Build Longevity Tips

[N9]

I did Ground Support cabling when I was at JPL (I also did Flight Systems PCB Assembly and was a Polymerics Technician (Epoxies, also known as "Goop" for the "booteek" crowd) ), and all the cables I made were with stranded wire.

Edit:
And bookmarked.
During the certification course at JPL (A fairly basic 2 week course) they gave us a book that had basically whats in that link.
But that was in the later 80's and the graphics weren't as purty
It was all black and white with crappy image resolution.
All the PCB stuff I did was SMD and done by hand under a microscope.
But it was a hell of a lot easier than building those cables.

Out of curiosity, what qualifications are needed to take a course at that level? I'd love to learn how to do assembly at that level, but I doubt I'll ever be employed by JPL.

[CodeMonk]

I did Ground Support cabling when I was at JPL (I also did Flight Systems PCB Assembly and was a Polymerics Technician (Epoxies, also known as "Goop" for the "booteek" crowd) ), and all the cables I made were with stranded wire.

Edit:
And bookmarked.
During the certification course at JPL (A fairly basic 2 week course) they gave us a book that had basically whats in that link.
But that was in the later 80's and the graphics weren't as purty
It was all black and white with crappy image resolution.
All the PCB stuff I did was SMD and done by hand under a microscope.
But it was a hell of a lot easier than building those cables.

Out of curiosity, what qualifications are needed to take a course at that level? I'd love to learn how to do assembly at that level, but I doubt I'll ever be employed by JPL.

I would guess your employer paying them for you to take the certification course.
When I took the certification course, there were several people that were sent there for certification by the companies they worked for.
I'm guessing those companies built stuff for the Aerospace industry.
Whether or not just anyone could pay to take the course, I don't know.

As for me...
I had been doing Electronic Assembly for about 7 years.
Living in Southern California, many of the Electronic Assembly jobs were with military contractors and subcontractors.
There were many others doing commercial stuff, but doing the stuff for the military contractors paid better.
Unfortunately, when the contract was finished, people got laid off.
At some point, I decided to sign up with several different temp agencies, which was the best way to stay consistently employed.
Contract finishes, get laid off on Friday, have a new job the following Monday or Tuesday.
Then I went and signed up with another agency who sent me to JPL to "Audition".
I passed and got hired.
They then had me "practice" for two weeks while they tested out numerous other people, about 50 or so, before they had someone else that passed the "Audition".
JPL was looking for 2 people so they had me practice until they found a 2nd person so they could send us both to the certification class at the same time.
The class was held on site at JPL.
They have several buildings there (100+) with at least one of them dedicated to education and certification.
Its a HUGE place.
Check it out on Google Maps (Just type in "Jet Propulsion Laboratory")
And just for shits and grins:
34.201328, -118.171748  (Co-ordinates for the school/certification building)
34.201706, -118.167938 (Co-ordinates for the the building I worked in, most of the time. Occasionally I would work in other buildings)

Technically, I wasn't hired by JPL, it was more like I was a subcontractor and working for the temp agency and JPL "hired" the temp agency.
My paychecks came from the temp agency.



This may have helped me too:

...
But I was trained in Electronic Assembly by a retired NASA technician (At West Valley Occupational Center in Woodland Hills, Ca.) in 1981, and then later had to take a certification course at JPL when JPL hired me (through a temp agency),
...

Oh, and they have certification courses for EVERYTHING.
They even have a "Connector Mate and De-Mate Certification".
You know those parallel and serial connectors like the kind you have on a computer?
You need to take, and pass, the certification course to do that for NASA. On NASA built equipment anyway.
I think that was like a 2 hour class.

[N9]

Many thanks for the detailed response! High-reliability systems and assemblies are a big interest of mine; I always enjoy getting further insight!

To keep the discussion going, here's another resource which folks may find useful: "Guidelines for Inspection of Aircraft Electrical Wiring Interconnect Systems"
http://www.assistdocs.com/search/document_details.cfm?ident_number=277535&StartRow=1&PaginatorPageNumber=1&doc_id=MIL-HDBK-522&status_all=ON&search_method=BASIC

There are a tremendous number of other military standards handbooks and manuals available from the DOD ASSIST database. The quickest way to access documents is through www.assistdocs.com/.

If you like, you can also create your own account on the main database site (https://assist.dla.mil/online/start/) for more in-depth searching.
Creating an account is something of an obtuse process, and the security is strict (complex password mandatory, must change password every few months, etc.), but I found quite a few interesting documents there.

[stallik]

Probably the wrong thread for this as I'm sure that im doing just about everything wrong but it does help me populate the entire board with different height components and trim the leads before soldering. My main problem is that I'm using too big a drill and the solder has to create a bridge between the lead and pad.

Foam

Same size backing

Bulldog clips

Trim

Solder

Everything held in place

[R.G.]

Longevity tips?

(1) if it moves, either make it move easily, or restrain it so it doesn't move at all
(2) if it moves easily, make the ends of the motion be damped and the end points not be stress concentrations
    (1a and 2a) if it's a wire, fix the ends so that *ab*normal use does not break them at the end of the insulation
(3) never fix a fragile component (PCBs and solder joints are) in more than one plane
(4) solder is not glue
(5) solder is not a mechanical component; it creeps, cold flows, and cracks
(6) if it gets even noticeably warm, go find out HOW hot it gets and why; then compare that to what it can do without degrading
(7) if you come near the absolute maximums on a component, expect some of them to fail even if you don't go over
( anything sticking out of the box *will* be kicked, stomped, crushed, twisted and pounded upon; and your customer will tell you that it's YOUR fault
(9) the box will be kicked, stomped, crushed, twisted, drowned, dried, pounded, hammered and hacked
(10) customers will apply any convenient source of electricity, from a guitar pickup to a speaker output to a 220VAC power cord to any connector that they can access; expect it

[karbomusic]

(7) if you come near the absolute maximums on a component, expect some of them to fail even if you don't go over

LOL, I always obviously adhere to that rule without exception. Except for a booster/OD that began with 2 OPA2134s (rated at 36 volts). I subsequently ran it at 18V for headroom purposes. Down the road, I changed one of those to a TLC2272 because it was rail-to-rail. I had forgotten the 2272 is only rated at 16V. Two copies of the pedal have been running in production so to speak at 18V for months now. I'm wondering if they'll die before I take the time to swap 'em out. 

[rousejeremy]

> Using the correct amount of solder in a solder joint. Seems like a no-brainer, but it isn't.

+1

Many of the joints seen here (all DIY) make me shudder. You should be able to see the underlying joint through the solder, not a blob covering who-knows-what.

As a full time guitar tech, I'm startled by the soldering inside guitars coming from the repair shops all over Toronto. Threads like this have made my work look damn good.

[Quackzed]

(1) if it moves, either make it move easily, or restrain it so it doesn't move at all
(2) if it moves easily, make the ends of the motion be damped and the end points not be stress concentrations

big yes to all points ,but these 2 especially!
i tend to think, if it can move - it can break.
and also-  where is the strain and how can I remove it... i want everything to be solid, yet comfy and relaxed...
i mean, how would I feel with something tugging on ME my whole life!

[R.G.]

LOL, I always obviously adhere to that rule without exception. Except for a booster/OD that began with 2 OPA2134s (rated at 36 volts). I subsequently ran it at 18V for headroom purposes. Down the road, I changed one of those to a TLC2272 because it was rail-to-rail. I had forgotten the 2272 is only rated at 16V. Two copies of the pedal have been running in production so to speak at 18V for months now. I'm wondering if they'll die before I take the time to swap 'em out. 

It's a question of percentages. In fact, each semiconductor wafer is different, no matter how hard the fab houses try to make them different. It's likely that those pedals got chips that by chance were higher than the minimum voltages. The next build might get chips just over the 16V line. Or that they have some power supply leakage through the substrate diodes, but it's not bad enough to make it fail, or not bad enough in lower than the highest temperatures.

It was always thought to be a good way to lose your job to spec a part at or near the AbMax and have the first production run work, then have the second - and third, and fourth! - runs suffer high fallouts from getting more "normal" chips, or getting a batch from another manufacturer, all of which met the spec.

[bluebunny]

(10) customers will apply any convenient source of electricity, from a guitar pickup to a speaker output to a 220VAC power cord to any connector that they can access

Yep.

 

[N9]

(10) customers will apply any convenient source of electricity, from a guitar pickup to a speaker output to a 220VAC power cord to any connector that they can access; expect it

This is a very good point! I struggle to imagine how to properly protect a stompbox from high-energy inputs, however.
Well-built multimeters are the only small devices I can think of which can reliably survive really ugly input overloads, and they devote a heck of a lot of engineering to it.

[PRR]

> 2272 is only rated at 16V. Two copies of the pedal have been running in production so to speak at 18V for months now

My experience with 26V chips run at 35V is that I got a decade out of each, twice.

As R.G. says, your experience will vary.

In my case I may have been pushing electromigration. The Aluminum starts crawling across the chip due to the high voltage concentration.

In particular, some "5V" chips will catch-fire if the rail hits 7V (the breakdown for no-special-spec junctions).

"18V" CMOS in fully-driven digital duty can sometimes go higher but if biased Linear will burn-up (not breakdown) at those levels.

There is usually a compromise between gain and breakdown. Max hFE of BJT goes about inverse of square-root of voltage rating. (The super-Beta inputs of LM308 are 2V devices.) In today's 3.6V world, designers will optimize for lower voltages to get their gains up. Going from a 36V process to an 18V process is good for about 1.4X more gain. Not much alone, but in a 3-stage amp that's 2.8 times better performance, which sure may edge-out a competing product.

[thehallofshields]

I do all my soldering under moderately high magnification, then several complete checks of every joint and trace with magnification. This has three lenses and only cost $7.00.

Please tell me what that is and where I can buy one!

[R.G.]

There are several brands, but it's hard to beat these for value per dollar:
http://www.harborfreight.com/magnifier-head-strap-with-lights-38896.html
I've posted the link to this several times here.

[karbomusic]

Please tell me what that is and where I can buy one!

One of the most useful tools I have ever purchased. I grabbed mine off of Amazon; It has the "one-eye" magnifier plus an additional flip down magnifier for both eyes giving me three levels not including various combinations of the three.

[karbomusic]

Probably the wrong thread for this as I'm sure that im doing just about everything wrong but it does help me populate the entire board with different height components and trim the leads before soldering.

I actually like that idea. I usually add components from lowest to highest but the above still looks darn handy.

[rousejeremy]

Probably the wrong thread for this as I'm sure that im doing just about everything wrong but it does help me populate the entire board with different height components and trim the leads before soldering. My main problem is that I'm using too big a drill and the solder has to create a bridge between the lead and pad.

Foam

Same size backing

Bulldog clips

Trim

Solder

Everything held in place

This is a great idea. Thanks.

[duck_arse]

erm, I've found when using run-of-the-mill foam, as opposed to proper, high temp foam, that it is better with a bit of card betwixt parts and foam, as the soldering heat will often melt the foam onto resistor leads.

[bluebunny]

^^^
Try using a sausage, split lengthwise (DIYSB mandatory accessory).  It's got a bit of "give" to it, holds the components nicely, and if you go a bit OTT with the iron, you get the bonus of a small snack.  (Rob would approve.) 

Yeah, I know: bloody poms have no idea how to do a proper barbie...   

[stallik]

+1 on the sausages 
But more seriously on the foam, I usually tear the scourer off a kitchen pad and use that foam and it works ok. If the components are getting that hot, it's taking too long to do the soldering