Just to clarify some things about ITRI/Soldertec data that has been quoted;
The dissolution rate graph shown in Davids document on wave soldering is a result of wire dipping tests designed to simulate, for instance, dissolution of coatings in molten solder joints - a fairly static situation rather than for wave soldering. Such data will obviously be dependant on the relative wire/bath size and saturation effects.
The drossing rate data quoted below was obtained in a small bench-top machine. The tests were performed only to provide an indication of whether sigificant differences between tin-lead and lead-free alloys may be expected. There was an indication that the SnAg and Castin solders produced slightly less dross and the SnCu slightly more but it was also noted that "Experimental variance was high and the standard deviations show that the spread of values for each solder overlap." The full report describes other factors such as the level of black powder in the overall oxide/metal dross mixture, for example, that SnCu produced a more powdery dross than Castin.
Drossing levels cannot be readily compared between different production conditions. As with SnPb, impurities may have other effects - these were not measured.
regards
Kay Nimmo
+++++ Visit our lead-free.org website +++++
Soldertec at Tin Technology Ltd, Kingston Lane,
Uxbridge, Middx, UK, UB8 3PJ
tel: +44 (0)1895 272406 fax: +44 (0)1895 251841
[log in to unmask], http://www.lead-free.org
*****************************************************************
Soldertec is a membership based research, test and information provider available to any organisation
*****************************************************************
This information given in this e-mail is for guidance only. It should be reproduced only in full, with no part taken out of context without prior permission. We believe the information provided in this statement and any attachments is reliable and useful, but it is furnished without warranty of any kind from the authors. Potential users should make their own determination of the suitability of any information provided and adopt any safety, health, and other precautions as may be deemed necessary by the user. No licence under any patent or other propriety rights is granted or to be inferred from the provision of the information herein. In no event will Tin Technology or any of its affiliates be liable for any damages whatsoever resulting from the use of or reliance upon this information.
The information contained in this e-mail is intended only for the individual or entity to whom it is addressed. It may contain privileged and confidential information and if you are not an intended recipient you must not copy, distribute or take any action in reliance on it. If you have received this e-mail in error, please notify us immediately by telephone on 01895 272406. Please also destroy and delete the message from your computer.
-----Original Message-----
From: David Suraski [AIM] [mailto:[log in to unmask]]
Sent: 19 November 2002 15:38
To: [log in to unmask]
Subject: Re: [LF] Lead-Free Wave Soldering
Hi Keith,
First of all, thanks for your reply. Personally I think it's nice to see a technical discussion now and then on the listserv.
Your reply is well written and thought out. There are a few things I'd like to clarify. Maybe some others can chime in, as well.
Regarding dross rates, you state that "Some, particularly those containing silver, can have a dross rate twice that of 63/37. Others have a dross rate about half that of 63/37." Could you direct me to the data for this? In our field experience and data review, Sn/Cu alloys appear to be the worst for drossing. ITRI in the UK (now SOLDERTEC, I believe) performed tests that show silver-bearing alloys (both Sn/Ag and especially Sn/Ag/Cu) offer reduced drossing rates as compared to Sn/Pb, whereas Sn/Cu showed approximately a 10% increase in drossing as compared to Sn/Pb.
Regarding the inventory management issue, I probably should have been clearer about this in my initial description. Of course, I did not mean to imply that someone might accidentally utilize paste for wave soldering or a solder bar for SMT- that would be some trick! We are referring more to rework. Two alloys on the same board means two cored wire alloys for rework. Managing which wire is used where can lead to real headaches, as anyone that has managed a hand soldering operation where wire solders of different flux types are in use can attest to. The fact that Sn/Ag/Cu flows better than Sn/Cu will very likely result in its "accidental" use in many cases, just as RA and water soluble wire solders are often "accidentally" used instead of the no-clean wire that should be used for certain jobs.
At any rate, the above is not the only reason that we are not so keen about utilizing Sn/Cu in the wave. The poor wetting of the alloy and inferior fatigue characteristics also come in to play. Of course, this is only based upon what we have seen in various manufacturing environments and published technical data.
That being said, if our customers want Sn/Cu, then we give them Sn/Cu. It's not our place as a solder company to try to dictate which alloys are used, but instead to provide reasonable data and advice on several viable solutions. It makes OUR inventory situation difficult, but hey, what can you do?
Regarding dynamic vs. static test methods, forum members may be interested to know that what spurred this study was a manufacturer contacting us because they had recently implemented a lead-free alloy for wave soldering and their solder maintenance quickly became out of hand due to copper contamination. Prior to this we witnessed similar problems with a Sn/Ag/Cu alloy used in a HAL process. The solder very quickly became contaminated with both copper (from the pads) and iron (from the pot itself). I understand the argument of lab vs. real world studies, but in this case the real world issues have come first and the lab work was performed to understand the dynamics by which they occur.
We also were not surprised that the Sn/Ag/Cu alloys stabilize at around the same copper concentration. What makes us nervous, however, is the fact that the excess copper cannot be easily "frozen out" of the lead-free solder like it can be with Sn/Pb alloys. This to me, combined with the fact that there are manufacturers out there experiencing excessive copper build-up in their lead-free wave pots, is the crux of the issue.
To be candid, I have never seen any hard data on the nickel-stabilized tin-0.7copper eutectic solder to which you refer . Maybe this is a good solution, I do not know. Are there any references that you can supply that provide data on its Cu dissolution rates, dross rates, throughput, fatigue characteristics, etc.? A comparison with it and the "straight" Sn/Cu eutectic? Also, am I correct that this is a patented alloy? Are there plans to make this product available to the industry so that other solder companies can sell it, or has this already been done?
Regards,
David
----- Original Message -----
From: [log in to unmask]
To: [log in to unmask]
Sent: Tuesday, November 19, 2002 8:39 AM
Subject: Re: [LF] Lead-Free Wave Soldering
Hi David,
You have hit on what is certainly a key issue in lead-free wave soldering- the management of the valuable asset of several hundred pounds of solder in the pot. The good news, however, is that the apparently bleak scenario that emerges from your laboratory studies with static solder pots need not necessarily be the one that will apply in the coming lead-free era. Ways of managing a lead-free wave soldering line for periods of years without needing to scrap the contents of the bath as frequently as you suggest have been developed and are in widespread use in commercial production.
Unfortunately the long term management of a wave soldering pot has not always been a factor that has been taken into account in the considerations of the various industry committees that have been trying to identify a lead-free solder that could be adopted as a standard for the electronics industry. And in the attempt to standardise on a single alloy for both wave and reflow soldering the significant differences between these two processes have often been overlooked.
It should not really be necessary to describe those differences in a forum like this but, just in case, it might be worth highlighting that in reflow the solder system that has to be managed is confined entirely to the tiny droplet of solder formed by the reflow of the paste printed on a pad and the (usually) two surfaces in which the solder is in contact. Given the range of finishes that can be present in that system and the fact that it could be above the melting point for perhaps a minute or more there is some fairly complex metallurgy that can occur in that tiny system. However, the consequences do not extend beyond that fillet. The next pad and the next board have the benefit of a fresh deposit of virgin solder paste On the other hand, in wave soldering the joints are formed by a shared pool of solder that carries in it a bit of anything on any board that has been passed over the wave that is soluble in the solder al! loy. It is a bit like the difference between a personal spa bath and a public swimming pool!
The experience of managing copper contamination in tin-lead solder should have alerted the industry to the potential problem that your study has highlighted. It is the tin in tin-lead that dissolves the copper but this tendency is dampened by the diluting effect of the tin which has almost no tendency to dissolve copper. It would be expected that when the lead is eliminated the tendency of the resulting high-tin solder to dissolve copper would increase and that is what you have confirmed.
The dissolution of copper is even more significant in lead-free soldering because whereas in tin-lead the copper is only an impurity, albeit a rather annoying one, in most of the lead-free solders the copper is actually an ingredient whose level has a significant effect on the properties of the alloy, particularly the melting point. You are thus faced with the possibility of one of the key solder alloy constituents varying in an uncontrollable manner.
The results you report in your study confirm one key factor in managing a wave soldering pot. The rate of dissolution of copper is related to the copper content of the solder. You report that the alloy with 0.5% copper dissolves copper at almost twice the rate of an alloy with 0.8% copper. That is the saturation effect- if you have already dissolved three spoonsful of sugar in your coffee it is more difficult to get a fourth to dissolve. There is some evidence that silver in the solder alloy accelerates copper dissolution so that the fact that the alloy that you found dissolves copper more quickly contains 3% silver while that which contains 2.8% silver dissolves copper at a lower rate could be significant.
It is worth noting here that the consequences of copper dissolution at this rate have already shown up in commercial wave soldering with tin-silver-copper solders. In addition to the problem of keeping the copper content within specification, tracks and pads have been eroded to less than the minimum acceptable thickness.
The fact that after 24 hours both alloys studied in your experiment ended up with about the same copper content is not really surprising- the copper content will increase until it reaches saturation at that temperature and the two solders tested are close enough in composition that it could reasonably be expected that the saturation limits would be similar.
This is not really relevant to practical wave soldering, however, because the equilibrium that the system reaches is dynamic rather than the static one simulated in your experiment. In wave soldering the printed circuit board is in contact with the solder for only a few seconds and there is constant turnover of the solder in the pot. Every board carries out a few grams of solder with it and solder is periodically removed from the bath when the dross is skimmed. Fresh solder has to be added regularly to replace these losses. Some copper (and other impurities) are carried out in the solder on the board and in the dross. A simple mathematical model can predict the equilibrium level of contaminants and the rate of approach to that equilibrium. Interestingly the equilibrium level of a contaminant turns out to be independent of the size of the solder pot although the size of the pot does affect the! time taken to reach that equilibrium.
By using a solder that has a copper content higher than 0.5% and which does not contain silver, and by topping the solder bath up with an alloy of lower copper content it is possible to establish a dynamic equilibrium with a copper content that is within specification. With proper process control that equilibrium can be maintained for months and years without ever needing to scrap even part of the contents of the solder pot.
The dynamic nature of the solder pot management in wave soldering also means that the problem of sluggish flow which you found in your laboratory study can be avoided. As you noted that sluggishness is the result of a build up of crystals of copper-tin intermetallic precipitating out of a saturated solution. If the solder bath composition is managed to keep the copper content below saturation at soldering temperature this precipitation will not occur and the solder will remain fluid and provide stable bridge-free soldering. Similarly the other problems which you foreshadow, clogging of baffle plates with intermetallic, can be avoided. (A more likely problem with baffle plates is erosion if you use a solder containing silver. But that is another story.)
As you suggest, your study would seem to lead to the logical conclusion that a solder based on the Sn-0.7Cu eutectic is the ideal solution for wave soldering. I believe that your grounds for rejecting that conclusion require further consideration. Inventory management is not really a major problem. Solder bar and solder paste are very different products that are managed separately even if they are the same alloy. I do not think that there are many companies who dump scrapped 63/37 solder paste into their 63/37 wave solder pot and there is no reason that they would resort to that practice when using lead-free solders. And if a board reflowed with a tin-silver-copper alloy is subsequently wave soldered with a tin-copper alloy there is unlikely to be any direct contact between the reflowed joints and the wave solder. There can be problems of remelting of joints reflowed with an alloy that ! melts at 217C when the bottom side is wave soldered with a solder with a melting point of 227C but in practice this seldom seems to occur unless there is some lead contamination in the reflowed joints, e.g. because of reflow onto tin-lead coated pads or terminations. With a nickel-stabilised tin-0.7copper eutectic solder bath temperatures need be only in the range 250-260C to achieve bridge-free soldering and good through hole penetration and fillet formation. At these temperatures it is rare that reflowed joints on the topside will exceed the melting point of the solder.
Your conclusion that a move to lead-free will necessarily mean a major switch to OSP finishes, with consequent greater copper exposure, seems to be premature. This possibility has been canvassed because of the belief that there is no viable lead-free process for HASL finishing of printed circuit boards. Several printed circuit board manufacturers in Europe are confirming every day in commercial production that lead-free HASL finishing is an attractive and viable process. Because a reasonably stable intermetallic layer has already been formed in this process further dissolution of copper from the boards during wave soldering is slower than it would be from OSP coated copper. And where very flat finishes are required on fine pitch pads there are a variety of immersion and electroplated finishes available.
In any case, OSP does not seem to have been as much of a problem as you foreshadow. Probably the majority of the boards that have been wave soldered with lead-free alloys have had an OSP finish and if the right wave soldering alloy is used copper levels have been manageable.
In regard to the gold-over-nickel finish which you suggest as an option to reduce copper pick up from boards, this finish will certainly have a place in lead-free era although, for the reasons I gave earlier, it is unlikely that gold would build up to levels where it could have detrimental effects. The worst case equilibrium level of gold contamination could be calculated by assuming total dissolution of all the gold on each board soldered and plugging this, together with solder top-up rates and the solder bath capacity into the mathematical model.
An issue that you did not raise is dross rate and it is worth noting that experience so far indicates that there can be a factor or up to four in dross rates between some of the lead-free alloys being considered. Some, particularly those containing silver, can have a dross rate twice that of 63/37. Others have a dross rate about half that of 63/37. This is clearly a significant consideration when selecting an alloy.
In summary, a conclusion based on the practical experience of more than 200 wave soldering machines in commercial production for up to four years, is not as pessimistic as the conclusion you draw from your laboratory experiments. As you have highlighted, lead-free wave soldering certainly presents new challenges for the process engineer but if the right approach is taken it is a manageable and economically viable process.
Keith Sweatman
Nihon Superior
Best wishes
Keith -------------------------------------------------------------------------------Leadfee Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Leadfree To temporarily stop delivery of Leadree for vacation breaks send: SET Leadfree NOMAIL Search previous postings at: http://listserv.ipc.org/archives Please visit IPC web site http://www.ipc.org/html/forum.htm for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 -------------------------------------------------------------------------------
-------------------------------------------------------------------------------Leadfee Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Leadfree To temporarily stop delivery of Leadree for vacation breaks send: SET Leadfree NOMAIL Search previous postings at: http://listserv.ipc.org/archives Please visit IPC web site http://www.ipc.org/html/forum.htm for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 -------------------------------------------------------------------------------
-------------------------------------------------------------------------------Leadfee Mail List provided as a free service by IPC using LISTSERV 1.8d
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Leadfree
To temporarily stop delivery of Leadree for vacation breaks send: SET Leadfree NOMAIL
Search previous postings at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
-------------------------------------------------------------------------------
|
