Internet commentary
A heavy summons lies like lead upon me, and yet I would not sleep[1]
Over the past few issues, my major concern has been one of security. Even though the Sasser and SomeFool worms are new – and dangerous –issues as I write this, I'll discuss Web site design, this time round. Regular readers from a few years ago will remember my concerns about the speed of downloading Web sites. They may also be able to recall that I promoted the use of meta functions to improve search engine performance. Things have changed, but not that much.
Let us keep in mind an inescapable fact: only about 5 per cent of users connected to the Internet have broadband connections. As a case in point, I do not have one, although I requested it nearly 3 years ago. DSL hardware, at the exchange level, is simply not available where I live. The same applies to many countries; for example, my daughter in Switzerland cannot obtain an ADSL connection, even though she lives in a dormitory village just outside a large city. The same applies to cable systems. So, I appeal to everybody who designs a Web site, or who orders a Web site to be designed, to remember that the vast world uses telephone line modems which are limited in speed and bandwidth. Furthermore, in some developing countries, these may be operating at much smaller rates than you may think. It is certainly not unknown to have connections at 14 kilobits/second or even less. I'm lucky, in one way, that the cost of online work is low, at CYP 0.20 hour (GBP 0.22, USD 0.35); my daughter is less so at CHF 4.80 hour (GBP 2.10, USD 3.30) during working hours. So, don't forget that your viewers may be paying through the nose for the privilege of looking at your site.
Of course, if you can be absolutely sure that every single one of the persons you wish to view your Web site has wideband connections, then please feel free to ignore my remarks! What I would recommend to all the others is to check your Web site out with a telephone line modem; if you find that the download times are excessive, then you know that something has to be done if you do not wish to antagonise those who come to your site. Bear in mind that the patience of surfers is limited and that they will often abort a visit if they cannot see what they wish to see within a matter of 15 seconds or so.
This general message has not changed over the past years, although the reasons behind it are slightly different. Please allow me to give a few hints how to attract more people to your site.
A couple of years ago, I ranted against those with enormous numbers or sizes of graphics, especially on the home page. This is still valid and I recommend a total graphics size of less than 30 kilobytes. As an illustration, just open Google at www.google.com. No matter how slow your connection, it will come up rapidly, simply because the page has been well designed. Yet the title is in the form of a graphics. What else do you notice about the page? It has no unnecessary advertisements, while the links are in text form.
The important thing with graphics is to minimise them. Some Web site authoring programs will allow you to physically reduce the size of a graphics and then automatically reduce the file size commensurate with the physical size. This technique can make dramatic improvements to the download time. Another possibility, where a photograph or a detailed drawing is required to be shown,is to simply have a thumbnail image linked to the full-sized one. This can also be done automatically with some authoring programs.
There is nothing new about what I have said up to now. What is perhaps new is what I detest most! This is the home page with a flash image. I would guess that I abort downloading a site nine times out of ten, if I see a flash image starting to load, without a link to bypass it. I always know that I'm in for a long wait if I let it go on. My advice is to avoid flash like the plague; if you must have it, make absolutely sure that your viewers can disable it. The same applies to video techniques.
Another problem which appears very regularly is that many designers like to show off their prowess by doing everything in enormously long scripts. More often than not, there are simpler methods of achieving the same ends. In fact,it is probable that, if a home page has more than a few kilobytes of script, it is very badly designed.
My faithful readers will remember that I used to rant about not having meta keywords and descriptions. These are somewhat less important today because the spiders used by search engines for cataloguing their links are much more sophisticated, by parsing the real meaning of the text in the pages of a Web site. It is still useful to have them, although far less essential.
No matter what, a good Web site is always legible. This means that the text should appear with a good contrast against the background, have an easily read type face and be neither too large nor too small. Personally, I dislike white text on a black background, or anything similar, because I do not find it easy to read. The only excuse for a dark background is to show off a gallery of photographs, but certainly not for text. I'm not over keen on pure white backgrounds, either, because they can be somewhat dazzling; a muted light shade is better. If you have a pattern or a watermark in the background, please make sure that it does not interfere with the legibility of the text.
What should the home page contain? The most important thing is, of course, a menu of some description, which will link you to at least the important pages of your site. Text should be fairly short, half a screen at the most, giving a very brief description of what the site is all about. If it is for commercial purposes, then I strongly recommend that the full name and address, telephone number, fax number and general e-mail address be added to the home page. The telephone and fax numbers should include the international dialling codes preceded by a plus sign, such as +357 22 532 762.
Other pages can relax these rules somewhat but, if a page is excessively heavy, it is a good idea to warn users of the fact in hyperlinks, especially if the link goes to a file, such as PDF documents or streaming video. Talking of video, the recent fine imposed on Microsoft by the EU, for their media player may be a two-edged sword. Much as I detest Microsoft's policy of imposing their applications technology as part of their operating systems, it must be admitted that their Window's Media Player 9 is good (perhaps apart from the hideous“skins”, designed, I think, to appeal to 5-year old kiddies) and a distinct improvement on the real and quicktime competition for streaming. Unfortunately, the WMV format for the video is not yet universally cross-platform, as the others are.
May I make one last appeal? On your Contact us page, even if you have a form for doing so, please, please, please, also give the e-mail addresses of your key personnel or, at least, departments. Filling out forms, often requiring a lot of irrelevant information, is much less “user-friendly” and time-consuming than a quick e-mail; worse, it doesn't leave a copy in the sender's Sent e-mail box, so he has no record of what he said or when he said it. Also,make sure that every message received, whether by e-mail or via the form, is answered (not by an automatic acknowledgement, which only wastes bandwidth)within one working day or two, at the most. There is nothing more frustrating than to send a message requiring a response, and not to receive one – and its very bad for your corporate image (I've blacklisted suppliers and potential ones for just that). If you don't have the answer immediately at hand or you cannot give a valid reply don't hesitate to say so – at least, the receiver will know. Do not just ignore an e-mail. And don't, for Goodness' sake,do what a supplier has just done to me; he acknowledged my order via his secure Web site by ordinary e-mail, complete with my full credit card details, for anyone to use.
My theme for the review section, this time, is lead-free hot air solder levelling (HASL). A search produces a handful of sites of varying merit. Let's be quite clear from the start; this is a controversial subject as there are many who consider that lead-free will sound the death-knell of HASL, already tottering because its uneven surface profile, even with tin/lead eutectic alloys, is deemed incompatible with some modern components. The poorer flowing of some lead- free alloys may acerbate this, plus there are some who do not like the extra excursion to a higher temperature, well above the glass transition temperature (Tg) of even the most robust substrates. On the other hand, there are many who believe these disadvantages are minor and the simplicity and low cost of the finish will pay dividends. So, let's see what the Internet has to tell us, shall we?
This is a massive file, taking me 11½ minutes of downloading with a good connection. It contains 43 pages of a presentation, generally on lead-free solder issues. There is relatively little on HASL, although it's mentioned several times, in comparison to ENIG, OSP and other finishes, with pull testing as the criterion. The author concludes that SOIC components on a HASL finish have a significantly higher pull strength than on the other finishes, but this did not seem to be the same with QFP components, lending one to think whether other factors were in play. The authors made a few mistakes of fact, such as stating the entry-into-force date of WEEE (not RoHS) was 2008, even though the presentation was made in September 2003, well after the definitive date was announced.
Short and sweet, the IPC lead-free site is not over- informative, with just two succinct paragraphs. The first is a single sentence: Even though lead-free HASL is available for PWBs, some manufacturers may choose to move away from this process if required to produce lead-free product. The second one starts by almost contradicting the first, stating that HASL will most likely work well and will wet faster than alternative finishes. Concern is expressed about the higher temperature and its effects.
This is a commercially oriented, but very interesting, five-page paper,specifically on the subject under discussion, in relation to a single make of horizontal levelling machine. Although the results are not definitive, it would seem that they are very promising under the specific conditions described. However, one word of warning is in order; these tests were done over 10 years back and more moder data may be available.
This is another commercially oriented PDF file with a 9 minute download time with a good connection. It contains many microsection photos of assembled lead-free solder joints discussing different alloys, atmospheres and component types. Most of these were on OSP-finished boards, although there was the odd photo with a lead-free HASL finish. However, one of the conclusions suggested that HASL may even be preferable; Potential concern – Voids. May require more aggressive Flux and/or No lead HASL finish.
This was another presentation set with 15 slides, happily fast-downloading,only about 15 seconds! This discusses various finishes and gives pull- strength results. This also seems to tend towards the conclusion that lead-free HASL is good, if the uneven coating can be tolerated, but that horizontal HASL is distinctly better, on all counts, than vertical HASL. This one is worth the detour!
This is an IT trade organisation's report on no-lead assembly. There is actually little mention of board finishes, as if they were of small importance. The authors claim that it is virtually interchangeable with tin/lead HASL with no, or only slight, changes to the production process. I feel this is a rather optimistic view which detracts from the credibility of the report.
This is a full IMAPS technical paper, entitled “Lead-Free Solderability Preservative Coatings of PCBs”. From the opening paragraphs, the author(from Poland) is clearly of the opinion that HASL has had its day and must be replaced with other lead-free finishes. Unfortunately, his criterion for comparison is a tin/lead HASL finish. Nowhere does he entertain the notion of lead-free HASL. However, the paper does provide a valuable comparison between the other candidates and is therefore likely to be very useful for those who wish to leave HASL behind.
This is a magazine article which dismisses lead- free HASL in very succinct words. It states categorically that it is not in production. It may be better if the four authors, all from reputable companies, checked their facts before pressing a single key on the keyboard because this kind of glib statement is a way of losing credibility. I'll grant them that the article in question was published in 2000, but some of the other commentaries in this review, making comparisons, date from the same epoch.
This interesting 1999 paper compares eight lead-free finishes with respect to various humid and dry aging protocols. HASL, as may be expected, comes out consistently well with both wetting balance and wave soldering tests. Interestingly, OSP and silver came out better in the humid tests, compared with the dry tests. I don't wish to go into the significance of this in terms of practical aging, but some thought may be given to the wisdom of packing boards with silica gel sachets.
This is another commercially-oriented presentation, with a 2 minute downloading time. On the 14th slide, out of 32, it states, “Lead-free HASL– OK if there are no planarity issues, large installed base”, which sums up much of what I've already said. In the conclusions, one of them stated,“OSP finish was the most difficult to solder and all of the metalized[sic!] finishes performed better and approximately the same”, which contradicts an earlier statement that OSP is problematic!
The High Density Packaging Users Group is, as the name implies, a group set up to study the effects of HDIS on manufacturability. This is an interesting paper by ten members, some of whom are our industry gurus, from various companies. The paper itself studies the effects of such things as microvias in pads and lead-free soldering, over a range of surface finishes, of which SnCu HASL was one. On visual inspection, it came out very well. Even though the test medium was HDIS, implying small, tightly-packed components, there was no mention of planarity problems.
As may be expected, the UK NPL have published this document, entitled “The Reality of Lead-free Soldering” by Brian Richards. There was not much information on our subject, but I can quote, “Trials on lead-free HASL(hot air solder levelled) boards have been carried out, and it appears that SnCu will prove a suitable alternative to SnPb.”.
This useful information from the viewpoint of a PCB fabricator, although it would appear that, at the time of writing, he hadn't taken the step of going lead-free on HASL and preferred other finishes. However, it is suggested that this may be possible. Where this commercial paper is useful is the comparison of the other finishes, with some indications of relative costs.
It may be unlucky to have just 13 review sites, but they reveal that the summons that the EU RoHS Directive, with just 2 years to go before implementation, has not provided a firm answer yet as to whether HASL is or is not likely to be the preferred choice, at least for “punch and crunch”boards. Like Lady Macbeth, I fear this summons would not yet allow me to sleep,were I in a position to choose.
Brian EllisCyprusb_ellis@protonique.com
Note1 William Shakespeare (1601), Macbeth, Act 2 Scene 1.
