? Fixing Burned Flatbed Scanner Power Resistor

[Alec S.]

Hi,

Apparently a recent thunder-storm has burned out my flatbed scanner (no thanks
to my fancy-schmancy mid-size APC surge protector).

I checked to see if it was the power-adapter, but that's fine. I opened the
scanner and looked at the circuit board and it is indeed burned. It seems that a
large resistor next to the power connector has fried and the coating crumbled
onto the board. The board in that area is a little discolored as well. I've
attached a couple of snaps to show the damage.

Anyway, I am hoping that I can salvage the scanner by replacing the resistor,
but I cannot determine what value it is since the coating is gone. Can anyone
give me some advice on this?


Thank you very much.

 

[Arfa Daily]

Hi,

Apparently a recent thunder-storm has burned out my flatbed scanner (no
thanks
to my fancy-schmancy mid-size APC surge protector).

I checked to see if it was the power-adapter, but that's fine. I opened
the
scanner and looked at the circuit board and it is indeed burned. It seems
that a
large resistor next to the power connector has fried and the coating
crumbled
onto the board. The board in that area is a little discolored as well.
I've
attached a couple of snaps to show the damage.

Anyway, I am hoping that I can salvage the scanner by replacing the
resistor,
but I cannot determine what value it is since the coating is gone. Can
anyone
give me some advice on this?


Thank you very much.

You can't just 'attach' pictures on this type of newsgroup. You need to put
them up on some webspace somewhere, and post a link. For what it's worth, I
think it unlikely that the burnt resistor, which may or may not be your
problem, was caused by storm damage. If the pcb is discoloured under it,
then this resistor is just a hot-running device as a normal situation, which
given the location that you suggest for it, is quite likely. It is not at
all uncommon for the cement coating of such thermally stressed resistors, to
crumble away as you describe, leaving a very distressed looking component,
which often is not electrically faulty at all. Are you quite sure that the
power adapter is "fine" ? Have you measured its output voltage with it
connected to the scanner ?

Arfa

 

[Alec S.]

Arfa Daily wrote (in

You can't just 'attach' pictures on this type of newsgroup. You need to put
them up on some webspace somewhere, and post a link.

I don’t know if attachments are blocked in this newsgroup, but I forgot to
attach them, so they weren’t there anyway. Here’s the pics:

http://img153.imageshack.us/img153/4051/img3525dm0.jpg
http://img153.imageshack.us/img153/8240/img3526xv5.jpg

For what it's worth, I
think it unlikely that the burnt resistor, which may or may not be your
problem, was caused by storm damage. If the pcb is discoloured under it,
then this resistor is just a hot-running device as a normal situation, which
given the location that you suggest for it, is quite likely. It is not at
all uncommon for the cement coating of such thermally stressed resistors, to
crumble away as you describe, leaving a very distressed looking component,
which often is not electrically faulty at all.

Are you quite sure that the power adapter is "fine" ?

Not really, when I did the “quick test”—aka licked it to see if there’s juice—it
didn’t have the same kick as the third party adapter. By fine I meant the DMM
said it had the right voltage.

Have you measured [the power adapter’s] output voltage with it connected to

the scanner ?

No, I measured it on its own. It is a 12V, 1A brick. It read ~13.3V, and slowly
rising (I got tried ~14V and stopped). I also tried an 800mA variable adapter
set to 12V (which read ~14V). Neither one causes the LED power indicator to
light up or the scanner to be recognized by the computer.

I just tried to read it while connected, but connecting it caused that resistor
to get REALLY hot; it burned my fingers. I checked the resistance across the
power connector’s terminals to see if it’s shorted. There’s three terminals
(which I don’t understand, there’s only two conductors in the wire). I get 65
Ohms between front and back, 0.5 Ohms between front and side, and 65 Ohms
between back and side (in the picture, the “back” is facing us).

 

[Arfa Daily]

Arfa Daily wrote (in

You can't just 'attach' pictures on this type of newsgroup. You need to
put
them up on some webspace somewhere, and post a link.

I don’t know if attachments are blocked in this newsgroup, but I forgot to
attach them, so they weren’t there anyway. Here’s the pics:

http://img153.imageshack.us/img153/4051/img3525dm0.jpg
http://img153.imageshack.us/img153/8240/img3526xv5.jpg

For what it's worth, I
think it unlikely that the burnt resistor, which may or may not be your
problem, was caused by storm damage. If the pcb is discoloured under it,
then this resistor is just a hot-running device as a normal situation,
which
given the location that you suggest for it, is quite likely. It is not at
all uncommon for the cement coating of such thermally stressed resistors,
to
crumble away as you describe, leaving a very distressed looking
component,
which often is not electrically faulty at all.

Are you quite sure that the power adapter is "fine" ?

Not really, when I did the “quick test”—aka licked it to see if there’s
juice—it
didn’t have the same kick as the third party adapter. By fine I meant the
DMM
said it had the right voltage.

Have you measured [the power adapter’s] output voltage with it connected
to

the scanner ?

No, I measured it on its own. It is a 12V, 1A brick. It read ~13.3V, and
slowly
rising (I got tried ~14V and stopped). I also tried an 800mA variable
adapter
set to 12V (which read ~14V). Neither one causes the LED power indicator
to
light up or the scanner to be recognized by the computer.

I just tried to read it while connected, but connecting it caused that
resistor
to get REALLY hot; it burned my fingers. I checked the resistance across
the
power connector’s terminals to see if it’s shorted. There’s three
terminals
(which I don’t understand, there’s only two conductors in the wire). I get
65
Ohms between front and back, 0.5 Ohms between front and side, and 65 Ohms
between back and side (in the picture, the “back” is facing us).

I pretty much agree with all Jeff L says in his reply. You can ignore the
fact that the socket has 3 terminals - they generally all do. The barrel
connection is a 'sprung' one which not only serves to 'anchor' the plug when
it's inserted, but also forms a switch to the remaining terminal, when there
is no plug inserted, which allows it to spring back and make contact with
that third terminal. It allows for the batteries to be disconnected on
portable gear, when an external power supply is used. This is important to
prevent dry batteries being 'charged' by an external supply. You can also
use a diode to isolate batteries, but that loses you around 0.6v of the
battery's terminal voltage, so the switched socket is generally preferred.
In the case of your scanner, it makes no odds, as you will see, if you look
at the print carefully, the side and back terminals are just shorted
together, which is why your meter reads 0.5R. It's the best it can manage,
to display what is effectively the resistance of the meter leads and probes.

A reading of 65R both ways across the 'active' connector pins, does seem a
little low for a scanner in an idle state. Assuming a 12v supply, that
equates to getting on for a 200mA draw. I agree with Jeff that although you
can't actually see where the burnt resistor goes, it probably is straight to
that 3-term regulator at the back. That being the case, I would not expect
to get an ohms reading that low between its input pin and ground. What is
that device's number ? 7808 or 7805 perhaps ? If so, you need to measure the
voltage between its tab and its left hand pin, to confirm that you have the
same voltage as on the side of the resistor remote from the input socket,
and then measure between the tab and the right hand pin, to see if you have
whatever the numbers on the reg say (78- 05 = 5v, 78- 08 = 8v). If you
do, then the problem is downstream of the regulator. If you don't, then the
problem may well be the regulator itself. These regulators do have overload
foldback protection on them, but you have to be up at over an amp of draw,
before that kicks in. Another possibility for the problem, could be C44
being leaky. It appears to be connected straight across the incoming supply.

Beyond checks finding a fault at this level, it is unlikely that you will
get to the bottom of the problem if it's elsewhere. Once you are past the
regulator, you're into the realms of faulty LSIs ...

Arfa

 

[Franc Zabkar]

Once you are past the
regulator, you're into the realms of faulty LSIs ...

If you look at the print around L5, there are at least two sections of
the PCB that see the full unregulated input. IMHO that's potentially a
big concern, although it could be just supplying the motor(s) and
lamp.

- Franc Zabkar

 

[Alec S.]

You can ignore the fact that the socket has 3 terminals - they generally all

do…

Thanks for the explanation. I actually figured out the anchor reason shortly
after posting, but the other one makes sense too.

A reading of 65R both ways across the 'active' connector pins, does seem a
little low for a scanner in an idle state. Assuming a 12v supply, that
equates to getting on for a 200mA draw. I agree with Jeff that although you
can't actually see where the burnt resistor goes, it probably is straight to
that 3-term regulator at the back.

Yes it does. I’ve posted an overhead shot of it here:
http://img204.imageshack.us/img204/5918/img3570nz1.jpg The resistor goes from
the + terminal to the input pin on the 7805, apparently passing through an SMT
resistor or something (B32—is there a site with a list of standard PCB text
identifiers?)

That being the case, I would not expect
to get an ohms reading that low between its input pin and ground. What is
that device's number ? 7808 or 7805 perhaps ? If so, you need to measure the
voltage between its tab and its left hand pin, to confirm that you have the
same voltage as on the side of the resistor remote from the input socket,
and then measure between the tab and the right hand pin, to see if you have
whatever the numbers on the reg say (78- 05 = 5v, 78- 08 = 8v). If you
do, then the problem is downstream of the regulator. If you don't, then the
problem may well be the regulator itself. These regulators do have overload
foldback protection on them, but you have to be up at over an amp of draw,
before that kicks in.

It seems that the input voltage pin is shorted to the ground pin and tab at 1.8
Ohms—the ground pin is connected to the tab at 0.3 Ohms. The power adapter that
came with the scanner does indicate a current of 1000mA and it is an
always-connected/on device (the power LED is always on, but the light is only on
when in use).


So it’s probably the voltage regulator that died? I don’t have another 7805 on
hand (all I’ve got at the moment is a 7905, oh so close). I’ll keep my eye open
while scavenging and try it out.

 

[Alec S.]

Any particular manufacturer or model scanner?

It’s a UMAX 2100U (with a PowerVision PV6098F IC).

Methinks there's more blown than just the resistor. Judging by the
devastation, it got very very very hot over a fairly short period of
time. To me, that means a short or high current draw somewhere else
on the board. My guess(tm) is that the burned resistor is in series
with the input of the nearby 3 terminal regulator. It's intended to
reduce the heat dissipation of the 3 terminal regulator (because
there's no heat sink). One lead goes to the power supply connector
hot +12V pin. No clue where the other end goes.

It goes to the 7805’s input voltage pin
(http://img204.imageshack.us/img204/5918/img3570nz1.jpg) through SMT component
B32—which seems a little crispy. When I connect the power adapter, the resistor
does get extremely hot. Also, a 470µF, 16V cap towards the bottom of the board
also gets really hot. When I pull the power, the resistor cools down quickly,
but the cap stays hot for quite a while. The voltage regulator does not get
discernably hot.

Even if the resistor were replaced with the proper value, my guess(tm)
is that it will smoke again. The most obvious shorted culprit is a
fried 3 terminal regulator, but I'm guessing.

It very well could be. It is only 1.8 Ohms across the input voltage to the
ground (and tab).

Anyway, judging by the internal (spiral) construction, my guess would
be a 3-5 watt resistor, with a value of less than 1 ohms. All the
overheating has done is crack off the ceramic coating. The resistor
is just fine and the correct value will be whatever it measures. It's
fairly difficult to blow up such a big resistor.

I’ll try to find another 7805 to replace this one and see if that works. All I
have at the moment is a 7905. Doh!

 

[Alec S.]

Regulators are cheap. It might be better just to replace them.

I’ll try to find another 7805 and try that.

If the unit is pulling "too much" current, a shorted input cap or regulator,
as has been suggested, is likely the problem.

Is there a fuse anywhere?

No there isn’t. I’ve uploaded a couple more (~1MB) snaps of the board:
http://img204.imageshack.us/img204/5918/img3570nz1.jpg
http://img207.imageshack.us/img207/4535/img3558sc8.jpg

 

[Franc Zabkar]

Yes it does. I’ve posted an overhead shot of it here:
http://img204.imageshack.us/img204/5918/img3570nz1.jpg The resistor goes from
the + terminal to the input pin on the 7805, apparently passing through an SMT
resistor or something (B32—is there a site with a list of standard PCB text
identifiers?)

I suspect that B32 is a ceramic (b)ypass capacitor connected between
the input and ground pins of the regulator.

In a similar fashion, B24 appears to bypass the unregulated +12V
supply to ground after it passes through L5.

- Franc Zabkar

 

[Alec S.]

DaveM wrote (in

Is the regulator itself (the 7805) getting hot?

The voltage regulator itself does not get hotter. I have tried using both the
12V, 1000mA adapter that came with it and a 500mA variable adapter set to 12V.
The stock adapter causes the resistor to get really hot, really quickly, and the
capactor warms up a bit. Using the 500mA variable adapter causes the resistor to
quickly heat up a lot and the capacitor gets much hotter than with the 1000mA
adapter. Yes, I let it cool down between.

I’ve taken a couple more snaps to show the the whole left (right?) side of the
PCB from top and bottom. The capacitor in question is C1 at the top of the first
pic.

http://img388.imageshack.us/img388/2324/img3587vh1.jpg
http://img83.imageshack.us/img83/3206/img3588xy1.jpg

Your statement that the 470uF capacitor gets hot is your big clue here.
Capacitors in a linear power supply shouldn't get hot. I'd wager that the
capacitor is shorted (or very leaky). If you trace the PCB wiring, you might
find that the capacitor is the input filter from the wall wart.

It seems to be a multi-layer PCB, so I can’t identify all the paths. I don’t
have DMM with cap-testing ability, but I did the simple resistance test and it
seems to check out.

If it's connected to the input pin of the regulator, it's definitely the input filter.
Replace the capacitor!!!!

Do you suppose it could be capacitor plague? It is only slightly bluging at the
top; it doesn’t seem to have vented at all. Here’s a closeup of the cap:
http://img388.imageshack.us/img388/6310/img3590qt9.jpg

 

[Alec S.]

Franc Zabkar wrote (in

I suspect that B32 is a ceramic (b)ypass capacitor connected between
the input and ground pins of the regulator.

In a similar fashion, B24 appears to bypass the unregulated +12V
supply to ground after it passes through L5.

Thanks. I tried some Google queries like “d1 u1 l1 q1 r1 c1 resistor capacitor
inductor diode transistor” and got a fairly good list of PCB labels. I still
wish I had a proper list of standard labels. I’ll have to check a couple of
electronics engineering texts later, maybe they give them in the schematrics
sections.

 

[Franc Zabkar]

Do you suppose it could be capacitor plague? It is only slightly bluging at the
top; it doesn’t seem to have vented at all. Here’s a closeup of the cap:
http://img388.imageshack.us/img388/6310/img3590qt9.jpg

I don't understand why you are having such a hard time of it. You have
measured a dead short at the input to the regulator. Desolder the
regulator's input pin and test for a short again. If the short remains
after desoldering the regulator, then desolder other connected
components until the short goes away. The last component will be your
culprit.

- Franc Zabkar

 

[Jim Yanik]

Is the regulator itself (the 7805) getting hot? If not, the problem
is before the regulator.
Your statement that the 470uF capacitor gets hot is your big clue
here. Capacitors in a linear power supply shouldn't get hot.

nor should those in a switcher supply.

 

[Alec S.]

Franc Zabkar wrote (in news:[email protected]):

I don't understand why you are having such a hard time of it. You have
measured a dead short at the input to the regulator. Desolder the
regulator's input pin and test for a short again. If the short remains
after desoldering the regulator, then desolder other connected
components until the short goes away. The last component will be your culprit.

Sounds good, thanks. I don’t have an electronics bench, so I can’t _just_ do
that. I have to get out my iron and wait for it to heat up and so on. It’s a
whole production, which is why I am saving up steps/tasks to do together
(including getting replacement components). I’ll try your suggestion the next
chance I get.

 

[Alec S.]

DaveM wrote (in

Forget about trying to identify a component by its reference designation.
Those designations are for schematic and parts list identification only; they
have absolutely no bearing on the component's value. The only "standard"
thing about them is the use of Cxx for capacitors, Rxx for resistors, Uxx for
ICs, Qxx for transistors, etc.

That’s what I meant.

And even that's not standard... As I said, it's up to the manufacturer.

Yes, unfortunately that’s what I’ve read. I was trying to find a list of common
lables because a lot of surface mount components look identical, so it can be
difficult figuring out what a component is. Even bigger components look alike.
For example on this one, L5 looks an awful lot like a resistor to me.

A capacitor doesn't have to burst open and ooze its guts all over the place to
be defective... If it's getting hot, replace it. If it's bulging, replace it.
If it's bulging and getting hot, REPLACE IT!!!

As Franc suggested, start unsoldering components and check resistance/voltage
until the problem goes away. Replace the component that made the problem go
away. If I were you, I'd start with C1.

Really? C1 is all the way down there on the other end of the board. It doesn’t
look like it’s directly connected to the power area. The negative terminal is
connected to pin 11 of the IC U2, and I can’t tell where the positive terminal
is connected to (it connects on one of the middle layers).

Anyway, I’ll get break out my iron and check those three main ones (the cap at
C1, the regulator a Q2, and the resistor at R58) tonight or tomorrow. Hopefully
it is indeed the capacitor because I’ve got a ton of spare caps, so it shouldn’t
be too hard to find a replacement.

 

[Alec S.]

DaveM wrote (in news:[email protected]):

A capacitor doesn't have to burst open and ooze its guts all over the place to
be defective... If it's getting hot, replace it. If it's bulging, replace it.
If it's bulging and getting hot, REPLACE IT!!!

As Franc suggested, start unsoldering components and check resistance/voltage
until the problem goes away. Replace the component that made the problem go
away. If I were you, I'd start with C1.

Okay; I’ve removed the capacitor, voltage regulator, and resistor. There still
seems to be a short between the I and G pins of the voltage regulator’s place on
the board (the component itself is not shorted, just its holes on the PCB). Here
’s a close-up of the area:
http://img292.imageshack.us/img292/9303/img3593rj7.jpg

To me, it looks like the short is at B32 (either that, or caused by the residue
of the resistor). I was also wondering if the short could be between the layers
(in which case I guess I’m SOL).

Any ideas? Is there a good way to clean this up? I was going to try some rubbing
alcohol/nail polish remover, but figured I’d wait.


Thanks.

 

[Alec S.]

Franc Zabkar wrote (in news:[email protected]):

You have measured a dead short at the input to the regulator. Desolder
the regulator's input pin and test for a short again. If the short
remains after desoldering the regulator, then desolder other connected
components until the short goes away. The last component will be your
culprit.

Okay; I’ve removed the capacitor, voltage regulator, and resistor for now since
they seem to be the primary suspects. There still seems to be a short between
the I and G pins of the voltage regulator’s place on the board (the component
itself is not shorted, just its holes on the PCB). Here’s a closeup of the area:
http://img292.imageshack.us/img292/9303/img3593rj7.jpg

To me, it looks like the short is at B32 (either that, or caused by the residue
of the resistor). I was also wondering if the short could be between the layers
(in which case I guess I’m SOL).

How can/should I clean this up? I was going to try some rubbing alcohol/nail
polish remover.


Thanks.

 

[Arfa Daily]

Franc Zabkar wrote (in news:[email protected]):


Okay; I’ve removed the capacitor, voltage regulator, and resistor for now
since
they seem to be the primary suspects. There still seems to be a short
between
the I and G pins of the voltage regulator’s place on the board (the
component
itself is not shorted, just its holes on the PCB). Here’s a closeup of the
area:
http://img292.imageshack.us/img292/9303/img3593rj7.jpg

To me, it looks like the short is at B32 (either that, or caused by the
residue
of the resistor). I was also wondering if the short could be between the
layers
(in which case I guess I’m SOL).

How can/should I clean this up? I was going to try some rubbing
alcohol/nail
polish remover.


Thanks.

Well, before you go *any* further, you need to remove the solder bridge that
you now have between the output pin's pad, and ground. You need to get to
work to tidy up the print and get the holes clear, using some soldawick or
whatever.

B32 may very well be the culprit. I have had these little surface mount
block ceramics fail on many occasions. Just remove it and see if the short
goes away. If it does, any old value of 0.01 to 0.1uF will do. It's only an
HF bypass to quieten any tendency of the regulator to hoot, which they
readily will if not decoupled properly. It is unlikely in my experience,
that you will have any kind of inter-layer short. It looks like it's only a
top and bottom layer anyway. From what I remember looking at the earlier
pictures, now that the resistor is removed, I don't think that there is
anything other than B32 still connected to the input of the regulator, so
it's a fair bet that is the shorted component. These little caps also have a
tendency to go leaky, so it may have been in such a state for a while,
stressing the resistor, which is then heating the track, and thus B32,
serving to accelerate the failure, to the point where it finally went short.

Arfa