Inferno #07 31 мая 2005

Железо

Repair - Faults Pentagon 128 + and their repair.

     Failure ... Pentagon 128 +


   When you turn perfectly cleared, chickens
sor runs through the options, but at the entrance of the BASIC
1948 system hangs, the screen is blank, not painted
is called (not trying to lose). Small
warming - and it works. You can enter from
BASIC 128 (User SPECTRUM), but the USR
0 hangs. Warms up - works.

   All devices are functioning properly (me).


   Once served a Pentagon 128 +,
in Tower'e. Propayal, obkonderil, put ku
Chu sockets (at MAGIC and printer), not dopayal
how detaley.Pri test comp worked
well, BUT! After turning on bismuth at the entrance
in BASIC 48. Reset perfectly, options ne
reklyuchalis, but when you log in BASIC in 1948 screen
not turned black, and remained empty. Through time
screen was black, but the system hung - in general,
Computer heated a few minutes and
began to work normally. Interestingly, the
at the entrance of the BASIC 128 (via user
SPECTRUM) BASIC 48 worked.

   First suspected ROM (remembering that
had, like, everything worked immediately,
inclusion), but he doubted it. Decided
check. Included a computer, came from the B128 to
B48. Overwrite the contents of ROM code, the logic
RESET, did USR 0, making sure that
Comp still buggy (the screen is not even turned black, the 
system theme once hung). When you understand the codes with

horrified to find that the code is correct, and reset
should work! But he did not rabotal.Komp with
warmed, and I found in the normal state
Rom. Repeated the experiment several times.

   Horror! Rom is stable, normal program
Nye, but the reset does not work! Through the options and
through USR 0. Under Suspicion processor and
appeal to the DOS. Check something, and another
(Changed the processor, called the scheme) - all normal
mally. Changed the ROM, AY and took out the SH-shku not
pomoglo.Podozrevayu logic (broken mikruha)
but obviously did not catch on for anything. Decided otsle
dit address test failure, a USR.
Consistently inkrimentiruyu (the code, for
instructions) the address and start ... Until the screen
does not even turn black, comp just hangs. Found it!
The test clearly starts with the same address - and yes
more. There seems to be, with instructions LD I, A Work
em, and before that with LD A, 67 - hangs. Analysis
ruyu address bits: like, to anything without taking
tsya, of course, logic can fail. Dialed in
Memory prog: LD A, 63: LD I, A: RET. Start
- Hang! A damn yourself! Try different values
radiation Register A, for some all DC
for some immediately hang, under certain
not all at once (not with the first call)! Explicit 
bitzakonomernosti not see, but there provoka function fails for 
some values ​​of accumulators torus. Triger blocking DoS 
(arrogantly close the entry) - and all DC Find a breakdown:


   Resistor from AG3 to TM2 (470 ohms) is not under
veshivaet entrance Triger - a cliff. But when
testing device - a little tucked - all soybean
Dinh, not pridereshsya.V entrance catches up
interference that are sufficient
For some reason when the corresponding value of re
Giustra I. Invisible to the eye crack doro
LCD on the board. Propayal trim, fat
tin snot. System is perfectly zarabo
tala. When you turn on every time was a
BASIC 48, and he was discharged well and work
Tal. Case closed, but a direct link between
failure and execution of processor instructions
not found, perhaps a special case
indirect effect.

KSA-7G
April 4, 2003


    Malfunction. Pentagon 128 + (my)


   The system is not reset, the on-screen garbage
and, possibly, any spontaneous melteshenie.
Border color can be changed, if discharged into
B.128. After heating may attempt
reset distorted print menu, blackening
screen when you reset B.48.


   It was about six months ago, in autumn
2002. I do not remember why, but I'm a week
did not include the Speccy (in any case, remember
long period of time before sex
s). So, I decided to include, cut in, and the
screen garbage instead of the menu (I have always
when you turn on reset is activated.) Especially
surprising that before any recognition
Cove, predicting the fault was
ie PC works fine and should have been pa
bot and now! But it is broken ...

   I opened the computer, meryu food, potryahi
vayu board, fan, draw in the ROM sockets,
but nothing works. The screen is ...
that in the DEMO can not be found - the spontaneous
texturing, the multicolor individual
familiarity ... Climb gloomy thoughts about the
swept memory (it's still good) or
some small logic that the answer
directly in the short (and just did not check)
period of time.

   But then I notice a spot on the board. At zapy
Universe board it looks like a better
and clean place - starting with the series of memory and
a little further to the multiplexer, just
in the middle of the board. Spilled something, probably
not wet to the touch (already dried up), prinyuha
lsya ... KOSHATINA!

   But how he managed through the body ... Osmatri
vayu cover, on which four stacks of words
my wife diskety.U standard, grayish,
square, plastic body (base,
cover, back strap, side unit with circuit breakers
Thelema, the front two panels - solid and
cut under FDD). In the mid-caps rose
Lena black plastic logo. Rising
Lena neplotno.I through her CAT Cologne
leaks inside the company, and got on the board. Bu
bottom marks on the inside of the lid, and
on the outside, where they were discs, visible band of
jet.

   Kotyara, of course, a sniper! He managed
splash between stacks floppy just settlement
middle of kryshki.Zhidkost little raspolzlas
from the center and wet (total) 3 diskettes.
One, still damp, I then thoroughly washed,
others lightly soaked konverty.Dalee Ms
dkost leaked through the logo, and yet
fell on platu.Elektrolity certainly includes
in the organic matter plus water, which is a group
Lew forms a very conductive environment. Set
set of relationships can not quickly
switch logic, t.k.energiya pulse
spent on the electrical dissociation of liquid
bone, and this process (because of inertia
ions) can not pass quickly (often with
of the system). As a result, the signals are stretched
in time and acquire additional
indirectly dependent on one another. These and
due to absurd behavior.


                  Repair


   No alcohol and water, this infection is not otmo
eat. Just have to wait until she
dried, a process that is very not a quick.
Cat rare marks, but for how long.

   To speed up you can turn on the system, and
blow. She warmed up, and to establish those
ploobmen SYSTEM-liquid-air interface. But! When
Moreover, under the influence of the currents will be formed
t.n.provodyaschie channels. Ions can vtravli
vatsya to the tracks that will worsen the isolation plate
you are as a whole and make it more chustvitel
plane perpendicular to the external moisture. And since by 
evaporation fluid leaves a relatively large number of energy

energy, 7 watts boards as a heater for laughter
Similar. What is included, there is - the difference is small
visible and obvious damage to the motherboard. So I re
shil wait.

   Still, something did. I had
a few small bubbles with clear
field of magnetic heads, and later I bought one
large bottle, and this little thing (bubbles) I
no longer required. I decided Prom Dresses
it (seemingly) isopropyl spirtom.Ved
if you reduce the concentration in one koshatiny
Mr. infested, place the board (on blur
larger area), then weathered this stuff
rapidly, and its concentration on the board will
weaker at safer. Of course, it will affect
more tracks, but with less
force.

   I was not sure that the alcohol will dissolve it
infection, but decided. I pull no fee
steel, and poured directly into the body content
on board, carried the little brush in hand
chkoy, lift the board and wiped down.
Vbuhali in this case, probably 5 or 6 of the bubble
Cove. Then half an hour blew through the fan,
for fun included, but the complex did not work;
slightly adjusted ROM (suddenly disturbed concentration
cycle) and a payment processor (AY, joystick,
Reg. volume and Z80).

   Each day included, tested, the system does not
worked. Passed, probably a week, when
memory was properly cleaned, but the menu
B.128 not vyvodilos.Ya snachalo worried
Is not broke anything? But in another
Day two brought the system menu. And even entangled
disco! But then freezes ... Attempts
dumping steel sokraschatsya and half an hour
there was only rubbish on the screen. It is hard to you
Sail the liquid from the chips ... C ka
zhdym day increased reliability, system
the system worked all longer (up to the first gluon
Cove, and then I turned off). Why glitches manifest
were determined not immediately? A small amount of moisture
reduced the concentration of active ions, but
system is heated, the activity-increasing
(With temperature), and a conducting channel
Nala outsourced ions, increasing
electrical circuit and signals. In
particularly when the system is gradually arrive at
Deal in un buggy state, but discharged,
observed abnormal superior printing
On the menu B.128, with the edge (on the left, it seems) from 
the tsentra.T.e.glyuchila memory in the system

facil variables.

   In total it took 10 days (no more than
two weeks), since the system is reminiscent
signal glitches on this story. But it is clear that
Now it is more Sensitive to moisture, because
salt remained, and only their solutions,
to cause a new failure. I filled emblem
several layers of colorless TSAPON Lac,
but kotyara may splash from the side, through pro
resident. Of course, this is unlikely, but there
he was able to casually (just passing) on
jaws on the lid of the system, not above or below.
Patch can not be slotted, the system should
ventilated. You can put a burn
DKI to directly to the board did not
fell, but so far the way it left.


                   Total


   If the system suddenly stopped working
it is quite possible (at least 30% if
system is well-collected), this is not the
It (or rather, not her fault) and the outside
shnem factor. Especially if it left
unattended for long periods of time.

The archive repair
KSA-7G
May 3, 2003


    Malfunction. Pentagon 128 + (my)


   System buggy may drive ports,
do not run the program, or hang ...

   Then do not reset computer, and when about
heats up - dumped B.128, and (switching
tea on a bit 4/7FFD) - full B.48, but no
resets (hangs). B.48 incomplete (without
CAPS).

   Sometimes when you reset B.48 flies into DOS,
type USR 15616.

   A reset can increase the stack on the screen,
counter-ray.

   Features: How long did not include dust,
mud system MAGIC'om.


   Broken off the day before yesterday! Arrived (was not
week), the first day of the complex is not included, then
turned and began to run the discs. One ispor
Teal, put a strip of protection. Continued,
programs are not loaded, hang in the middle,
Even MAGIC (BATTY) in 48 mode: boots,
the ball hit the brick, he disappeared, and the system
Theme hung, spun the disc ... Through bore
How many minutes (ў5) unsuccessful attempts to triggering
ka system suddenly normally earned ...

   Just in case (maybe horrible:
dust particles of solder ...) opened the system and
slightly blew, shook his fee. Since
system ruined the disc, looked connector
(FDD A) and lead to an external FDD. On conduction
de dangling tubing and some were exposed
torye ends - put kembriki.Moglo and coronal
It is noteworthy long time there have not tested it, but doubts 
wasps talis - yet not so easy to spoil

disc closure plug, it was hardly room
complex circuit.

   Any not correctly identify the fault
sequence probably still be apparent ...

   The next day, the system not a dump
LAS! Hanging cellular debris, and revolves FDD
A. My first thought - flew the SH-scale. Although, with
Why? Maybe once burnt, or
protection circuit (diodes) are dead, flew away? Not
discarded and 128, and 48 BASIC. I have ne
switch is suspensory output bit 4 /
7FFD to positive, to reset once in BASIC, 48
with the open pages. While wandering thought
that the controller is in order.

   Inspects fee prinyuhi
vayus ... Feline spirit smells, not the word
- BE!. However, examination of the body, saw no
Fresh signs of cat spray, hence the smell
old - half a year has passed, BO! a persistent
y Murzik feline LABEL. I opened the Speccy,
check the protection of SH-shki (DC), shoot the cake
(Z80, AY, KJ), shoot the SH-shku, insert the Z80,
include ... The system still does not work.
Garbage on the screen, RESET increases stack
accidentally gets on your screen. Somewhat
Torture does not yield a result, taking the scheme. Reconcile 
their ROM-shki signals: a signal ROM 2 (27 N). from

Port 7FFD by blocking DOS is, the signal
signal-DOSEN (1 N.) arrives, the foot 20 should
be in the red, meryu, and suddenly the system sbra
described by (periodically shed it myself)
almost perfect match, but the contact
20 feet is excellent. In general, the system is warming up
las, reset brings up a menu 128.

   Instinctively switch as B.48, sbra
describe the - Wisit! Switch back and forth carrying
How many times, drop, eventually interesting
movie ... B.128 reset, B.48 - No, but
if you log in from B.48 B.128 - everything is normal
(USR0 - O.K.). It is interesting that a full reset
(Press CAPS) operates as part-time - hanging
sometimes flies into DOS (type USR 15616). On
rank to check the choice of ROM.

   Signal-DOSEN (1 N). Almost immediately after
reset switch to 0. Sometimes B.48
sbrasyaetsya, and then-DOSEN stable soda
rzhit 1. So, pinpoint the problem,
check the formation of-DOSEN.

   It is formed on the stem 11 D83 (LL1) in
Trigger-state DOS and addresses
A15, 16. A little confused about the logic, then
I specify:-DOSEN = 0 (which is not true when you reset
B.48, the fault my computer), only
However, if working memory (which is logical, because
RESET! ") AND! included DOS! (Which should not
be at fault).

   Obviously, something is included with DOS Throw
ce B.48. This may be incorrect Adre
relaxation of the sample DOS, or MAGIC. MAGIC easier
check. Meryu leg 17 (-NMI, Z80):
small short-term (tens of mV)
Drawdown 1 at reset (several times about
believing / shed) indicates the start of
MAGIC'a, which should not be. Interference? Dis
kovod I took long ago, when the complex opened, and
on the screen nothing changes, so MAGIC
was not evident, probably also because
SH-shku I took).

   Check the condition of the launch: legs 1 and 2
D78 (AG3) when meryu leg 2 (signal you
Birky ROM addresses (A15, 16), the system starting
is reminiscent to work - the device reduces the noise on the
line, slightly prosazhivaya RF signal (tens of
MHz). Well, I think everything has found a bad contact
signal to the leg 2 bad comes ... About
a note an interesting feature: the system
theme starts and when checking the knife
ki 1 (-MAGIC), which feeds the emitter
KT361 (my scheme of MAGIC to 0 - with a total mi
cone). The system also works if the measure
base KT361 (this is my-MAGIC with buttons
through a resistor).

   Button checked, not short. Gradually
system is warmed and became more frequently discharged
tsya (partial discharge B.48). Remain som
equation with a two foot (D78). Turn off the system,
check the signal - loss! FOUND!

   Disconnect and take out a fee. It is in the dirt,
wool dust. Regret not having brought with him
alcohol, should be washed, but there is nothing, at least go!
On the reverse side of the creeps moth flew away,
bastard! Likes warmth as well as a cockroach, and often
settles in a computer ... Creature!

   Full termination signal is evident! I myself
it created when something - a track cut.
Because logic element (D77, legs 1-3)
is faulty, I took the signal from its analog -
D65, legs 4-6, 6-foot jumper thrown.
Ping (Z80 - D65 - D78), I convinced
lsya proper operation. So where, then,
fault? Perhaps in obveske D78 (re
Zyuk away). Once upon a time I had all the fees
propaival, maybe that came off ...

   Obveska in order! But the resistor R10, with
feet 1 (D78 AG3) at plus another nomination
LA - not 1 ohm and 10 k! The scheme must work
thief (before after work). But now the plate
that more dirty than before, and possible
leakage currents that are commensurate with such a large
resistance as 10 ohms. When I incorporation
ival MAGIC-0 (KT361), I have not checked the results
transistors, thought corresponds to the scheme, it was necessary
then clarified. In addition, the base is actually
hung in the air and catching interference. On my,
obkonderennom, comp them a bit, and all work
TALO, although I remembered the case when on the outside
shnego inclusion (davnoo. .. it was a unit
lnik seems to be included) worked MAGIC.
Obviously, the scheme has become more muddy and less
stable.


                  Repair


   I hung the base KT361, rather, the very CCW
PKU (the base of the button through a 10 kOhm) to positive
at 1 ohm, resolder R10 1 kohm and indie
fication latch port (not rezyuk - LED
to minus, as before, but from the port of shunt diode
induces LED with rezyukom, suspended for
plus). clean, pochertil space needle
in between the findings of AG3, removed some of the chlorine
pya pyli.Postavil fee, the SH-shku, drive.
Turned and checked the reset, ran the program
We - all perfectly working! Turned off the system
subject, put the cake (see above), all closed.
Everything works fine.


                   Total


   Failure per se was not. Ver
She was - but not systemic, and biological
Kaya, and temporary.

   When the system heats and draws
insects. They carry the moisture that
at pollution constitutes a false contact, and
unsolicited communications. Warmed System
vaporizes a tiny amount of moisture, which
PoE goes into it from insects. Themselves in
insects do not sit on the board tracks, and Paz
tsya under the card or in the vicinity, they are also reluctant 
to get under the current, albeit weak, nizkovol

tnye ... However, if the system is then permanently
remains off (as I was leaving), then the population
komye can safely wander on board
extending from it moisture, which dissolved
satisfies the dust settled on the computer (it is on the 
floor), and forming a salt-conducting electrolytes.

If you then turn on the system, it will flow
spurious currents. Usually they are weak, and the signals
one logical exit not squander
others through a false connection that
can behave like a container-loading
current logic output. Sensitivity to noise
Indeed inputs rather than outputs. But if the limit
dpolozhit that all inputs are connected to vyho
DAMI, they are not susceptible to false Relations
- Each input is protected by its yield, which
ry compensates its output current of all
false, low leakage currents and parasitic
capacity. All would be fine, but there are no entries
outputs, and they are so well protected. This
entrances, waiting for the signal.

   Inputs such as keyboard or Joist
ka, or MAGIC. The latter is the most dangerous -
as Interference may interrupt the
the desired program. The keyboard is securely protected
schena BASIC'om (more on the 48th, with the problem
Contact the keys, I was convinced of the reliability
Poll KEYBOARD'a, buggy only at the very
poor contact). Joystick for serious
operations is not used, and it should be significant
with respect to position (cursor). Of course,
bend are crooked written, but these glitches
interface is usually obvious. But MAGIC -
here how it goes, it is not a priori preduga
give. Either circuit is soldered to the mind, in good
state, and will work, or wait syurp
Rizov.

   So why are the inputs to the Sensitive
Interference? It's simple - their entire input current
ma small. He's made up of current entry
integrated circuits (Series K555 up to 0.5
mA, and the K1533 could barely be measured, the K155 can
be more than 1mA, but there are exceptions) and current
resistor, which hangs in - usually
but to plus. Rezyuk usually no more than 10 ohms.
Like, on the 48th on the clave are. Take
mikruh K555 series, then interference current = current
Input + current rezyuka. Hence: the current noise = 0.5 +
+ Current rezyuka. A current rezyuka: the ratio of the UNE
masking power to his opposition. Current
noise = 0.5 mA +5 V / 10 ohms. It is equal to 1 mA
- And it's maximum! Enough noise by 20%
more slaboy.A when you consider that the false interactions
imosvyaz can behave like a capacitor,
the bond may be weaker for it
ku, as when you switch to a single state
false charge capacity, and when switching
Research in another state recharge, SFO
rmirovav short and strong enough
(Using the energy of the previous charge), they
pulse interference. That may be enough for a
Start the same univibrator AG3 system
MAGIC, and I have happened! Interference was
from a sample of addresses (A15 OR A16), and POSCO
lku reset is performed differently, the conditions
interference (the change of address
core processor to memory) is different, and faster
it was addressed to the RAM, at partial discharge
B.48, the more a hindrance, which vyzy
Shaft MAGIC.

   If the input is normally connected to the output, rather 
than hanging, like management or MAGIC, even substantial 
overcharging by bellows - a capacitor can not affect the

work t.k.tok recharge flows from the exit
and can reach much higher ve
quantities (when switching) than the current hanging
input; intermittent overcurrent you
move quickly recharge the interference at a
short vremya.Fakticheski output level
only slightly prosyadet from interference, but not to
stignet Switching value and logically
will not change.


                  Output


   People who do not feel sorry for alcohol ... not for its
insatiable womb, its you and the other rubbish
be able to ease off. Just occasionally,
at least once every two years, rinse the board.
Avoid the hassle, save time
and nerves. Liter of industrial alcohol in our
RADIOPROFI worth a total of 55 pe. He calmly
(Not as acetone) evaporates, take a simple
thuja brush, open the window (now
spring, as you know, there's still summer and late
Fall ...) or log on to the balcony. Dunk
brush into a bottle with alcohol (often, Poch
alkaline ... do not lick!) and clean clothes.
So that it is all wet was visible, with two
parties. Then wait for 5, well
let 10 minutes, the alcohol evaporates, and eaten
dirt will collect in the form of small lumps.
Simply shake a fee, you can brush
it obmahat. You'd be surprised how little alcohol
spent ... You are not so many FSUs
melitsya ...

The archive repair Speccy
KSA-7G
May 2, 2003