Build Longevity Tips

[karbomusic]

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.

[Crowella]

-Clip leads before soldering unless you plan to reflow the joint after clipping the lead. You create microscopic fractures in the joint when you clip after soldering.

Hmm, I might try that. I thought heat dissipation would be an argument against it, then I remember SMD is a thing.  

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.

I look back at my soldering and shudder.  I always wonder if people would be better choosing a lower gauge solder...

[Jdansti]

-Clip leads before soldering unless you plan to reflow the joint after clipping the lead. You create microscopic fractures in the joint when you clip after soldering.

Hmm, I might try that. I thought heat dissipation would be an argument against it, then I remember SMD is a thing.  

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.

I look back at my soldering and shudder.  I always wonder if people would be better choosing a lower gauge solder...

I've been happy with 0.032" for almost all stompbox circuits and connections.

[bloxstompboxes]

I just spent the past few days in a soldering training course. I've been soldering for a couple of decades now, but wanted to learn IPC standards for soldering and learn what it takes to reach the level of quality required in aerospace and critical medical devices. It's incredible the difference a few new techniques will make. All of the little soldering mistakes most of us make are actually considered pretty big defects in those fields even if the product works.

I've gone back to the drawing board on soldering:

-Clip leads before soldering unless you plan to reflow the joint after clipping the lead. You create microscopic fractures in the joint when you clip after soldering.
-Always use additional flux and clean it off after you create the joint...even if it is "no clean". If the soldering iron will touch it then flux should be on it.
-Using the correct amount of solder in a solder joint. Seems like a no-brainer, but it isn't. There are rules for different parts packages.
-Create a "heat bridge" by laying the solder on the joint then bringing the iron to it. It heats the joint faster and approximately dispenses the correct amount of solder for the joint.

This stuff doesn't matter so much in stompboxes, but it really matters in a rocket or a life support machine. That's the level of quality I want to strive for even if it is a fuzz box.

The project I am supporting now at work is considered class 3, I believe it is, since it incorporates some of safety features of the vehicle. Therefore the IPC standards we have to adhere to are leads no more than 25% off the pad and so forth. I think that's the correct amount. I had my recertification this past year, but all that stuff sort of runs together after a while. I just do a good job and then I don't have to worry about it. Besides, I don't even have to solder if I don't want to. I have a dedicated rework person for that. She loves that I do most of it for her though. lol. I'm not big on through hole desoldering with some of the gnd planes that stuff has so I do most of the SMD stuff because I like it.

[CodeMonk]

I've been happy with 0.032" for almost all stompbox circuits and connections.

I use 0.032 or 0.025 for almost everything.
The smaller sizes make it much much easier to do a quality solder joint.
Although I generally use 0.050 for soldering to the back of pots.
I got several spools of each for free a few years back (A friend of a friend died and left him all his stuff, I helped empty the house and he let me keep all the electronics stuff), so I'm good for a few more years.

-Clip leads before soldering unless you plan to reflow the joint after clipping the lead. You create microscopic fractures in the joint when you clip after soldering.

Clipping after soldering also exposes the copper in the component leads.
During my training, that was the main concern.

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.

Yeah and I see these $300+ "booteek" pedals with horrendous soldering on them and I just face palm at the audacity of someone that would charge so much for something with such poor build quality.
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), so I tend to be VERY anal about the quality of solder joints.
Although I may not always practice what I preach when building a prototype or something that isn't going to leave my possession, its certainly a lot damn better, IMO, than a lot of those $300 pedals.



And I always shrink sleeve wire solder joints.
Not only does it provide electrical insulation, it also ands more strength to the solder joint.
Hot glue for strain relief.

[mremic01]

> 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.

Is there a reason why? I've repaired amps where joints like that cracked and became loose. I usually clip the lead first (so even if it's oxidized, the clipped part isn't), bend it down so that it should have a good connection to the pad even without solder, then try to get a nice round blob. Works every time and I feel like a well shaped blob is less likely to crack that thin layer of solder where you can see the shape of the lead underneath.

[Thecomedian]

> 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.

Is there a reason why? I've repaired amps where joints like that cracked and became loose. I usually clip the lead first (so even if it's oxidized, the clipped part isn't), bend it down so that it should have a good connection to the pad even without solder, then try to get a nice round blob. Works every time and I feel like a well shaped blob is less likely to crack that thin layer of solder where you can see the shape of the lead underneath.

https://learn.adafruit.com/adafruit-guide-excellent-soldering/common-problems

Too much is as prone to problems as too little.

[CodeMonk]

> 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.

Is there a reason why? I've repaired amps where joints like that cracked and became loose. I usually clip the lead first (so even if it's oxidized, the clipped part isn't), bend it down so that it should have a good connection to the pad even without solder, then try to get a nice round blob. Works every time and I feel like a well shaped blob is less likely to crack that thin layer of solder where you can see the shape of the lead underneath.

NASA and Military electronics manufacturers would disagree with you.
And I am quite sure that they have spent more time and money studying proper and reliable soldering practices than everyone on this forum combined.
For them a bad solder joint could mean a 5 billion dollar piece of equipment becomes scrap or people die.
I know, I've worked for both.
And I've been doing this type of work (to earn a living and as a hobby) since 1981.

That link that Thecomedian posted is a good guide.
Although in that first picture only one of those solder joints would pass inspection.
Fourth one from the right.
The second one from the right still has to much solder.

If you have had solder joints like that that cracked and became loose, its likely another issue and not the amount of solder.
More likely poor technique somewhere during the manufacturing process.
Or parts with corrosion, oxidation or some other form of contamination.

[Crowella]

I was going to say, the second from the right is okay but could be better. For our applications, I honestly would not stress *that* much if there was the odd few like that... totally guilty.  

After a few attempts, I learned some better soldering techniques from my uncle who is a mechanical engineer in the medical industry. I've had to become better at it probably to appease him as a result, even if I turned to the dark side of computer science... Still not perfect (severe anxiety = shaky hands) but yeah, not going to worry about any of the joints now.   I actually went back and reflowed an entire pedal I made. What shock, it works flawlessly afterwards and probably took longer than the original soldering job, as well as a fair bit of solder wick.   A nice adjustable iron with interchangable tips also makes the job so much easier.  

The honest thing at the end of the day is practice. Are you going to get it right 100% of the time in a stomp box? No. Don't expect to get 100% but aim for it and you'll be golden, or volcano silver preferably. A good job early on saves hassles down the line.

Also, PCB cleaner and a toothbrush. I use it to help remove anything that might hinder inspecting joints to ensure they are all good. Always inspect your board after soldering. Magnifying glasses!

Yeah and I see these $300+ "booteek" pedals with horrendous soldering on them and I just face palm at the audacity of someone that would charge so much for something with such poor build quality.

Yeah... like this? $550 later...
http://i.imgur.com/pvROT7Q.jpg

[Brisance]

-Create a "heat bridge" by laying the solder on the joint then bringing the iron to it. It heats the joint faster and approximately dispenses the correct amount of solder for the joint.

A technique I sometimes use for smaller stuff is just applying solder to the iron tip and then heating up the joint with it. Once the joint gets to temperature, the solder will flow into place and voila.

[CodeMonk]

I was going to say, the second from the right is okay but could be better. For our applications, I honestly would not stress *that* much if there was the odd few like that... totally guilty.  

After a few attempts, I learned some better soldering techniques from my uncle who is a mechanical engineer in the medical industry. I've had to become better at it probably to appease him as a result, even if I turned to the dark side of computer science... Still not perfect (severe anxiety = shaky hands) but yeah, not going to worry about any of the joints now.   I actually went back and reflowed an entire pedal I made. What shock, it works flawlessly afterwards and probably took longer than the original soldering job, as well as a fair bit of solder wick.   A nice adjustable iron with interchangable tips also makes the job so much easier.  

The honest thing at the end of the day is practice. Are you going to get it right 100% of the time in a stomp box? No. Don't expect to get 100% but aim for it and you'll be golden, or volcano silver preferably. A good job early on saves hassles down the line.

Also, PCB cleaner and a toothbrush. I use it to help remove anything that might hinder inspecting joints to ensure they are all good. Always inspect your board after soldering. Magnifying glasses!

Yeah and I see these $300+ "booteek" pedals with horrendous soldering on them and I just face palm at the audacity of someone that would charge so much for something with such poor build quality.

Yeah... like this? $550 later...
http://i.imgur.com/pvROT7Q.jpg

Ugh
Yeah, not the best work I've seen, but not the worst either.

I don't always get to NASA levels of perfection in my own work, but I do try to at least get closer to that end of the spectrum than the other end.
Some of my prototypes are pretty damn ugly.

And the 2nd from the right isn't totally horrible. Would most likely still pass for general commercial work.
4th from the right would probably be good for NASA, but not so much for Military. But NASA also has an issue with weight, and when you have a a few hundred thousand solder joints, it adds up.
Military would prefer something between the two of the "OK" ones.

But hey, we're making pedals here.
No ones life depends on it.
But its nice to see it when someone take pride in the quality of their work.
If they don't seem to care about the quality in one aspect of a product that they charge hundreds of dollars for, how much do they care about the other aspects of the product?

-Create a "heat bridge" by laying the solder on the joint then bringing the iron to it. It heats the joint faster and approximately dispenses the correct amount of solder for the joint.

A technique I sometimes use for smaller stuff is just applying solder to the iron tip and then heating up the joint with it. Once the joint gets to temperature, the solder will flow into place and voila.

I do that myself as well.
It heats things up faster, lessening the chance of over heating components.
Very little solder, just enough to give better heat transfer, then I can apply the rest of the needed solder faster.
I'll also add some flux beforehand to keep things nice and clean.
Of course its a bit more of a mess to clean up later.
Its especially helpful with ground planes and the like that act like heat sinks.

[karbomusic]

Yeah... like this? $550 later...
http://i.imgur.com/pvROT7Q.jpg

One of the reasons it cost so much "should" be due to taking the extra time to solder properly but I'm sure many grossly underestimate real costs and inevitably raise the price in the name of mojo and reduce the quality in the name of not losing money. To be fair, quality and costs are always pushing up against each other but that isn't reason to let quality suffer. One of the reasons I started the thread was to improve build quality from the perspective that if it doesn't break, I'm spending less time repairing builds and rebuilding trust.

[PRR]

> Are you going to get it right 100% of the time in a stomp box?

If you solder a mainframe computer 99% right, that's about 100,000 failures you have built-in. Each failure becomes many hours of on-site high-paid technician time, plus the customer's lost revenue. So back in the day, computer makers were very-very fussy.

In Kary's case, 99% perfect in 50-joint pedals means about half his output is liable to come-back as an angry customer, shipping costs, sour reputation.

One goal for low-low failures is "Six Sigma". "A six sigma process is one in which 99.99966% ... are ...free of defects (3.4 defects/million)" G.E. worked to this goal, "although...this defect level corresponds to only a 4.5 sigma level."
http://en.wikipedia.org/wiki/Six_Sigma

Most of us will never have the massive data to show duds/million. I suspect that few of us will approach 100,000 solder joints in our life. if a dinosaur like G.E. can do 3 duds/million, is it possible that we can have *NO* solder failures in our entire career? (I admit that I am probably over 1/10,000, even though I HATE failed solder and blame myself bitterly for my failures.)


Too-thin solder joints should not be a problem. If they are, it is dirty surfaces or crappy mechanical support. Solder is weaker than copper (though not really weaker than trace-to-PCB bond), so parts which may shake-and-break should have mechanical support. A good pull-and-cinch will absorb much abuse. SMD parts are either sized so the solder will support the mass on the shaker-test, or a dot of glue is put under the package.

[Brisance]

On SMDs, I ordered a bunch of SMD resistors for experimenting, and what I noticed is that I can barely see the darned things, let alone solder them, those things are damn tiny!

[karbomusic]

> Are you going to get it right 100% of the time in a stomp box?

I got a call a while back from a friend that knew someone who was looking for someone to design and build handmade pedals for him. I already had a design I had been working on so I built him a branded prototype and took it over. He looked over the exterior, then popped it open and immediately threw on the magnifying head gear. A couple minutes later he said... "You are obviously aware of the importance of both presentation and the underlying build quality that goes with it", let's talk. Yay!

But... I don't have 1000 pedals in the wild so my work has really just begun because I can't know what I don't know, thusly go beyond hoping they'll be reliable. I have to proactively attempt to be ahead of the game because the moment serious players were paying several hundred dollars instead of just building them for my band mates, that game suddenly changed. That's fine, most of what I have done my entire life, has to been enjoying riding that line and being successful under pressure, so....

100% must be the goal regardless. The result of that goal should arrive at an acceptable level eventually. By eventually, I mean it is tough to beat the benefits of iterative refinement so any failure should be treated as an opportunity to prevent it from happening again. That can't happen without building a lot of stuff and having failures to improve upon, and that can't happen without at least some customers whom you care about experiencing failures. "Those are great pedals but they break all the time" is unacceptable so the failure rate must be low enough to prevent that.

I don't think I will ever be able to match the longevity of boxes made by a major manufacturer who has spent decades if not millions in making their products road worthy. I'll likely always be some distance behind that 8-ball, and I'll have to make up for it elsewhere just like a $5k Taylor guitar isn't as "road worthy" as a $300.00 acoustic made from plywood. However, I still have to constantly be aware of what could/does break and go to great lengths to prevent it without turning it into a $2000.00 pedal. Being overly aware of this helps keep us on our toes.

On a side note, I just threw a completed build on the scope and no worky. Also noticed it was pulling around 100ma when it should have been 22. A brand new pricey electro cap was bad out of the box causing the circuit to short. I'm glad I found it now instead of later.

[CodeMonk]

Yeah... like this? $550 later...
http://i.imgur.com/pvROT7Q.jpg

One of the reasons it cost so much "should" be due to taking the extra time to solder properly but I'm sure many grossly underestimate real costs and inevitably raise the price in the name of mojo and reduce the quality in the name of not losing money. To be fair, quality and costs are always pushing up against each other but that isn't reason to let quality suffer. One of the reasons I started the thread was to improve build quality from the perspective that if it doesn't break, I'm spending less time repairing builds and rebuilding trust.

I have actually given this some thought in the past.
It should actually take you less time to do a proper solder job than to do one with excessive solder.
With a proper solder joint, you are feeding it less solder so you should be spending less time feeding the joint its solder.
Its like if you have a 12 ounce coffee mug. Putting 10 ounces of coffee in it takes less time than topping it off with a full 12 ounces.

> Are you going to get it right 100% of the time in a stomp box?

If you solder a mainframe computer 99% right, that's about 100,000 failures you have built-in. Each failure becomes many hours of on-site high-paid technician time, plus the customer's lost revenue. So back in the day, computer makers were very-very fussy.

In Kary's case, 99% perfect in 50-joint pedals means about half his output is liable to come-back as an angry customer, shipping costs, sour reputation.

One goal for low-low failures is "Six Sigma". "A six sigma process is one in which 99.99966% ... are ...free of defects (3.4 defects/million)" G.E. worked to this goal, "although...this defect level corresponds to only a 4.5 sigma level."
http://en.wikipedia.org/wiki/Six_Sigma

Most of us will never have the massive data to show duds/million. I suspect that few of us will approach 100,000 solder joints in our life. if a dinosaur like G.E. can do 3 duds/million, is it possible that we can have *NO* solder failures in our entire career? (I admit that I am probably over 1/10,000, even though I HATE failed solder and blame myself bitterly for my failures.)


Too-thin solder joints should not be a problem. If they are, it is dirty surfaces or crappy mechanical support. Solder is weaker than copper (though not really weaker than trace-to-PCB bond), so parts which may shake-and-break should have mechanical support. A good pull-and-cinch will absorb much abuse. SMD parts are either sized so the solder will support the mass on the shaker-test, or a dot of glue is put under the package.

I probably surpassed that 100,000 mark every few months back in the day.
But when you have a soldering iron in your hand 8+ hours a day, 5 days a week, for several years, its gonna happen.
I hated getting any failures in my work as well, but it happens.
I think I averaged about 3/1000 rejected solder joints.
But that was all surface mounted ICs and inspected under a microscope.

[N9]

This thread raises some very good points about workmanship.

One resource I've found useful is the NASA reference pages for electronic inspection:
http://workmanship.nasa.gov/insp.html

I often wonder about the relative durability of solid core wire compared to stranded. I've always preferred very flexible stranded hook-up wire, but that's simply for convenience.
Perhaps someone here could shed some light on what's most accepted in high-reliability assemblies?

[GGBB]

It should actually take you less time to do a proper solder job than to do one with excessive solder.

+1

But only if you are already at the point where you are competent enough to not have to think how to do it right. Doing it however you normally do it, good or bad, is usually fastest. So for many DIYers, myself included, doing it properly does take a little more time because we have to slow down and concentrate on doing something we have yet to fully get the hang of. I'm sure I'll be much faster once I get good at it. I sure hope so. Good thing I don't have commercial quotas to fill. 

[CodeMonk]

It should actually take you less time to do a proper solder job than to do one with excessive solder.

+1

But only if you are already at the point where you are competent enough to not have to think how to do it right. Doing it however you normally do it, good or bad, is usually fastest. So for many DIYers, myself included, doing it properly does take a little more time because we have to slow down and concentrate on doing something we have yet to fully get the hang of. I'm sure I'll be much faster once I get good at it. I sure hope so. Good thing I don't have commercial quotas to fill.  

Well, I guess when you do it for a living for several years, it becomes second nature.
And no one liked getting a PCB back from an anal retentive inspector telling you that you "done fucked up", here, here, here, here, and here.

This thread raises some very good points about workmanship.

One resource I've found useful is the NASA reference pages for electronic inspection:
http://workmanship.nasa.gov/insp.html

I often wonder about the relative durability of solid core wire compared to stranded. I've always preferred very flexible stranded hook-up wire, but that's simply for convenience.
Perhaps someone here could shed some light on what's most accepted in high-reliability assemblies?

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.

[Crowella]

One goal for low-low failures is "Six Sigma". "A six sigma process is one in which 99.99966% ... are ...free of defects (3.4 defects/million)" G.E. worked to this goal, "although...this defect level corresponds to only a 4.5 sigma level."
http://en.wikipedia.org/wiki/Six_Sigma

I have numerous issues with six sigma. I think it's a horrible system for anything other than highly precise systems such as medical/military where budget/time/manpower is less of a concern and not something I would ever expect from the small level, especially guys making stompboxes.   I always aim for best quality but have to accept that in my case, there is going to be letdowns just because I might slip up one one solder joint by accident, sadly human error. Thankfully the checks I do after each pedal means I get a second chance to ensure it all works as intended and will continue to do so. Nothing wrong with a reflow in our case.

When I said do you expect 100%, I mean 100% perfect, time after time solder joints that look splendid. I can't achieve it over 10 pedals consecutively as one or two may look close to a blob but that doesn't mean that less than 100% of my joints work. Naturally, practice gets me closer and quicker.