Lead vs. Silver solder

[ilimzn]

Originally Posted by TANDBERGEREN

Solder or not to Solder, thats the question.
(At least it was in the late seventies )

Yes, wirewrap was used a lot then. But the components and connectors were bigger and manual labor cheaper. Still, there are components in use today in various high-rel applications that are not soldered, but provide contact through similar means to wire-wrap, i.e. cold weld. Most mil and com spec connectors with a lot of pins are made this way, they are simply pressed into plated holes in multilayer PCBs, however the holes and the plating have to conform to strict specs, and the pins are really more similar to small knifes. (edges are very similar to hardened square pin edges on wire wrap components - not a coincidence, no doubt).

I have seen many rebuilds and fixes of wire wrap joints in old equipment by soldering over the joint - a knee-jerk reaction of 'servicemen' with no knowledge of wirewrap. This is likely to form a joint of lesser electrical quality, although at first it may look more mechanically reliable. It's not - the wrap normally uses rather brittle wire but the last few 'windings' offer a form of 'spring' preventing the force that would bend the wire from concentrating into a short lenght thereby breaking the wire. Not so when it's soldered over. The wire wrap joint produces a very good contact even when it's badly corroded from the outside, I've often had to unwind and resolder the wire because it would be seriously weakened from the corosion, and found nice shiny traces fot he cold weld on the pin and wire once I have done so. Ironically, I've had many cases where mechanical forces would actually crack the solder joint holding the wire wrap pin to the PCB

[jcarr]

When it comes to choosing solders, in addition to considering ease of soldering and reliability, one of my methodologies is use subjective (listening) bypass testing with RCA cables.

If you have 3 solder types that you wish to test, prepare 4 pairs of RCA cables of identical length and construction.

I suggest using really simple twisted-pair solid-core RCA cables, with the cheap types of RCA connectors that are available in multiple colors.

Have someone else do the preparation and assembly work, and ask them assign one color of RCA jack to one solder. They should keep notes as to what solder went into which colored cable, but they should not show you their notes. IME, it is OK if you know that one cable is different from another, but make sure that you don't know the contents of each cable - otherwise expectation bias will affect the results.

All of the cable pairs except one should be cut at approximately one-half the total length, and soldered back together using one of the solders. Make sure that every detail of each solder joint is as identical as your assembler can manage (other than the temperature differences required between leaded and non-leaded solders). If the cable lengths of the positive and ground leads are offset slightly, you will not need to use heatshrink tubing. But if the solder joints look visibly different (lead-free vs. leaded), you may still want to cover the joints with tubing or tape.

The uncut cable pair will be the reference, to which every soldered cable will be compared against.

Listen for as long as you like, and write notes of how each cable, focusing particularly on how it compares to the reference (rather than like or dislike).

Not all listeners are sensitive to the same sonic artifacts, so if other experienced listeners are available, give them the set of cables and have them conduct the listening tests, too. You may not want to tell them what is different in the cables, as some listeners mentally become incapable of hearing differences if you tell them that they are listening to things that they consider inconsequential, such as solder or resistors.

After all of the listening tests have been completed, determine if any differences were heard or not. If not, then use any solder that is convenient, reliable and easy to work with. If consistent differences were heard, go back to your assembler and find out which solder it was that sounded closest to the no-solder reference.

FWIW, by now I have probably tested a few dozen solders, and to my ears at least, not one has sounded like the direct connection. Among the solders that I tested, there were two solders that I felt sounded reasonably close to the reference (but not to each other). One was a lead-bearing industrial solder made by a large company, and the other was an unleaded solder sold by a small audio accessory company.

[gmphadte]

In the Tektronix service manual for CRT oscilloscope, it is mentioned that solder with 3% silver is to be used for service. There has to be some reason for it.

Gajanan Phadte

Edit: This manual was from pre-ROHS era.

[marce]

Simon7000, I would interested in your results and measurements. But you do state hundreds of joints! would the results have any relevance in a real world situation where maybe 10 solder joints would be the max a signal would pass through, before some component is in the signal path.
I have never known a good solder joint to adde any distortion to any signal of any sort, and would imagine that even with Simons data, single joints would have a figure so low that it is insignificant.
As I mentioned earlier the IPC has a hell of a lot of info on soldering and how to do it correctly, most electronics production companies will train there operatives to IPC-A-610. Nassa and trhe nation physics lab have lots of info on tin whiskers, CAF and other gremlins to keep you awake at night when you have to do high reliability products (commercial, current thinking is a MTBF figure of 10 years before tin whiskers etc may render your gear usless!).
To put soldering into perspective (and how critical it is to the elctronics industry) there are somthing like 6 billion passive components placed a day, thats 12 billion (minimum )solder joints per day, so there has been and still is a lot of studying going on, the main push being to find a solder that is as good as tin/lead eutectic (or tin/lewad/silver).

[qusp]

Originally Posted by gmphadte

In the Tektronix service manual for CRT oscilloscope, it is mentioned that solder with 3% silver is to be used for service. There has to be some reason for it.

Gajanan Phadte

Edit: This manual was from pre-ROHS era.

whatever their reason, its not to make it go faster.

[marce]

Ironic Facts: The best way to stop tin whiskers is to add a minimum of 3% lead to the tin!
And for info NASAs tin whisker page:
NASA Goddard Tin Whisker Homepage

To put the problem into some perspective for all the small electronic devices we have, including computers TV's etc, 0.5 and 0.4mm pitch BGA devices are becoming more common, so distances between conductive surfaces is decreasing, so the whiskers have less distance to grow. And with the ever decreasing operating voltages the chance of a whisker being fused decreases. But as they expect us to change our phones etc every year or so it is not seen as a big issue. Though again for expensive items such as decent camera lenses, figures of 10 years are being quoted.
These are a BGAache to lay out on boards.
http://www.ti.com/lit/an/spraav1b/spraav1b.pdf

[wintermute]

The solder I had been using for quite some years has nearly run out. It was a 60/38/2 sn/pb/cu blend.

I hadn't been able to find any to replace it but just did a search of element14's website today and found I can buy 250g 500g and 1KG rolls of it, as well as 60/38/2 sn/pb/ag solder in various sizes.

The copper blend is cheaper. any oppinions on pros and cons of using the copper rather than silver? The copper has a 183c melting point and the silver 179c melting point.

Tony.

[tinitus]

Originally Posted by wintermute

The copper blend is cheaper. any oppinions on pros and cons of using the copper rather than silver? The copper has a 183c melting point and the silver 179c melting point.

Tony.

might try a modest 'test'

solder...and resolder
and see what happens

my guess is, the one with high copper content looks like crap, and the one with silver doesn't

[qusp]

I use Cardas Quad tin/lead/copper/silver for just about everything, or Kester for the finer stuff. its VERY unusual for me to need the finer stuff even on fine pitch work, I use cardas organic flux as well because I like paste vs liquid.

I have 2lb of the cardas solder, i'll probably get lead poisoning before I run out

[5daudio]

Originally Posted by jcarr

Not all listeners are sensitive to the same sonic artifacts, so if other experienced listeners are available, give them the set of cables and have them conduct the listening tests, too. You may not want to tell them what is different in the cables, as some listeners mentally become incapable of hearing differences if you tell them that they are listening to things that they consider inconsequential, such as solder or resistors.

the most accurate post in this thread, at least you have listened in blind tests, its not blind speculation about things you think you understand but don't,
nothing makes me laugh more than having a dataman/meter watcher scratch his head in disbelief when you can identify blind 10 or 20 times in a row what they can't hear or measure.