then this sucker could run for extended runs. The museum TC folks may want to look into this. But again what have people done spark wise with a half-bridge? Regards, David T.
15) 6-10-4. Original poster: "Antonio
Carlos M. de Queiroz"
>Original poster: "Steve C. Hi Antonio, thanks very much for your input. The big problem is to guarantee good matching for any load, what seems
impossible. At first sight it does, but I think it's not really that bad. If the streamer
load is not heavy enough, the output voltage (and primary current) will just
ring up, lengthening and heating the streamers until they present the right
load. (as streamers get longer/hotter their impedance decreases, but this
impedance is inverted by the "matching network" so the impedance seen by the
inverter INCREASES)
Yes, I was also thinking that the system would "auto tune" to the right load,
the one considered in the design, but see below...
> So, if you design to a streamer length that is appropriate to the power/bang
energy you are running, I think the ring-up will level off at your target current
and it will all work out.
See below.
> You might argue that the system could get "trapped" in a state where the
current rings up but the streamers for some reason don't grow to limit it.
But I think this will only happen if you try to produce a streamer longer
than your inverter can handle- your IGBTs will explode before the streamer
reaches target length.
I was verifying what happens with my design when the output load changes.
The result is curious, and maybe very significant: If only the load
resistance changes, the input impedance -remains resistive-, and changes to
match the load. This happens because of the doubly tuned transformer, that
operates independently of the load. I verified then what happens if the
capacitive loading changes, due to streamer growth. The result is not very good:
The input current -increases- and becomes reactive, out of phase with the
voltage, for changes in both directions of the capacitive loading.
> Anyway this sounds like serious non-linear math, so time for me to throw away
the calculator and get soldering ;)
Some experiments with measurements would be interesting to see how real this
linear modeling is. I imagine that it is realistic. I have the materials to make
an experiment, but this will have to wait some time.
> excellent! I think I understand, in your equation, the voltage gain (n) is the
term we use to get the desired impedance match? And you just chose the bandwidth
arbitrarily? If you choose a much wider or much narrower bandwidth, what does
this do to the resulting component values? I suppose it would affect the
coupling and the tunings of both "matching networks".
Yes, the transformer gain, n, is what makes the impedance conversion. The
reactances are just to compensate each other and reduce the circuit to an ideal
transformer. The controllable bandwidth (at design time) is an additional
feature over a basic design as a cascade of L-matches. Its main effect
appears to be in the coupling between the coils.
> >Not very different from the usual for a capacitor discharge coil. Spooky
:)))) In fact I have such a coil (the Tesla-2) with similar parameters to the
ones you posted- but roughly twice the primary inductance and half the tank
capacitance.
All indicates that a conventional coil assembled as a solid-state coil will work
well.
> However I run my stuff on 377V DC here (240v AC mains) so if I understood you
right, the extra voltage would compensate for the higher primary impedance, and
the peak primary current would still be 100A.
Yes.
> (100/sqrt(2))A * (4/pi)*(377/2)V= 16.9kW flowing for 200uS= 3.3 Joules bang
energy, which is more than it had as a spark-gap coil 8-o
I get:
Rms current = 100/sqrt(2) = 70.7 A
Rms voltage = 377/2*4/pi/sqrt(2) = 169.7 V
Average power = 70.7 x 169.7 = 12 kW
In 200 us: 2.4 J
> I'll try to run this coil as an ISSTC and get some measurements. Thanks to
your equation, I can easily calculate what the streamer load impedance actually
is, by measuring the primary current.
It will be interesting to see what happens. I have implemented the design
equations in a little program. I will see if I add a simulator on it too.
Antonio Carlos M. de Queiroz.
Eastern Voltage, Full Bridge Power Section - 5-14-04
-
C1, C2, 2 pcs - 4.7uF, 500V, GE 41L switching cap. High performance metalized polypropylene, axial, GE #41L6471 is 4.7uF, 600V
-
CR1, 2, 3 & CR4, 4 pcs - 1N5819, Schottky barrier rect., DO-41 package - Fairchild 1N5819
-
Q1 - Q4, 4 pcs, IGBT, Fairchild HGT1N40N60A4D
-
R1, R2, R3 & R4, 4 pcs - 5.1 Ohm
-
R5, R6, 2 pcs, 1meg, 3W
-
T1, T2, 2 pcs, Gate Transformers, Toroid Cores, www.Fair-Rite.com 5978006401, paralleled primary windings
-
VR13 - VR16, 4 pcs, TVS, 15KP220CA Microsemi also by ProTek Devices & Littelfuse
-
VR17 - VR20, 4 pcs, TVS, 1.5KE200CA Microsemi also by Littelfuse
-
VR3, VR6, VR9, VR12, 4 pcs, TVS, 1.5KE33CA Microsemi also by Littelfuse
-
VR1, 2, 4, 5, 7, 8, 10 & VR11, 8 pcs - 1N4752, Zener 33V, 1W, Fairchild 1N4752A
Eastern Voltage Research, Base Current Feedback Self-Resonant Driver SC2045
-
C1, C2, C8, C9, C10, C11, C12, C13, C14; 9 pcs, 0.10uF
-
C3, 1 pc, 2.2Uf
-
C4 - C7, 4 pcs, 10uF
-
CR1, CR2, 2 pcs, 1N90, hi V diode, modern equivalent = Fairchild 1N4454 - hi conductance, ultra fast diode. 1N60 Germaniums can be substituted.
-
CR3, CR4, 2 pcs, 1N4002, general purpose rectifier, 100V, 1A, Fairchild 1N4002GP
-
CR5 - CR8, 4 pcs, 1N5819, Fairchild 1N5819 1A, 40V max. repetitive reverse voltage. See #2, above.
-
R1, R3, R4, 3 pc, 10k
-
R2, 1 pc, 2.5k
-
T1, 1 pc, 1T 8mH:60T 30mH
-
U1:A - U1:F, 6 pcs, SN74HS14N ?Hex Schmidt trigger? Tri-state octal buffer
-
U2, 1 pc, 555
-
U3 - U4, 2 pcs, TI UCC27322P, plastic dip 8, 9A, MOSFET driver
-
U5, U6, 2 pcs, TI UCC27321P, p dip 8, 9A, Power MOSFET driver with enable
Eastern Voltage, Low Voltage PS
-
C51, C52, 2 pcs, 4700uF, electrolytic?
-
C53, C58, 2 pcs, 0.10uF
-
C54, C59, 2 pcs, 10uF
-
CR51, CR52, CR53, CR54, 4 pcs, 1N4002, standard rect., see #5 above, FSC 1N4002GP.
-
D51, D52, 2 pcs, LED's +5V power
-
R51, 470
-
R52, 1.5k
-
T101, 1 pc, 115Vac prim., 24V CT, 2A secondary
-
U51, 1 pc, LM7805, 5V, + V reg., Fairchild, Nat. Semi LM340K-5.0 or their LM7805CK (obs), TO-3 pkg.
-
U52, 1 pc, LM7815, 15V, Nat. Semi LM340K-15 or their LM7815CK (obs) TO-3.
-
VR51, VR52, 2 pcs, TVS, 1.5KE15CA Microsemi also by Littelfuse
They are quite different in the way they work. The OLTC is a development of the spark-gap coil whereas the ISSTC is oscillator-driven like an SSTC.
What does this mean? Well if we use the "shattering a wine glass" analogy for resonance, the ISSTC sings very loudly at it, but the OLTC just hits it with a big hammer ::) Steve C.
5) June 20, 2004. Original poster: "Steve Conner">If you think I jest, consider Greg Leyh's heroic exploits in Z transform >pulse power at SLAC in trying to replace 30yr + vintage thyratrons with IGBT's.
Us high power SSTCers have been watching Greg Leyh's exploits with great interest. I remember reading Greg's papers and thinking "wow, if IGBT bricks can stand that kind of abuse, they should breeze along in an OLTC".
So it was partly Greg's work that encouraged me to build the OLTC II. And this turned out to be correct, the IGBTs can happily stand the high currents, as long as they don't see excessive di/dt.
The resonant primary SSTC also does a great job of controlling di/dt because of its zero current switching, and that is why Steve, Jimmy H. et al. get such great results, and don't blow any IGBTs (well not many). Steve C. http://www.scopeboy.com/
4) You mention that at one point you had to use a breakout point on your (smooth) 8" x 24" toroid. That interests me because, (when my SSTC was working), I didn't need one.
Original poster: Steve C.
Yes, I did use a breakout point. The OLTC is like a classic spark-gap coil, it
starts with a fixed amount of bang energy stored in its tank capacitor, and if
that isn't enough to cause breakout, then tough luck :( However an
SSTC will just keep pumping in energy until it either breaks out, flashes over,
or explodes.
Later, I was able to get it to break out without a breakout point, but I still
used one, because it gave me more tuning options. To get breakout, I had to tune
for maximum output voltage in the absence of streamers, and this is not the same
tuning as for maximum streamer length. (Because of detuning by the streamer
capacitance.) I tried various breakout points, from a tiny bump to a large
spike, and the spark length didn't change.
Again with feedback SSTC's you don't have to worry about all that stuff. The
inverter tracks any frequency changes due to streamer loading and just keeps the
power coming.
*** an aside
Personally I quite like the fixed bang energy of the OLTC. It makes the system a
bit tamer and easier to build/get working.
The OLTC does have an electrolytic filter bank that stores a lot of energy, but
it's connected to the IGBT's through a choke and a fuse, unlike in a big pulsed
SSTC where you tend to have several thousand uF of low ESR caps bolted directly
to your IGBT terminals.
With all that stored energy in a low impedance circuit, a failure of a pulsed
SSTC can be pretty spectacular, the transistors tend to blow apart and scatter
shrapnel and gorilla snot all over the place. (safety goggles
are a good idea here).
But then when a pulsed SSTC works, it's even more spectacular 8-9. Steve
C.
3)
Some of you might remember the OLTC II, the large solid-state disruptive
coil that I have been working on for almost a year now. It is basically the same
as a classical coil, but with an IGBT "spark gap replacement".
I just demonstrated it at the UK Teslathon and it worked great, giving strikes
up to 73". This was using a bang energy of ~8J at 400bps. It also produced a
couple of ground strikes (which we sadly never got pictures of)
and even struck its own primary coil, triggering the overvoltage protection
circuit. Luckily the protection circuit did its job, as changing an IGBT is a
half-hour job with a socket wrench, and cheap replacements are getting
a bit harder to find.
The IGBT's are currently running at 2500A peak each. This is twice the maximum
pulse rating but I expect they will take up to 8000A each. So it should be
possible to get even more output from this coil by just using a
bigger toroid and more tank capacitance.
But it will need a new primary circuit with lower resistance, as the one I'm
using just now, made from copper pipe and twelve coax speaker cables in
parallel, runs dangerously hot. The MMC bank, IGBT's, and secondary coil only get
slightly warm.
Here is a great pic taken by James Pawson
http://www.scopeboy.com/tesla/oltc2_6foot1.jpg
More pics of the UK Teslathon on Mike Harrison's web
http://www.electricstuff.co.uk/uktesla2004.html
And you can read technical details of the OLTCII here (look for Tesla Four)
http://www.scopeboy.com/tesla/
2)
First class work, Steve--on the OLTC and also on the Web-site. You
mention that at one point you had to use a breakout point on your (smooth) 8" x
24" toroid. That interests me because, (when my SSTC was working), I didn't
need one. I used a smooth 6" x 24" Landergren toroid on a 12" diameter
secondary. The primary was untuned, driven by a feedback system so as always to
be at the exact frequency of the secondaries Fr. It consisted of 2 or 3
equivalent turns driven by what amounted to a ~1000 V p-p square wave. (I'd
posted a photo of it doing its thing some time back on hot-streamer/temp).
I suppose, with the continuous excitation (lasting well after first
spark-breakout; there was no "notch", of course), the voltage just built up & up
until the breakout occurred. I never did get more than 30" or so of sparks out
of it, and I think that was due to the relatively slow rate of voltage
rise--over some 30 cycles, as I recall. Ken H.
>
> OTOH, the new DR and ISSTC's are capable of going "nuclear" when shorted out by an arc to ground. The low impedance at the resonator top transforms to a near infinite impedance at the base, and the tuned primary sees no load, allowing the primary current to ring up without limit.
>
> We get round this one by running them in pulsed mode, i.e. the inverter operates for say a 200uS burst every 10ms. Even if the secondary was totally shorted, 200uS is not enough time for the current to ring up to _really_ scary levels. But it can still cause a higher than normal current, hence not a good idea to totally short one of these coils out with an arc to a very close by target. (A long arc to ground has more resistance and can dissipate power safely.)
>
> The primary current can also increase without limit if breakout does not happen, but this is less of a problem, it usually just ends in a pri-sec flashover rather than IGBT-Ma-Geddon. After all, the secondary voltage increases without limit in this case too :)))
>
> Anyway, somewhere in between these extremes of "totally short" and "totally open" there is a streamer load impedance that leads to optimum power transfer. Richie and I are working on a design procedure that will let you start by specifying what length of streamers you want, then work back to find the primary L/C/k you need to match your inverter to that particular streamer load.
>
> It will also tell you the peak primary current, hence what size of IGBT's you need. And if that turns out bigger than the ones you've got, you start again with a smaller streamer length :)
When I spoke of difficulties, I was referring to the difficulty of defining a streamer impedance (which as we all know depends on current), rather than ensuring MOSFET survival etc. It may be that for air streamers, there is a reasonably well defined value (Terry and Greg came up with values some time ago). Perhaps the streamer tends to adhere to such values by lengthening, broadening into branches etc., a property of the discharge medium (i.e. air at about 1 atmosphere).
As far as defining the characteristic impedance of the resonator goes, a better option is to consider the topload as a termination rather a distributed part of the line. The Corum's did all this a long time ago and produced Smith Charts to go with their writings. A problem which occurred with their analysis when considering disruptive coils was that their output voltage calculations were applicable only to CW coils and invalid when applied to disruptive discharge operation with the coil being fed with a charged capacitor. Malcolm.
-
CM600HA-24H
-
74HC132
-
mmc 15uF
>Can you suggest a suitable part number or two for OLTC
IGBT's?
For a mini OLTC, IRG4PF50WD made by International Rectifier (get from Digikey)
for a big boy, the Powerex CM600HA-24H IGBT module, 1200V, 600A, 20 - 25kHz. It
costs $300 new but you can get four for $50 on eBay, and some of them might even
be functional, lol, (6-7-4 price for 10pcs thru distribution = 150.00 each).
Powerex also makes the CM600HA-28H, 1400v, 600A, 20 - 25kHz.
>Do you have a schematic posted?
Everything I know about OLTC's is on my website
http://www.scopeboy.com/tesla/ (look for Tesla Three and Four). Steve
C.