The AM Forum

I've been working on a couple of projects recently around the shack that involve PIC microprocessors.  Invariably, the features I want (say, USB and a CAN [controller area network] together in the same chip) are only available with surface mount packages.

Also, the number of pins is much greater with the surface mount devices which gives more functions.

I have virtually no experience soldering these devices.  I think the initial plan would be to solder the PIC do an adapter board that would allow connection to a breadboard for prototyping.

I have heard of methods using the home oven, heat guns, etc. for soldering such devices.  It would be completely impractical to try to manually solder a hundred pins, one at a time, particularly with .5mm center-to-center pin spacings!  Apparently, even though the device temperature reaches the solder melting temperature, the device is not damaged?

Please point me in the right direction  :)

Thanks and Regards,

Steve

Bacon WA3WDR posted his comments on SMT soldering one time here on the AM Forum and I pasted it to a Word document.  It is attached below:

I've not done much with SMT so far.

I'd want a wide field stereo microscope my self... that and a hot air "reflow workstation" I suspect.

Solder paste is all the rage I am told.

Doubtless there are a number of tutorials on youtoobe to check out.

A bit depends on the number of pins and the style of SMT, and how small a tip you have on your soldering iron as to what needs to be done... it is tricky because the alignment is somewhat critical because of the number of pins and the close spacing...

IF ur doing a one off, and NOT on a PCB, you can get away with "dead bug" style, mount the thing upside down, put short runs of wire wrap wire to each pin and then use them to go to where you need to go... but with a hundred pin uproc, I dunno if that's what I'd want to be doing.

Let us know what you arrive at Steve... I'll be interested since SMT has arrived and it looks like that's where all the parts are headed now. This discrete stuff is antique.

                     _-_-bear

Steve,
Maybe a special solderless socket for your prototyping application could be an approach which may work for you.

This company appears to have lots of tools, adapters, sockets, proto-boards, etc., etc.


http://www.emulation.com/


http://www.emulation.com/catalog/off-the-shelf_solutions/sockets/overview.cfm

Steve,
I'v been doing SM for 20 years and find it easier than through hole.
The first thing you need is a quality microscope and a couple good pairs of tweezers. Mark D. was selling a real nice one at Deerfield. I don't think he sold it. I almost bought it because it has a better stand than my microscope.
Then you need a good soldering station. I'd be glad to share some tricks when you get the basic tools. Ovens and solder paste are tricky to do at best.
Sockets usually sell for big bucks but that is a way around SM. At some point you can't avoid it.

Frank,

???  What sockets do you mean that are a way around SM ??   Chip sockets for SM?

T

Steve,
Another painless route, I think is buying a development board.  Lots and lots of companies make these for all the differnet micrcontroller makers and platforms.  We used one from a Canadian company using a Motorla uC; board came with surface mount IC already there installed, regulatior/LED, programming port connector, lots of room for wirewrapping sig. conditioning IC's , I/O was all ready for headers connectors or wirewrapping.   
It might be easier and save lotsa time than investing in the special solder equipment.  You can deveop your ideas, test them, than spin a board to a PCB house.   I also did some SM years ago.  Don't miss it.  Ugh.   :P

i.e.:

http://www.diolan.com/pic/demoboard.html

http://www.futurlec.com/T89C51DevBoard.shtml

http://www.futurlec.com/AT89C51RE2DevBoard_Technical.shtml

http://www.mikroe.com/eng/products/view/1/easypic5-development-system/

Hi!  Thanks for the responses so far.

There are some SMD components that can indeed be socketed  The best is the PLCC (plastic leaded chip carrier) which is a plastic, four-sided chip carrier, with a "J"-leads.  These leads can be surface mount soldered, or inserted into a socket.  The sockets are very INexpensive.  Unfortunately, few components are available in PLCC packages and they are very good packages because they can be either soldered or socketed.

(http://www.classeradio.com/plccpic.jpg)

Sockets for other types of surface mount devices are very expensive (in the order of up to $300/each and $150 or more is common), so are not practical for production work, but possiblities for prototyping.  However, there are only certain SMD configurations for which sockets are available, and they really are expensive.

Which leaves soldering the SMD to an adapter board (Futurlec has these for under a dollar) and then using the adapter board.

Very cool. Too bad they are not cheap/common for all SM chips.  Though, the socket itself must be soldered in, right?

My last HPSDR project had some 60+ pin? chips that wud be almost immpossible to replace if they went bad.  Frank knows what I'm talking about.

T

If you have a PC board ready to go and want to solder one of those types of chips, I simply get a chunck of very small diameter solder and cut it the same length as one side of the IC. Center the chip on the pins, make sure its dead on, then lay the length of solder across all the pins on one side. drag the iron across it slowly and the solder will automatically flow where it needs to, especially if the board has some flux on it already.

http://kd5ssj.com/solderpaste

Cash Olsen has a nice series of tutorials using various techniques.  He also repackages and sells solder paste in small syringes.  A great service since the parts houses want forty bucks for a full production size unit.

At work I routinely hand solder entire boardloads of components, 0805 size is most common but we've used anything from 1206 size down to 0201 size. One of our typical designs uses about 300 components, 0805 sized, plus a CPLD and various SMT chips and transistors.

All it requires is a steady hand and good view of your work. If you need magnification then you should use it.

It's much more trouble than it's worth to play with ovens and that sort of stuff because as soon as you try moving the board with unsoldered components upon it, you risk shifting everything.

   Geo,  

   I used to do the same kind of work back in the early 1990's. We went down to 1206 size resistors and caps. I was amazed how solder wick would take care of too much solder that flowed between the leads of an IC.

   I used to laugh at those old guys who needed extreme light, and magnification to work on SMD. I could do it all with the naked eye and average light. Now I am an OM, and I am just like those guys I used to laugh at.  :'(

   It gets scary to realize that even with good light and proper magnification, the image is still sometimes unclear. To some degree we compensate mentally for declining I/O to our brains. Then when I think I'm having a bad eye day, I observe someone else who is 15 years younger than I who cannot do the detail work that I still can muddle through with extreme effort.
  
Jim
WD5JKO

LAST year? there was a nice article in QST about SMT and making a hot air station AND using a video camera with a macro-focus to see the little stuff. For "under $200". He gave 'screen shots' of the video from the camera and I could see every thought of a PC trace on his work.
No use of toaster ovens, etc.

Fred

My understanding - limited as it is - is that either the slobber paste or another
ablative adhesive is used to hold the components in place so that they can not shift if the board is moved, such as in a production line oven environment or a prototype environment...

anyone?

Geez, we are old fogies out of touch, eh?

                          _-_-bear

I am an old fogie and out of touch.......damn it fees gooooood!

This utube video shows how to remove smd devices quickly and cleanly.
http://www.youtube.com/watch?v=7kyaz4Zrd78

  All kinds of ways to do it. I used to just shine up the board (bare copper) with ultra fine steel wool, clean, and put a thin film of liquid flux over the pads to be soldered. With IC chips, I would solder the leads on the corners (or just 1 set of corners diagonally) to hold chip. Then using regular 25w iron (or thereabouts) with medium tip (need some thermal mass), flow solder as you pass by the leads, say pins 1-7 then 8-14 on a 14 pin SOIC package. The solder will collect on the pads, and separate between them. Over do it, and just suck out solder with fresh solder wick (not tarnished old stuff). NO NEED TO SOLDER PIN BY PIN. With good eyes, and average equipment (nothing special), this is pretty easy. One thing I used the heck out of though was a set of wooden tweezers made from two pop-sickle sticks. I still have them and I use them today in RF work (like bridging a capacitor to see the effect on a hot circuit).

   I was working in a Lab, and we'd come up with something over lunch doodled on a paper napkin. By the end of the day there would be a full schematic, and the beginning of an Orcad PCB layout. By lunch the following day, PC boards were made, and by the end of the day they were assembled and being tested. We had an old HPIB HP XY plotter. Using ultra fine point Sharpies, we would write the PCB traces directly to the PCB. Using the thin 1 sided boards, we could also print the reverse image and then glue together to make 2 sided boards using tiny eyelets for the plated through holes. That was a lot of fun...

   The company I worked for was hired by BEL to develop the automotive laser detector since the cops were switching over from radar guns to laser guns (early 1990's). These guns pulsed the lasers 20ns on at a 400 hz rate. We had to develop a RF front end to detect the laser pulses. I must have built a hundred different boards that changed several times a week. They had to cost << $5.00 each, and trigger on a laser pulse over 1 mile away.

Jim
WD5JKO

NICE!!! Like liquid solder wick.
Thanks

After looking at the U-tube video Bob/VK7ZL posted and some of the other training video's it doesn't look hard at all. Looks like the hardest part is seeing in detail so you can spot problems. Having some liquid and paste flux on hand and very small diameter solder is the ticket. I'm going to look into that video camera macro focus idea.

Mike

Steve,

If you just need one or two of these soldered to a carrier board, you can send them over to me.

Otherwise, your best bet is to do some practicing on some cheap SOIC packages.

I use paste solder and a hot air gun for the real fine pitch stuff,
Like the 256 pin FPGAs etc..

You can use an iron with the paste solder too, just lay a bead right down the line of pins, on top of them..
Go a long the row with the iron to solder.
Use solder wick to clean out any extra solder between pins.

If you're quick about it, you'll have no troubles.

Paste solder for SMD has a much lower melting/flow point than any fine wire solder does....

But then you need good optics to see if there are and shorts between pins.
The effect is called theiving when a strand of solder bribges two close pins.

Honestly, even from a tinkerers standpoint, I'd rather play with SM components than thruhole. After you get yourself a couple cheap pieces of gear, it really is easier to deal with. There's no trimming or having to hold parts in place then flipping the board for slobbering, and as we already know, pretty much ANYthing is avail in SMD's that can't be had with thruhole. Those microscopes are really nice, and they carry a nice pricetag even used. I opted for a video microscope instead. It plugs into the USB on my bench puter and I can pop the image up on the 22" LCD and it works great. This is the one I use, it's a 20-200X cheapo job with a manual focus knob and some LED lights built into the end. It can be handheld for inspecting, but for working and needing both hands, I have it on an old mic boom using a rat shack spring clamp mic holder and a usb extension cable to make up the cord length. I think I paid around 25 bucks a year or so ago. Searched fleabay and here it is. http://cgi.ebay.com/20X-200X-1-3MP-USB-Digital-Video-Camera-Microscope-New-/280675910713?pt=LH_DefaultDomain_0&hash=item4159968439

A couple of other things to have around for convenience would be:

An alcohol dispenser (the type with the cup on top that you push down to get the denatured alcohol out. Cost is about $5) and a bigass box of Qtips. Cleanup after putting down a part or reworking an existing part is important.

Grab a couple flux pens from digikey. I like the Kester 2331ZX water soluble ones. Kester 186 is good too if your using leaded solder, but it's a harder to clean off. I kinda lean towards the 2331 because if I make a big mess with the flux, I can literally hold the board vertical and hose it down with a pump sprayer with distilled water and it comes out spotless. Dry it with a hair dryer or heatgun, and it comes out nice and clean. Any flux left over is gonna cause you problems. Also have some paste type on hand which is sometimes nice for sticking parts down to a board before tacking them in.

A good soldering station. Try not to use one of those iron that plugs right into the wall and regulates it's temp with a clixon or a big thermistor, or some other device. the temp control and temp swing is too wide and too hot. Use those outside for making antennas, and leave it outside. Use a decent station like a Hako or Weller, or whatever as long as it has a power supply with a temp control knob on it. Mine is a weller wd1001 with a WP80 hand iron. I chose it because it's digital for starters, has a timed temperature setback so if it sits idle for a bit, it lowers the temp because small tips burn up fast, and the replacement tips are super cheap, don't seize to the iron, and there's a million different ones to choose from. It's digital, and powerful (80watts) so you can go from soldering tiny little Litz wires all the way up to soldering the braid of .400 coax into a PL259 with ease or tinning some 12 or 10ga for a mobile install. Tips are about a buck each and last a while if you take care of them. (tin them with a touch of solder before shutting off the iron.)

Brass wool tip cleaner. Most good irons come with this, or atleast a sponge wetted with distilled water. The tip has to be real clean so that nothing but nice clean solder touches the SMD components. (Picture what an oddball strand of tiny wire or junk left over on the tip would do if it gets across the chip legs and you don't see it.)

Solder braid. We all should have rolls of this crap around already. If you really wanna trick out your setup, pick up a vacuum desoldering tool. Not a squeezy ball type, or for god sakes, not one of those plunger type trace breakers) but a real powered vacuum type gun with a pump. Put the hot tip on the pads or whatever you wanna clean up and push the button to start the pump. It'll suck up any solder in sight and does an awesome job.

Just realized the time and have to drag ass to work so heres my quick and dirty on a typical QFP chip lets say and using my flooding method.

If the pads aren't already tinned, then flux them with the pen and put a dab (droplet size) bit of solder right on the iron tip. Drag the iron over the pads. They are only going to hold just so much solder, so run the iron over them till they look full. Looking thru the camera, line up the chip getting all the legs to sit right on top of it's own pad. I use a little wooden stick thing for this but anything that won't melt and not made of metal works fine. When it's lined up, hold it down with a bit of pressure from the stick, and I'll touch the cleaned iron to to tack down one pin in one corner. Then final align the chip by nudging it around till all the pads are lined up. Hold it and tack a pin on the opposite corner. (diagonally)

Put a little more flux on all the other pins. Now turn the board upright at maybe 45deg or a little better and facing you. Hold it however you can make means to do so. I use this little plastic bench vise. With the iron around 475 and a 3-5mm wedge tip in it, put a droplet of solder on it and while pointing it downward, drag it up and down along a vertical row of pins and add some solder to the tip while it's in contact with the pins as the solder gets used up. After every pin in that row has solder on it, wipe off the tip and look closely with the camera. You should see a nice little clean kerf between the leg and the pad with no thieves anywhere. If there are any bridges, just hit the pins with a dab of flux and run the iron over them again. You don't want to remove too much of the solder you just put down, just enough to suck up the bridges. Run a qtip with alcohol over the pads and inspect them up close with the camera.

If they look good, then do the vertical row on the other side of the chip the same way. When that's done, rotate the board 90deg and do the other two sides of the chip.

Clean everything up with the alco and qtips. If the board doesn't have anything else on it that can get ruined by water, the take it to the sink and use a pump sprayer filled with distilled water and blast it and maybe scrub around the area with a tooth brush. (Usually not needed because the water soluble flux rinses clean. Dry it with a hair dryer and ur done.

Go and ask the guys at a puter store if you can have a couple old dead puter motherboards and get some practice done. It's not bad to get the hang of and after you do a few, an 80 legger will take maybe 3 minutes to put down.

Just my .02 on this. I learned this method watching utube vids of all things. just search SMD rework on the tube and theres lots of vids with different ways to do the same. The description above id the one that works for me.  

    

Hi Frank,

Thanks for the post. I also watched everything I could find on UTube. One of the technical collage videos showed a digital video microscope which I thought would be perfect for the job. I don't think I could work very long looking through a scope.

Fred mentioned earlier a QST article that suggested using a home video camera with macro focus. That got me thinking, I have an older analog video camcorder that dose fairly good close up (or so I remembered). I pulled it out and gave it a try. I also have a modern lipstick analog video camera so I tested them both. The lipstick camera was a disappointment. Fairly good as an area cam but ugly closeup. The camcorder did better but still lacking the kind of detail you would need.

How far away from the board can you mount that eBay microscope? I would guess the closer to the work, the more magnification. If it could be mounted, say, 6 inches or so away from the work and still see good detail than you might be less apt to bump the camera with the iron and other tools as well as see what your doing in real time.

Mike

Mike usually 3 inches works well. It has a manual focus knob and the magnification is in the software.

Typing this on a phone and for.some reason I.can't see the box so im not sure what the hell im typing haha ;D

I have that trouble when I'm driving and posting.  ::)  I'm installing a 27" monitor on the dashboard so I can see what I'm texting.

I'm going to order one of the video microscopes and give it a try. I can think of a few other uses for it and the price is right. If I have time before the next job I might post some pictures here.

In this Youtube video http://www.youtube.com/watch?v=7kyaz4Zrd78 (http://www.youtube.com/watch?v=7kyaz4Zrd78) (posted earlier) the guy obviously is selling a kit of paste flux and low temperature melt solder? I would like to find out what kind of solder he is using. Can anybody point me in the right direction?

The solder he uses is hard and thick but the benefit of keeping the regular solder molten longer so the part can be removed would be a big help on other projects as well. Thanks.

Mike

I work for a semiconductor company and we solder surface mount components every day. If I convert .5mm to English that translates to almost 20 mils? Gee, that's a lot more than the 5.5 mils we regularly "play" with. I think in mils because that's what we use in-house even though we are a well-established international semiconductor company. Our spec sheets use both measurement systems. While there will be myriad expert opinions offered, the ones by W1ATR are very similar to what I would suggest. If you are new to SMT, then I honestly think you may be a little late in the game to start with this particular project you have in mind. You might try "warming up" so to speak on a smaller, less complicated project. On thing to remember is that virtually all the components offered for sale today are ROHS certified. That is to say, meant for use with lead-free solder. There will be those who lament the politics involved, the world is going to hell in a handbasket, etc...etc. This is nevertheless a fact that has to be reckoned with and it can be easily addressed: pre-tin the part with lead solder. Otherwise the quality of the solder connection, particularly if the solder pads are small, will be at best tenuous. Cleaning the board and IC before starting will also address another troublesome variable. You need to start with clean surfaces to help ensure good solder connections.

Here's a juicy little tidbit for the hive. If anyone is interested in the chipquick method, the alloy wire they use, then same alloy that is approx $100 for 8ft, is whats called Cerrolow-136. It's been around for ages in the milling and casting industry, and I think it's only been in recent years that someone figured out it's eutectic and mixes well with electronics type solder. It lowers it's melting point so low that after heating, it stays liquidy (<is that a word?  ::)) and workable for a pretty long time.

A buddy of mine brought me home a small block of it years ago from the mill where he worked and I used to carve off short slivers and use it while playing around under boatanchors. You know how annoying point to point wiring can be when you have to take something out and theres 4-5 wires and leads all twisted up thru a single eye connection on a tube socket or wiring bridge. Id flux the connection, heat it, and feed in some of this stuff, then suck away all the solder from the connection. the true beauty was that the remaining slobber would stay molten for like 30-45 seconds so I would have loads of time to unravel the mess from the eye to get to the part I wanted.

I lost that little block a long time ago  :( then just got lazy after that and go in with the dikes and cut away everything haha.

Googled up an MSDS below. Maybe someone knows a source for wire.

http://www.chemcas.com/msds112/cas/2834/7440-69-9_7439-92-1_7440-31-5_7440-74-6.asp  

I hate it when a solder connection goes round and round and when she stops nobody knows. That was when it was thought solder could not be used to secure a connection. You had to wrap wire securely around a terninal and then solder. Then solder got better and they got smarter. A serious problem we sometimes encounter is trying to unsolder a SMT chip from a multi-layer PCB with some of the layers being 2 ounces. 1 ounce copper is more common but 2 ounce copper is used when you're carrying a lot of power. For the curious, 1 ounce copper is 0.00135 inch thick and 2 ounce copper is twice that. Really sucks up whatever heat you put on the part and it takes a very long time to remove. It's especially horrendous when the solder that was originally applied is lead-free because lead-free solder has a higher melting temperature. I end up using 2 hot air guns, one on top and one on the bottom while the board is held in a clamp. When I see solder melting, I grab the part with a pair of tweezers. It's not getting any easier.

Thanks for all of these good suggestions!  I'm still working on the PIC software at the moment using a development board (it has the parts already mounted).  However, soon I'm going to have to mount a PIC on my own board, and this is where the surface mount soldering is going to be necessary.

I'll report back when I do some actual soldering!

Thanks and Regards,

Steve

I picked up an HD video, USB, microscope and have been testing it out. I pulled out a board from a 1987 Compaq (286) laptop computer and took a few stills to show the scopes ability.

The first shot shows a row of small objects. The top brown component (a cap I think) measures .058" square X .120" long. The two black components (marked 470 and 330) I believe are resistors and measure .030" square X .060" long. The object in the foreground (out of focus) is .060" by .135".

The second shot is an electrolytic cap. At .175" square X .250" long they are the largest components on the board not counting the microprocessors. I didn't catch this problem with my magnifying visor. I checked the cap with an HP 4261A and an ohm meter. It's shorted and possibly the reason this board failed.

Both shots are 1280 X 1024 resolution and about 1MB if saved as JPEG's. The second shot was saved as a bit map image and converted to JPEG for viewing here on the BBS. The scope was just a fraction of an inch from the board for these shots but the scope will work at 4-6 inches from the work by adding more light. You can view your work in real time @ 1280 X 1024 full screen and take a snapshot any time by pressing the button on the LED light control located on the USB cord.

I like to sharpen knives and think I do a descent job. I use a visor often to make sure the edge is consistent. I took a look with the scope and can see I need more practice.

 Hey Mike,
  The micro shots are pretty cool ! We do quite a bit of micro surface mount stuff here at work using the binoc visor  (30 years ago would have been easily done with the naked eyes !) The USB scope really make for a sweet look at the final product. If not used for the actual intitial work it would still be a great inspection tool.

  A note on the "Cerro" alloys: There are many different Cerro products in use. We used them a a filler for moldings and tubular parts in metal fabrication for many years. The Cerro products were pretty much made history  here due to the fact many of the alloys contain substantial percentages of Cadmium. Just adding this a a caution. Know what you are working with and the safest handling practices !

One thing to consider with the micro optics is "depth of field"...

                      _-_-bear