quick question about off board components

[Arcane Analog]

This post wins the day.

Notice in these two photos how the soldered leads are just lying against each other with no mechanical connection before soldering. Not Cool:





This shows a couple of caps hanging out and at risk of the leads breaking:



These photos show mechanical connections before soldering:







These photos show lightweight components with short leads. These are much less likely to break off than the ones above that have one lead soldered to a lug and the other one to a wire.

[R O Tiree]

AA - without emoticons this thread would have gone nuclear a long time ago. With them, it is obvious that RG is trying to keep it light-hearted and civilised, as always. Sadly, you have chosen to view it as RG laughing at you. He has already said that your work is good and that PTP has its place.

Why can you not accept his points that this approach is not for everyone, that it needs careful thought, knowledge of the stresses involved, attention to detail and is clearly not for the inexperienced. Then we can all move on.

[R.G.]

Posting a picture of poor work is simply pointing out your observation that the builder has poor skills. This does nothing to prove your point other than offering your disdain for this builder.

Actually, it does. It shows clearly the differences in good and bad work. It makes clear that simply saying "mount components off the board" is not sufficient. And it shows that without further clarification, it is entirely possible to have unreliable construction with flying components.

The bottom line is that done correctly you can fly caps and/or resistors and they will be as secure as any PCB mounted component. In fact, if you attach the part with a good mechanical connection that joint will be stronger than any joint to a PCB trace could ever be.

I really can't understand why you keep coming back to this. Let me quote you, then myself from the fourth and seventh posts in this thread:

There is nothing wrong at all with mounting off-board components as long as you plan ahead and observe proper soldering technique.

I will conditionally agree with that as long as you underline, boldface, and italicize "as long as you plan ahead and observe proper soldering technique."
Your picture shows some decent PTP technique. However, in my experience, it's notable because it's the exception. My experiences with resistors or cap soldered in series with one end of a wire are legion.

There are good and bad PTP techniques. I pointed that out. I illustrated bad ones. You seem to be in violent agreement with me again, but can't let go that there are any bad PTP techniques. There clearly are. There's a picture of some.

It's possible to do bad work. Some practices are so fraught with opportunities to do bad work that it's simpler to use a different technique. But with enough effort planning, time, patience, practice, and a modicum of skill, you can thread through even a minefield. Or a daisy field.

Beginners ought to stay in the daisy fields until they know what they're dealing with.

[PRR]

> a good mechanical connection. The solder joint is an electrical connection and should not be relied upon to {resist mechanical stresses}

I need to tie a jack to my plow. Mechanical load. What metal is strongest?

• Steel
• Bronze
• Aluminum
• Zinc
• Lead (and tin/leads)
  

Steel is obvious. Bronze is good but currently expensive. Aluminum can be strong enough if I use a greater thickness (maybe less weight). Zinc would have to be awful thick.

NObody would use lead to tie a plow, build a bridge, etc. Aside from weight, it's FAR too weak. Weak short-term, and no-strength long-term because it "creeps" under sustained stress.

Lead is structurally useless, and tin/lead hardly any better.

OTOH....... but but but.....

> the soldered leads are just lying against each other with no mechanical connection

In fact that's not a terrible joint. I've had joints like that serve decades in mobile equipment. The key is that there's actually a LOT of solder for the mechanical stress. That cross-section of steel could stand 300 pounds, lead in our lifetimes maybe only 3 pounds. But the part is 2 grams, times a 50 Gee impact is 100 grams, is 0.05 pounds impact force, or 66:1 safety margin. It will break some day, but you may not be around to see it.

BTW: the Western Union splice is 100 years out of date and gross overkill unless you are stringing 100 feet of wire in the air without support.

My objection to stuff-on-pots is that the lugs on many pots are crap, barely last when gently soldered, and sure don't need other junk hanging and vibrating. That said, I done it plenty of times with few failures.

[greaser_au]

In my time on the road servicing (relatively modern) pinball machines and video games, inadequately supported components (whether tag or PCB mounted) were the most regular casualties by far.  Long leads are worst for this. Typically it was things like:
T5 & T10  bulb sockets mounted only by their (wide & flat) leads;
1N4004 diodes mounted to switches  (google 'switch matrix' if you're keen to find out why they are switch mounted -  it saves *miles* of wire in a pinball);
cement wirewound resistors;
...even board mounted LEDs with inadequate support in high vibration areas used to pull straight out of their solder. ("The Getaway"'s  supercharger emulated a small railgun with leds/phototransistors for timing control).

It is typical in avionics for taller board-mounted components like 'block' ceramic caps (at 5mm (0.2") square), and (even axial) electrolytics  to be fixed in place with a bead of RTV silicone after conformal coating (which 'glues'  the smaller parts to the board also BTW).  In mil-std design, the torque moments of EVERYTHING are considered...

Properly supported PTP work is likely all well and good with old-school components and normal-g environments, but leads are getting thinner and thinner these days...     That said, I have seen 3" of 20SWG tinned copper bus wire with a big sag in it in a piece of test gear that came in in for repair.

david

[teemuk]

Suppose I have ten or more components bundled up together like this and I only need to replace one.

     


Thank goodness for PC boards.

[LucifersTrip]

thank you Amplifier Corp of America:



http://www.luciferstrip.com/fuzz/fuzz-king-front1.jpg

[tubegeek]

So, in my day job, I teach people how to solder.

I've seen all sorts of horrendous work, (the "bad" photos in this thread aren't even CLOSE) but at least it's being done by folks who don't know any better. Yet. (They would get MUCH better at it FAST if they would only frickin' LISTEN when I tell them how, but that's another issue entirely.)

One of the things I often explain is that it is possible to make a solder joint that not only IS good, but is OBVIOUSLY good: there's no need for me to yank on a joint with my pliers to see if it is a cold joint when I can tell by looking it is clearly not.

Also, I explain the difference between a device that works TODAY and a device that works FOR QUITE A LONG TIME.

My goal is to take them part of the way (in 18 sessions) from folks who have no idea how to QC their own work to folks who have some idea how to. Stuff always breaks and I think it's important to be able to sift through the more and less likely ways it might be broken, and to be able to tell bad from better from decent.

Relative to this thread, what I'd say is this: some ways of connecting components are inherently more likely not to work for "quite a long time," and are harder to construct in such a way that they are obviously well-made. Building like that usually makes devices harder to troubleshoot because you have to check over many more likely-to-be-broken spots. Those ways of making connections, I hope we ALL can agree, are easiest to just avoid, whether we are mass-producing or making one-offs.

As I like to say, "Even my time is worth something."

I was just reading a thread in another forum about a MASSIVE power supply in a modern console which was blowing rectifier diodes at turn-on. The tech who posted several troubleshooting iterations and who got all sorts of intelligent advice and who spent several sessions investigating several extremely annoying and time-consuming "what-ifs,", finally reported that the problem was a small hole in a length of heat-shrink tubing that caused a hidden intermittent short. The designer of that power supply needs to be taken out back and put out of everyone's misery - what a stellar example of exactly the sort of problem we're talking about here. This is in a multi-tens-of-thousands-of-dollars production piece.

One other thing: repairing pinball machines seems to me like it might be the best job ever.

Finally, I'd like to include a gratuitous emoticon for no particular reason. This one appears to me to be drooling, so I'll go with it:

[PRR]

> it is possible to make a solder joint
that not only IS good,
but is OBVIOUSLY good

+1

[greaser_au]

One other thing: repairing pinball machines seems to me like it might be the best job ever.

Don't you believe it, mister...    I've had a beard since I was 20. I started fixing pinnys at 26, and my beard was going very grey at 29 because I was looking after five company-run arcades (300 pinnys/videos), the sales set-up sites, and doing about 15-20 customer calls a week.

The level of STUPID problems I used to see...  I have to admit that I developed a reputation for a 'reverse midas touch'* - especially with intermittents, if I could play it for 15 minutes without the fault coming back, it was probably fixed... Story: Most pinnys use a switch matrix for rollover/target switches, and pulse 12V (brake-light) bulbs with the 50V coil supply for bright flashers. One of my teammates had a 'wide-body' pinny in for repair that would randomly reset during play. He pronounced it fixed,  and I said I wanted a game (I hadn't played one of these). I managed to break it again.  He pronounced fixed it again, and then ASKED me to have a game on it - this cycle repeated at least five times until he found the switch with the loose mounting screw that would occasionally touch the flasher socket at the exact instant it flashed - and BTW this guy was *THE BEST* pinny tech in Adelaide.  It was nice to be wanted for what was otherwise a curse.

Also apropos of the thread,  the CD jukebox that came in that was 'fixed'  4 or 5 times - but would skip every time I used it  turned out to be an inadequately secured earth braid on the CD transport.

* this 'skill' is why I have no time for half-ar$ed builds or solutions. I've worked in IT for 17+ years, and have suffered because of ' the cowboys'.  I believe only best-practice solutions are good enough.

david

[tubegeek]

My wife has the same skill with cellphones. Perfectly good ones die or go berserk whenever she owns them.

[Jdansti]

>My wife has the same skill with cellphones. Perfectly good ones die or go berserk whenever she owns them.


I used to have a bookkeeper who could thoroughly destroy any computer she came into contact with. She blamed it on some kind of unusually strong electromagnetic radiation given off by her body. That didn't explain the soft drink residue I once found coating a motherboard!  Maybe if her computer had been wired PTP with no motherboard, the drink syrup might not have caused as much of a problem.

[Ice-9]

>My wife has the same skill with cellphones. Perfectly good ones die or go berserk whenever she owns them.


I used to have a bookkeeper who could thoroughly destroy any computer she came into contact with. She blamed it on some kind of unusually strong electromagnetic radiation given off by her body. That didn't explain the soft drink residue I once found coating a motherboard!  Maybe if her computer had been wired PTP with no motherboard, the drink syrup might not have caused as much of a problem.

Yeah and because of the size it would have been it would have been in a different room to her, problem solved. 

[Jdansti]

Possibly...  It would depend on the field strength of her computer-destroying electromagnetic radiation.  Maybe if the room were lined with lead. 

[mistahead]

I've regretted almost every "mechanically strong" solder join I've ever made at some stage down the line. Personal experience should have taught me to stop doing it... but I sculpt nonethless.

In my previous life as a customer facing geek I used to have to explain to people that the metal box did the thinking and if they were going to physically attach a computer to attack that part, you'd be amazed how well that worked - back then HP small form desktop PC's were HARD metal and the "temper tantrum" incidents became significantly less (though I'm sure the medics loved me for it). A similar "all tech hates me - it all just breaks" customer I revealed to just be a lazy arse who was actually frighteningly savvy about ways to destroy her desktop PC to avoid actually working, I could write an essay on the techniques and skills of this customer to blow shit up...