There is no such thing as a Subtle Fault

[ruffrecords]

There is no such thing as a subtle fault has been a mantra of mine for more than 40 years. I first coined it when a colleague was having great difficulty getting an RS232 link to work. Sometimnes it would work for an extended period then it would suddenly start receiving corrupted data. He scratched his head for days looking for software bugs and logic faults without success. I checked his connector wiring and found he had the TX and RX and RTS and RTR and frame ground all connected as the should be but he had forgotten digital 0V. Once that was connected it worked 100% of the time.

Intermittent faults are the worst kind to find. I was reminded of this this week when a Mark 3 tube mixer module I was commissioning would work perfectly sometimes but if I pulled it out to tweak something and then put it back, sometimes the level would drop by about 60dB or more. You could still make it go up and down using the gain control but the level was way out. To cut a long story short I tried cuting out sections of the circuit in an attempt to find what was wrong. As often as not this would appear to cure it then a few minutes later after another tweak it would happen again even though the 'faulty' section had been bypassed. I just could not nail down the location of the fault. Then, almost in desperation I decided to check for dry joints and dodgy cables. Within five minutes I had found the fault. A crimp connector had not had the crimp pushed fully home into its connector housing. Sometimes it connected with the mating half, often it did not. I just pushed the offending crimp terminal fully into its housing and the problem disappeared. Tweak as much as you like but the fault never reappeared.

So bottom line is, no matter how puzzling a fault may seem, in the vast majority of cases the cause is something very simple.

Cheers

Ian

 

[Ike Zimbel]

My take on this is: "After you have eliminated all of the smart things, it's going to be something stupid".

 

[Tubetec]

The explanation is almost always simple , finding the cause can cause lots of headscratching though. Intermitant faults can become time consuming and sometimes even time spent cant be recouped , to some degree simpler easier to detect faults have to pay for the jobs you end up headscratching on . Still rewarding though after you've torn out whats left of your hair , through dogged determination you get it back up and running .

I had one Marshall amp in , sounded like crap , turned out the pre phase splitter Dc coupled cathode follower resistor measured 1meg instead of 100k yet DC readings were spot on . Another time a buddy scored an ex-stock JCM-800 from a music store about to shut its doors, it was downstairs about 25 years , clean channel worked great , lead channel only worked the odd time , I spotted one heater section of an ECC83 didnt glow , turned out the bridging between the adjacent heater pins wasnt soldered correctly from the factory , still made contact the odd time though and wasnt until you looked very closely and poked around the fault became visible , Ive fixed more faults with my eyes than a multimeter by a factor of ten and maybe another factor of ten where the scope was useful, well the ancient tube based scopes I used have invariably blew up/or melted down on me , a nice ,relatively modern analog Tektronix I could live with alright .

 

[pucho812]

this reminds me of years ago when I picked up a Gibson amp on the inexpensive. The pentode and triode switch on the rear did the following. IT basically is a 15W/30W switch and when in 15W mode, would squeal something awful. This came that way from the factory and rather then fix it, it got sold to me on the cheap because the amp is unusable in the apartment. Anyway I get it on the bench and discover that the wiring on that switch was flipped making a negative feedback become a positive feedback. once I flipped the two wires which were on connectors and easily moveable. no more squeal.

 

[Tubetec]

I got an Ac30 style home built tube amp in once from an Aussie touring band , fecking thing oscillated like crazy at anything above whisper volume , turned to complete mush , squealed like a f'ing pig . I found a couple of very obvious constructional/component issues , once sorted it actually amplified signal . The guitarist was a little miffed all the distortion/RF howlround on his sound was gone , I documented the changes I made on paper and included notes to the guy who built it back in Aus , might have taken the band a little time to re-adjust to the guitarists new loud sound , im sure it ended up better on the longer road though .

Pucho the Indefatigable , no stone left unturned

 

[Brian Roth]

Reminds me of "fun" 40 years ago when I was hired as Chief Engineer (as well as chief cook and bottle washer...lol) at a new studio. I did all the tracking as well as maintained the gear.

The multitrack was a MCI JH-16 24 track that was maybe two years old...purchased new in 1979 just before I was hired. The audio electronics cards were pretty much clones of Ampex AG-440/MM-1000/MM-1100.

After a year or two there, one track began to cause me fits....don't recall if it was Record or Repro problems. Most of the time it worked, but then sometimes......The guys were getting irritated with me!

The Ampex design and MCI clone audio electronics used a daughterboard with the EQ, etc. controls visible from the front. That card plugged onto the main audio board for that channel/function.

Long story short...the through-hole components on that daughter card were never soldered to the PCB traces at the factory in Ft. Lauderdale! The leads were bent over onto the PCB pads and made enough contact for operation for several years before they became intermittent.

 

[pucho812]

I got an Ac30 style home built tube amp in once from an Aussie touring band , fecking thing oscillated like crazy at anything above whisper volume , turned to complete mush , squealed like a f'ing pig . I found a couple of very obvious constructional/component issues , once sorted it actually amplified signal . The guitarist was a little miffed all the distortion/RF howlround on his sound was gone , I documented the changes I made on paper and included notes to the guy who built it back in Aus , might have taken the band a little time to re-adjust to the guitarists new loud sound , im sure it ended up better on the longer road though .

Pucho the Indefatigable , no stone left unturned

I looked up the word Indefatigable, thanks for the vote of confidence.

 

[RSRecords]

Steve Sadler of MCI tech fame taught me that “intermittency means connection issue”. And that holds true 99% of the time (especially with mci’s)

 

[tomas.borgstrom]

A local tech I know recieved a pair of PA speakers that made scratching intermittent noises. Turned out a mouse was inside one of them.

I quite often get gear that work as it should. People pay a lot of money for vintage gear and have extreme expectations.

 

[Ike Zimbel]

The Ampex design and MCI clone audio electronics used a daughterboard with the EQ, etc. controls visible from the front. That card plugged onto the main audio board for that channel/function.

Long story short...the through-hole components on that daughter card were never soldered to the PCB traces at the factory in Ft. Lauderdale! The leads were bent over onto the PCB pads and made enough contact for operation for several years before they became intermittent.

Ah, "no-solder joints"! I've seen lots of them over the years. I had an 1176 come in once. The owner told me he'd purchased it surplus from a radio station and when it came in, it had a "hangar queen" look about it...no signs of daily wear, but had been raided for a few parts (meter bezel, for example). After I got it back together it would work (ie: Compress) one day, but not the next, happily passing signal the whole time. Eventually I traced the fault to the white wire that runs between the Attack and Release pots...wrapped around the terminals, but never soldered, at either end...
I could imagine that piece spent its whole service life going back and forth between the rack and the repair bench.

 

[ruffrecords]

My take on this is: "After you have eliminated all of the smart things, it's going to be something stupid".

I think it should be before you postulate a bunch of smart things it isn't going to be,

Cheers

Ian

 

[Ike Zimbel]

I think it should be before you postulate a bunch of smart things it isn't going to be,

Cheers

Ian

Yah, but that would be good advice! I think we all get sucked into the rabbit hole of over thinking these things...sort of a "What's the worst thing it could be?" mentality. I've been trying to re-program my thinking to "What's the best thing it could be?" so as to look for the simple things first.

 

[Tubetec]

One of the bands I did sound for had a Peavey desk and speakers , the xovers had lightbulb limiters across the horns , the desk was around 600 watts per channel so it could easily light the bulbs on peaks , the light was visible through the ports on the speaker boxes . We'd often get people coming up asking what the ghostly glow emanating from the speakers was , an explanation often helped to allay the fears of the more superstitious types.

 

[rogs]

I recall a strange fault way back in the day when I was a service tech working on home organs. We had one charming elderly lady who lived 'out in the sticks' as we say here in the UK (out of town, in the countryside).
She complained that the organ would sometimes go 'out of tune' - but only on Tuesdays.
Needless to say we didn't believe her, but it turned out she was right.
Seems she played hymns on the organ every day - around lunchtime.
On Tuesdays, the lady next door only cooked at lunchtime - on Tuesdays.
The AC supply to these out of town houses was very iffy , and the voltage would drop way out of spec if you drew too much current - like cooking.
So the regulators would not have a sufficient input voltage to keep the organ working properly (all linear supplies in those days.)

Power company fault of course, and it would have happened any time the volts dropped....
But for this little old lady, the organ was only faulty on Tuesdays !

 

[ruffrecords]

Yah, but that would be good advice! I think we all get sucked into the rabbit hole of over thinking these things...sort of a "What's the worst thing it could be?" mentality. I've been trying to re-program my thinking to "What's the best thing it could be?" so as to look for the simple things first.

It is good old Occam's razor. The simplest explanation is the most likely.

Cheers

Ian

 

[pucho812]

When we fix something that is easy we often think we are stupid, when we fix something complicated we feel proud. It’s one of many reasons we miss the simple things trying to find the more difficult explanation

 

[Trouble Report]

I think it should be before you postulate a bunch of smart things it isn't going to be,

Cheers

Ian

Sounds like a permutation of the Sherlock Holmes quote "Once you have eliminated everything else, whatever is left the truth."

 

[Ike Zimbel]

When we fix something that is easy we often think we are stupid, when we fix something complicated we feel proud. It’s one of many reasons we miss the simple things trying to find the more difficult explanation

Too true! I haven't had this with electronics so much, but with things like networking and computers in general I've often found myself, after a long struggle to get something to work saying "That?! That?! That's what I needed to do to make this work?! That's the stupidest f**king thing I've ever seen!"