#video test chart

generator
part 2: VGA mode, S-VHS and VGA
.~.~.~
,'••••••
••• ~==~
!I:i••••••
•• ~••••• ~ As explained in Part 1 with the de- though the line frequency is fine, the

•••••
••••• •
il\!••••••

••••••
•••••
~, scription of EPROM codes, the data for
each pixel (R/G/B) consists of bit
raster freguency is a little too high for
most monitors. That can be changed,

•••••
•••• 11
.11 •• 11
••••• I
11•
.
triplets D2/D1/00 and 06/05/04.
Bit D3 is the disable bit for all colours,
a.11owing pixel bits to be used (as byte
however, by a larger number of lines
per raster. Each line reguires astarage
capacity of 284 plxels x 4 bits =
••••• 1 __ 11•• 11. codes) for contral purposes, With one 1,136 bit (142 bytes); so, 354,432 bits for
•••• .,111111111I111111I, J~I•••• exception, D3 is available in all byte 312 lines. The RESET address for TV
•••• I ••••• codes. Only XINS (80Hl is not allowed non-interlaced mode is then 354A3218
••• :,;l •••••
•••
~..
..... ••••
---e·
to disable the colours, because RlGIB
has to be inserted at WSS in
line 23. TI1e hexdump of a
= 44,304 or AD10H. The available
memory for a 16-bit address range is,
however, 524,288 bits (512 k), allowing
vertical sync pulse (Pig- a maximum of 524,288/1,136 = 461 full
ure 1) shows the arrange- lines to be encoded. When the frame
ment of the individual sync pulse is set ta the start of the
byte codes. FH2 (half hori- memory, RESET to address A010H
zonta.l &equency) is always only for TV (88H) and VGA RESET
switched on the falling edge (89H) to 461 lines x 142 bytes = 65,462
of SHV (cornbined horizon- (FFB6H), then the VGA picture has
tal/vertical sync), in other 461 lines at FH = 31,250 Hz. That re-

•1"•
p
words, also on each equaliz- sults in a raster freguency of 67.79 Hz,
ing pulse and raster pulse, which should be okay for any modern
but not when DO is set (de- VGA monitor. Provided the picture
....0;;;;;

••
••
.IJ
tail1). In the VGA modes,
SHV may only contain steady
line sync codes, and no vertical
pulses with equalization pulses. In
content is cleverly arranged, for exam-
pie, colour bars wirheut a circle as in
test chart 2 (see table), then the transi-
tion from 312 to 461 lines is hardly vis-
.11 VGA mcdcs, 04 (detail 2) ensures that
SHV is omitted. In IV mode, D5 has
ible. In VGA mode, TV test charts COI1-
taining a drcle are, of course, reduced
.11 the same function. 03, without 04 er in height about 1.5 times, and com-
•• 1111 1r••

III~ ~II ••
•• 05, causes SHV to appear in TV as pressed. TV test charts 5 and 6 have
well as in VGA mode (detail3). Bast- 625lines per frame, or 312.5lines per
cally the same applles to SV (vertical rastet; and are, therefcre, interlaced.
sync). 06/D3 sct (detail 4) means SV in The TV RESET at 625 Jines x 142 Bytes

•••••••••••••••••
11... 1111.== .... 11111111
both modes. 01 also set means SV
omitted with VGA. The same for 02
and TV mode. The bits allow you to
= 15MH is also applied as VGA·
RESET (89H), the total memory re-
quirement (17 bit) being 1,048,576 bits.
adapt the sync pulses to the relevant With two vertical sync pulses as with
The main subject of last mode in any way you like. RESET TY, we get FH = 31,250 Hz and FV =
should be added: 03/07 works only 100 Hz, 625 lines interlaced (VGA test
month's instalment was with TV mode, and with 00 also set, chart 5). If the raster sync pulse for the
the structure of the PAL with both rnodes. second raster is suppressed, FH is not
As already mentioned, the circuit changed, although FV is halved to
signal supplied by the uses the same EPROM data for the 50 Hz, 625lines non interlaced (VGA
test chart generator. In VGA test charts and the TV tcst charts. test chart 6). That allows the vertical
The essential difference lies in the dou- synchronisation of a multi-sync moni-
this second and final in- bled clock frequency at which the tor to be checked from 50 Hz to
EPROM is read. None of the other 100 Hz. By adapting the number of
stalment we tackle the sub-circuits has to be switched sepa- picture lines, any raster frequency may
VGA mode and a number rately, however, because the 8.86-MHz be programmed.
clock source (front which a11frequen- The line freguency, Ioo. is only Iim-
of non-theoretical matters cies are derived, with the exception of ited by the smallest pixel duration of
the WSS PALpius signal), is doubled to 112 ns. The desired line time is n x
including construction 17.7 MHz with the aid of an XOR gate 112 ns, where n should be an even
and practical use. inside the EPLO. The doubled ad- number, taking into account the byte
dressing rate with VGA test charts code for RESET (89Hl. lf the number of
causes a horizontal (line) frequency of pixels per line is reduced, the pixels
15,625 x 2 = 31,250 Hz, and a vertical become wider, as da any vertical lines,
Design by W Foede frequency (FV) of 50 x 2 = 100 Hz. Al- of course. When the generator is only

Elektor Electronics 10/96

24
 HEX EPROM-CODE 2
451ns/8yle
22Sns/8~le {VGAI
\TV I '"' r, I~I~ 0 1'_1'"

•• •• •• •• •• ••
58 58 58 58 58 58
59
58
58~58
SB-58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
.YXXXXXXXXX
XXXXXXXXXXXXXXXX
I~ 11 ~
SKvTV~
58 58 58 58 58 58 58 58-58 58 58 58 58 58 58 58 XXXXXXXXXXXXXXXX
58 58 58 58 58 58 58 SB-58 58 58 58 58 58 58 58 XXXXXXXXXXXXXXXX
58 58 58 78 78 78 78 78-78 78 78 78 78 48 48 48 XXXx:<xxxxxxxxHHH
I ,.-."'~ 111 11
SVTVlVGA~" .... .. ..
48 48 48 48 48 48 48 SB-58 58 58 58 58 58 58 58 HHHHHHHXXXXXXXXX
58 58 58 58 58 58 58 SB-58 58 58 58 58 58 58 58 XXXXXXXXXXXXXXXX
58 58 58 58 58 58 58 58-58 58 58 58 58 58 58 58 XXXXXXXXXXXXXXXX I 111 11
58 58 58 58 58 58 58 SB-58 58 78 78 78 78 78 78 XXXXXXXXXXxxxxxx SiiV VGA
" ..~ "," ,.
78 78 78 78 59 58 58 SB-58 58 58 58 58 58 58 58 xxxxYXXXXXXXXXXX
960076· T6a
HEX EPROM-CODt:
CKAR
10
, .
" :'1:1 ,• ., 51

Figura 1. The hex

used to test computer monitors, the Apart from the preprogrammed ICs dump of averticai
frequency of crystal X2 may be in- for this project, a diskette is available sync pulse provldes
creased to about 15 MHz, or even which not only contains the standard useful information on
higher depending
EPROM will tolerate.
on what the test charts as shown in Table 1, but also
data for 20 more test charts. The rele-
the structura the
individual byte codes.
0'
vant files contain the data for a half-pic-
THE TEST CHARTS ture (raster). Using the instructions also
The Altera EPLD type EPM7032 is sup- provided on the disk you should be white areas are available in the rest of
plied ready-programmed through our able to build your own test charts. The the screen, monochrome areas are not
Readers Services under order code disk also contains a number of basic pat- present in the colour bars. The vertical
966507-1, or through your kit supplier. terns such as sync pulses without pic- line at the transition from green to vi-
The functions programmed into this ture content, a circle, a rhombus, and a olet covers the unavoidable interfer-
Ie are indicated in Figure 2. The grid, as well as precise indications about ence caused by the quarter-line offset
27C040 EPROM containing the data picture start, pieture centre, picture end, which is inherent to PAL. The pieture
for the test charts listed in Table 1 is WSS, and mcre. The test charts are pro- is programmed using only one bit per
also supplied ready-programmed grammed in Pascal because circles and colou.r. Consequently a green bar is not
(order code 966507-2). characters are not easily produced man- possibJe, unless the colour contrast is
Test charts 1 through 4 occupy ually from just pixels. reduced to nil. The replacement is the
512 kbit each. numbers 5 and 6, 1 Mbit The following text discusses test dot raster for focussing. The micro-
each. The EPLD is automatically reset charts numbers 1 (Figure 3a) and 5 rhombuses in the corners of test chart
at power-on, so that test chart 1 ap- (Figure 3b), which are universal charts, nO.5 serve the same purpose. To en-
pears in the first instance. TI"e order really. The foundation is a white grid able you to run a check on proper
and the possible memory space for on a black background, for checking helper-Iine prccesstng. the helper
each test chart is preprogrammed in the dynamic convergence and picture range in test charts 5 (Figure Sb] and 6
the EPLD. A 3-bit counter, ACTR4, al- gcomctry The verticallines in the grid (Figure 4b) has a text insert which
lows you to step through the test have the required width of 225 ns. The reads 'PALplus'. It appears as a white
charts. Depending on the state of this outer bounds of the grid are marked ident in the grey bar in test chart 5,
address counter, A16 of the EPROM is by white rhombuses. When the geom- and in the black bar in test chart 6.
driven either by AIS and ACTR3, or di- etry is correctly adjusted, these 'dia- When not decoded, the text 'Palplus'
rectly by ACill4 to effect the test chart monds' shou1d just fall outside the vis- appears in blue in the helper onJy. Test
selection (see Figure 2). ible area. Because sufficient black and chart 6 also shows you which hel per
lines originate from the basic pieture.
The centre cross helps you to check
static convergence. Next, the multi-
burst has the smallest pulse width of
,,, PAl, COOSA
225 ns (TV) or 112 ns (VGA). The or-
1

1 '
ange-coloured line raster has alternate
red and yellow lines which serve to
I. sirnulate at least the hue according to
'.' 1
G- Y (approximately). Next comes a
I'
,
, BIT

' ..
f----------L--~~
.,,-+ -5HV/SH
black pulse on white. The Une rasters
beside the red/yellow transition
line-wise blue/yellow, and operate just
are

f-------------~.+5HVI'~V
1 like the usual non-ecloured areas. With
the right phase positions of the gener-
ator and receiver colour subcarrier fre-
quencies, these areas should be as

Flgure 2. S/oe/( dla-

gram of the pro-
grammed EPLD. Apart
from two quartz oscil-
lators, thls IC con-
talns ma/nly counters
In comblnation wlth
FH2 FlESET some control logic.

Elektor Electronics 10/96

25~
Figure 3. Test chart 1 colour-free as board is available ready-made through switch. However, parts 05 and K3 are
(a) Is the basic, unI- possible. The our Readers Services. Mind you, the then required.
versal, TV test chart. structure of the final design 01 the board has a few lay- Whichever supply is used, the
Test chart 5 (b) Is the black/whi te out changes with respect to the first ground plane 01 the peB and the
PALplus version. rhombus and prototype shown in the photographs screening at the inside of the enclosure
the entire pic- in last month's instalment. should be connected by at least one
ture in general The board should be populated mounting screw.
is such that the circle covers as little with great predston because errors are The test switch on the modulator
white area as possible. Immediately not easy to trace later. With ver-
after the vertical pulse of the first tically mounted parts, short-cir- Table 1. Test Charts
raster, all test charts have two test lines, cuits may be prevented by cov-
a colour bar and a multiburst up to ering the lang terminal with 150-
No. Pattern Application
4.43 MHz. By selecting TV raster trig- lanng pieces of wire sleeving.
Non·;nterlaced (NI)
gering, and operating the oscilloscope Take care to observe the polarity
1 Standard TV test chart Universal
in 2-channel mode, it is possible to of diodes and electrolytic capac-
2 Cotour bars Faultfinding
analyse the colour stgnal and the video itors, and the orientation of the
frequency response, independent of voltage reguletor and all other 3 Micro diamond Focussing
the aetual test chart used, and without Ks in the circuit. TI1e Ks shouJd 4 Red Cotour processor
a second timebase, Because both lines be fitted as the last parts, and it
are sure to occur in the blanked f1y- is recommended to use Je sock- Interlaced/Line jump (I), PALpius
back period, they also allow the func- ets, 5 as 1; 16:9 with PALplus ident
tion of the cut-off to black during the Two options are available for 6 as 2; 16:9 with PALplus ident
flyback (picture blanking). VGA mon- the power supply: either a small
itors are not particularly strang in this pes mount transformer is used, VGA test charts:
area! Fortunately, that 1Snot a problem OI an external adaptor which is 1 as TV 1; FH ~ 3125kHz, FV ~ 100Hz NI
because the video signal is usually capable 01 supplytng 9 Vdc at a
2,3,4 as TV 2,3,4; FH ~ 31.25kHz, FV ~ 68Hz NI
nought during this period. current of about 150 mA. If the
5 as TV 5; FH ~ 3125kHz, FV ~ 50Hz NI
trans former is used, the pes
CONSTRUCTION support pillar in the corner near
A double-sided through-plated printed terminal block K4 must be a plastic Figure 4. Test chart 2 (a)
circuit board (Figure 5) was developed type. The alternative is not to secure shows colour bars and Is
for this project to keep the construc- this corner of the board at all. If the sultable for faultflndlng.
tion as simple and compact as possible mains adaptor option 1S chosen, the The same for test chart 6
despite the relatively high frequencies following parts are not required: TrI, (b), which shows the
used in the circuit. The printed circuit K4, BI and, of course, the mains on/off colour bars In PALplus
format wlth an ident.

Elektor Electronics 10/96

26
5
,0) K2
960076-1

0
r-evooee

G
0

COMPONENTS LIST C5 = 100J.lF 16V radial .Miscellaneous:

C6,CS,C9,C12-C17,C20,C21,C24, Kl ::;::15-pin sub-O cormector, VGA,
Resistors: C25 = 100nF high-density, angled pins, PC6 rnount
Rl = 100kQ Cl1 = 15PF K2 = 2-pin PC6 terminal bl.Qck, pitch
R2,R19;R26 "" 1kQ C22 = 1000J.lF25V radial smm
R3,RS = 10kQ C23 = 22pF trimmer K4 = 2-pin PC6 terminal bleck, päch
R4,R7::;:: 2MQ2 C25 = 47pF ceramic (se~Hext) 7.5mm
RS,R6,R14,R16,R20,R22,R23, S1 = pushbutton, debounced (Conrad
.R24;;; 7SQ Semiconductors: o/tl 70 76 00.a8)
R9 = 47kQ 01,02,03 = lN414S Xl = 5 MHz quartz crystal
Rl0,R13,R15 = 390Q 04""LEO. X2 =..S.S67238MHz quartz crystal
R11 =iOOQ , OS ee 1N4002 (see text) Trl == Mains transformer, 6Y, 2.4VA,
R12 = l5kQS 1% 61 = 6S0C1500 (round) Hahn.
R17 = 27kQ ICi= EPM7032LC44·15 (Altera) (order UHF modulator (Conrad oln 192791-88
R1S,R21 = lS0Q code 966507-1) or similar)
~5 = .. x 15QQ SIL array IC2 == CXA1645P (Sony) Screened ease 1S0xSOx45rnrn (Ham-
103 = 27C040 100ns (order code mond 1591 OGY or 1591 OS6K)
Capacitors: 966507'2) Isolation ~t for Tl, T2
Ci,C19= 100pF IC4 = 74AC4040 Printed clrcuit board, programmed ICs
C2,C3,C7,Cl0,C1S = 220J.lF 16V radial IC5 = 7805 (966507-1 and -2) and diSkette
C4 = 220nF 966011-1: Set order code 960076-C.

,., ..."P.... .......

Figure 5. lTaek layouts
(750/0)and eomponent
mountlng plan of the printed
eireuit board. Note that the
final layout dlfferssllghtly
trom that of the prototype
shown In last month's Insta/-
ment. Double-sided.
through-platedboard. aßiI ..
able ready-made through the
Readers Services.

unit is converted into a TV/VGA mode

selector. Details of the conversion are
shown in Figure 6. First, the non-used
supply pin for the internal RF ampli-
fier is 'freed' by cutting the track to
pin 5 of the modulator at the under-
side of the module panel. As indicated,
a wire connection is made from pin 5.
Because of the wire jumper, the pole of
the switch is permanently at ground
level so that the test mode is no longer
available. The third free switch contact
now switches the VGA signal applied
to pin 5 via the isolated wire.
The non-used RF input socket on

Elektor Electronics 10/96 27

 connect
the modularor is turned into an audio
signal input. This requires 'freeing' the
centre pin by a second track-cutting
operation, and connecting it, via a
length of isolated wire, to the audio
Input, pin 2. The external signal is in-
jeeted via R4 (2.2 MQ). The irnpedanee
of the modulator AF Input is high
enough to a1low address signal All to
produce an auctible reetangle wave of
abou t 550 Hz. When an external AF
sound source is connected, the inter-
nal signal is practically quenched be-
cause of the low impedance. The AF
input sensitivity is about 500 mV at -'"
o
47 kil, allowing cassette decks, CD
players and the like to be eonneeted.
~
:;
The internal test tone may be muted o
by short-circuiting
the AF input Flgure 6. As de-
socket. serlbad In the
Flgure 6 also text, the modula-
shows an addi- tor panells
.tional ceramic ca- sllghtty modlfled •
paeitor of 47 pF !:ItiUUfti· zu
which appears as
CONNECTION terferes with the picture) when the TV
AND ADJUSTMENT is fine-tuned in the + directian. Be-
cause the modulator produces a dou-
Apart from the antenna output of the ble-sideband signal, it is possible to
modulator, a 15-pin sub-D output tune to the wrong sideband. The
socket is available to connect a VGA sound subcarrier is probably pre-ad-
monitor. The eVBS and Y stgnal are justed to 5.5 MHz and may require
~ output on non-used pins of the sub- tuning to 6.0 MHz as used in the UK
~ D socket. The C signal is available on The test tone may SQw1da little harsh
pin 11. With VGA, the colour subcar- because it is derived from a rectangu-
rier is switched oft leaving this pin Iar signal - this is normaL The modu-
o available for IDO. Having a level of just lator bias (C) is also Iactory-adjustedIf
200 mVm/47 kQ the AUDIO signal is you want to fine-tune this setting, note
hardly likely to cause interference at the colour yellow, the sound and the
AUDIO
the IDI pin (15). The remairring pin as- synchronisation. Yellow is a good in-
[BI'"'
U liGA o
IN/OUT
signment corresponds to the VGA
standard. Table 2 shows an overview
dication of the video signal being
clipped at the peak-white level.
of the most commonly used video con- (960076·2)
nectors and their pinning.
960076·19
The test charts are se-
Table 2 VIdeo connector pmout overwew
lected by pressing 51. TI,e
onlyadjust- Video generator SCAR7"E" SCART SCART DIN HOSIDEN
Figura 7. Connections rnent point DC15HD (VGA) SNHS cvas RGB AV' S·VHS

C25 in the parts list. and adjustment points on the Pin 1 +R 15

This capacitor must of the UHF modulator. board is Pin 2 +G 11
be fitted with the trimmer
shortest possfble lead C23. lt is ad- Pin 3 +B 7

length at the indicated Iocation. With- [usted for a colour sub- Pin 4 NG (102)
out this decoupling device, the modu- carrier frequency of Pin 5 +FBAS 20 2
lator picture suffers from sound carrier 4.43361875 MHz in TV
Pin 6 GNOR 4, 17 4,5
interference, which is particularly an- mode. lf you have an ac-
noying in the deeoded PALpius signal. curate frequency meter, Pin 7 GNOG 14 4, 17 9, 13 3 1,2
The modifications should be carried the signal may be mea- Pin 8 GNOB 18
out with great precision, and checked sured at pin 6 of IC2. Pin 9 +Y (NC) 20 3
with an ohmmeter. Mind you, that's Without this instrument,
Pin 10 GNO
less of an effort than having to remove you may also adjust the
the modulator from the finished trimmer for the best pos- Pin 11 C (100) 15 4
board. sible colour reproduction Pin 12 NC (101).
The five connecting pins of the in the test charts. The ad-
Pin 13 -SHV(-SH)
modulator are bent at right angles. The justment points of the
modulator is then fitted Flush on the UHF mod ulator are Pin 14 +SHV(-SV) 20

board using four solder pins at the cor- shown in Figure 7. The Pin 15 AU (103) 2,6 2,6 2,6 4,5
ners. In addition to the connecting UHF ehannel (A) is ad- GNO screening 21 21 21 3 5
pins, four mounting pins at the top jus ted such that the test
RF Audio/Out AudiollN
corners are soldered to the board. chart does not become
blurred (or the sound in- * links pins 1 and 5 in plug

Elektor Electronics 10/96 29

64-channel logic analyser
(May 1996, p. 35-43, 960033)

Constructors using the ready-made print- 960033-2d

ed circuit boards lor this project should
note that capacitors C25, C35 and C45
were not included in the circuit diagrams,
PCS layouts and parts lists as printed in
the magazine. These capacitors afford
additional supply decoupling, and should
have a value 01 10!1F,16V.
Channels 48 through 64 (probe D) are
not available because IC40 can not be
selected. The problem is simple to solve
by connecting pin 28 (CSTI)and 32 (CSl)
01 IC40 to ground (see drawing). The cir-
cuit diagram on page 38 should be cor-
rected likewise.
Finally, on the main board, copper
tracks run very close to the board mount-
ing hole near pins 49/50 01connector K4.
Care should be taken notto cause short- Video test chart generator Oscilloscope prescaler
(October 1996, p. 24-29, 960076) (November 1995, p. 28-34, 950115)
circuits here by PCB spacers or screws.
The S-VHS output may oscillate. This A number 01 readers have reported tim-
problem may be solved by litting a ing problems with the RAMs used in the
330-pF ceramic capacitor between circuit. For these RAMs, a short period
Matchbox BASIC computer junction R16/C3 and the ground con- appears to be necessary between the
as data logger 'address stable' and 'write enable low'
(September 1996, p. 18-21, 960065) instants.
Dwing to a conversion mistake in the elec-
tronic page layout process, all underscore Two solutions are available:
characters (_) have disappeared lram the 1. Use the type GM76C2BA-10 lrom Gold-
listing on page 19. Readers wishing to star in position IC13. This RAM chip
obtain a Iree copy 01the corrected listing was also used in our prototype.
(on paper) may apply to our Customer 2. Modily the PCB as lollows:
Services department in Dorchester. - desolder the socket for IC9;
- cut the connection between pins 1 and
2 of IC9 at the component side 01the
board;
SIMM tester - fit a new IC socket;
(February 1996, p. 18-26, 960039) - connect a short isolated wire between
II last SIMMs are tested, a bus conllict pin 2 01 IC9, and pin 10 011C6.
may arise, causing a latch-up situation
and an incorrect message stating that the The latter solutlon causes a delay 01
SIMM is laulty. This ~ happen because 23B ns on the WE line, enabling the cir-
buffer IC13 uses the RD signal to reverse cuit to work with RAMs having a specifi-
its direction, while the SIMM does so cation other than tas= 0 also.
using the WR signal.
This problem may be solved as lollows: nection 01 C11 (at the underside 01
a. Disconnect pin 1 01 IC13, and connect the board, see drawing).
itto pin 12 01ICB (a non-used inverter). The value 01 capacitor C14 has to be
b. Disconnect pin 1301 ICB, and connect increased Irom 1OOnFto 470nF.
it to pin 2 01 IC7 (WR). The modulator case has to be soldered
c. Disconnect pin 19 01 IC13, and con- to the ground plane 01the PCB. This may
nect itto pin 2 01 ICB. be achieved by litting solder pins near the
corners 01the modulator (drill additional
holes), or by removing the protective lac-
Quer in these locations, and sold er the
Keyboard swap for PCs modulator case directly to the ground
(June 1996, p. 40-43, 950126)
plane.
Secause resistors R5 and R6 may lorm
a too large load lor IC1, the PC may not
receive any1hing although the LEDs indi- U2402B ballery charger
cate that one 01the keyboards is active. (April 1996, p. 10-15, 950120)
This problem may be solved by increas-
ing the value 01 R5 and R6 to 1kQ. II the In the circuit diagram on page 12, the
LED intensity is reduced too much, high- switch idenüfiedas S2a (near R22) should
efficiency LEDs should be used. be S2c.

Elektor Electronics 12/96

86