The latest on tin whiskers in lead-free soldering?

Traceability?

--
"Design is the reverse of analysis" 
                   (R.D. Middlebrook)

All that work for free? No.

Cheers

--
Syd

IOW, you can't...

--
"Design is the reverse of analysis" 
                   (R.D. Middlebrook)

Well, I'm not fundamentally dishonest, and I don't have a vested interest. I have never once seen a tin whisker under my 20x stereo inspection microscope in many years of using lead free solders. I think the problem is greatly exaggerated and probably insignificant compared to other soldering issues. I suspect that, because lead free is touted as 'green' or 'European' it provokes a knee jerk negative response in many.

But I'm not going to write a book. Have you seen any real tin whiskers yourself?

Cheers

--
Syd

Syd Rumpo wrote

Having done a lot of reading, I think the industry has got away with it, due to a number of factors:

1) The whiskers need an elevated temp to grow well, say +60C

2) The whiskers need a reduced pressure to grow well (aircraft and satellite apps)

3) The whiskers won't grow on component leads unless *bright* (smooth) tin plated, which most aren't (but some definitely ARE)

4) Ref 3) what really helps is if the surface is under compression (concave curve e.g. the *inside* of a transistor case)

5) SAC solder doesn't seem to have a big problem with whiskers because the surface finish is very dull

6) IC lead pitch miniaturisation stopped c. 0.65mm pitch (went to BGA for high lead count packages)

7) Very few products are going to fail within say a 2 year warranty

8) IF SOMETHING FAILS IT GETS CHUCKED AWAY AND NOBODY LOOKS FOR WHISKERS!

9) The vast majority of electronic mfg is consumer stuff which gets chucked away after a few years at most

So there you are...

I *would* be concerned about elevated temps and TSOP packages, in industrial products which are expected to work for years. Conformal coatings can be used there, or just use leaded solder and nobody will notice ;)

The conformal coating I mentioned earlier is difficult to store (6 month shelf life and a crazy price) and extremely hazardous if warmed up.

Does anybody make SMT solder paste with say 2-3% lead? I guess it must exist because the military demand 2-3% lead.

You are totally full of shit. Totally!.

We have used SAC solder since RoHS came in. All are stuff goes to industry and military so any failures are chased down to the root cause. We have not found a single failure due to whiskers. We never conformal coat.

Possible reasons.

SAC solder.

We have only recently moved to ICs with a lead spacing of less than

0.8mm. Many new parts don't have a large body version so we now have to use small pitch devices. I have been designing in 0.65mm pitch parts for about 2 years and 0.5mm pitch for a few months.

Luck?

Sorry, but I can't imagine any device that needs to have its connections (component pins) removed before use. Specific examples would be useful. In any case, I don't see this as a sufficient advantage to justify a change to RoHS solders.

From my warped position, I find the higher temperatures to be more difficult to work with, more destructive on the iron soldering iron tips, and produce far more "cold" solder connections if moved during cooling. It's also more difficult to inspect as a normal and "cold" solder joint have a very similar surface finish.

I presume you mean accidental lead contamination. Mixing RoHS and leaded solder creates a dull and very brittle joint. I sometimes run into the problem when repairing 2000-2005 vintage PCB's, that may user either type of solder. I try to use the same type but sometimes get them mixed up. When that happens, I have to clean and re-tin the solder tip, use a vacuum desoldering station to remove as much of the contaminated mix as possible, and then resolder with the correct type of solder. Not my idea of fun. Sometimes, I wish the EU had required a warning label for users of RoHS solders. The RoHS logo on the PCB is usually a good enough clue, but I've been fooled a few times by counterfeits.

In any case, intentional contamination of a solder joint, in order to facilitate component removal, also doesn't strike me as sufficient justification to intentionally switch to RoHS solders.

Look for something that is plated with pure tin or zinc. You will need a microscope or preferably a USB microscope and some really good lighting. Look for tin plated connectors from the 80's that have been sitting on a shelf for at least 10 years. The whiskers are sufficiently fragile that even walking across the room will blow them away. Therefore do not move the PCB in open air very quickly, and shield the PCB from and air flow. I ruined a good picture by simply closing the office door. That's also why I like using a USB camera, where I don't need to move the PCB. Lighting is important, but I find that my pocket Maglite or LED ring on the USB camera are good enough.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Raveninghorde wrote

I have not found any good data suggesting that silver stabilises the solder, one article saying it has no effect, but I suspect that it

*does*.

It certainly makes it flow much better.

However the flux used also makes a *dramatic* difference (to hand soldering, at least). The Almit solder I mentioned earlier has 3-3.5% of flux which is relatively high and clearly this helps too. We solder at 360C - same as with 63/37.

Possibly...

0.65mm is common enough - has been for over 20 years. 0.5mm is much less common.

The "problem" is that if you take my hypothesised list of why everybody has got away with it, and cross off each item which might have "helped" you, it's easy to see how IF there was a problem it might go undiscovered for quite a while.

Example: About 10 years ago we placed a few hundred 2.2uF 25V tants backwards. They were biased to 5V. I found it only weeks after we shipped the product, all over the world - by accident: I took a unit for some +60C tests, and it failed within about an hour. Another one did too. I quickly found the problem. We recalled the lot, giving people DHL account numbers to return them on, and shipping new ones in advance. Cost us a fortune. We got all but about 20 back. Those 20 never came back... Either they never failed (at say +30C they would have lasted probably years) or they failed and got binned... So, many little things can conceal a problem. At the extreme end of things, if you sell a product cheap enough, say under ?50, into the industrial market, you won't get returns anyway - unless you happen to ship 100 to the same customer and they all pack up! (I replaced that tant with a ceramic, right away ;) Now we use tants only where we need the specific range of ESR).

The space and military people are not taking any chances anyway...

IOW, I see no evidence that whiskers don't happen.

We actually find we can hand solder with the Almit stuff at 360C (same as before) and we can reflow solder with some other brand of SAC (we use a subcontractor for SMT) with a temperature profile which is within the specs of the old non-ROHS parts.

What one has to watch especially is miniature electrolytics. We had hundreds of Nichicon 50uF 50V ones go "dome shaped" after the contractor put a 1000 circuits through a very slightly higher (ROHS) temp profile. We changed to another Nichicon P/N and it's OK now.

I think soldering non-ROHS components with SAC solder works OK, because very few components have lead on their leads. Or is this wrong? The Japanese went unleaded c. 2002.

Anyway, soldering ICs with 63/37 coated leads, using SAC solder, works fine. The final product is not strictly "ROHS" compliant of course but nobody can tell by looking at the board visually because SAC solder joints looks crap compared to 63/37 ones anyway ;)

OK I can see that. Rework is a problem.

The ROHS logo probably means the company had their arm twisted by some big customer, but they continue to use 63/37 ;)

Everything made in China in last 10 years has ROHS on the carton.

Same with the stupid CE mark...

I saw the DB connector pictures, with massive amounts of whiskers. But surely if the connector is mated, the whiskers cannot short anything out. They cannot do anything useful on a female connector; it is only on an un-mated male that you can see them, and as soon as the male is mated with a female they will be swept away.

A LOT of cheap DB connectors that have the earthing fingers (and thus have a high mating force) are often tin plated. The higher quality ones, and the milspec ones, always without the fingers, never use tin AFAICS.

No, I mean that the non-ROHS part has its legs or pads double dipped in lead free (a 'dirty' bath followed by a 'clean' bath) to remove the lead as far as practical. It's not a matter of ROHS compliance, rather to make joints which will be reliable at high temperatures. Nowadays, that's rarely done as nearly everything is now lead free.

Sn/Pb is easier to work with, I agree, but the evils of ROHS are overstated. You get used to it, you carry on, the products work.

One side benefit of lead free solder which I've observed, is that you can send any manual assembly staff on a soldering course without upsetting them :-)

Cheers

--
Syd

Recommended practice is up to 300x magnification, with oblique illumination. Ring, "shadowless" illuminators are no good. The NASA site has photographs taken both ways. The difference is pronounced.

Not personally, but I've seen the effects of the several hundred amp metallic-vapor arcing that they can instigate. Whole busbars, just not there anymore.

--
"Design is the reverse of analysis" 
                   (R.D. Middlebrook)

Thank goodness airplanes won't fall out of the sky and nuclear reactors won't blow up if my stuff croaks. So, I don't need traceability. Since I'm a one-man shop, I have to do EVERYTHING here, from buying parts and designing boards, to building, testing, shipping and customer support. I really can't handle any more levels of paperwork.

Jon

Tell me about it. I recently was able to turn my production testing over to the fab house that makes my boards reducing my work to paperwork. But my customer has moved production to Mexico so I am doing international shipping now. That is *such* a hassle if you don't know the ropes. Even knowing the ropes it is not so much fun. I spent a bunch of time today dealing with not having written down my password for the government AESdirect site. Their requirements for passwords are not only far, far beyond what any web site I've ever used includes, it isn't even documented properly on their web site. I probably spent two hours total just trying to get onto their site over the last few days. Of course that was my fault for trying to *remember* the insanely complex password. I should have known better, they make you replace it every two months anyway and if you don't use it for a month they deactivate your account... :(

Then there is tax time...

--

Rick

What? They've changed TV standards once in some 60 years... that is "ever more often"???

--

Rick

The original set-top box:

.. from that to 8K 3D TV, there have been a few changes.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Really? From the time they introduced color and UHF I don't recall any standards changes until we had HDTV and digital transmissions. What did I miss?

--

Rick

UHF and color are already two changes. cable boxes, digital cable boxes, HD and digital HD cable boxes, VCD, DVD, Blu-Ray, 3D, 4K, 8K. RCA->Component->HDMI, internet (with Hulu, Netflix, Slingbox, Chromecast etc.) Lots of changes. Though they did succeed in preventing analog HD from ever being adopted.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

You not read too well...

He said the window between UHF's introduction and HDTV's first standard.

There was no hulu or any of that crap. It was just as MPEG-2 was being formed that the HDTV standard got mapped out too.

ALL satellite transponders were still being fired with standard 4:3 SCPC and each 6MHz slot still only carried one channel.

All the cable boxes (companies) did was expand the channel fill to include the rest of the already existing band allocation. At ONE channel per 6mHz slot. After MPEG-2 and As HDTV was being formed, there came MCPC and a new compression schema which allowed those same 6MHz wide channel slots to carry more than one channel. That was when we saw our channel counts shoot up and satellites that were already in place multiply their capacity overnight. But the NTSC video standard that constructs the image... not much change there.