OMG KISS Trouble Shooting

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

iomegaman

Well-known member
GDIY Supporter
Joined
Dec 22, 2008
Messages
867
Location
Tucson, Az
Just spent about two months off and on trouble shooting a Sony Mixer I have that I really like the EQ's on...

I have replaced a lot of the opamps with BB stuff, recapped the PSU and generally done some house keeping on the unit...

About 8 months ago it started developing this noise in the master bus...so I decide to track it down...

Right side of the LED meter all over the map as this AC erratic noise starts growing and growing...

Tear it down...1 million tiny screws, cables up the wazoo...sort of a "lets play sleuth when I want to be distracted" kind of project...no biggie....

Until I start getting pissed...opamps are good, trace down noise to the ground plane somewhere...take apart each and every board looking for bad caps/etc...nada...

Weeks...turn into a few months...this thing on my bench mocking me calling me names in Japanese "Big stoopid American you no fix me!"

"You hecka stoopid no smart electrons in you dumb boy"...sh*t like that every night...

I replace sliders, trade out caps, look for obsolete transistors and pull my hair out...

Ignore it, throw a blanket over it's mocking guts taking up space on my bench...

I even buy new probes for the scope thinking I'll get inspired...nada, nothing, zip,..

No help online...

Finally...tonite...I am using a jewelers lens going over it inch by inch in hopes of a light going off...and I notice an ever so faint looking maybe possibly a cold solder joint...I jumper it, solder it just to make sure...

WHAT???

A friggin cold solder joint after 20 years? (It was made in the 70's)...

1 solitary solder joint...hells bells batman...

Keep It Simple Stupid...

You ever out think yourself into a hole?

My bench is clear now...

I have a G-Pultec to build...finally.
 
> A friggin cold solder joint after 20 years? (It was made in the 70's)...
> 1 solitary solder joint...hells bells batman...
> Keep It Simple Stupid...


Similar case. CRT monitor, not new, odd erratic symptoms. Finally found a cracked/dry solder joint on a key point. Monitor served perfectly ever after.

Yes, it is often the stupid things.
 
One thing that always confused me was how customers so highly value hand built gear, dismissing mass produced gear as inferior. The quality of machine soldering and assembly is generally far superior to hand builds.

Over the decades I have found many bad solder connections, some that looked like they should be good. Back when I had my kit business, I used to see many remarkable examples of untrained soldering. One customer installed the components on the wrong side of the PCB.  ;D ;D

Just like when you are looking for your lost keys and find them in the last place you look. The bad solder connection will also be the last one you checked.

JR
 
I debug for a living, electronic hardware as well as software.  The fix for almost every problem is a simple one.  Often the big problem is finding it!

Regards,
Jeff

 
There is an important distinction between debugging a new design and troubleshooting proven existing designs. When debugging you do not know for a fact that the design is correct and should work, while troubleshooting generally involves a failed part or a bad connection in a product that was once working.

Both are very good mental exercise, but different and technicians can repair designs without knowing completely how they work. Debugging generally requires a more comprehensive understanding, while cut and paste cook book application notes allows some degree of success without the expertise.

JR
 
> quality of machine soldering and assembly is generally far superior to hand builds.
> repair designs without knowing completely how they work.


This bad joint was bad machine work/design. It was the one large wire in a dip-soldered PCB. The dip was hot enough for all the small leads, but this one didn't heat-through.  A good hand-solderer would have done it right. Some re-design would have reduced the mass difference. An aggressive flux might have helped.

While I knew some of what the board did, it wasn't reverse-engineering insight, but dogged look-look-look that found the problem.
 
PRR said:
> quality of machine soldering and assembly is generally far superior to hand builds.


This bad joint was bad machine work/design. It was the one large wire in a dip-soldered PCB. The dip was hot enough for all the small leads, but this one didn't heat-through.  A good hand-solderer would have done it right. Some re-design would have reduced the mass difference. An aggressive flux might have helped.



Not sure what dip soldering is. I've seen small boards dipped into a solder pot to solder while that is far from a precision process or what I consider machine soldering. Well managed wave soldering can be extremely reliable, of course oxidized or dirty leads on an IC could cause an issue , but those are rare in large scale manufacturing. In fact if one pin was consistently not taking solder, the preponderance of evidence stacked up in the rework area would*** cause the design to be adjusted

For wave soldering mechanical design matters a bunch. Pad sizes and resist areas can be adjusted to accommodate unusual pins.

JR

** Should,,, sometimes such feedback from the factory does not make it all the way to design, and corporate politics could cause some to resist taking blame for the cost of another cut to the PCB, but precise cost management of things like factory rework will raise any such weak sisters up to get appropriate attention. Rework is expensive and over reliance on rework to back up weak design is a crutch that does not promote good design discipline. 
 
Mental exercise is a great way to describe it.  I do both new and working designs and it is true they are a bit different.  The most difficult problem solving is when you don't have a single issue but have multiple issues presenting itself as one problem.  This is especially true for software.    I'd much rather fix hardware. 

Systematic problem solving is the key and its hard to teach.  I find that in problem solving you often go down the "wrong road".  Its important to remember this path though since it may come up again in your search for a fix.  Another roadblock to solutions is assumption.  This can get the inexperienced as well as the experienced in trouble. 

I usually enjoy the challenge except when there's a drop dead deadline and the problem is not revealing itself!

 
deveng said:
Mental exercise is a great way to describe it.  I do both new and working designs and it is true they are a bit different.  The most difficult problem solving is when you don't have a single issue but have multiple issues presenting itself as one problem.  This is especially true for software.    I'd much rather fix hardware. 
While hardware is easier to troubleshoot (generally), I find digital hardware harder than analog because you can not always see the fault in real time. Software is easier to  manage but can be a bear to troubleshoot/debug especially with indirect addressing and obscure firmware faults. I've seen some doozies. 
Systematic problem solving is the key and its hard to teach.  I find that in problem solving you often go down the "wrong road".  Its important to remember this path though since it may come up again in your search for a fix.  Another roadblock to solutions is assumption.  This can get the inexperienced as well as the experienced in trouble. 
Yes, a key principle in troubleshooting is to divide or isolate the problem as much as possible to reduce the number of variables you are dealing with.
I usually enjoy the challenge except when there's a drop dead deadline and the problem is not revealing itself!
I had my share of hard deadlines around trade shows where the new design absolutely positively had to be finished and working. The trouble with pulling too many all nighters in a row is that your brain can get tired. Sleep, not only recharges the brain's energy level, but in my experience the brain keeps thinking about problems while we sleep, so sleeping on a problem sometimes leads to a solution.  I have been refreshed and recharged by catching a few hours of sleep in the middle of a hard push.

JR
 
Sleep, not only recharges the brain's energy level, but in my experience the brain keeps thinking about problems while we sleep, so sleeping on a problem sometimes leads to a solution.

Its funny you say this because its absolutely true.  Whenever I just can't seem to find the answer, I try to walk away from it for awhile.  I don't know how many times I've solved problems while not directly working on it.  If you get away and clear your mind, take a new approach or sleep on it,  sometimes the solution presents itself in that "aha" moment.   

Regards,
Jeff
 
deveng said:
Sleep, not only recharges the brain's energy level, but in my experience the brain keeps thinking about problems while we sleep, so sleeping on a problem sometimes leads to a solution.

Its funny you say this because its absolutely true.  Whenever I just can't seem to find the answer, I try to walk away from it for awhile.  I don't know how many times I've solved problems while not directly working on it.  If you get away and clear your mind, take a new approach or sleep on it,  sometimes the solution presents itself in that "aha" moment.   

Regards,
Jeff

I would jokingly call that my "overnight design team"... working on problems while I sleep.  8)

Back before my knee let me down, I could work on difficult software problems while jogging. Now that i'm covering the same roads on a bike, things are happening much to fast, requiring me to pay attention to real time hazards (like traffic or road surface) preventing me from zoning out and problem solving.  :'(

Another perhaps personal quirk, while I code and enter schematics at a computer, I literally need to move 10' away and sit in a comfortable chair using pen and paper to do high level design. like circuits or program architecture. 

JR
 
So to sum up trouble shooting in a general way:

1.  Understand the circuit (best but not always necessary)
2.  Isolate.  Reduce as many variables as possible
3.  Document (either on paper or mentally, paper better)
4.  Don't assume
5.  Learn to move from global to local, local to global thinking (avoid tunnel vision)
6.  Tenacity
7.  Step away and think about it

I'm sure there are a dozen more!

Regards,
Jeff
 
Good stuff.
deveng said:
So to sum up trouble shooting in a general way:

1.  Understand the circuit (best but not always necessary)
You can go a long way just understanding how basic semiconductors work. Valid voltage relationships for diodes and transistor junctions. Basic relationships between op amp input voltages and output state.
2.  Isolate.  Reduce as many variables as possible
amen...
3.  Document (either on paper or mentally, paper better)
4.  Don't assume
This is a big one.. like don't ASSume your test equipment is always telling you valid data.
5.  Learn to move from global to local, local to global thinking (avoid tunnel vision)
This is a variant on ASSume... while troubleshooting is all about narrowing the scope of the search, do not narrow it so much you miss the fault.
6.  Tenacity
You haven't failed until you decide to give up.
7.  Step away and think about it
Don't ignore asking for advice. The internet has lots of anecdotal reports about problems with products. Manufacturers with service departments have seen many common failures.

Sometimes just talking about a problem with a non-technical friend will triggers a new idea or open up a new vector to investigate.
I'm sure there are a dozen more!

Regards,
Jeff

Forums like right here are a good place to get information about fixing stuff...

JR
 
Two examples of  my past experiences.

First one was a vintage 1176 that belonged to a friend. He said suddenly LF and MF disappeared and all it was passing was fizzy HF. I always start trouble shooting with checking the board under the magnifier. The solders were fine but that was exactly where the trap was. Obviously during the assembly he/she cut the secondary pins of the output tx too short They were flush with the PCB surface. Normally if it was a plated through board it would have been fine. But on a single sided the solder was just about holding onto the pins and it passed the test. But after so many years they broke away with just a hairline gap. The capacitance was sufficient for the HF to pass through but nothing else. Extending the pins and re-soldering solved the problem.

The second one was a bit more nightmare. I had designed a control electronics for an industrial dish washer for Asda (which  were bought by Wal-Mart).  We had ten prototypes for field testing.  As usual my brother and I ended up working through the night, fitting the electronics. About five o'clock in the morning all finished and we started to test. All worked except when we got to the tenth one an erratic behaviour.  A nightmare that you don't want at that time of the day.  By about lunch time I found the bastard. It was one of the wires that did not crimp properly.  So, don't drive and crimp when tired.
 
Very good advice, thanks!

This particular Sony has really great documentation, (funny how that used to be SOP in quality stuff in the day)...each sub board had exact component layout and schematics, down to which colored wire went where...even though it's an old scanned manual it was still helpful...I narrowed it down to the PGA board several times and was not thinking squarely each time I departed from that board to go back to others...(I had actually unplugged each board and powered it up at each removal with the noise still showing up DUH!)

I knew I was getting some AC over voltage to one leg of a particular opamp but could not figure out where it was leaking in...turns out is was a 100k resistor to ground from one leg of a transistor...

I gotta hand it to Sony though the wiring work and documentation was immaculate, from screw sizes and exact component list to a blok diagram of cable hookups...I suppose no one really does this kind of manual anymore...they actually give you an order of tear down to take it apart as well as all the specs and procedures to calibrate meters/gain levels etc...
 
Sounds like the story of my Wurlie 200A piano. Crackling noises that took me a couple years to find, after replacing all the wet caps and a fair number of resistors.  Finally figured out that the big power resistors had cooked their solder pads off the board, but not enough to look loose. Hairline cracks where the pads met the traces... a little bit of wire along the traces and pads and it's been perfect ever since.

One of my troubleshooting maxims is "look first".  Spend a little time looking for parts or traces that have gotten hot, or even just the obvious black/broken resistors.
 
Scodiddly said:
Sounds like the story of my Wurlie 200A piano. Crackling noises that took me a couple years to find, after replacing all the wet caps and a fair number of resistors.  Finally figured out that the big power resistors had cooked their solder pads off the board, but not enough to look loose. Hairline cracks where the pads met the traces... a little bit of wire along the traces and pads and it's been perfect ever since.

One of my troubleshooting maxims is "look first".  Spend a little time looking for parts or traces that have gotten hot, or even just the obvious black/broken resistors.

When looking at solder connections for overheating, look for a shiny (good) or dull (bad) finish. I've seen voltage regulators that had thermal shut-down that protected the regulators from failure but the solder connections managed to become intermittent  over time from the heat.

Note: Some new lead free solders don't look shiny when new..  :'(

JR
 

Latest posts

Back
Top