BUILD A NEGATIVE ION GENERATOR

I F YOU'VE SEEN ADVERTISE -

m en ts for n e g ativ e i on
ge ne rat o rs in mail- o r de r
catalogs a n d wo n de re d h ow
they work. this project is for
you. The s im ple vers ion desc ribed h ere provides insight into their theory a n d
app lications, a n d is both in formative a n d en te rtain ing.
S om e of th e d em on strations you ca n do wi th .i t will
a m a ze y ou a n d your
fri ends.
Many cla ims have b een
made for the b enefioial att r i b u tes of n e gati v e at mospheric ions on human
and plant life , es pecially by a
Dr. Albe rt Kru eg er of t he
University of Ca lifo r nia .
S tu d ie s h a ve s h ow n t hat
negative ions promote physical and m en tal alert ness
a n d w ell-b eing, whil e
pos it ive at mosp h er ic ion s
(su ch a s in pollu t ed air )
ca u se d isc o m f o rt an d
lassitude.
Ce rtain n ega tive ion properties ca n be demons tra ted .
For exa m ple, the s u r round ing air a fter a th u n de rstorm
sme lls clean a n d fresh. due
to g e n e rat io n of n egative
ions from lightning. T he
n egative ions att ach to
s m o ke. dust. a n d p ollen
particles, bringing them to
the g r ou n d to di s charge ,
lea vt n g fr esh , c lea n a ir.
Th at's why a coo l room wit h
a breeze is invi gora tin g .
com p a re d wi th one that's
stiflingly h eated . Coo l a ir is
ge ne ra lly n eg ativel y ioni z ed,
whereas h ea ted a ir is ge ne rally
posit ively ionized.
Negative ions a re a ir molecules
with one or m ore excess electron s , produced in th is case by a
low-power. 9 to 14-kilovol t DC
su p ply. The positi ve terminal is
gro un ded, and t he other (the
em itter ) is a n eedl e exposed to a ir.
(To ge nerate positive ion s . the polarities would be revers ed .) Extra
elec tro ns on the em itter's s urface
produce a high loc al elec tric field
owing to its pointed s hape . The
elec tron s ex it the em itter n eedl e's

Build this negative

ion generator and
put some charge
in your life.
ANTHONY J. CARISTI

WARNING!! This article deals with and

involves subject matter and the use of
material s and substa nces that may be
hazardous to health and life. Do not attempt to implement or use the information
contained herein, unless you are experienced and skilled with respect to such
subject matt er , materia ls , an d substances. Neither the publisher nor the author make any representation as for the
completeness or the accuracy of the information contained herein, and disclai m
any liability for damages or inj uries,
whether caused by or arising from the lack
of completeness, inaccuracies of the information, misintepretations of the directions, misapplication'of the information, or
otherwise.

s u rfa ce due to the polarization of surrounding air molecu les between the emitter
n e edle and g r o u n d. Th e
elec tron s collide with the air
mol e cul es a n d produ ce
n egative ions .
A common misconc eption regarding high-voltage
coron a or a rc in g is that electrons are "overcrowded" on
the tip, and forced off by
mutual repulsion. What ac tually ca u s es corona is the
hi gh electric field at the tip.
which is directly proportional to the voltage. and is
en h a n ced by sharpening an
electrode tip to a fine point.
Th e hi gh electric field
s trains t h e air molecules.
polarizing them b y a phenomenon called dipole po la r ization. If the electrode is
positive. electrons are litera lly ripped off, creating
positive ions. or if the elect ro d e is negative . they're
forced to accept electrons.
creati ng n egative ions .
High-voltage supplies like
those in TV's need careful
design, s o no undesired discontinuities l i k e shar p
points or ed ges cause arcing. Howev er. for the negative ion generator d iscussed
h ere. the goal is to generate
coron a , not to avoid it. and
the high electr ic field in the
vicin ity of a discontinuity
more readily polarizes and
ionizes air molecules.
The reason for the high
electr ic field at the tip of Gl
n eedl e is due to its s m all localized
radius of curvature. A sphere has
a much more uniform electric
field at its surface, due to its cons ta nt and much larger radius.
The audible hiss caused by a
hi gh-voltage discharge is called
"co ro na wind. " and is often heard
in olde r large-s creen color TV's .
es pecially on humid days. when
the breakdown potential of the
s u rrou n d ing air is reduced. making the flyback transformer arc.
You can often also feel it . if the
corona is strong en ou gh .
The negative ion generator de-

e;

z
j;

CD

41

scribed h ere is low-cost. easy to

build. and uses a high-voltag e
flyback transformer from a n old
bla ck-and-white TV. It ge nerates
hi gh voltage . but at very low cu rrent. How ever. the s afety precautions taken for a ny high-voltage
device s h ou ld be observed h ere.

(/J

o
Z
o
a:
IoW

....J

Ci
~
a:

42

The circuit
The s chem a t ic of the negative
ion gen er a tor appears in Fig. 1:
T2 is a TV flyback transformer
with an open ferrite core partially
enclosed in an aluminum bracket. The only original connections
used are one low-voltage tap as
ground. and the lead connecting
to the flyback transformer output winding (at right).
The other low-voltage taps
aren't used. and the new feedback and bifilar primary windings are wound on the ferrite
core. The bifilar primary winding
goes to the collectors of an astable multivibrator made up of Ql
and Q2. with the center tap driven by the + 3-5 volts DC from
ICI. The astable energy-storage
elements are the inductances of
the new feedback and bifilar primary windings.
The term bifilar means a pair of
transformer windings optimally
coupled by having been wound in
the same direction. either adjacent to or. preferably. superpos ed
on top of one another. The feedback winding goes to the bases of
Ql and Q2 for positive feedback.
with the center tap driven by the
+ 3-5 volts DC from ICI through
R3. The transistors are then forward-biased by the opposite ends
of each half of the feedback winding.
When power is applied. the current through Ql and Q2 is unequal, du e to difference s in
dopin g . la yer thickn ess . and
base-emitter (B-El turn-on voltage. That's what causes oscilla tion ; if all the astable parts were
perfectly balanced (nearly imposs ible), it might not oscillate at all.
Whi ch ev er transistor ca r r ies
higher current s a tu r a tes due to
positive base feedback. a n d the
other cuts off.
Wh en t he sharp increase in
transistor current in half the bifil ar primary winding is m a ximiz ed . t h e induced vo ltage
revers es . so the second transistor conducts and the first one
cuts off. The collector wav eform

1- 8,-1..,1+- - - -,1

(USEEARTH GROUND
IF BATTERY-OPERATED

AL~L1NE

(4
"D"CELLS)
lEOl

S2:\,

R4

5600
BATTERY. :
OPERATED :

'ft f::~.i
01

lM317T

TIP318

~ ~

AOJOUT IN

B C

ERS 1 0~

9VAC-

'---

EMITTER
NEEDLE
ISEETExn

......-

--'

Cl

1 H4004
1000flF
C4
C3 '----+---....-..001
.001
-),H---~Ih

C5
.001

09
SK3067/502

02

R2

.-_ ---J 2200 +

C2
100~IF

---'

07
SK3067/502
03
1N4148

010
SK3067/502

C8
.001

04
1H4148

FIG. 1-THE SCHEMATIC OF THE NEGATIVE ion generator. T2 is a TV flyback transformer; only its output winding is used, and new feedback and bifilar primary windings
are wound on the ferrite core. The bifilar primary winding goes to the collectors of astab le
01-02. Its square wave induces high voltage into the flyback transformer output winding,
boosted followed by a ladder voltage-tripler - 9 to - 14 kilovolts.

02

~-~vr----'

TO FLYBACK TRANSFORMER
FERRITE CORE

FIG. 2-THE PARTS-PLACEMENT DIAGRAM for the main PC board ; 01, 02, and ICl have
U-shaped 1 x l-inch copper heatsinks, with silicone grease for heat transfer; each goes
on the main PC board without insulators. In the prototype, Sl , S2, Fl, and NEl were
omitted, and T2 was salvaged from a small-screen TV; note the new feedback and bifilar
primary windings. Pin 7 had the highest resistance relative to the flyback transformer
output lead, so it becomes ground.

fu n d a menta l fr equ en cy was

23. 26 kl-lz, alt h ough higher harmonics greatly exten d th e to tal
b andwid th . Also, t he p eak-topeak b iftlar prima ry winding
voltage is fou r times that of the
supply.
The as table s q uare wave induces voltage into the flyback
t r a n s fo r m e r output Wi n d i n g,
proportional to the tra n s for m er
turns ratio . A flvb a ck transformer bifilar primary winding is

normally quite s m all, while the

flyb ack tr an sformer ou tpu t
winding norm ally has about
2000-2500 turns. inducing - 3
to - 4 kilovolts.
Since the n egative ion generator n eeds to produce - 9 to -14
kilovolts DC, a ladder voltage-trtpIer with six diode-capacitor
rungs is u sed to half-wave rectify
and multiply the flyback tran s former output voltage. Its operation is best understood in

PARTS LIST
All resistors are Yo-watt , 5%, un less
otherwise noted.
R1-1000 ohms, PC-board mounted potentiometer
R2-220 ohms
R3, R4-560 ohms (the former for the
astable-flyback transformer combination, the latter optional for the batteryoperated version)
R5-R6-200 - and 40-megohm series
h igh-vol tage focus div ide r, RCA
SK3868/DIV-1 , used for an optional
high-voltage range extender for a conventional high- impedance (10 megohm) voltmeter (see text)
R7-2.7 megohms
Capacitors
C1-1000 tJ.F, 25 volts, electrolytic
C2-100 tJ.F, 16 volts. electrolytic
C3--C8-0.001 tJ.F, 10 kilovolts, ceramic
disc
C9-0.001 tJ.F, 500 volts, ceramic disc,
optional for aluminum can/neon bulb
experiment (see text)
Semiconducto rs 01, 02-1N4004 silicon diode
03, 04-1N4148 silicon diode
D5--01o-RCA SK3067/502 high-voltage
diode. 12 kilovolts PIV
LE0 1-l ight-emitting diode for the battery-operated version
01 . Q2-TIP31B NPN transistor
IC1-LM317T adjustable voltage
regulator
Othe r components
F1-Q.5-amp slow-blow fuse with holder
NE1 -120-volt AC neon-bulb assembly
with 47-100K built-in series resistor for
the plug-in version (Radio Shack
272-712)
NE2-neon bulb, type NE2 (not the part
number). optional for experiment (see
text)
S1-SP..sT toggle switch
S2-SPST toggle switch
T1-18-volt center-tapped transformer
(Radio Shack 273-1515)
T2-standard TV flyback transfo rmer
(see text)
M is cellaneous : Plastic case
(7.5 x 4.25 x 2.25-inches, Radio Shack
270-224), enamel magnet wire, threewire line cord, emitter needle (made
from either straight pin or sewing needle), RTV silicon rubber, heat sinks, four
alkaline "0" cells (optional for battery
operation), 2-liter plastic soda bottle,
and a sewing needle.
NOTE: The fo llowing are available from
Antho ny J. Caristi, 69 White Pond
Road , Wald wick, NJ 07463: Two
etch ed and drill ed PC boar ds (one
each for the main and voltage-tripler
sections) for $15.95, IC1 for $3.25, 0 1
and 02 fo r $2.75 each. Please add
$2.00 for post age and handling with
each order; NJ resid ents please add
7% sales ta x.

r ~ C3

INPUT FROM
FLYBACK
TRANSFORMER
(MAIN BODY
WINDING
CARRIES HVDC)

;:[
4 C5

: 07
:
:

rJ

:f
I

11~
I

C6

08

i 0:

i, ':"

""!

10

r'

C7
C8 ~
HVDC OUTPUT 0 ..l

FIG. 3-THE PARTS-PL ACE MENT d iagram fo r the high-voltage PC board. The
parts shown include high-voltage ceramic
disc capacitors C3--C8, and selenium rectifiers 05--010. 05 , 07, and 0 9 go on the
component side (solid lin es), while 06,
08 , and 010 go on the foil sid e (dashed
lines) .

segments . The AC waveform

from the transformer is coupled
to 05 and 06 via C3, charged
through 05 acting as a clamp.
The peak-to-peak magnitude
of the AC waveform at the anode
of 05 equals that of the flyback
transfo rmer outp ut Winding,
with a n ega tive DC component of
half the peak-to-peak value. The

EMIITER__-l
NEEDLE

CASTING ~
RES IN

AC output waveform from the

flyba ck transformer is coupled to
the anode of 07 via C3 and C4,
where 07 ch a rges C4 . The action
is repeated again via 09 , which
charges C5 . The DC potentials on
C3-C5 add, tripling the voltage
from th,e flyback transformer
Winding. The anode of 010 is the
output of the negative ion generator, and should be at - 9 to - 14
kilovolts DC.
Power is supplied by either a
standard full-wave rectifier followed by variab le voltage regulator ICI , or four series "0" cells
producing 6 volts DC. Although
SI, S2, FI, and NE I are s hown in
Fig . 1, they were om itted in the
prototype . ICI co ntrols the DC
voltage fed to the oscillator via
Rl. Since the negative ion generator output must be - 9 to -14
kilovolts, and the exact flyback
transforme r turns ratio is normally unknown, adj usting RI is
mandatory. Once RI is set, the
negative ion gene rator output
voltage will be stable.

-I

-------.

FLYBACK
OUTPUT
(RUBBER
TEST LEAD)

FINAl
HVDC
TR1PLER
OUTPUT
(RUBBER
TESTLEAD)

RlV
SILICONE
RUBBER

e,
):>

2
C

FIG. 4- THE NEGATIVEION GENERATOR, showing the case, T2, plastic soda bottle, and
the top of th e high-voltage PC board , wit h 05, 07, 09 , and C3-C8. Both the plastic soda
bottle and high-voltage PC board are held in place with RTV. The neck is sealed with
casting resin to wit hin Yo-inch of the top; cover the needl e with a cork when unused.

~
-<

-.I

z
~

o
~

ui
o

is

C?
44

Construction
Next, we' ll di s cu s s co nstruction , focusing on th e ma in
and vo ltage-triple r PC boar ds,
the plas ti c cove r for the voltagetrip ler PC boar d m ade from the
top th ird of a 2-liter plastic so da
b ottle , the em itter n eedle , a n d
the cu tout in the lid of the plas ti c
case for the flyb a ck transformer.
Figu re 2 shows the parts-pla cem ent diagram for the m ain PC
board , and Fig . 3 the ladder voltage-tr ipler section . Remember to
ad d S I , S 2, FI , a n d NEI off the
m ain PC b oard, a s s h own in Fig.
1, if you want them .
The m ain sect ion includes the
regulated s u p ply, oscilla tor, a n d
flyback transformer, a n d goe s on
the s ingle-s ide d PC board menti oned in the parts lis t , with the
fo il p attern s h ow n h ere. Th e
h igh-voltage PC board contains
t he ladder voltage trtpler, co n n ected to the main PC board by
two leads , a n d the em itter n eedl e
is a b ove th e hi gh- vo lt a g e PC
b oard. In Fig. 2 , QI , Q2 , a n d ICI
n eed heatsinks ; you can buy
them, or m ake th em from 1 x linch copper b ent in a "U," with
s ilicon e grease for h eat transfer.
The h ea t s in ks go on top of the
PC board without insu la to rs .
Neon lamp assem b ly NEI acts as
a n indi c a t or , and h a s a
47K- 100 K resi s tor in ser ies wit h
i t . Fo r b a ttery op era ti on , u s e
LEOI in ser ies with R4; u s e alkaline b atteries, not ca rbon -zinc.
Us e a 3 -wire cor d with gro u n d , or
a n ear th gro u n d in battery-opera ted versions. The voltage-tripler
sect ion is built to avoid a rc ing
a n d shock. Cut the top th ird off a
2-li t er pl a sti c soda b o ttl e , as
s hown in th e lead photo, or u s e 4in ch d i am e t er pl a s ti c PVC
plumbing pip e.
Use 18- gaug e rubb er- co ated
test leads to co n nect the highvo ltage PC b oard t o b o th the
flyback trans former output and
the n eedle, to withstand h igh
voltage, but n ot until indicated
(m ore b elow). The h eavy voltaget r ipler output lead is solde re d to
t he emitter n eedle , wh ic h goes
upwar d through the plas tic so d a
b ottl e a n d proj ect s o ut of t he
n eck , which is filled wit h casting
resin. Don't fill the n eck wit h res in all the way to the top; leave Y4
inc h of the n eck ope n, so you can
cap the n eedl e with a wine bottle
co rk when n ot in u s e.

08
FLYBACK
OUTPUT

06

FINAL
HVDC - - - - - if---oiEOUTPUT

UNDERSIDE OF
VOLTAGE-TRIPLER
PC BOARD

GND
---WIRE

FIG. 5-THE BOTTOM OF THE HIGH-VOLTAGE PC board, showing 06 , 08 , and 010, and
the attachment points of the ground lead and both high-voltage leads.

The volta ge t ri ple r s hown in

Fig. 3 cons is ts of ceram ic di s c ca p aci t ors C3 -C8 , a n d high-vol tage selen iu m rectifi ers 05-010 .
Note that 05 , 0 7, a n d 0 9 go on
the co m p one nt s ide of the PC
b oa rd , while'Dfi. 0 8, a n d 010 go
on the foil side. They're com m on
to most TV's , s o most TV repair
s hops should s toc k them; see the
parts list.
Figure 4 s h ows a photo of the
n egative ion ge nerator, wi th the
clos ed case , flyb ack transfor mer,
plas ti c s oda b ottle, a n d top of the
high-volt a ge PC board s howing
05 , 07 , 0 9 , a n d C3-C8 . The
hi gh-voltage PC bo ard is h eld in
place ins id e the plastic s oda b ottle top using RTV s ilicone rubber,
and the plastic s oda bottle goes
on the top of the cas e the same
way, for a water-tight seal th at
avo ids con de nsati on, a nd h a s
good high-voltage attr ib utes.
The markings a n d s izes of the
high-voltage TV d iod es used in
t he vo ltage tripl er di ffer fro m
t hose on sta n da r d di od e s.
They're much larger, a n d h ave a
very high p eak-invers e volta ge
(PIV) rating. Als o, on a con ven ti onal diode, the cath ode is n orm all y m ark ed with a b and ,
whether a wh ite band on a black
b od y for s uc h low-vol ta ge recti fiers as t he IN4 00I , or a red
b and on a clear glass b ody, as
with the IN41 48 or IN914 .
In high-volt age TV diodes, the
b ody' is ge ne rally eit her wh ite,
bla ck , or see-t hrough. In Fig . 4,
05 an d 07 a re wh ite wi th a b and
of bla ck" + .. s igns on the r ight
e n d , while 0 9 is bl a ck with a

dashed band of wh ite" - " s igns

on the ri ght en d. The en ds with
the bands are the cathodes , for
both d iod e body types ; the reas on
for the" + .. s igns on 0 3 and 05 is
that circ u it p owe r is normall y
cons idered as b eing ext racte d
from the cathode en ds in TV, a n d
is called B + . In Fig. 5 , 06 , 0 8 ,
and 010 all h ave bands of "+"
s igns on th e cathodes.
A 3 x 3-inch h ole was made in
the case top for the flyb ack transformer; adjust you rs as n eed ed .
Dori' t con nect the voltage-t r ipler
PC bo ard to the m ai n PC board
un til t he final check out b elow ;
put it aside a n d let the RTV dry
overnight . The so da bottle a n d
the high-voltage PC board a re
h eld in place u sing RTv. The rubber-coated test leads con nect the
flyb a ck-transformer ou tpu t to
the voltage tr ipler a n d the voltage
tr ipler to the em itter n eedl e; the
th in wire is the Q:ro u n d .
A flyback transformer is also
ca lled a tu ned trans former, a n d
co ns ists of a ferrite co re s u rrounded by a m et al bracket, with
a large coil of windings coated in
pl a s ti c o r c er ami c with a n
extrem ely high di electric breakdown voltage. Often , as wi th the
flyback transformer u s ed in the
prot otyp e , there are a d d itional
low-volt age output win d ing taps
(seve n , in t hat ca s e ) p laced
a rou n d a plas t ic ring on one s ide
of the transformer coil, that h ave
eq ually h igh-voltage breakdown
character is t ics .
That's all we h ave ro om for.
Next month we'll fini sh up th is
project.
R-E