Reasons for/against circuit integrated with 3PDT PCB?

[Arcane Analog]

There appears to be some very heavy shielded cables dangling off those jacks and footswitch. Maybe it's the photo, but I can't see anything tying them down? If that's the case then you have a serious reliability issue- those cables are going to vibrate all over the place, and the stress is directly on the solder joints.

...says the fellow with two unprotected 12AX7s sticking out of the top of a stompbx.  

Bottom line is that Military Spec - which is the standard for a quality build with durability in mind - incorporates wired leads. If PCB mounting was more secure/reliable/durable the US military would have used that technique.

You got the number of the spec you follow?  I'd be interested in looking at that.  

Sorry - direct me to where I said I built to mil-spec and perhaps I can go from there.

This has simply got out of hand. I guess this is what I should expect from disagreeing with someone that everyone has become acustomed to blindly hoding as the authority on all things stompboxes.

Edit: My apologies to RG and the forum for ever doubting his wisdom and expressing a contrary opinion and pushing for a direct answer. I know that is not going to be enough so continue on with tar and feathering me.

Enjoy!

[R O Tiree]

Bottom line is that Military Spec - which is the standard for a quality build with durability in mind - incorporates wired leads.

Present tense... implication that it still is the required spec. That deserved comment, I felt.

Of course the military uses different building practices now.

I've quoted you exactly...

...every time someone misquotes, implies I said something or complete fabricate a statement (see the two posts above) you leave me little choice but to clarify or correct you.

That's all a little disingenuous, isn't it? Particularly since it was you who opened the lid on mil-spec, leading to a fairly natural assumption on posters' parts that you adopted at least some of the techniques therein? Personally, I think those plastic-body DC jacks are more than adequate for our use - you'd have to work fairly hard to mess them up.

OK, that aside, let's examine the forces involved... It takes roughly 6 lbf to actuate one of those switches (I just measured it). Your leg weighs, what, 40 lbf? Perhaps a bit more, depending on your general size and build. Plus the fact that your leg/foot can be moving at quite some speed as it hits the switch, especially if you have an "Oops" moment as you realise that you need that effect RIGHT NOW! and you apply an even bigger proportion of your weight to it. As the plunger is depressed, the internal spring compresses. As soon as the force on it exceeds 6 lbf, the switch clicks over and the plunger hits the end stop. Let's adopt your (apparent) argument that any excess force gets transmitted through the switch and into the PCB... Well, what would happen if there was no PCB? Surely the switch would disintegrate? Clearly, it doesn't do so, so there is something else at work. OK, so the plunger hits the end stop, which is actually machined into the metal barrel. The excess force is therefore distributed from the metal barrel, via the nylon washer and the nut, into the top face of the enclosure, down the sides and thence to the floor. The enclosure is more than capable of dissipating this force. The excess force never makes it any further into the switch than the end stop in the barrel. So far, so good...

What about the remaining 6 lbf? Well, the little metal "fishtails" that secure the plastic body to the metal top are more than capable of absorbing this, many, many thousands of times. What happens to that spring pressure? The little lever arm inside that flips the contacts over is only about 0.15" long each side of the fulcrum, so that's 0.9 lbfin, although it snaps over fairly fast. But the plastic body is very sturdy, so the impulse is absorbed by the body itself and then re-transmitted via the fishtails into the metal barrel and thence to the enclosure, yada, yada.

Vibration is a whole different story, however. A dinky little board with only a couple or three opamps plus associated components will not worry the switch unduly. Larger boards will, though, as I said above, and some additional support or alternate method must be adopted in this case.

Hopefully that all explains why mounting the switch onto a typical stomp-box sized board is highly unlikely to do any harm to your PCB in normal use. As I've said before (a few times, on various threads) I've only ever had one pedal returned due to a broken switch. That was a result of the customer trying to disassemble the pedal (presumably to reverse-engineer it). Because he didn't realise how I'd assembled it, he broke it - the fishtails holding the plastic body to the metal top were twisted and deformed, the LED leads (also board-mounted) were twisted around and the LED body was cracked (although it still worked). Clearly, quite some torsional force had been applied to do that much damage which would never be experienced in normal use (which only ever involves longitudinal, straight-line force). Interestingly, there was absolutely no damage to the PCB pads or traces around the switch area or the LED mounting. Those milled slots of mine could well have saved the day? Who knows? Certainly any shear forces were transmitted over only a couple of thou into the large-ish solder-blobs on the pads, which are quite wide, giving maximum copper adhesion area onto the FR-4 substrate...

Whatever... bottom line, if you do it right, it works very well, even after abuse like that.

[Perrow]

...via the nylon washer...

[CodeMonk]

Saying something was built to "Mil Spec" by itself doesn't mean shit.
I could say that my car was built to "Mil Spec" means it might be a Lamborghini.
It could also be a KIA.
The words "Mil Spec" cover so much ground that the words by themselves mean nothing.

[Arcane Analog]

I found this last night and forgot to share. I especially find the "Pros and Cons" chart to be interesting.

http://www.geofex.com/article_folders/pt-to-pt/pt-to-pt.htm

Carry on!

Edit:

[aron]

Ok keep it cool guys.

[R.G.]

I found this last night and forgot to share. I especially find the "Pros and Cons" chart to be interesting.

I'm glad you liked it. It's from an article I wrote about why automatically assuming that PCBs in guitar amps are bad for tone is incorrect.

The common wisdom among techs used to be that PCB based amps are automatically both bad sounding and unreliable. This is, of course, incorrect. There are better and worse construction techniques for both point to point and PCB in amps, as well as the tagboards that are bandied about as "point to point", but are not.

I think the paragraph before and after that chart are worth looking at too, so I've reproduced them here. Here's the one before:

So - PCB based amps, from what we've seen so far, have poorer reliability and techs don't like to repair them; but that is because they were poorly designed from a reliability standpoint in the first place. The tone is not necessarily poorer. Here are some comparisons

And then after the comparisions:

Most PCB amps that have been produced *have* been poorer than tube amps, for reasons having nothing to do with the PCB's. Like plastic bobbins in transformers, this does nothing to change the tone. The *other* poor practices that go with a cost-cutting attitude that were introduced at the same time may, but PCB's are unfairly indicted.

I've added emphasis to the parts I think are relevant here. There are better and worse ways to design PCBs, or point to point, or whatever-the heck-else in terms of reliability. If someone wants things to be reliable, need to take the time to know what the stresses are and come up with ways to deal with them.

[R O Tiree]

...via the nylon washer...

I knew someone would bite at that one but, aside from aesthetics...?

[Arcane Analog]

I've added emphasis to the parts I think are relevant here. There are better and worse ways to design PCBs, or point to point, or whatever-the heck-else in terms of reliability. If someone wants things to be reliable, need to take the time to know what the stresses are and come up with ways to deal with them.

I am advocating that wired leads are more durable (sturdy) and easier to repair (withstands repair surgey well) than the same circuit with PCB mounted external components. I also took the position that PCB mounting offboard components are not as easy to repair, are more prone to damage during repair and are much more inexpensive to produce which is clearly the reason large manufacturers are using the technique.

Your artcile, while certainly about the tone of amplifiers, still address the current topic directly as pros and cons of each build technique. You include sonic and practical pros and cons. You specifically list "durability" and "ease of repair" as a pro for PTP. You specifically listed "fragile to repair" and "extra design work just to avoid mechaincal problems" as a con for PCBs. Your article directly relates to this discussion and reinforces my position. Of course each technique must be planned out and executed properly. I have not said a poorly wired lead is a good lead. The same can be said for a poor PCB. However, no matter how well the PCB is designed/engineed, the above points, as summarized in your article's chart, still hold true. Granted, many builds are impractical and simply too expensive to produce for tag/turret and PTP builds hence the widespread use of PCBs but that is not what is being debated presently.

Edit: As an aside, I would also submit that turret or tagagboard is more durable (sturdy) than PTP assuming we are comparing true PTP that is absent terminal strips, etc. Tag/turretboard has the advantage of strong mechanical connections associated with PTP and the advantage of using a board (PCB) for added stability of the components. Tagboard or turret is the best of both worlds on the durability front.  

[R.G.]

I am advocating that wired leads are more durable (sturdy) and easier to repair (withstands repair surgey well) than the same circuit with PCB mounted components. I also took the position that PCB mounting offboard components are not as easy to repair, are more prone to damage during repair and are much more inexpensive to produce which is clearly the reason large manufacturers are using the technique.

Ah. That explains a lot. And thank you for writing down something that can be discussed.

I am advocating that details matter. I think that if you know the environment and usage of the thing you're designing you can meet the needs of that use.

I think that knowledge of the various factors that would degrade a circuit's function lets you design to meet that, and that there isn't necessarily an advantage or disadvantage to PCB-mounting things IF IT'S DONE CORRECTLY FOR THE APPLICATION. "Durable" to me means "withstands the foreseeable environmental and usage degradation, plus any reasonable random degradation". "Easier to repair" to me means "the person doing the repair can do it quickly and easily".

I think a lot of "ease of repair" depends on the point of view of the person doing the repair. I personally prefer fixing PCB mounted stuff unless it's really poorly designed. I understand and in some cases sympathize with people who have difficulty with this - it takes some learning.  For ease of repair on a PCB, I like to cut the leads off the faulty component where it sits, remove the leads one at a time, then insert the replacement.  Done. It's even incredibly EASIER if you have special tools like solder-tweezers or a hot-air station. The solder tweezers are only $30-$50, but the hot air stations the board repair techs used to use on the manufacturing line are expensive. It was amazing to watch a tech remove and replace 48-64 pin quad flat packs with one of these. It took under a minute in most cases. But the tech had the tools and training.

I personally *hate* cutting and stripping wires. And I use a PTS-10 thermal stripper to do it. Still hate it.

As an aside under the heading of "durability', stripping wires is one place where unskilled work can introduce failures. Unless the repair person uses thermal stripping tools (which mil-spec used to require, back when mil-spec applied) there is a high chance of nicking the wire with the stripping blade, building in a stress point to start a break. Unless that nick is covered with solder, and perhaps even then, the chances of a break there go up hugely over the chances of breaking the wire itself. My views on this are probably unduly influenced by working on Thomas Organ Vox Amps, which I love the sound and look of, but which have built-in reliability problems.

The Thomas Organ amps are built on one large PCB with flying wires off to the controls and switches. Should be fine, right? Well, there are issues. I challenge you to find an amp tech that will work on one. Such techs do exist, but boy, there are not many. The problem is the wiring. They used solid-core wires, mechanically stripped. The wires break at the stripping points. And there are yards and yards and yards of the wires. The bundles are properly laced with nylon lacing tape; that doesn't prevent them from breaking. I'm sure that components do fail on the PCBs, but I've only ever seen that happen twice, other than the electrolytic caps simply getting old and failing.

I'm bringing this up because the Thomas Organ designers did the layout and implementation for the off-board wiring very, very poorly FOR A MUSICAL INSTRUMENT AMPLIFIER that would be moved around a lot. I'm sure they used standard practice for the electronic organs they made as a normal product, and where that was fine.

But back at the idea that PCB mounted switches and pots are inherently more fragile. I disagree with your flat statement that there's anything inherently worse. I will happily posit that someone who does not understand how to keep stresses off a PCB or mount and support them properly will have higher failure rates, just like someone who does not understand how to prepare a wired terminal before soldering will have a higher failure rate; and that poor application and layout can make it a nightmare to fix. I will also happily posit that without training or skill, a repair person may inadvertently make a mess of a PCB during repair - just like they can make a mess of putting wires back in place, making splices, stripping and soldering on wire-to-terminal stuff.

In my mind this translates to "if you don't have the training or experience, the results are not going to be as good". There are some fairly simple rules for applying PCB mounted controls and such, some of which I pointed out. I don't think that blaming the parts for being poorly applied is fair. It makes a lot more sense to say that whomever designed this layout with PCB mounted parts (or flying leads!) did a poor job of keeping stresses off the predictable weak points.

As an analogy, a bicycle is easier for a typical human to use than an automobile. It takes more experience to use an automobile well. And there are situations where a bicycle is a more appropriate tool. The list goes on.

And yes, there is a whole lot of unskilled, poorly-thought-out design work done. But we ought to try to do better, right?

[pappasmurfsharem]

I think a lot of "ease of repair" depends on the point of view of the person doing the repair. I personally prefer fixing PCB mounted stuff unless it's really poorly designed. I understand and in some cases sympathize with people who have difficulty with this - it takes some learning.  For ease of repair on a PCB, I like to cut the leads off the faulty component where it sits, remove the leads one at a time, then insert the replacement.  Done. It's even incredibly EASIER if you have special tools like solder-tweezers or a hot-air station. The solder tweezers are only $30-$50, but the hot air stations the board repair techs used to use on the manufacturing line are expensive. It was amazing to watch a tech remove and replace 48-64 pin quad flat packs with one of these. It took under a minute in most cases. But the tech had the tools and training.

RG have you thought about one of these


http://www.amazon.com/gp/aw/d/B009PP36MK

[R.G.]

Ooooh! Nice! I hadn't realized they were down that low.

I have and use these:
http://www.mcmelectronics.com/product/21-8230
They make removing both SMD parts and ICs and DIP parts on regular through-hole boards a breeze.

[pappasmurfsharem]

Ooooh! Nice! I hadn't realized they were down that low.

I have and use these:
http://www.mcmelectronics.com/product/21-8230
They make removing both SMD parts and ICs and DIP parts on regular through-hole boards a breeze.

How does it help with DIP parts?
I can understand single lead but how would you use that to remove a DIP8?

[R O Tiree]

Wires introduce problems all of their own. Vibration vs solder joints has already been mentioned. Solid-core wires can be made to dress very nicely but, if you nick the core as you strip the insulation off, then that introduces a stress-raiser which will cause the wire to fracture either under vibration or repeated flexure. In extremis, this can happen after only a few bends. (RG already said this while I was composing) Stranded wires are better at withstanding repeated flexure, as the strands are much thinner, but they are also individually weaker. Long wire runs lead to increased parasitics and can also require additional support (also covered in RG's article). Lots of wires bundled up in a rat's nest can get trapped as you nip pot, jack and switch nuts up tight and also take up valuable space in the enclosure, let alone introducing still more unpredictable parasitics, which could be critical in inducing oscillation in higher-gain pedals (again, alluded to in RG's article). If the wires are kept as short as possible (good for so many reasons) then you're probably going to have to disassemble the whole thing in any case as stage 1 of any putative repair in order to get at the item in question, in which case, if it works and if it's appropriate, why not PCB mount? About the only thing that I would never consider PCB mounting is a jack socket - been there, done that with a commercial pedal that some drunken buffoon trod on at a gig, and that was the end of the show. I managed to salvage it (fiddly repair, stipulated), but I never gigged with it again (eventually, one of my cats decided that it was hers and peed on it and that killed it for good).

Given that the preferred/recommended method is to make a fairly solid mechanical connection before you ever show it a soldering iron, this can make it necessary to replace each wire, rather than salvaging, depending on how good the mechanical connection was in the first place and how short the wires are.

Don't get me wrong - I'm not saying "No wires at all!" as they clearly have their place, but...

If someone wants things to be reliable, need to take the time to know what the stresses are and come up with ways to deal with them.

...which neatly paraphrases what people have been saying all along, really.

If your methods work for you, then fair play. If you can afford to turn down work simply because an item has PCB mounted components, then fair play. But if people say, "From a purely engineering stand-point, this solution has its place," then one should, perhaps, desist from the "You're all doin' it wrong!" mantra and figure that they might have a point.

[R.G.]

How does it help with DIP parts?
I can understand single lead but how would you use that to remove a DIP8?

Hot air tools give you a jet of air hot enough to melt solder. It melts solder over a whole area, all at once. What you do is to heat the entire area of pins. DIP-8s are easy. You hold the PCB in some kind of fixture so you can work on the component and the solder at one time. With SMD parts, there is only the one side to contend with, but with through holes, you hold the PCB vertical so you can heat the solder side and pull on the component from the other side.

To remove a DIP, I would put some component-puller tweezers
http://media.digikey.com/photos/OK%20Industries%20-%20Jonard%20Photos/EX-1.jpg
on the component side, hooking under each end of the IC. While I held that in my left hand, I would gently apply hot air to the solder side with my right, while ever-so-gently wiggling just a little on the IC with the puller. With just a little practice, you can feel it when the solder goes liquid, and the IC pops out entirely.

I have in the past done a cruder version. I used a propane torch for the hot air source. This is too hot to use, really, but you can generally get the ICs out without damage, even if you do set the board aflame sometimes.  This was back in the day when 64K (yes, "K") memory chips were $10-20 each, and you needed them eight at a time, so it was worth the time to pull them off salvaged boards. Granted, this is one situation where the board was not being repaired, but was entirely expendable. But well-set-up hot air tools do a good job of heating a small area to solder temp and not much beyond.

They work very well indeed for ICs where all the pins are about the same mass. I would definitely try them on PCB mounted switches and pots, because if you could make all the leads go molten simultaneously, you could just lift them out like an IC. But not having done this, I don't know how much room you have between making the solder go molten and damaging the board.

On the other hand, a solder pot would make this a breeze. Get a good hold on the board with one holder in one hand, and on the component with another holder in the other hand and set the solder side on the solder surface. The solder's temp is not as extreme as the propane torch gasses, so you don't have so much overtemp to worry about, and the solder's thermal mass melts the PCB solder semi-instantly, so the part lifts out. Notice that this process and all processes using molten metal in quantity is as dangerous as cuddling your pet rattlesnake - solder pots scare me. Probably a leftover from my days running a liquid type-metal pot for a small newspaper in my teens.

In all cases, the idea is to make the solder on all pins of the part go molten simultaneously. That's also what the solder tweezers are for.

Failing that, snipping off all the leads and removing them one at a time is a good sidestep.

[Perrow]

"withstands the foreseeable environmental and usage degradation, plus any reasonable random degradation"

eventually, one of my cats decided that it was hers and peed on it

Don't think R.G's statement included that one 

[pappasmurfsharem]

How does it help with DIP parts?
I can understand single lead but how would you use that to remove a DIP8?

Hot air tools give you a jet of air hot enough to melt solder. It melts solder over a whole area, all at once. What you do is to heat the entire area of pins. DIP-8s are easy. You hold the PCB in some kind of fixture so you can work on the component and the solder at one time. With SMD parts, there is only the one side to contend with, but with through holes, you hold the PCB vertical so you can heat the solder side and pull on the component from the other side.

To remove a DIP, I would put some component-puller tweezers
http://media.digikey.com/photos/OK%20Industries%20-%20Jonard%20Photos/EX-1.jpg
on the component side, hooking under each end of the IC. While I held that in my left hand, I would gently apply hot air to the solder side with my right, while ever-so-gently wiggling just a little on the IC with the puller. With just a little practice, you can feel it when the solder goes liquid, and the IC pops out entirely.

I have in the past done a cruder version. I used a propane torch for the hot air source. This is too hot to use, really, but you can generally get the ICs out without damage, even if you do set the board aflame sometimes.  This was back in the day when 64K (yes, "K") memory chips were $10-20 each, and you needed them eight at a time, so it was worth the time to pull them off salvaged boards. Granted, this is one situation where the board was not being repaired, but was entirely expendable. But well-set-up hot air tools do a good job of heating a small area to solder temp and not much beyond.

They work very well indeed for ICs where all the pins are about the same mass. I would definitely try them on PCB mounted switches and pots, because if you could make all the leads go molten simultaneously, you could just lift them out like an IC. But not having done this, I don't know how much room you have between making the solder go molten and damaging the board.

On the other hand, a solder pot would make this a breeze. Get a good hold on the board with one holder in one hand, and on the component with another holder in the other hand and set the solder side on the solder surface. The solder's temp is not as extreme as the propane torch gasses, so you don't have so much overtemp to worry about, and the solder's thermal mass melts the PCB solder semi-instantly, so the part lifts out. Notice that this process and all processes using molten metal in quantity is as dangerous as cuddling your pet rattlesnake - solder pots scare me. Probably a leftover from my days running a liquid type-metal pot for a small newspaper in my teens.

In all cases, the idea is to make the solder on all pins of the part go molten simultaneously. That's also what the solder tweezers are for.

Failing that, snipping off all the leads and removing them one at a time is a good sidestep.

I understand the hot air.

I'm curious how the heated tweezers would bee useful to remove a DIP 8. You only have two points of contact and you still need a solder sucker. I'm sure there is some key element I'm missing

I assume the tweezers are basically soldering iron chopsticks?

[R.G.]

I'm curious how the heated tweezers would bee useful to remove a DIP 8. You only have two points of contact and you still need a solder sucker. I'm sure there is some key element I'm missing

"Tweezers" is a misnomer, too. Think of two soldering irons, with a hinge on the cold end so you can pivot the hot ends together. The hot ends have removable tips that are both replaceable and of varying widths so that you can heat all the pins on one side of a DIP at the same time with each side of the "tweezer".

You're right, if they didn't contact all pins at the same time, it would not be much help.

[Ice-9]

I'm curious how the heated tweezers would bee useful to remove a DIP 8. You only have two points of contact and you still need a solder sucker. I'm sure there is some key element I'm missing

"Tweezers" is a misnomer, too. Think of two soldering irons, with a hinge on the cold end so you can pivot the hot ends together. The hot ends have removable tips that are both replaceable and of varying widths so that you can heat all the pins on one side of a DIP at the same time with each side of the "tweezer".

You're right, if they didn't contact all pins at the same time, it would not be much help.

And the larger the component the more heat is needed as it takes a lot to heat to keep it hot enough for larger DIP packages. Is this thing up to removing something like a 28 pin DIP ? If it is at that price its a bargain.

[R.G.]

The largest tip sets are 30mm wide, which means they are 1.18" across the DIP contact area. That's enough for 11 pins on 0.1" centers, so no, a 28 pin DIP is out of the running based on the available tip sets.

I don't think that heat supply is an issue though. My unit goes from room temperature cold up to soldering temp in about 45 seconds. I'd be more worried about it overcooking something. I'm always pretty careful to get the device out fast before it has much time at high temperature because of that.

I've written about this before, but removing parts is an exercise in deciding what you have to end up with.  If the part (IC, pot, switch, whatever) has a ready replacement and you can afford to trash it getting it out, things are easy: snip off all the leads toss the body in the trash, and then lift out one pin at a time and clear the holes. In this case, you're doing your best to not damage the PCB.

If you are trying to salvage the IC and can trash the board, you can use techniques which destroy the board but take it easy on the IC. The propane torch comes to mind - I didn't mind a little flames, smoke, blistering and general disaster on the solder side of the board if the IC came out easily and cleanly.

The big problem is when BOTH the board and the IC (or whatever ) are too valuable to lose. Then you have to work one pin at a time, clearing solder from the holes one at a time and (usually) breaking any remaining solder threads by wiggling pins in holes - with luck this can be done without shredding up any through-hole plating. I have removed and replaced parts on computer motherboards with inner ground planes and had the thing work afterwards. This is NOT easy. The parts in question were big electro caps from the industry-wide capacitor debacle. They were installed flush with the board on top and I could not get at the leads to clip them off without chancing collateral damage to the motherboard's 6-mil-wide traces. So I did it the slow way. Shockingly - to me! - it worked when I got the new caps in. It took about 30-45 minutes per cap. The inner planes ate up the heat I was trying to apply from the outside that I was afraid of how hot I had to heat the board on the outside to get the inner plating hot enough.

It's much easier if you can trash one or the other to get a replacement done. Or have a solder pot and nerve, or a hot air repair station.