[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Magnifier topics



Here are a pair of spice subcircuits to represent Antonio's
3:4:5 system.

L1,L2,C2, modelled in 29 segments:

 http://abelian.org/tssp/acmq345a.cir

 .SUBCKT acmq345a N1 N2 N3 N4

N1: hot end of primary
N2: cold end of primary
N3: hot end of L2/C2
N4: cold end of L2

This subcircuit contains the lumped C2 (using 70pF not 62pF)

L3,C3, modelled in 72 segments

 http://abelian.org/tssp/acmq345b.cir

 .SUBCKT acmq345b N3 N4

N3: hot end of L3/C3
N4: cold end of L3

Here is an example spice model which uses them:-

ACMQ 3:4:5 System

.OPTIONS NOMOD NOPAGE
.INCLUDE acmq345a.cir
.INCLUDE acmq345b.cir

* A very high load resistance on the topload
RL 1 0 50000000K

* Tertiary coil
X2 1 2 acmq345b

* Primary/secondary plus C2
X1 50 0 2 3 acmq345a

* A voltage source in the secondary base just to
* plot base current with
VB 3 0 DC 0 AC 0

* Primary cap
CPRI 50 52 5.08n ic=10KV

* Voltage source across gap for impedance sweep
VP 52 0 DC 0 AC 1

.END

To plot the transient response, include

.TRAN 50nS 10uS UIC
.PRINT TRAN I(VP) I(VB) V(2) V(1)

And the impedance sweep seen looking into the gap, include
something like

.AC LIN 2000 1K 1000K
.print AC I(VP)

These two spice circuits contain about 1000 and 5500 components,
representing the distributed self and mutual reactances of each
unit.

They work fine with ngspice under Linux.   Not sure what other
flavours of spice will make of them.   The subcircuits contain
resistances to represent very roughly the losses.   These are
not even remotely accurate.  They would need to be frequency
dependent resistors to do a proper job.    For now they are just
set to a value based on skin and proximity loss at about 250kHz
or so.   Always these models underestimate the real losses.

The number of segments is larger than necessary to get realistic
terminal voltages.  I use this many steps in tssp just to get
smooth V/I distributions within the coils.

It would be interesting to hear if these work in other spices
and if they are useful.  GeoTC could be programmed quite easily
to generate these.
--
Paul Nicholson
--


_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla