Satellite Workshop 22
It is the lightning season as I write this. My workshop has been
filled with "lightning damaged" receivers, most of which had merely
refused to start up after being unplugged from the mains power for a
while.
However, the MSS100 was a different story. It worked perfectly but
there was no "E to E" (terrestrial UHF loop-through). This had the
effect of producing a very grainy, colourless TV picture, although
satellite channels were perfect. However, the owner was not satisfied
with only 250 channels. He wanted the other five as well!
The aerial input amplifier, U201, is a surface mount TDA8725T
(909-0872501). I had to replace this as well as the 39 resistor, R219
(940-3900501) and 10nF capacitor, C227 (950-1035621). While I was
putting the screws in, I had a phone call from Wossname up Church
Street.
"Gorranaarreff mod fer an MSS100?" he enquired.
I explained, gently, that the MSS100 does not have a separate RF
modulator but he insisted,
"Course it 'as." Where d'you think you stick the aerial plug, you
great oaf?"
I gave up at this point and suggested he ring CPC or HRS. Let them
deal with it! I just hoped that the MSS100 owner would not bring it
to me after Wossname had "looked" at it.
Meanwhile, the one on my bench was now reassembled and working
perfectly.
Another MSS100 which arrived a few weeks ago had me stumped for a
while. It announced "No Signal" on a blue background. Now, this is
often caused by a dry solder joint on D18 or, occasionally, a broken
solder joint on the tuner input pin. In this case, however, it was
neither. Also, the LNB voltage was present, so obviously it was not
the usual trick of selecting "LNB power OFF" in the menu!
I turned off the blue screen to see what was happening, by pressing
"F" then "store." There was a marked difference between the noise on
the picture with the LNB connected and with it disconnected. That
implied that the tuner was probably all right but was stuck on an
out-of-band frequency.
The design of the board is such that a tuner with serial data control
could be used. Currently, however, the receiver still uses analogue
voltage control with a divided frequency feedback loop. This
operation is handled by the frequency synthesiser U100, an LM7001 as
used in SS9xxx series receivers. The tuning voltage is determined by
FET Q100.
A quick measurement revealed that the 30 volt supply to Q100 was
present. However, Q100 was permanently off. Before jumping in and
changing Q100, I used my 'scope to check U100. Interestingly, I
discovered that crystal X100 was not oscillating! Changing this 5.625
MHz crystal restored normal operation.
The MSS1000 which arrived by carrier had fared less well. Not only
had the power supply been killed by the surge, but the damage had
been increased by the owner who decided to have a go at fitting a
kit. Even time-served bench technicians have problems with power
supply kits so I was not surprised to see the mess beneath the board
on this one. Several copper tracks had been peeled away and dark
brown flux was interspersed with solder splashes.
Allow me to digress while I educate those of you who feel inclined to
attempt the fitting of a kit. Firstly, the idea is to leave the unit
looking exactly as it did before you took the screws out. You should
bend the legs of resistors and diodes at right angles so that they
match exactly the hole pitch and don't rip off the pads when you
insert them. Bend the legs towards each other, just as
the factory machinery did. Bending them outwards can create short
circuits. Do not leave any component standing proud of the board.
Vibration could fracture the tracks. This applies especially to the
heavy switching transistor. Even some "engineers" leave this on
stilts. Wrong! It should be pressed down until the shoulders of the
legs touch the board. Some designs rely on the copper tracks to take
away heat so you must keep the legs short. Leaving any component
standing proud will allow it to move and break the tracks. The only
exception to this is the pluggable resistor, which is designed to
stand up, and also some high dissipation components whose legs must
be kinked to secure them before soldering.
Lecture over. The MSS1000 had suffered horribly. The optocoupler had
been fitted the wrong way round. The wires of the TL431 adjustable
zener had been used to bridge the gaps between the holes and the
tracks because no solder pads remained. I replaced all of the badly
fitted parts then began my round of measurements. Three surface mount
resistors were open circuit: a 10 , a 1k and a 470 . The silicone
insulator behind the SSP5N90 switching MOSFET had been torn.
Finally, when I was happy that no damaged components remained, I
plugged the unit in to be greeted by a blank screen containing only a
row of zeros. The front display also contained a row of zeros. This
is a classic symptom of a fault on the I2C bus which the
microcontroller uses to transmit and receive serial clocked data from
other I.C.s in the circuit. It's usually indicative of a faulty
EEprom or MSP3400 audio processor. However, on this occasion, it was
a faulty "Nicky" video I.C. next to the tuner.
Since the EEprom had become corrupted, I used my Pace Link computer
system to upload the standard programming. Finally, I upgraded C11
and C12 as recommended by Pace. (856-1082750 and 856-1583750
respectively). These are ultra-high reliability types and C11 is
increased in value to 1500 F. C216 1000 F/63v should also be
replaced, in my opinion, as it is highly stressed.
As I write this, it would appear that Grundig has handed over its
spares operation to CPC, just as Amstrad did last year. It will take
a while before they can decide what to stock, consequently many spare
parts are currently unavailable. For example, my workshop contains a
GRD300 awaiting a vacuum fluorescent display, a GRD280 awaiting a
microcontroller and various "Minerva" and "Matsui" receivers which I
have part-exchanged and rob for spares since there was no chance of
repairing them at an acceptable price. Hopefully the situation will
be corrected by the time you read this. Willow Vale, of course, will
still supply Grundig parts for out-of-warranty items.
In the heat of summer, a jovial gentleman skipped into the workshop
with a PRD800 tucked under his arm.
"Ho ho ho," he chuckled. "Doesn't like the heat. Keeps scrambling the
pictures. Ho ho ho."
He passed the offending receiver to me and I almost dropped it
because I was distracted by his attire. He was wearing nothing more
than a pair of shorts, a baseball cap and a large beer belly. A
horrible sight made worse by the fact that he was red as a beet
root!
He chuckled his way through the ticket completion procedure and then
took himself back, presumably, to his sun lounger.
Dutifully, I went through the motions of soak-testing the receiver
but, after six hours, it had not shown a fault and the day was
becoming cooler. I finished off my tenth repair of the day then
phoned the jovial one.
"Ho ho ho," he explained. "It does it on QVC, mostly. Her indoors
doesn't get out much nowadays so she likes to watch that."
I left the receiver on QVC overnight and, sure enough next day, the
picture was scrambled. I unplugged the receiver and let it cool down
for an hour before powering it up again. The picture was there but I
noticed tiny little speckles - not "sparklies", which are those
tadpole-shaped spots you get with a poor signal, but tiny dots on
bright colours. These were limited to the centre two-thirds of the
screen. As the receiver warmed up, the speckles became worse until,
finally, the picture scrambled. However, the other Sky channels
remained perfect for several hours before they, too, scrambled, one
by one!
Without boring you with the details of how many I.C.s I swapped and
how many curses were uttered, I traced the cause to the link wires,
J1 to J4. These are factory set by cutting certain combinations - a
sort of fine tuning for the decoder "cut and rotate" points.
Apparently one link had been cut by mistake and this was the
result.
I phoned the jovial gentleman to tell him the news. He didn't chuckle
at the price.