Poll: to reflow or not to reflow, that is the question.

[Jefft91]

Hi all,

I'm posting this in solid state. Hopefully this is where it belongs. Just for fun, I added a poll question below. It's the first time I've done a poll. I cannot see it in the preview, so hopefully it shows up correctly in the post.

I'm of two opinions on reflowing solder joints during restoration of vintage gear. In the past, I selectively reflowed joints that looked dry, were visibly cracked, or were connected to large devices subject to mechanical stress. However, I always finished the job concerned that I may have missed a bad solder joint that could cause problems.

The technique I used which is described in many posts, is to heat the cold joint to re-melt the old solder and add just a bit of new solder/flux. The challenge is to not get too much solder in the joint, and to not overheat everything in the process.

Yesterday, I took a board (F-2589, CD-4 & EQ board from a Sansui QRX-9001) that is very densely populated and replaced aged components. When I was finished replacing the components, my visual inspection and spot reflowing left me feeling that I most likely was missing some joints that were compromised. I thought there might be a "better way". So I decided to do an experiment.

Technique:

1. Coat the entire board in liquid flux. I used a small brush to spread the flux until all solder joints were wetted. I use MG Chemicals RA flux # 835.
2. With my (cheap) iron set to max heat, I proceed to re-solder every joint. I did not need to add solder unless the joint wicked solder and left the joint starved. Then I would add just enough to form a concave fillet.
3. I work fast, only heating each joint long enough to melt the solder and move on.
4. I work methodically across the board hitting every joint as I proceed, back and forth (like mowing a lawn).
5. When finished, I clean the board thoroughly with flux remover (my preference) or anhydrous isopropyl alcohol and let dry thoroughly before reinstalling.

I found with the added flux that the solder joints melted very quickly (much faster than dry solder), as one might expect. I had no lifted pads, and on inspection, all of the joints were shiny and well formed. I had to add solder to maybe three joints that wicked solder when heated. I think these were probably hollow and likely to fail, but didn't look abnormal before re-melting the solder. After flux removal, the board looks like new (see attached photo). Sorry, but I cannot find the before photoof this board, so I'm showing another board (that I didn't reflow) for comparison. Note how dull the old solder joints are compared to the shiny re-soldered joints.

So, I'm hoping this will spark a discussion on the pros and cons of the full reflow approach and feedback on the process that I described. If you're so inclined, please respond to the poll.



- Jeff

 

[HPMguy]

Kenwood and Sansui are both known for solder that doesn't age well, especially the earlier wave soldered boards. I've redone a few boards with lots of cracked solder joints, I figure if more that one or two are showing signs of failing there are a lot more needing attention.

Liquid flux is a good idea there.

 

[TapeForever]

Based on my experience with electronic test equipment design over a period of almost 40 years I'll be the "contrarian"
1) If there are bad solder joints (cold solder etc.) the cause is poor mfg process and the process needs to be fixed. And those failures will usually show up in months or year or so, not decades later. If you have a board with bad solder joints, yes they need to be repaired and the rest of the board examined carefully.. If there are a few obvious, that's an indication of bad process for that board and the whole board need to be examined. you likely will find more.
2) If soldering was done properly and solder joint integrity is good, IMHO resoldering a good joint is serving no useful purpose and may actually be degrading components. due to heat.
3) In almost 50 years working in electronic design I can probably count on two hands the number of bad solder joints I've ever encountered. None in electronic audio, most in automotive electronics. The rest in someones attempts to solder when they didn't know the proper techniques.
4) Failed solder joints as a cause or involvement with field failures was extremely low, very very rare. And our products were often subjected to rapid temperature changes from around 0F to turn on and operating in a temperature environment that was on many occasions above 120F.

Of failure modes that showed up, the early surface mount electrolytic cap premature failure from seal failure was one that caught us and many other mfg by suprise. it was early surface mount electrolytics, thru hole leaded parts no issues even long term. So I'm not an advocate of replacing electrolytics willy nilly either. Only if they are diagnosed as out of spec and causing a performance issue.

that's my opinion based on a long time test equipment designer and mfg mgr.

 

[JohnT123]

I've worked on some really quality gear where the solder joints were in really bad shape. Usually it was because of being used in a bad environment where it was subjected to temperature swings far higher than normal home conditions. Gear used in unheated shops and garages seemed to cause problems as did frequent swap out of connected peripherals. Most I/O connectors are soldered to main boards or headers and repeated disconnect/reconnect cycles can result in cracked solder joints.
For anything that's had a rough life I'll do the entire board. It's usually less than a second for each joint. I set up, put on my magnifier and just go.