开云体育

Re: Hacking a digital camera

Removing the IR filter won't make much of a difference. The objects have
to be *very* hot in order to have any appreciable IR emission at
wavelengths short enough for a silicon-based camera. With a bandgap of
1.2eV or so, you aren't going to see anything with greater than 1um
wavelengths. Think of how hot something has to be to glow red, and back
off a little. If you have parts that hot, you won't need a camera to
tell you that!

--Cheers
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070


On 5/6/2022 18:13, James55 wrote:

Hi fellas, thanks for the input, it all helps.

@ Harvey;

Nicely outlined. You've helped clarify a few things which weren't so

clear. The feeding a 50% duty cycle signal shall have to wait. My intention
right now is to nail the technique in finding shorts by analyzing the
millivolts

One more question, would be, "how might one find out if the tester has

a high impedance?" There is no mention in the manual.

@ Aldue;

Yes, an infrared camera is definitely on the horizon, in fact, I do

have one back in Europe.

Has anyone ever tried hacking a 'normal' digital camera by removing

the IR filter lens to see if the components glow when hot? That might be my
best option here.

@ Ed Breya;

I'm currently in Rio, so obviously have access to lots of LIA's...

What color were you thinking?

No... apologies for being silly, but don't have access to a Lock-in

Analyzer, tbh I have never heard of one. That said, I do appreciate the
input and shall look further into them. My goal is ultimately to fault-find
as quickly as possible, regardless of which technique or apparatus is used.

I'm still a hobbyist so still accumulating tools and testers. The

reason for asking about the millivolt method in particular was that
following a prior thread, it became very clear that it was something quite
basic which was missing from my repair 'armory'.

Like Jim Ford however, I am intrigued as to recommended makes and

models.

Appreciations to you all.

James

Hi fellas, thanks for the input, it all helps.

@ Harvey;

Nicely outlined. You've helped clarify a few things which weren't so clear. The feeding a 50% duty cycle signal shall have to wait. My intention right now is to nail the technique in finding shorts by analyzing the millivolts

One more question, would be, "how might one find out if the tester has a high impedance?" There is no mention in the manual.

@ Aldue;

Yes, an infrared camera is definitely on the horizon, in fact, I do have one back in Europe.

Has anyone ever tried hacking a 'normal' digital camera by removing the IR filter lens to see if the components glow when hot? That might be my best option here.


@ Ed Breya;

I'm currently in Rio, so obviously have access to lots of LIA's... What color were you thinking?

No... apologies for being silly, but don't have access to a Lock-in Analyzer, tbh I have never heard of one. That said, I do appreciate the input and shall look further into them. My goal is ultimately to fault-find as quickly as possible, regardless of which technique or apparatus is used.
I'm still a hobbyist so still accumulating tools and testers. The reason for asking about the millivolt method in particular was that following a prior thread, it became very clear that it was something quite basic which was missing from my repair 'armory'.
Like Jim Ford however, I am intrigued as to recommended makes and models.

Appreciations to you all.

James

From: "Bruce Atwood" <CCDman1@...>
To: "tekscopes" <[email protected]>
Sent: Friday, May 6, 2022 3:47:16 PM
Subject: Re: [TekScopes] Proper Way to Pull Single-Ended Connectors?

I also had a blue one disintegrate to dust on my 7904. I put shrink tubing
(blue) on every other pin. The wire to wire bonding in the flat cable keeps
them in order but I drew a diagonal line across the group just to be sure.

From: "EB4APL" <eb4apl@...>
To: "tekscopes" <[email protected]>
Sent: Friday, May 6, 2022 10:51:46 AM
Subject: Re: [TekScopes] Proper Way to Pull Single-Ended Connectors?

Been there. You can replace the broken connectors with DuPont ones. They
are cheap, widely used, i.e. in Arduino projects and in various pin
counts. They are always black.

Ignacio


El 06/05/2022 a las 16:22, David C. Partridge escribió:

Yes, It's always the blue ones that crumble.

D.

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of n4buq
Sent: 06 May 2022 14:41
To: tekscopes <[email protected]>
Subject: Re: [TekScopes] Proper Way to Pull Single-Ended Connectors?

Regarding those harmonica connectors, when I pulled one of them from one of the
circuit boards last night (P1306 on the A22 board), it crumbled in my fingers.
Interestingly, it was a blue connector. The other colors were okay. I have
noticed that the blue connectors have crumbled more consistently in some of my
plugins and I'm wondering what a specific color would have to do with this.
Presumably they're all mixed with a dye but odd that blue is affected more
than the other colors. Anyone else noticed this?

Thanks,
Barry - N4BUQ

--
El software de antivirus Avast ha analizado este correo electrónico en busca de
virus.

I'll see if I can summarize the concepts that I understand.

The basic concept is that you measure the voltage drop across each
suspicious component and look for the lowest drop. Another concept
would be that the shorted part would be hotter than the rest, and would
be detectable on a thermal camera.

The caveat is that the power supply that is feeding the voltage to the
board should be current limited to avoid burning out series filters, and
actually blowing up the shorted part. The supply should be voltage
limited to the nominal supply voltage being simulated, and current
limited (IMHO) to about 100 to 500 ma, depending on the original supply
current. Limiting the supply current also keeps from damaging traces on
the board.

The wiring on the board can be in two forms, one is a single supply rail
with all parts hanging off it, another could be a divided set of feeds
from some central point, or a combination of the two. This depends on
board geometry and trace routing, none of which will be indicated in the
schematic.

The concept of coloring in the parts on a copy of the board is simply to
show you which ones to check and which to ignore. You might consider
putting all the capacitors in one color, and all the chips in another.

Considering the straight line model, think of a long extension cord with
lamps plugged in every few feet. Since at a low voltage, the parts are
not likely to draw much current, from the end of the string to the
short, most voltages will be the same. At the short itself, you'll get
the same voltage as down the string. At the next part in the string,
the part sees across it the shorted part /with the added resistance of
the traces to that part, both ground and supply/. This means that the
voltage across a part upstream of the shorted part should be higher.
That voltage ought to increase up to the point where the supply is
connected.

Since these resistances are quite small, the use of a millivolt meter is
important, and with a high input impedance. That high input impedance
minimizes the effect of a slightly inconsistent connection.

It is also important to measure the voltage drop across the part itself,
not with reference to a ground. Since you don't know where the ground
goes, you don't know how much a ground trace, ground flood, or ground
plane contributes. Best to measure across the part itself.

In the case of a split supply, one branch should show the fault
(assuming there's a single fault).


There are other techniques. The heating test requires more current and
a thermal camera.

Another technique would be to feed the power supply with a 50% (or so)
duty cycle square wave voltage limited to the supply rail voltage and
current limited to about 50 ma. or so. HP made a non-contact current
probe that was designed to find a shorted node to ground. It could be
used to trace the supply track from the main input supply terminal to
the shorted part. I think I've tried that and it works within limits
(pulses will be shorted out by the capacitors, but by how much?). That
current probe is rather rare and expensive, which makes the millivolt
meter technique the best one.


Harvey


On 5/6/2022 7:15 AM, James55 wrote:

Hi all,

I'm looking for tips and tricks for millivolt fault finding.

I recently ran a long thread whilst finding the solution on a TEK 466

power supply section, and in that thread various members suggested using a
millivolt meter to locate the shorted components, and I know the details
are already in that thread, however it would be good to have a short
compacted thread with all the relevant information.

Apart from the minor detail of not having a miilivolt meter at the

time, I struggled to visualise how to know that the component under test
wasn't part of rail without any issues? The 466 interface board is a sea of
components with six different voltage rails intertwining as the circuits
weave their way to the other boards. For example a zener diode from the +5v
might sit side-by-side on the board with an identical zener diode from the
-8v rail.

Anyway, there is now a 0.xxx range millivolt meter in the toolkit and

I am up for another learning curve, so it would be seriously appreciated if
we could clarify a few points.

The fact that any shorted or failing component will show the lowest

voltage across it is understood.

What is not so clear is whether people differentiate between rails (by

colouring-in components on a print, for example?) or just measure across
every component on the board? Thinking beyond Tek scopes, some components
measure low anyway (a 0.12Ω resistor would be one), although they would
also be marked as such, but surely there must be short-cuts to avoid
measuring 500 components on a packed board if the rail at fault only has 50
in total?

Therefore if anyone can add anything to speed up miilvolt fault

finding it would be great.

Thanks in advance,

James

Hi all,

I'm looking for tips and tricks for millivolt fault finding.

I recently ran a long thread whilst finding the solution on a TEK 466 power supply section, and in that thread various members suggested using a millivolt meter to locate the shorted components, and I know the details are already in that thread, however it would be good to have a short compacted thread with all the relevant information.

Apart from the minor detail of not having a miilivolt meter at the time, I struggled to visualise how to know that the component under test wasn't part of rail without any issues? The 466 interface board is a sea of components with six different voltage rails intertwining as the circuits weave their way to the other boards. For example a zener diode from the +5v might sit side-by-side on the board with an identical zener diode from the -8v rail.

Anyway, there is now a 0.xxx range millivolt meter in the toolkit and I am up for another learning curve, so it would be seriously appreciated if we could clarify a few points.

The fact that any shorted or failing component will show the lowest voltage across it is understood.
What is not so clear is whether people differentiate between rails (by colouring-in components on a print, for example?) or just measure across every component on the board? Thinking beyond Tek scopes, some components measure low anyway (a 0.12Ω resistor would be one), although they would also be marked as such, but surely there must be short-cuts to avoid measuring 500 components on a packed board if the rail at fault only has 50 in total?

Therefore if anyone can add anything to speed up miilvolt fault finding it would be great.



Thanks in advance,

James