dbx Model 3BX

3 band dynamic range enhancer

5er(//c.F MANUAL

I %

PUWfB fUMOTi
RMS Detection Attack and Release Times
The 3UX uses a detection method that acts on the RMS The expander must decide how last to react to changes
(Root -Mean-Square) value of the input signal. RMS de- in program level. The length of time between the detection
tection is different from either peak or average detection. of an input signal level, and its expansion is known as the
An RMS detection circuit will not over-react on musical ATTACK TIME. After the expansion of an input signal,
transients or noise spikes, yet it responds precisely and the expander allows the input signal to return to its normal
accurately to all changes in input level, and produces level. The amount of time to return to normal is known as

natural sounding expansion. In fact, the human ear judges RELEASE TIME These terms also apply (or compressors
.

sound levels by their RMS values, which means that the and limiters. In general, an expander's detection method,

RMS detection circuit in the 3BX

electronically parallels and its attack and release times are not directly related.

the way the human ear hears music. Until recently, how- For example, an expander that uses peak detection may
ever, RMS detection was very complex and costly, dbx detect an increase in program level very rapidly, yet not act
has pioneeied die development ol moderately pi iced RMS on that increase (by expanding tt) until much later. Thus,
detection circuitry, and has led the iiulustiy in applying peak detection does not necessarily imply fast attack lime;
RMS detection to expanders, compressor/limi teis and tape the expantler just cited wuuld have a slow attack time. Con-
noise eduction systems.
i
versely, an expander using average detection might have a
very fast attack time. While such a unit might not detect an
Linear dB Expansion increase in input level until the actual level had begun to
Once the signal has been "detected," the expander decrease again, the expander could react quickly, expanding
knows when to increase or decrease its level. The circuit the program upward at a point where it should be expanded
that actually performs this level change is known as a downward. These are extreme examples, but they illustrate
"voltage controlled amplifier" or "VCA." The "AVC" the distinction between attack and release times and detec-
(automatic volume control), and "ALC" (automatic level tion method. Even art expander using an RMS detection
eontroilon many cassette recorders are examples of scheme may not have optimum attack and release times for
voltage controlled amplifiers, as are the level changing every type of musical program. The reason is that the attack
circuits inany modern expander, compressor or limiter. and release times ol most expanders (and compressor/
The voltage horn the detection circuit increases or decreases limiters) are fixed (unchanging). In those units where the
the gain of the VC A which increases or decreases the level attack and release times are adjustable via front panel
of the program. While some expanders may increase or controls, the times are still "fixed" at the control settings”
decrease the program level by a lixed amount, the 3BX fixed since they do not change as the program requires.
increases or decreases the level of the program on a "linear Diffeiem types of music require different attack and
decibel" basis. This means that the output dynamic range release times. For example, a smooth classical string quartet
and the input dynamic range are linearly related by the number might sound best when expanded with a slow
"expansion ratio" over the entire dynamic range (as attack and slow release time, where a punchy rock and roll

desciibed earlier) for a smooth, natural sound.

number might sound best when expanded with a fast attack eliminating the possibility that a strong note in one band
Other programs might sound best with a will cause audible effects in another frequency band. After
and release time.
fast attack and slow release time. The point is that attack
careful consideration, dbx decided that three bands

and release times should be allowed to vary according to represented the optimum number of frequency divisions

the program content for the most natural sound. The 3BX for an expander.

does just that. The 3BX's attack and release times auto-
matically and continuously follow the rate of change of The 3BX
Devices such as "noise gates," "correlators" and
the "envelope"* of the program. In fact, because they are
"horizontal filters" may be able to subtract noise from
not fixed, the 3BX'$ attack and release times are specified
certain types of programs, but they do not significantly
as rates which change in response to different program
envelopes. The result is a smooth action that does not alter
improve the dynamic range of the program in any other
the character of the music as dynamics are expanded and
way. In addition, such devices may be misled by certain
types of music, causing undesirable changes in the sound,
noise is lowered.
or they may change the frequency response of the pro-
Single vs Multi-Band Expansion gram, or even remove parts of the music.
dbx produces three other models of expanders, the 1 17, The 3BX represents a new state of the art in expansion.

1 18 and 1 1 9, and a combination expander/dbx tape 1 1 With its three band linear dB expansion, RMS detection,
noise reduction system, the model 128. These expanders and program-controlled attack and release rates, it avoids
are single band units that expand the entire program at all the midrange coloration of pre-emphasized units (so called
frequencies at the same time. A single band expander is less "weighted" expanders), the "breathing" of single band
expensive than a multi-band unit, and this method is satis- units, and the "noise modulation" of units that divide the
factory for most types of program material. In fact, the 117, spectrum into many bands. The 3BX does not subtract
1 1 8, 1 1 9 and 1 28 outperform every expander
marketed any frequency components from the music. Instead, it
except the new 3BX. However, in some cases, when the reduces its gain during quiet passages and increases its gain
program has a very heavy bass line, single band expansion during loud passages, achieving 35dB or more of natural

can cause audible "breathing" as it also raises high fre- sounding expansion.
quency noise when a loud level, low frequency note is
detected. Fortunately, the high frequency content of most
program material masks this high frequency noise so that
the breathing effect is usually unnoticable. A multi-band
expander has the advantage of being able to control the
expansion of different frequency bands separately.

* The envelope is a graph of the program level versus time; see the

Glossary for a further definition.

 INITIAL FACTORY TEST S ALIGNMENT PROCEDURE

1 UJ T« ^ c '
DBX MODEL 3BX '

/<£/: 2,/4 MCOJ-Z*’

UC W (

]/s 1 J
Vt=/W v Mir; &.C. *>**-
TEST EQUIPMENT REQUIRED

QUANTITY ITEM TYPE OR EQUIVALENT

Signal Oscillator Kron-Hite 4200, Heath 1G1272

1

Digital Multimeter Dana 3300, Simpson '464

2
Heath 1M2202

Oscilloscope Any (Covering Audio Rangel

1

Wide Range DB Meter DBX SK-C-1026 , HP427

1
Heath 1M5238

Harmonic Distortion DBX SK-B-1070, Heath 1M58

l
Meter HP 330, GR1500, 1900

+ 3 Db Meter (REF 1 VRMS) DBX SK-C-1071

Remote Simulator DBX RS-1

Tone Burst Generator GR 1396B

GR Double Banana to 2
Phono Plug, Shielded Cable

GR Double Banana to 1
Phono Plug, Shielded Cable

= 1.000 VRMS AC
Note: All Input/Output levels referenced to 0 DbV

Test Set-Up

A+»lO ,

3BX
Osc.
iiMTV ts oj
—
T?cmc5te
1

c)C OO o o C>9
1

Sc©F>e
>OETB.FS
ETC
 3BX SEQUENCE OF STEPS

1) Power Supply & DC Tests

2) RMS Adjustments

3) VCA Level Adjustments

4) VCA Distortion

5) Tracking

6) Noise

7) Transition Level Range

8) Relative Level Set

9) Relative Level Response

10) Frequency Response

11) Remote Simulator

12) Maximum Output Level

13) LED Display

14) Output Clamp

15) Tone Burst Response

16) Wrap-up
 1) Power Supply & DC Tests

1.1) Plug 3BX to be tested into appropriate power (120 or

240 VAC) and ensure that the power and remote LED light
*
in the correct sequence.

1.2) Connect a DVM from the plus side of C51 to ground and
adjust R135 for + 12.00 vdc.

1.3) Connect to the minus side of C58. A reading of between

- *
-11.88 to -12.120 should be noted.

1.4) Connect the DVM directly across R145 and short circuit
X the plus supply to ground. Voltage reading should not
exceed .450 vdc.

1.5) Remove the short from the plus supply and connect to the
minus supply. With the DVM directly across R138, this
reading should also not exceed .450 vdc.

1.6) Remove the short and check for resumption of normal

operating voltages. (Given in steps 1.2 & 1.3)
Install jumpers Y1 & Y2 and check for normal operation.

1.7) Operate the remote switch in and out of mode, while

( checking that the gatesof 01 through 04 change polarity.
(Typically from +.4 to -6 vdc)

1.8) Install RMS & VCA in their respective connectors

(215) (210). RMS in J7 8, 9. VCA in Jl, -6.
,

2) RMS Adjustments

\ 2.1) With the "remote" switch out, the "source" switch

pushed "in" and both the expansion and transition
controls fully to the left of their travel, feed
100 Hz at -15 dBv (.178 VAC) to the 3BX inputs 1 & 2

(which are marked from pre-amp output.)

2.2) Low Band Adjust ;

Connect the scope to pin "7" of 0A12 (Anode CR2) and

adjust the symmetry control located on the RMS module
215 (nearest the rear panel) for a symmetrical waveform,
(equal peaks)

* Note: For tests 1.4, 1.5 and 1.6 jumpers Y1 & Y2 should
not be in board.
 2.3) Connect the DVM to pin "l" of 0A12 and adjust R58 for
a reading of 0 volts to +. -01VDC.

2.4) Mid Band Adjust :

M
Connect the scope probe to pin l" of OAIO (Anode CR4)
with an input of -10 dBv (.316 VAC) at 1 khz and adjust
the center RMS module for a symmetrical waveform.

2.5) Connect the DVM to pin "l" of 0A11 and adjust R76 for
0 vdc .01 vdc.

2.6) High Band Adjust :

Connect the scope probe to pin "1” of OA9 (Anode CRB)

leaving the input set for -10 dBv at lkhz and adjust
the remaining RMS module for a symmetrical waveform.

2.7) With an input of -27dBv (.045 VAC) at 10 khz, connect

the DVM to pin "1” of 0A8 and adjust R98 for 0 VDC
+ .01 VDC.

3) VCA Level Adjustments

3.1) Set the input to 0 dBv (1 VAC) at 100 hz, and while
observing channel 1 output, adjust R69 for unity gain,
(input s output)

3.2) Switch the oscillator to 1 khz, and adjust R90 for unity
gain.

3.3) Switch the oscillator to 10 khz, and adjust R109 for

unity gain.

3.4) Repeat the last three steps, readjusting if necessary.

3.5) Switch the oscillator back to 100 hz and while observing

,

Channel 2 output, adjust R70 for unity gain.

3.6) Switch to 1 khz, and adjust R91 for unity gain.

3.7) Switch to 10 khz, and adjust R110 for unity gain.

VCA Distortion

>

'v?*, 4.1) Reset the oscillator to 0 dBv at 100 hz and adjust the
low band VCA (J 4) symmetry control (the trimpot pro-
truding from VCA PC Board) for minimum second harmonic
distortion. (Must be below .03 %)
4.2) Set the oscillator to 1 khz and adjust the Mid band
VCA (J5) for minimum distortion.

4.3) Set the oscillator to 10 khz and adjust the high band
(VCA) (J6) for minimum second harmonic distortion.

4.4) Repeat steps 4.1 through 4.3 for Channel 1 adjusting

Jl, J2 and J3 respectively.

4.5) Check that the distortion readings agree with chart

LUW .

FREQ- 1,0 e*Pt 1.5* EJCP.

a ni < JOS % <. 1 %

tOO HZ
3 rd £.. 03 % <. 1
%
•a."
4 -
<. 03 % <. 1 %
\ KHZ 3 rd. <. 03 % <- 1 %
2 .
n<l <.03% < . 1 %
lO KHZ.
3 T* d -
<.03% < - 1 %
5) Tracking

5.1) With the remote switch disengaged and expansion control

set to 1.5, adjust the transition control for a level
match with an input of -10 dBv at 100 hz.

5.2) Change the input level in 10 dB increments from + 10

to -40 dB. The output should change 15 dB for every
10 dB input change.

5.3) Repeat at 1 khz and 10 khz. (Re-set level match for

each band.) (Note: At 10KHZ it is only necessary to
step from +10 to -30DB)

6) Noise

6.1) Set the expansion control to 1.0, and short both inputs.
Output reading for both channels should be -85 dBv or
better.

7) Transition Level Range

expansion con-
7.1) Set the input to -10 dBv at 1 khz with the
trol at 1.5,

7.2) Sweep the transition level control through it's complete

travel, ensuring that a 20 dB output range is obtained.
 7.3) LED DISPLAY BOARD CALIBRATION

Short the input to the 3B?. (Note: The input can be

shorted by placing all the DB step switches on the
Kron-Hite 4200 Oscillator to their "out" positions.

Place expansion control fully to the left (1.0) Rotate

R14 fully CCW.

Rotate RIO CCW until red led lights; note position of RIO
wiper.

Rotate RIO CCW until yellow led lights; note position of RIO
wiper

Position RIO wiper equally between previously noted position

wax RIO and R14.

(Note: On some led boards it may be necessary to rotate R14

slightly CW to get the yellow led to light)

'RG'n 1
\_E:t> TSE.G1MS
SET ToT. 'Vtu.ow lel> Begins

UGHT hOito-WM
T*.£lw£EN
 8) Relative Level Set

and expansion
8 . 1) With an input level of -15 dBv at 100 Hz
transition level control so that
control at 1.5, adjust the
At this point, the control should
a level match is obtained.
be at approximately mid-travel

8.2) Repeat for -10 dBv at 1 khz and -27 dBv:at 10 khz

9) Relative Level Response

dBv at khz, and

Set expansion to 1.5 with input to -10
1
9.1)
adjust transition control for level match.

compare level
9.2) Check the output at 100 Hz and 10 khz and
Response Chart
readings with those provided in the Level
specified levels.
Readings should fall within + 1 dB of

Check both channels.

9.3) Repeat with a -30 dB level match.
OvTp'vr qv/tput
Xni»ot *t\ — *t iooh^ KT K3 KHZ

—8.5 db* — Q..O dU

— lO dtv at \ K.H2.

- 30 «itv <xfc \ K.H4. -2S3 dbv -ai.sdw*

1

10) Frequency Response

10 . 1 ) With an input of 0 dBv at 1 khz,

set the expansion control to
20 Hz to 20 khz. The
1.0, and sweep the oscillator from
output should remain flat to within +_ h dB.

11) Remote Simulator

11 . 1 ) Plug the simulator into the Cinch

Jones connector located on
to
the rear panel. Place the remote switch on the front panel
remote.
i

Phase should shift

11 . 2 ) Check operation of the bypass switch (S3)
+ 5
180°. Return to "remote" and output should increase

11.3) With S2 in the 1.5 position and

an input of -10 dBv at 100
to -5 dBv in one position
Hz, SI should shift the output level
and -15 dBv in the other.
the output signal to 10 dBv
11.4) In the 1.0 position S2 should return

Repeat steps 11.3 and 11.4 at 1 khz and 10 khz

11.5)

11.6) Repeat for Channel 2.

 12)

Maximum Output Level

12.1)

With the remote switch on 3BX in the "out position,

< set the oscillator to 1 khz
<a + lOdB and increase the
output level of the test unit (full expansion ) by
adjusting the transition level control until clipping of the
^

on the oscilloscope. A "7 volt RMS

observed
waveform is 13)
occurs,
20V P-P) output should be achieved before clipping
and the waveform should clip symmetrically -

L.E.D. Display

13.1) With an input of + 10 dBv at 1 khz and expansion control

at 1.5, move the transition control from left to
right
and ensure that the leds turn on in sequence. With
transition level fully to the right, all the red leds
should be lit. At no time should two leds turn on at the
same time, nor should a red and yellow led be lit
simultaneously. (Some yellow leds should be lit in both
the high and low band and occasionally in the midband with
the threshold control fully to the left.)
both
13.2) With the transition control fully to the left and
expansion control from 1.0 to
inputs shorted, move the
14) the yellow leds turn on in sequence.
1.5 and observe that
all three
As the expansion control is moved to the right,
bands should have their yellow leds light approximately
at the same time. With the expansion control fully to
right, all
the right and the transition control fully to the
Check to make sure they are
the yellow leds should be lit.
15) With a shorted input and both the
all the same color.
and expansion controls fully to the left, no
transition
leds should be lit. Adjust if necessary.

Output Clamp

should
14.1) When the power switch is activated, a brief pause
observed on the oscilloscope, before the output climbs
be
to it's correct level.

Tone Burst Response

the
15.1) With a tone burst input of 8 cycles on, 128 off and
expansion control set to 1.5, check that the output as
observed on the oscilloscope resembles the drawings
both
provided for 200 Hz, 700 Hz, 3 khz and 7 khz for
control may be adjusted to get the
channels. Transition
correct wave shape.
 16) Wrap Up

16.1) Install tie wraps on all modules.

16.2) Wax all trim pots.

16.3) Fill out route tag and place in Burn-in rack with the
appropriate power applied (120 or 240 VAC) for three
days.

(
 SEQUENCE OF STEPS FOR FINAL TEST

1) Check Power and Remote LED'S and LED Display

2) VCA Level and Distortion

3) Relative Level Set

4) Relative Level Response

5) Tone Burst

6) Noise

7) Clamp

8) LED Display

9) Clipping
 rzio
ZED f ^IcK
Ca 04>9 ye,i-_i2u
"
(^1 »OPuT_£ff£>
ttP 5if Q OEHCE
CR. G>

_J.1I ^o-
<CZ.°t '

Ctl'to
IMSU>Lfi7 C.O
M AXTC?

0 @.V
+ !
t Y.Yi
<3MD i- J>

-=
1 +ZTI/ I

-Hi, 4a*/ IMPJT

 POWER
TRANSFORMER
 t

SCHEMATIC
ItKJW

RHTO

IOK
OAI7 OAt 2.

. M A 04.1
Mil t

1 IOOK p*
C»i 4:
*
!
NLBQ
v .0J» ;
too
-l?v
-?t*i - -1 tt

c ss tmvl I to NlGM <MWO

11,1 •{* filll-C I I OA }2
T ooT
niom
_ 4* 7 |Mfco<>« no oisRtAr

InrM - ,2V
*93 I 1 l ipansioN
JUOK^ *»8 FAC TON
Uo*ow»
Y i BIOS
m
!

RIOS ,

4.TK) t?GM CAHO -

A.19K
HO* 1

if v( L SET ifv *>;{* 1%

AIK -1 2V AlOf
C»t £ CSA,*.** 1 ^SU-COKtC)
,o» 4- toon
JOOH
co cmv*
ixv* MOB MIX _ .

HIGH pi ^- *
w -t 0 £PT 2)
• AND » 3JOH • OOK
CAIN CM
AOJ.
•

(Uti) ftA.e)

CAJ X S*
(NIMOTlJ
«WCO^ ^WCA
CSS
, Ti 1 — .» a0
c
?^rl»4
Mil [TlO/tSV 2W/X5T
.1
,

> ll EL .

-SXL
4AVf3V HO/iTr
a ft
M/X CSO 68* v<5

+ 24V C 44 CS3 (XV T IOO !

(

mei
,

:#**)
TO UO TA. *' 2IOOA <
tOK
OIJRIAT **y*sv 4 10/2 IV 0*13 COiOl

tc **2*
14*012

^
if]
*»*
Vo'S
vtXT*& C
*DJ, i. unless OTMfftwtff XPicirito:
A. ALL RtMSTOA* Afif llWfSHO
IH 0«MS AND AM l/4W,»».

«. Alt OIOOIS AM • 1400X1.

0*11 -OAlt, 0*2 C.OAtT, AM ALL
**«4«tSI
C.
o«i,a»i,o«.a»B,a..o.o»«,A« .uNM.umi
*i CM
o.
p ,
*141

!
«'*
&
¥ 14000*
40121
p22/ttV
1 TA
C.
3 ( LI CT EO PEN tAOOIfe,
QA«, 0*4, 0*«» # OA«,CIAI| # CLAIl^ AAI ALL
***** *M4««*I * MO

silccieo rt« icooo*.

3 JutLx * AMO W. ALL CAPAC»TOA» AH* f IPMIIIO IN MICNOAAAAO*.

It a» *
142)21
liK

art
CD »0S
I. IWITCHI3 AS DRAWN CORMSPONP
mooes lis c o below:
S«: MWtA o*r
10 THE OPERATING

st: remote Off (if- main unit control)

n»o/r
XI S3A,ft: SOURCE
NTMTAj OO* 3 AA 6: XriTCM BTAASS
°~*'o o/r
« INDICATES MCAT3IMM* ONOCVlC*.
Manufactured under one or more ol (he
3.

following U.S. patents: 3.681,618;

3.714.462; 3.789.143; 4,101,849; 4.097,767.
Other patents pending.
 < i t )

SCHEMATIC

Ict4 :«t 7 l ,

,oot .iook r~*

IT GMt,
CO c*7 4r
'
’***
1

«*.eo
ioo
.J
V
I
J-C6S A*
;" mt U io% ;

/3 ^ |
4 se
(Riwote)
« | >*> C 35 C
'T“
«*'« c hi
J .
«« ID M*6H 0**0
T oo? r iOAtK
I CO MIOM 4*7 l»*404t LEO DiSPUr
Rt7S
tORJX -—
4
T CMV
GMV
-'VA- R»a
i*v (DM
high who 1 RIOS ]

level st t -I iV *•»»* !

ttt ± C3A,s.*«r
.o» 4
CO GMT”
4IEV*
HIGH |MI0
•*«0 COK
GAIN CM.fc
AOJ.
• •

Cmt) ti&.i)

s*
?)«*« (REMOTE)

*1*4 > 4»/t»V ITO/tS*} *134 1

3 a,

pGMT
4*Vf5V
C44
ft
4VO//7V
ft
CSS
-*<l
f V)
mzf
GO*
mn
TOO* COK
*7CV»3V 4IO/ESV
EL ~ ft

I. UNLESS OTHERWISE SPCCiritOi

A. Act RESISTORS ARC EKPACSSfO IN OHM* AND AM l/AW, JS,
8. ALL 0I0DES ARC *140031.
C. OAS -OA!t,OA** # OA«7# ARE ALL *I4«*S»

1 1% ^R*4I

,K
«* «**
140*21 140001
<
O. OAI,

C.
OAS, Q AS, OA0 ,0*10,0*11, ARE ALL NPR 4*411*1 ARO
SELECT CD PER 1400*6.
QAt,OA4, QA #*,Q AQ Q Alt , C1AI 1, ARt
,
ALL P«P * 14*141 AND
aLJLuJ l SELECTED PER 16000*.
E. ALL CAPACITORS ARE EXPRESSED IN MIC 140 AAR AOS.

t. SWITCHES A* DRAWN CORRESPOND TO THE OPERATING

MOOES LISTED BELOW:
St! POWEROAR
S*; REMOTE Off <tC. MAIN UNIT CONTROL
S1A,B: SOURCE
°S>orr sea, a: its t m btrass

Manufactured under one or more of the 3. m INDICATES MEAT SINUS ON DEVICE .

following U.S. patents: 3,681,618;

3,714,462; 3,789,143; 4,101,849; 4,097,767.
Other patents pending.
 (CH.I
TO m(AMP
upi op
MOHtfO*
tMPUf i

too