Hi. This was the best place I could think of to post this curiosity query as there may be older engineering types here who would remember working on valve equipment which it pertains to.

I'm a user of valve guitar amps. There is a musical industry belief, (I have always used the word 'myth'), that PCBs in guitar amps are prone to dry joints everywhere as they are not up to the standard of 'hand wired' circuits. Almost the first response to anyone saying they have a problem is 'reflow all of the soldered joints on the PCB'. They seem to think that PCBs just grow dry joints. I have always maintained this is bunk and still do from my own industrial experience in electronics but recently I have had an experience which has made me look for more information on this aspect, creditable information in an engineering sense that is.

The idea that PCB technology is flawed and ultimately all PCBs are going to descend to dry and broken joints is a fallacy but I have actually experienced a situation where an amp of mine suffered a bad joint in one of the valve bases. It had been absolutely fine for a long time and was not obviously damaged even under magnification. I could locate the valve at fault by simply putting slight pressure on the valve tip and the reflow of its base solved the problem immediately and permanently. I thought nothing of it, just an odd statistical occurrence. Then at a later date I had occasion to do a little more mild modding and curiosity made me take the opportunity to reflow all of the valve bases even though I had no obvious problem. I was in there so it might as well be done. The amp seemed to come to life in a musical way even when set up in a configuration which did not use any of the modded modules.

I am a firm disbeliever in the idea of changes being identified by ear with something like an hour between hearing the before and after. I have run double blind listening tests with 'mojo gurus' on this sort of belief in the past with good engineering results, (a couple of magic tone capacitor type replacements produce a 'new amp' myth debunked). And of course there is the fact that a couple of other mild mods were carried out at the same time to confuse the issue so all is not transparent even though I do know the amp well from long time use. So, does anyone have any experience which points to the idea that the high voltages and perhaps temperatures experienced in valve circuitry do have an effect on solder joint integrity? I would have thought it more likely that this would be known in RF circles where the voltages would be considerably higher than in my 100W guitar amps.

 

Don't know what you mean by HT, but...

Solder joints that experience significant thermal cycling can develop cracks. I have had to reflow solder in the lamp control module in one of my cars a couple times. The connections carry more than 5 amps and it's a known problem area for people who do some of their own maintenance.

 

HT = High Tension?

There's always one enemy and that's flux residues. If the flux is not removed, the joints will corrode when you add a little bit of water/humidity/temperature. Electrolyte + dissimilar metals = battery.

Old school says, the joints have to be mechanically made first, then soldered. Then cleaned.

Oils from your hands could be an enemy them and an enemy now.

 

Hand wiring is definitely the best, because there is enough slack to absorb movement, whether from heat or general jostling. A solder joint is NOT intended to be the mechanical support for parts - only the electrical. If it is used as a mechanical support, then it will crack over time. High temp solder (HT) is even worse than leaded solder as for breakage, since it is more brittle to begin with.

 

I'm going to assume "TH" through hole..
And the cold solder joints becoming an issue is not bunk at all and a real problem.. Typically thermal cycling (turning product on/off) is what causes/accelerates the failures after time as stated in post #2.. But it could also be from mechanical stresses,etc...

Not all solder joints will fail... Some will though and thats a fact and reality that happens everyday..
There are many variables involved in getting a "for life/proper" solder connection and there is almost no way to 100% prevent it.. It is real though.. Nothing bunk/fallacy about it..

 

A couple of you have stated "it is real" but what do you actually mean by that? The fact that a joint can go crystalline or simply fracture is of course always a possibility so yes, it is real, but the idea that all PCBs are doomed to become riddled with dry or cracked joints is obviously a joke. Hand wired is not necessarily better in any way except for historical accuracy. If they are designed correctly, there are billions of boards with hundreds if not thousands of joints on each in use in equipment of all types which go on working day after day in hostile conditions for decades without problems. What we see in terms of visible problems is the most minute statistical fraction of what is successful out there.

I maybe should have said earlier, I'm not new to the process of soldering and I'm not a hobbyist. I have worked in electronics design for over 40 years and have designed for demanding environments like seismic surveying. I have worked on projects with high voltages 500V+ and immense current pulses 300A+ producing ultra high intensity mechanical pulses without any problems to the nearby control electronics on both through hole and surface mount PCBs, (seismic boomer sources for example). I have attended regular update courses on soldering theory and technique over the years so I am familiar with ideas like the mechanically sound joint is an electrically sound joint.

The truth of that, is that the parts should be in good clean mechanical contact before soldering and the solder simply surrounds that joint, removing air and preventing corrosion of the direct connection between the parts and mechanically stabilising it. It is not primarily intended to be a medium where current passes from contact1 through the solder to contact2. Unfortunately that is inevitable with most pcb joints and at normal impedances will make no difference at all. However at low impedances and higher currents it is a factor. This is all in line with the modern crimped joint which is far superior to the solder joint when made correctly, as the aircraft industry for one is quick to point out.

The valve arena is different however with its mix of high voltage and high thermal content. I wanted to know whether there was something in HT (yes, high tension, it's in common usage), and heat which had an automatically deleterious effect on the joints. It does produce its own odd effects like the notorious "black jam" which grew on places like the backs of cathode ray tubes in TVs and displays. The guys to ask about valves in depth are the old RF/radar engineers, not our amplifier guys. It isn't an accident that the goto books for real valve advice are still the Mullard and RCA books from the '50s. They tested properly in those days, not just propounded "common knowledge" as is all too frequent now.

Frankly, there are some glaring holes in the amplification designers' knowledge in a lot of high profile cases and a large number of guitar amps are designed with bad flaws in them. (The common Standby switch configuration is one, the DC Coupled Cathode Follower in that Standby setup is another.) The "theory" of valves and their perceived strengths and weaknesses in the musical field is a mess, full of myths and old ways of doing things the wrong way from when the people designing the first amps did not really know what they were doing in most cases. I just wondered if there was anyone who had that older experience of ultra high voltages and higher temperatures as found in old radar work. They would know if there was any inherent problem with the soldered joint under those conditions having had first hand day to day experience of the issue in spades.

I suppose we could always weld! ;-)

 

A couple of you have stated "it is real" but what do you actually mean by that? The fact that a joint can go crystalline or simply fracture is of course always a possibility so yes, it is real, but the idea that all PCBs are doomed to become riddled with dry or cracked joints is obviously a joke.

My "real" statement is that there is always the possibility for a faulty solder joint on a circuit board and it only takes one to render the circuit useless and in need of repair.. It is fairly common in any field/design..
Some of your post seemed to be implying a general statement that joint failure just doesn't happen at all.. and that a suggestion to reflow is useless.. That would be wrong as you seem to have recently stated..
I guess I didn't assume enough about your post to see that you were only intending that comment to be specific to a very specialized application/field only..

In general there are certainly stresses involved in soldering and its joint and heat/thermal cycling and mechanical vibration can cause those stresses to exceed a certain level and cause failure of the joint..
I am not aware of any phenomenon related to voltage levels causing solder joint failure.. "high" is relative though.. 500V is low voltage to many..

 

Some of the older circuit boards used rivets for the 'thru-holes'. many times the components ended up soldered to the rivet but not to the underlying copper. Then the joint between the rivet and the copper corroded and became an intermittent.

 

Voltages in the range of 400-600 V are common as dirt these days. Zillions of flyback switchmode power supplies are in use. In the "240 volt" part of the world, this means rectified mains is around 340 V. Flybacks reflect the output voltage to the "primary" of the "transformer" by the turns ratio (this is when the flyback inductor actually is behaving as a transformer - when you would rather it didn't), and stack that voltage onto the input DC, so voltage of 500-600 V is very common. In higher-power supplies, active power factor correction (called harmonic filtering in some places) typically boosts to about 400 VDC. Variable frequency drives for motors that rectify and filter 600 VAC input and produce 600 VAC output are everywhere. Some of the components are through-hole, many are surface mount. Many of these devices, if built by reputable companies, have mean time between failure specs of many hundreds of thousands of hours and many will run continuously for 10 years (about 90k hours) without failure.

I've seen more solder joints that have "failed" because the semiconductor melted its own solder than because of actual soldering problems.

The trend in recent years has been to "no-clean" fluxes. These are considerably less aggressive than the old activated rosin fluxes, which means it is more important that the things to be soldered are clean and free of existing oxides. Proper soldering of circuit boards is demanding and some of the equipment used is very sophisticated, especially when you get into boards with surface mount parts on both sides and through-hole parts.

The signal to noise ratio in the audio opinion world is very low. There is much unsupported and unsupportable yappage about what makes their openers for canned noise superior. There is a steadfast refusal to do double blind testing. I blow raspberries in their general direction. I'd cheerfully blow the pestilence that is subwoofer to kingdom come.

 

"My "real" statement is that there is always the possibility for a faulty solder joint on a circuit board and it only takes one to render the circuit useless and in need of repair.. It is fairly common in any field/design.. "
Absolutely agree there with your observation that a single fault makes the unit US mcgyvr. My questioning of the "real" implications in peoples' original posts was only that it seemed to somehow imply a degree of being 'common' which I don't really think it is. It may be common for a tech/engineer to have to address and fix bad joints but that is only because the ones brought to him would be the damaged faulty ones, the other possibly millions of units would still be functioning normally outside of his experience. I only questioned on the matter of scale.

"Some of your post seemed to be implying a general statement that joint failure just doesn't happen at all.. and that a suggestion to reflow is useless.. That would be wrong as you seem to have recently stated.."
Most certainly, solder joint failures do occur and I have had occasional experience of them myself. But in a long career working with gear of all types it has been a rare, I would even say very rare, occurrence. If it happens more commonly in other fields I will certainly accept that but would then, out of curiosity, have to ask why? What is there in that field which makes joint failure a more common happening? That in fact is what I am doing here in asking about high voltage and temperature specifically in valve circuitry. As I said before, while working in an area where there are high voltages reasonably in excess of the usual guitar amp levels, at very high currents way in excess of any experienced in music, and with incredibly high mechanical shock and vibration over seriously extended periods of time, (it isn't extraordinary to have these devices pounding away every half second continuously for over 48 hours), I have never experienced a single fault we could put down to a solder joint on a PCB. (Maybe we were just better at our design and implementation than I knew? ;^) ) I don't question the need to reflow some parts known to be susceptible to fracture but I do question the validity of the frequently seen advice which recommends reflowing every single joint on a board because an unidentified fault has occurred because it is one of the most likely causes.

""high" is relative though.. 500V is low voltage to many.."
Aha! Right on the button there. That is why I was particularly interested in the early radio/radar field where the voltages in transmitting devices can be very significantly in excess of that and powers are generally much higher too.

Ylli, your point is a good one and one that I have not considered. I wonder if this is from the era where maybe the idea of the "found everywhere" faulty joint originated? I can't help but feel that both that and KISS's original point about flux would have been more relevant then than now.

"There is a steadfast refusal to do double blind testing."
As Shylock said, "A Daniel come to judgment! yea, a Daniel!" Yes, let's never let something factual get in the way if it might disprove the mojo. There should be an insistence on it. But that wouldn't sell more components or pull in work for the workshop and techs. While I have noticed a marked increase in the number of engineering companies who are publishing online relevant results of well engineered testing of some of the myths, I haven't come across anything on this topic as yet.

"I blow raspberries in their general direction."
And I would have used the word "fart" and add to that... Their mother was a hamster, and their father smelt of elderberries. But I guess you are generally more polite than I am. :^D

 

Its certainly not as common as it was.. Tighter process control via datalogging/sensors/automatic wave soldering/selective soldering,etc.. has taken quite a bit of the process variables problems out...
I will say though that it still happens.. Much of it goes undiagnosed as many times people simply throw away the device now... The electronics world due to automation/low labor rates in countries like China,etc... have turned this into a throw away world where you don't repair stuff you just trash it and buy another..
I will say though that I've been a user of LED lights since they first started and every single failure I've had has been because of a cold/cracked solder joint and a simple reflow has repaired them..
Up until recently (I'm buying Cree now because of clear quality differences) I would say that the failure rate of LED bulbs for me was at like 90% .. I have not had a single Cree fail yet though.. But plenty of other cheaper ones..

 

I see solder joint wiring posts (properly mechanically wrapped and then soldered with a good eutectic alloy) or PCP connections that carry potentials of up to 250,000+ volts, 7/24 for 20+ years in a clean room environment. I've never seen a problem caused just from high voltage being there. Solder is a electrical connection, not mechanical, so it will eventually fail if stressed. Most long term failures I see are thermal cracks driven by stress on the solder joint or vibration. I normally use a silver alloy electronic solder for critical electrical connections where connectors are plated but I prefer crimps over solder for just about everything on wire terminals.

 

I have not had a single Cree fail yet though.. But plenty of other cheaper ones..

I've had 4 Cree failures and one Philips. I contacted Cree regarding the first 3 failures and they promptly shipped replacements; haven't contacted them regarding the 4th, I'm waiting for more failures.

Philips never contacted me after two queries. But they started spamming me about other Philips products. I contacted them for a single failure because I paid $25 for that bulb; I think it was the first, or among the first, available.

 

I've had 4 Cree failures and one Philips. I contacted Cree regarding the first 3 failures and they promptly shipped replacements; haven't contacted them regarding the 4th, I'm waiting for more failures.

Philips never contacted me after two queries. But they started spamming me about other Philips products. I contacted them for a single failure because I paid $25 for that bulb; I think it was the first, or among the first, available.

I'm not saying Cree is perfect.. But as you can see their customer service is second to none..
There was a news story around here about a guy that bought some for his house and a few had separated at the glass to base interface (back when they were coated glass globes).... Cree stepped up and sent boxes of them to replace every light in his house including numerous can replacements/trim,etc... It was thousands of dollars of bulbs I believe.. Now of course thats news report coverup stuff too but they still came through..

I really like their current 4 flow bulbs.. Working great for me..