[NTLK] Tantalum caps

Frank Gruendel newtontalk at pda-soft.de
Sat Mar 26 13:56:55 EDT 2011


> I got my hands on a dead 2100 for the case and some other plastic goodies.
> But when i took a look at the board, there was an obviously blown cap
(C147).

C147 is one of the components that tend to most eagerly express their
annoyance at having to deal with a wrong power supply.

Apart from that, using a wrong power supply frequently leads to what Forrest
describes here:

	
<http://www.anasazi4st.com/NewtonPhoenix/Newt_Tech/Entries/2010/11/21_Franke
nNewton-Reviving_A_MessagePad_2100.html>

If you want to fix this particular problem, simply read on...

Frank

-- Newton software and hardware at http://www.pda-soft.de



Warning: This post is a little long. And occasionally a bit technical. Then
again, maybe the info is interesting for you. I do think that even if it
isn't, it might be nice to read.

The following production has cost me one solid week's worth of nights...

##### Cast listing: #####

Frank			Slightly dumb hobby service technician who earns his
living by pretending that he is a programmer
Paul			The incredible Guyot, technical adviser
Dale			Mainboard provider
Andreas		Mainboard provider
Martina		Amazingly patient wife, who has every right to expect other
activities at Easter

##### Crew listing: #####

Misc.			Apple Newton hardware developers who could have
prevented all this
Microscope		Located in Florida. Bought from Germany in an eBay
auction, sent to a colleague's sister's house in Florida, picked up there on
the occasion of a holiday trip and lugged back to Germany
Weller WSD 50	Frank's wonderful soldering station
Voltcraft 3650	Frank's multimeter
BeeOne, BeeTwo	9 Volt batteries for multimeter
BeeThree		9 Volt battery, because two aren't enough for such a
project
Reddy and Blacky	Tiny test probes bought for this very project at EUR
15.00 each (ugh...)
Gossen Model 1	Frank's decades-old adjustable laboratory power supply


##### Awards: #####

None


##### Home video releases: #####

None


##### Plot summary : #####

(Spoiler warning: This will unravel one of the most mysterious Newton
mysteries).

Set in 2006, the story begins with Frank's getting yet another Newton on his
workbench that shows a classic defect: The Newton tries to power up, chimes,
changes his mind, switches off again and starts over. Usually this is
repeated twelve times, then the Newton gives up.

The defect in question used to be repaired by David Watson, also known as
Dr. Newton. Unfortunately, David seems to have dropped off the surface of
this planet, making it impossible to get at information that would be
necessary to fix this problem.

Up to the day that marks the beginning of the story, Frank has never been
able to fix such a board. He has always had to replace them. The result is a
small cardboard box containing 5 Newton 2000 / 2100 mainboards that all
suffer from the same disease.

Frank briefly contemplates making the 6th one join its 5 unfortunate pals in
the box, but changes his mind. Preparing and accomplishing a large-scale
expedition to fix one single Newton mainboard would of course make no sense
whatsoever. Frank could make much more money in much less time by pretending
to be a programmer.

On the other hand, preparing and accomplishing a large-scale expedition to
fix SIX Newton mainboards would make no sense whatsoever, either. Frank
could still make much more money in much less time by pretending to be a
programmer.

As a consequence, Franks starts preparing a large-scale expedition into the
depths of Newton hardware. Comparing every single component on a dead board
to it's respective component on a working board isn't exactly a professional
way of doing service, but with no technical information whatsoever at hand
it is the only way available to dumb service technicians.

The next day sees Frank putting a Newton mainboard on his employer's copy
machine and experiment with different settings until he has a decent
photocopy of each side of the board, enlarged to a size of about 0.4m by
0.3m. After all, even a dumb service technician is aware of the danger of
forgetting which of the hundreds of components has already been tested and
which hasn't, so he wants to mark the tested ones on the copy.

On his way home Frank secures himself to BeeOne, BeeTwo, BeeThree, Reddy and
Blacky.

0.24 square meters of table space not being available close to the
workbench, Frank tacks the copies to the wall, resulting in his wife's
tendency to frown with the facial expression of someone suffering whenever
she passes them, which happens approximately twelve times per hour.

To save time, Frank decides that, while he is at it, it would make sense to
not only compare two boards, but three, the third one being one that has
another standard defect: The Newton does not start up. Brainwipe works, but
the Newton never starts up after it. Frank has an inkling that these two
defects are only two different results of basically the same problem.

Equipped with three 2100 boards and a pen, Frank fires up his trusty
multimeter and sets out on his leap in the dark. For a start, he decides to
measure and compare only standard semiconductors like transistors and
diodes. Resistors die far less often, the death of electrolytic capacitors
is usually visible and smellable, small capacitors hardly ever die at all,
and the sheer amount of pins of integrated circuits is likely to make him
give up before he started.

Two and a half nights later all diodes and all transistors have been
compared. Unfortunately, nothing out of the ordinary has been found.

The logical choice would now be to start comparing the integrated circuits.
But the amount of pins dejects Frank, so he decides to have a go at the
resistors.

He strikes gold at about 3am during night #3. R120, which should have a
resistance value of 36kOhm, has a resistance value of 70kOhm. Replacing it
fixes this particular mainboard, which will make cast member Dale G. very
happy because Frank has promised to fix his Newton for free if his mainboard
happens to be involved in a successful outcome of Frank's crusade.

Overwhelmed by his success, Frank decides to call it a day and sleep a bit
for a change.

As of then there's one mainboard less to compare. The bad news is that
there's still one left, because R120 is perfectly healthy on the boards with
the repeated boot attempts, of which Frank still has six.

However, one leg of R120 is connected to a bit of circuitry around a Maxim
MAX1771 chip, which is an adjustable high-efficiency DC-DC controller. The
design around it is in exact accordance with one of Maxim's application
notes for this circuit, which allows Frank to determine that the Newton
developers have intended the output voltage of this circuitry to be 5 Volt.
This voltage is missing. Unfortuntately, it's missing on the working board,
too. It doesn't seem to be THAT important. Hgrmpf...

Now the incredible Paul Guyot enters the picture. Having always admired his
amazing knowledge of the internals of the Newton OS, a long time ago Frank
has asked him if he has an idea what exactly happens at the end of the
startup chime (remember that the Newton always powers off immediately after
the chime). Back then the incredible Paul Guyot had answered that the Newton
is checking the presence of a voltage required for the PCMCIA card slots.
Back then Frank had compared the slot connectors of a dead and a working
board and handn't found any difference.

But...

5 Volt is a voltage that IS required for some types of cards. So Frank
checks if there are any PCMCIA specific ICs on the board that might handle
the slots' voltages. This is a rare streak of genius for a dumb technician,
because he finds one SI9712 chip for each PCMCIA slot. This is basically a
switch matrix that routes 3.3 Volt, 5 Volt or 12 Volt to the card slot pins.

Frank measures that the 3.3 Volt and 12 Volt are present, but the 5 Volt
aren't. Unfortunately, they aren't present on the working mainboard, either.
Hgrmpf...

By another rare streak of luck Frank has another look at Maxim's application
note and notices that the MAX1771 has a shutdown pin. Since the Newton
usually shuts all unneeded voltages off to save power, it would make perfect
sense to shut card-related voltages off until they are needed. Which leads
to the third streak of luck, namely Frank's measuring this voltage while the
Newton is powering up instead of when it has done so.

Lo and behold, on working Newtons the 5 Volt supply is turned on for about 2
seconds (probably for allowing the Newton to check that it is there), and
then turned off again. On the dead mainboard the 5 Volt supply is NEVER
turned on.

The question now is whether this is because on the defective board the
Newton doesn't switch the MAX1771 on, or because it or the circuitry around
it doesn't do its job anymore.

Drunk with happiness Frank decides to be unable to concentrate any longer
and hits the sack. This is the end of night #4.

The next day Frank does something that should only be done if... if... well,
actually, it should NEVER be done. He sets crew member Model 1 to 5 Volt and
connects its output to the output of the comatose MAX 1771. After having
made sure that the fire extinguisher is still where it used to be, he
switches both the Newton and Modell 1 on at the same time.

Nothing happens. No sign of life whatsoever. No smoke, either, but a totally
dead mainboard. F*ck... This is the end of night #5.

One night later. Frank notices that there's no ROM board in the mainboard.
He had removed it for his soldering and completely forgotten about it.
Slightly ashamed of himself, he puts it back and repeats last night's
experiment.

Lo and behold, the Newton powers up. No more cyclic boot attempts anymore,
it just powers up like it should and smiles at Frank. It even stays on and
works hunky dory if Model 1 is disconnected after the powerup (unless,
probably, although Frank doesn't try this, cards are inserted).

This, quite obviously, is a major breakthrough. Although it doesn't fix the
defect, it allows to power the Newton up in a way to backup one's data,
which so far had never been possible.

This is the exact time when Frank decides to change his schedule at short
notice and insert an hour of celebration in company with a couple of bottles
of decent Single Malt scotch whisky. Since this hour somehow becomes a bit
more than just one hour, this is the end of night #6.

The last roadblock now is, of course, finding out WHY this voltage is
missing on the defective mainboard. Frank decides to approach this problem
in a somewhat pragmatical way by simply replacing the MAX1771. By now he
firmly believes to be kind of entitled to a little success that hasn't cost
him nights. Greedy as ever (and because it's Sunday and there's no chance to
buy a replacement), Frank cannibalizes a MAX1771 from an organ donator board
and solders it on the defective board. He then makes sure that the ROM board
is where it should be.

Biiiiiiingooooooo...

This is the end of night #7 and the beginning of another celebration
session.

Final scene. Frank, humming under his breath, has just finished fixing
mainboard #6. Every single one of these six fellows has been fixed by
replacing U34.

Heard it? U34!! MAX1771!!!

Boy, one of these days I will start accepting donations...

Frank

-- Newton software and hardware at http://www.pda-soft.de




More information about the NewtonTalk mailing list