FOUR/87 caibhsen
Audio Amateur
THE JOURNAL FOR AUDIOPHILE CRAFTSINDUCTANCE FOR YOUR
SUPPLY DESIGN
BY FERNANDO GARCIA VIESCA
fa the dawn of the elecuonics age,
the rectified sine wave was usually
Filtered with the help of inductors.
With the old technology, it was state-
ofthear to design equipment that
‘way. But as espaciter technology ad-
vanced, and with the trends toward
rminiatinization and light weight, hie
ductance has completaly vanished fom
(6OHz power supplies. Switching power
supplies, with their high operating fre-
quencies, (20 to 200kHz} are a case
apart. Nowdays, the odds aze 500 to 1
‘that any equipment users with prob-
Jems in their power supply will ty 0
solve them with more ond more
‘capacitance,
‘Are the trends correct? Yes, but the
areas of highcurrent power supply
design can benefit from some induc-
tance. There is a device in your bi-t
system that clearly fits inthis high cus-
rent category: the power amplifier.
(Capacitance vs. Inductance
‘Let's take a look atthe different advar-
tages and disadvantages of capacitor in-
‘pul fering, With capacitance you get:
+ Small size and weight in the fil
ter's components,
+ Enormous amounts of filtering c2-
pacity for your money,
«Higher average DC volts than AC
RMS volts
+ Low supply impedance.
+ High peak to average rectifier and
transformer currents,
«Very high in-rush currents,
With a conventional inductor input
filter, such 4s the Fig. 1 topology, you
get the opposite features. As we may
22, The Audio Amateur 4/37
[FIGURE 1: Conventional inductor fiter.a wasteful bleder resistor is employed, other
wise « very large inductor would result
FILTER
INDUCTOR
see
ibeage
[ie eer
FIGURE 2: Sugrested inductor filter topology. By moving the inductor out of the DC
current path, we enjoy greater jteedom in choosing an appropriate inductor valueWith 2 workedou example, you'll
cence this size restrition more cleaaly.
‘Suppose an amplifier whose peak cur-
rentdraw is 75A and the idling current
is ODA at 27V DC. To calculate Rit
4s safe to assume that:
well designed power tansform-
crdocsnot allow more than 25% dsop
fn is output voltage from idle ¢o fall
Joe
+ And that the transforner's series
FIGURE 3: Ripple reduction graph: Star-
sng with a 48 percent ripple ofe fll wave
rectifier, itmay be decreased with etther
‘apicitancs or inductance. Rlisthe equiv
alent full-load resistance.
see, with the exception of the two last
features, capacitor input filtering holds
the aces in its band.
‘Bat could we benefit from just a lite
tle inductance and therefore have the
best af hoth worlds? Yes, but an obscure
pitfall awaits whenever we use an in-
ductor in a DC current path. You ean-
not let it “run dry,” és when the in-
ductor current drops to zero. For that
reason there is a mbiniamuma csiticalin-
‘duczance that is inversely proportional
to the idling (minimum) output cur-
rent, A low idling carrent in your
amplifier would mean a very large im-
practical inductor. In. the past_we
customarily included a bleeder resistor
that never alowed a eartent drain £0
drop lower than the equivalent of 10
percent of maxinmam output current. In
«BOA suppl this would mean two
amperes of continaoasly wasted power:
Now you can see why inductors have
heen totally dropped from solid stare
supplies.
“The minim critical inductance re-
quired to avoid "sunning it dry” is
siven by the formula:
Le = (Re + Raa) + Sw
hero:
ww = 2xf = 377 (at CO)
Ry = series resistance, including trens-
former, diode and wiring resistances
Rane * Vine + Ile
resistance is the main component in Ry.
‘Therefore we may approximate the
series resistance:
By = Van X Nm = flaw ~ Ta)
We may assume tbat:
Vee = VDC + 1418
Ry-203 x 005 + (7.5-012|-0.1380
And our equivalent resistance at
mininimum curent draw would be:
Rew = 27 + 0.12 = 2250
Working with the above formula we
would get:
Le =[225+.135) + {8 > 377)=0.199H or
199mH
CClesly such an inductor would comsist
of an enormous chunk of copper and
iron. We could use, instead, « wasteful
bleeder resistor. We must provide
enough blesler current to complete
0.75A (10 percent of 7.54).
4 = 075 — 0.12 = 0.634
‘And the wasted power would be:
P= 0.63 x 26 = 17W
‘And recaloulating the critical indue-
tance we would still nee 38mH. No
wonder inductors went the way of the
dinosanr.
‘The saturation phenomena resulting
fom operating inductors with DC cur-
rent may be eccumvented if we operate
them from AC current. The sir gap re-
‘quired in a DC operated inductor, a
source of magnetic interference, is de-
leted in AC operation.
became aware of this in a rather
awkward situation. I once hid trouble
‘with apower aupply whose outpat volt
age was lower than calculated. Upon
IGURE 4: Load regulation: Output volr-
‘30 vanation sth changing Toad cuneit
‘re ploted in this graph, Calculitine your
RAR (max) and R/R, iin) will allowr
you to acutely pret tie fad et
hooking an oscilloscope, I found the
culprit a transformer with excessive
Jeakage inductance. I observed that
will the iower output wolage was ex
tainly « drawheck, the output voltage
txhibied a Tower ripple nd the peale
diode current was lowered. Isuddealy
realized that an extra indactor placed
45 shown in Pig. 2 would hivethe same
overall effect
‘We place the inductor before the
‘wide feoifier, and therelore the core
8 AC powered and reset every half
ove.
Core saturation is no longer a prob
Jem and we're free to include as much
or as litte inductance as the circuit
requires.
A Capacitor Input Example
1 will start with the conventional de~
sign of the same 27V, 7.54 power sap-
ply assuming thet only capacitance will
De used, and chat inductance will be in-
cluded later in emall increments, co I
could carefully tailor the value to min-
{mize drawbacks and maximize bene-
fits. All the following equations are de-
rived from the work of O.H. Schate,
subsequently reproduced in many
books
Its shown in Fig. 3, thet for low rip
ple, (less than 3 percent) the following
equation should be met:
WX Rina XC > 15
Solving for this inequality gives a
‘capacitance value greater than 132614
‘This results in a practical value of
‘The Ao Amateur 4/8723