Those who haven't tried it: do not be intimidated by surface mount soldering. It is a skill that requires a little practice and a modest amount of equipment is desirable but it is not that difficult a skill to acquire if you can already solder through hole components. And a lot of better components are simply not available in through hole these days. Kit designers are being held back by potential customers who are intimidated by surface mount.
Soldering the LQFP144_20 CPU was a challenge! With this kind of part it is basically impossible to solder to individual pins. But it can be done using just a good soldering iron, solder, and solder wick. Basically:
You really shouldn't dismiss soldering individual QFP pins as impossible. It isn't impossible. It isn't even that hard. If you can see it, you can probably solder it unless you have a really bad physical or visual impairment. And if you aren't actually legally blind, missing at least one limb, or completely quadrapelgic, and possibly even if you are, you shouldn't dismiss the possibility out of hand without trying some tests first. Even if you are under-equipped and under-funded.
I have soldered QFP packages (62.5 pitch instead of 50pitch so slightly bigger pins) pin by pin, and it was not hard. In fact, the first surface mount board I ever soldered (I also designed it) had a 112 pin 62.5 pitch QFP as well as all surface mount parts (except connectors and one voltage regulator) including TSSOPs, SOICS, transistors, electrolytic caps, tantalum caps, crystals, 0805 resistors/ceramic caps. I have also tried the brute force method you describe and found it was easier (and less likely to damage the board) just to do it right the first time. You might accidentally bridge a few pins and need to wick them but it is better than intentionally bridging them all. But I do have a 10X/15X stereo boom microscope. I also have a magnifying light and visor magnifier (sometimes the latter two were even used together) and I have a 230X USB microscope for inspection. For that type of package you will probably need to use at least one of these means of magnification. Depending on your vision, you may be able to do it with the magnifier and not need the microscope. Sometimes I think it is easier to use the magnifier lamp, possibly even on QFP/TSSOP packages and it is certainly possible. The microscope lets you see better but is slower to use in some other ways. I would prefer to have the stereo microscope available on QFP/TSSOP packages but it isn't absolutely necessary.
Again, my rule is "if you can see it, you can probably solder it". So make sure you have adequate light and at least some form of hands free magnification. If you don't have magnification good enough to solder by, you probably can not see well enough to properly inspect solder joints done by the brute force method, either.
Before going out and buying stuff, or throwing in the towel because you think you need to buy stuff, do a little test with what you already have available. Clean your magnifier and your glasses (if any). Take junk circuit boards (or even one which isn't junk - "take it apart") with surface mount components on it, put it under your existing magnification. Add a bright light source if not built into your magnifier. Take two very small jeweler's screwdrivers. Pretend one is a soldering iron and one is solder and practice touching each lead in turn under the magnifier. No animals or circuit boards are hurt in this experiment if you use proper ESD handling procedures. Just practice your eye hand coordination working under magnification. At first, you will feel like a bit of klutz because you are 1) making smaller movements than you are used to using and 2) the magnifier changes the loop gain between your hands and your eyes and this confuses your brain. Give yourself an hour to practice and you will likely find you CAN do it.
Now that you have an idea what YOU can do, try using your finest soldering iron and smallest soldering iron tip (cold). You might need a finer tip and/or finer solder.
For those unwilling to spend money on a kit until they know they can handle the soldering, or those who are anxious to get started before you even finish reading this message, you might consider taking some junk surface mount boards and unsoldering some parts with a heat gun, hot plate, or even a blow torch (used carefully, outside, likely to cause damage) and resoldering them. . While you want the board and part undamaged enough to use for practice, it is no big deal if you accidentally destroy them. Maybe even use a proper hot air rework solder/desolder gun as some of these can be had very cheap without hoods and you will eventually want one You can get old computer boards or cable or DSL modems for a few bucks at the local thrift store or hamfest (or just get on the local repeater and hollar) if you don't already have your own junk boards. Some of this old junk may be worth more for parts or soldering practice than whole. Or even some cheap new surface mount boards off of ebay. Buy something like a new USB RS-232, USB-TTL serial, USB Jtag, USB sound card, DC to DC converter, etc. and take it apart and put it back together. After a little practice, you might even be able to use it after you are done playing Dr Frankenstein - if you can keep track of what part goes where.
If you do find that you need a stereo microscope, some are available really cheap these days if you don't need a boom stand (good for working on larger boards) or trinocular port (for cameras). Entry level ones are about twice the price of a swing arm magnifying lamp. Just bear in mind that you will want to stick to boards that are less than the throat depth in either dimension unless you want to practice soldering leads from front, back, left, and right and twice the throat depth even if you do.
No-clean flux paste has been recommended. I agree. Having a syringe of flux helps. Tweezers are important. If you are desoldering components to practice on, you might also want a Xylotronic 490 IC popper ($5) or equivalent, though if you are just practicing or scavanging parts, instead of repairing, a sharp rap on the top edge of an empty open cardboard box will indiscrimently remove parts and blobs of molten solder into the box.
When soldering IC packages, tack solder the corner leads first to hold the package in place.
A 230X USB microscope used on low magnification (around 27X) has a long enough focus distance to use for soldering but note that it is not stereo so you lose depth of field, there is a short time lag, you need to remove the clear bezel (mine was glued but was able to force it off and still put in on/take it off), you want a good focusing stand, and it is plastic so it is vulnerable to melting from the heat of the iron so it should be protected by a small piece of glass.. Better for inspection than soldering.
Tip: If you do not have the artwork for your printed circuit board, scan it on a flatbed scanner or take a good high resolution digital image (macro mode up close but not closer than the camera can focus, camera steady). This way you will not only have documentation of traces that end up hidden under parts, you will also know which pins are supposed to be shorted together because in many cases these will form solder bridges that will drive you nuts if you don't know they are supposed to be there.
My first surface mount board was a microcontroller board which I designed. I wrote diagnostics to test almost every pin for opens and short circuits. You can also use the in-circuit debugger to the same end, manually. Some chips, including CPLDs, FPGAS, and some microcontrollers, have JTAG which can be used to test connections (though you might need to write your own software). These methods will not detect a poor connection but will detect shorts and opens. On the other hand, a processor with external flash and ram memory could require a whole lot of solder joints to be correct before the "brain" functions well enough to be useful in testing the others.
Here are some SMD kits I have seen recently that have some potential as SMD starter kits for people with various levels of ambition:
The Xprotolab kit ($35): Small, cheap, useful, reasonably challenging. Low speed oscilloscope/logic analyzer/arbitrary waveform generator/frequency counter that fits on a breadboard.
http://www.gabotronics.com/electronic-kits-modules/kit-to-build-xprotolab.htmhttp://www.gabotronics.com/development-boards/xmega-xprotolab.htmRelatively few pins have to be soldered correctly for the "brain" of the project to function and then it could be used to help you find faults in the other parts. If you have essential power/decoupling and at least one of the LCD, the USB port, or the in-circuit debugging port working, you should be able to exercise other pins using those.
Those who want to start simpler (no IC packages) could try something like this surface mount power supply kit for $11+ship.
http://www.ebay.com/itm/SMD-Adjustable-Power-Supply-Kit-2127A-/390564075772?pt=LH_DefaultDomain_0&hash=item5aef6ed8fcAnd for a real no-brainer, here is two versions of an SMD dummy load kit:
http://www.wa0itp.com/dummyload.htmlOr the QRPme RF Probe kit:
http://www.qrpme.com/?p=product&id=RFPFor $150, there is the Peaberry SDR transceiver kit (which is just about all surface mount except for the connectors, 1 TO-220, and some inductors) with built in USB soundcard.
http://ae9rb.com/index.php?main_page=product_info&cPath=1&products_id=1Finningly 80M single band SDR receiver kit, £16.50, surface mount kit similar to early softrock kits, external sound card needed:
http://www.kanga-products.co.uk/index.php?page=shop.product_details&flypage=flypage.tpl&product_id=17&category_id=6&option=com_virtuemart&Itemid=58AT sprint II is an SMD transceiver but it appears you need to be on the yahoo groups to see when batches become available.
BGA packages are another story, since the solder joints are hidden from view, but homebrewers have even had some success with those.
These days, there is a lot of info out on the net on surface mount soldering.
