0399o(2).pdf

(103 KB) Pobierz

by ALAN WINSTANLEY

the rectifier and smoothing ca-

pacitor voltage ratings).

Second, although makers

specify ratings assuming a

re-

sistive

load, in a full-wave recti-

fied power supply the load is

more complex than that

are placing rectifiers and a very

large smoothing capacitor

across the secondary winding.

The effect is that we find cur-

rents flowing in the secondary

circuit higher than those found

in an ordinary resistive circuit.

As a good rule of thumb,

the transformer secondary cur-

rent figure should be derated by

about 40 per cent in such cases.

A 5A transformer (for example)

This month’s roundup of readers’ queries and

comments follows up with more information on

electret microphones, and some pointers for using

variable regulators.

The question of using vari-

able voltage regulators to pro-

duce simple power supplies

crops up from time to time.

Here's a query from

G. A.

Wilsher

of Teddington, Middle-

sex, UK.

would be a suitable medium

on which to construct a reli-

able high current design. You

might consider assembling

with old-fashioned tag strips or

“turret board”, using heavy

gauge solid core wires for in-

terwiring, which should be kept

as short as possible. A simple

point-to-point design will cope

admirably with higher currents

as well as having a higher me-

Regulators sum it up

I have been waiting patiently

for a “decent” Bench Power

Supply unit to appear when I

came across the variable regu-

lator LM338 which will produce

an output from 1 2V to 32V at

5A, and this would give me ex-

actly what I'm looking for. Al-

though I could manage such a

design on stripboard I would ap-

preciate some guidance.

9 506

' 72 '

)8// :$9(

5(&7,),(5

$ 506

9 $&

0$,16

–

35,0$5<

6(&21'$5<

LO A D

/2$'

$ 0$;

O U T

60227+,1*

&$3$&,725

I happen to have a trans-

former with bridge rectifiers and

smoothing capacitors which pro-

duces 50V d.c. The LM338

specification states that the dif-

ference between input and out-

put voltages should not exceed

35V. What are the implications

of this? Second, I can see that

two resistors are required to set

the voltage. How does one cal-

culate their values?

Three-terminal regulators

are simple and convenient to

use “on paper” but there are one

or two drawbacks to consider,

especially at the higher volt-

ages/currents you are hoping to

provide.

I'm not sure that stripboard

Maxfield & Montrose Interactive Inc

Fig.1. Transformers must be derated when full-wave bridge

rectifiers and smoothing capacitors are used. Makers’ ratings

usually assume a resistive load.

chanical strength than anything

you could make with stripboard.

The output ratings of your

transformer are not quite clear,

so here are a few points to bear

in mind. Transformer manufac-

turers normally rate secondary

voltages and currents assuming

only a simple

resistive

load is

used at the maximum permissi-

ble current of the transformer.

This has two implications: first,

due to its regulation the sec-

ondary voltage will rise at

lower

currents, so allowances must be

made for this (when specifying

32:(5 75$16,6725

O UT

),;('

5()(5(1&(

92/7$*(

Fig.2. A simple pre-regulator

may help avoid excessive

input-output voltage appear-

ing across variable regula-

tors. D1 and D2 are Zener

EPE Online, March 1999 - www.epemag.com - 371

&LUFXLW 6XUJHU\

should only be expected to

safely supply about 3A total

load current in a full-wave

power supply, see Fig.1.

The term

root-mean-square

(RMS) is applied to alternating

voltage and current values, in-

cluding the ratings of transform-

ers. Root-mean-square values

are useful as a way of compar-

ing AC voltages and currents

against DC equivalents.

An alternating voltage of

say 12V RMS placed across a

fixed resistor will produce the

same heating effect (power dis-

sipation), as would a voltage of

12V DC. (Multiplying the RMS

value by 1 414, the square root

of 2, tells us the

peak

value of

the alternating voltage.)

This input-output voltage

rating is the maximum differ-

ence in voltage that is permitted

between the input and the out-

put pins of the regulator. A

bench power supply is likely to

be subjected to some rough

treatment, and it is quite easy

to overlook the implications of

this rating.

If the input voltage is known

to be, say, 30V DC maximum,

then if the power supply output

was shorted to 0V, the entire

voltage will appear across the

device. No harm will be done

at 30V and the device will cur-

rent limit.

If for some reason the input

voltage rises to say 50V or

more, then clearly the input-

output voltage would be ex-

ceeded and you risk destroying

the device. Hence you ought

not to directly use your trans-

former to produce a variable

design with the LM338. One

work-around is to use a form of

“pre-regulator” which clamps

the input voltage to a safe

fixed value.

A simple Zener diode with

an external pass transistor

might be suitable (see Fig.2)

provided the tolerances are not

too wide. You can combine

Zener diodes and also “jack up”

the voltage with ordinary

forward-biased rectifiers using

some trial and error if neces-

sary. The output voltage seen at

the transistor emitter (e) is one

diode drop lower than that ap-

pearing at the base (b). Re-

member to select a transistor

with a high enough collector

current rating.

just a derivation

of that for a potential divider. A

typical three-terminal variable

regulator is shown in Fig.3.

A reference voltage V

REF

developed across the resistor

R1, and by applying ordinary

potential divider theory we

can say:

REF

= (R1 / (R1 + R2) ) x

OUT

so V

OUT

= (V

REF

x (R1 +

R2) ) / R1

so V

OUT

= V

REF

(1 +

(R2/R1) )

A value of 1 25V is a nomi-

nal value for the reference volt-

age. A value of 240 ohms is

typically used (220 ohms will be

fine) for R1, which allows about

5mA to flow through the resistor

chain. A further 50

A (I

ADJ

)

flows out of the adjustment pin

0$;,0

80 ,13

2873

92/7$*

(

,))(5(17,$/

In the example of Fig. 1, the

transformer shown will have a

manufacturer's secondary volt-

age rating of 12V RMS at 5A

output. The voltage will tend to

increase when the load current

is reduced, and a value of

nearer 16V or so off-load would

not be surprising.

Furthermore, that 5A sec-

ondary current is only valid

when a simple resistive load is

used as shown. Ideally your own

transformer must be capable of

providing 8 3A on full load, but

to be realistic you are limited by

what you can buy off the

shelf or happen to have avail-

able anyway.

6(( 0$18)

$&785(5

6 '$

$'-

287

O UT

R EF

9 7<3,&$

RE F

[

5 5

OUT

5 5

OU T

[

R EF

RE F

Regulations

Looking at the regulator it-

self, a guaranteed way of de-

stroying such a device is to ex-

ceed the input-output voltage,

so well done for highlighting this

parameter! Incidentally my Na-

tional Semiconductor data book

specifies 40V as the maximum

input-output differential.

Maxfield & Montrose Interactive Inc

Fig.3. Calculating the output

voltage from a variable volt-

age regulator. A typical value

for Vref is 1.25V

to 0V, but for most purposes it

is not often necessary to con-

sider this any further.

One final aspect of your

proposed power supply is that

you should allow for a suitable

Doing your Sums

Your second question re-

lates to the calculation of the

resistor values. The formula is

EPE Online, March 1999 - www.epemag.com - 372

&LUFXLW 6XUJHU\

heatsink on the LM338, and it is

here where the plan may fall

down. Whilst the device is pro-

tected from thermal overload, it

does mean that if you are to en-

joy maximum current outputs

then some substantial heatsink-

ing will be required, otherwise

the device will thermally-limit in

extreme cases.

Unfortunately, it is not pos-

sible to calculate the heatsink's

thermal resistance until the

power dissipation of the regula-

tor is known. It is precisely be-

cause of thermal efficiency

problems with such designs that

switched-mode power supplies

are preferred.

ping a length of screened cable

connected to a high input

impedance amplifier will pro-

duce a noise on the output.

Electret material, which is a flu-

orocarbon polymer, can be

given an electrostatic charge

which lasts almost for ever.

If an electret material is

used as the diaphragm (with

one side metallised of course) in

a microphone together with a

rigid back plate then a capacitor

is formed with constant charge

Q. Sound pressure will cause

the diaphragm to vibrate and

thus the capacitance to vary.

Since the charge is constant

and as VQ/C the result is a volt-

age proportional to the sound

pressure.

The f.e.t. simply acts as an

impedance converter and this

must be (as you found only one

resistor) configured as a

common-source amplifier which,

given a suitable j.f.e.t., doesn't

need a source resistor if the in-

put signal is small. So, I would

guess that the resistor you de-

stroyed was a gate to source

resistor with a value of possibly

tens of megohms. I'd be inter-

ested if anyone can confirm this

surmise of mine about the f.e.t.

Professional condenser mi-

crophones generally do not use

electret materials. The di-

aphragm is a metallized polymer

film polarized by 48V d.c. via a

1G ohm resistor which results in

a virtually constant charge. The

polarizing voltage is fed to the

microphone via a “phantom”

power supply from the mixer

which also powers the built-in

impedance converter.

It is true that the “2N” or

“2S” is commonly omitted from

the nomenclature printed on the

device. My first thought was

also that the mysterious transis-

tor was a 2N or 2SK type, but

further research drew a blank

(as you found yourself). A de-

vice type KSK596 is listed by

Motorola, but there was no data

available. I rechecked the Na-

tional Semiconductor web sites

as described last month, and

this bounced over to the

Fairchild web site.

I finally obtained the data

sheet which states “N-Channel

RF amplifier” so I'm none the

wiser. In retrospect I would con-

cur that it's probably a junction

FET, but hopefully I can be for-

given my near miss!

Back to my groaning book-

shelves and my

Dictionary of

Electronics

defines “electret” as

the electrical analogue of a per-

manent magnet. It is a dielectric

body with separate electric

poles and a stable existence.

The metallic disk I alluded to is

a synthetic dielectric film, met-

allized on one face.

I agree that there is a dearth

of information about electrets,

which is what prompted the arti-

cle to start with, but a quick look

at several microphone manu-

facturers' web sites did reveal

interesting information including

more on “phantom” power sup-

plies. Try

www.sennheiser.com

and

www.shure.com.

I also found

an enthusiastic web page on

www.mpage.demon.co.uk/

electret.htm.

Thank you for

filling us in with the extra

information!

Electret microphones

Last month I described the

composition of a typical electret

microphone capsule, and went

off in search of the mysterious

transistor contained within. My

thanks to

Barry Taylor

who

adds further interesting com-

ments and describes the elec-

tret effect in more detail:

I would guess that the tran-

sistor labelled K596 is probably

a 2SK596 as I have frequently

come across devices from the

Far East that omit the first two

characters (esp. on TO92 pack-

ages). K indicates an n-channel

f.e.t. or MOSFET. You say that

Motorola suggest a 2N5484 but

according to my sources this is

not a MOSFET but an n-

channel j.f.e.t. I regret that I

can't find any listing of a

2SK596 though.[exp 98]

You make no mention of the

meaning of the word “electret”,

which could imply that the varia-

tion in capacitance directly pro-

duces the a.f. output. This is not

strictly true even though it is

easily demonstrated that tap-

Maxfield & Montrose Interactive Inc

Quick cap check

I would like to know how to

check a capacitor both out of

board and in-board, to deter-

mine whether the capacitor is

leaking or otherwise faulty.

EPE Online, March 1999 - www.epemag.com - 373

&LUFXLW 6XUJHU\

So asks

Siva Path-

manathan

(by email). It is

straightforward to check larger

capacitors just by using an ordi-

nary multimeter set to a high

resistance range. A moving coil

meter is possibly better for this

experiment.

By clipping the test leads to

an electrolytic capacitor, the re-

sistance reading will briefly be

seen to fall as the capacitor will

charge up through the meter's

internal battery. The meter's

black

lead is in fact positive (the

reverse is true on digital multi-

meters, which have a much

higher impedance), so by touch-

ing this to the capacitor's positive

lead, the needle “blip” will be

seen.

With lower value capacitors,

probably the best you can hope

for is to check for a shorted ca-

pacitor, which will show as a low

resistance.

Checking a capacitor in-situ

isn't reliable because other com-

ponents in the same circuit may

affect the readings. In any case,

definitely

do not

check while

power is applied to the circuit,

and

remember to discharge

the capacitor to start with, us-

ing a low-value resistor

for

larger electrolytic capacitors if

necessary.

Visit our Web site at www.epemag.com

In our library you will find:

♦

Alan Winstanley’s internationally aclaimed and

invaluable guide to soldering

♦

An amazingly useful article on LCD displays

♦

And much, much more!

Maxfield & Montrose Interactive Inc

EPE Online, March 1999 - www.epemag.com - 374

Plik z chomika:

fred1144

Inne pliki z tego folderu:

0399f(3).pdf (320 KB)
0399h(2).pdf (490 KB)
0399i(2).pdf (450 KB)
0399g(2).pdf (151 KB)
0399k(2).pdf (386 KB)

0399o(2).pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: