View previous topic :: View next topic |
Author |
Message |
jecottrell
Joined: 16 Jan 2005 Posts: 559 Location: Tucson, AZ
|
Hand Soldering QFNs |
Posted: Mon Oct 08, 2007 3:22 pm |
|
|
As I try to shrink some prototypes, I'm considering attempting to go from SSOP packages to QFNs. (I start my first 0603 board this afternoon, after working 1206s for a long time.)
Does anyone here prototype with QFN packages? What considerations or techniques do you find helpful?
My approach was going to be either, give myself longer pads to allow the iron to comfortably touch the pad along with the very fine solder and allow the solder to flow under the part. Or just toaster reflow the extremely fine parts onto the board first, and then everything else by hand.
Thanks,
John |
|
![](templates/subSilver/images/spacer.gif) |
libor
Joined: 14 Dec 2004 Posts: 288 Location: Hungary
|
|
Posted: Mon Oct 08, 2007 3:59 pm |
|
|
I have not soldered many of these QFN components (it was about a dozen of 20-pin QFNs), so the time it took was not an issue for me, My approach was using longer pads, I have put some solder on the pads before placing the components on them (used a suction iron to remove if too much) put the component on the pads (the height of the pre-soldering made some gap under some of the pins, this was my intention to let the solder flow under the component later. I held the component with a tweezer in postition while keeping the pad under one pin melted, when it was in right position , I removed the iron, this one pad was holding the component in place.
Then moving a very thin iron towards the component at each pin pushing a tiny wave of solder with the iron made the solder flow under the component. The iron tip must be about the same thickness as the pad.
(all this might have worked without the tedious pre-tinning the pads but I was going for sure,
there are QFN components having pads continuing a little beyond the edge on their side, these would be good if soldering them only from the outside, with no hassle about the solder flowing under them I think, though I had QFNs with the pads only on the bottom surface.)
A stereoscopic magnifier like this one here helped me a lot. (i have a plain vanilla looking-into version, btw: you are wrong if you think the lcd monitor version would be better, with two separate optics for your two eyes you have a real 3D perception needed for such things, you don't have this on the monitor, I have tried both.)
If making more boards regularly I would go for the reflow toaster solution. |
|
![](templates/subSilver/images/spacer.gif) |
newguy
Joined: 24 Jun 2004 Posts: 1903
|
|
Posted: Mon Oct 08, 2007 5:36 pm |
|
|
I've never soldered QFN packages, but I've done a LOT of fine pitch soldering of 0603 components and very fine pitch (0.5mm) FFC connectors and other fine pitch components. Here are a couple of things that I hope help you out.
1. Don't panic when you bridge pads. I've actually come to embrace solder bridges as the quickest, easiest means to the end. For instance, this is my standard procedure when soldering FFC connectors: http://media.digikey.com/pdf/Data%20Sheets/Hirose%20PDFs/FH34S_drawing.pdf. Once the connector is aligned and tacked down using the extra two alignment/mechanical pads on the connector, I make sure that all pads are liberally coated in flux. Then I solder all the connector's pins, again, making sure that each is soldered. They will get bridged, but that's okay. Once they're all soldered, the works is coated in flux again. Then I use solder wick (braid) to clean up the bridges. Works like a charm, and it's very quick as well. Also much less stressful than trying to solder pins individually under a microscope.
2. As has already been suggested, larger than normal pads on the PCB will help, as will the finest gauge solder and smallest soldering iron tip you can find. A slightly higher than normal iron temperature helps a bit too.
3. I'm sure you already know this, but use a quality flux-core solder.
As for soldering the QFN package, if you're lucky the QFN you are going to use will have very slight vertical pads at the edge of the chip. In that case, hand soldering will be easy to perform and visually verify. If not, soldering it down still shouldn't be difficult.
Align the chip, then apply the iron to one pad and gently press the solder into the pad. You'll be able to see the solder flow toward the chip. Don't use a lot of solder. Continue to hold the iron on the pad for a good 3-5 seconds once the (small) amount of solder on the pad hit (or seemed to hit) the chip. Remove the iron and gently use your finger to try and move the chip. It will quickly become apparent whether it is tacked down or not. If it is, tack another pin on the opposite side. If not, repeat the procedure until you tack the pin down.
With two pins tacked down, you can use the brute force solder bridge & wick to solder the rest of the pins, or you can go slow and try to solder each individually. Either way should work.
As an aside, the best way I've found to solder 0603 components is to put a very small amount of solder on one pad, then use the (unmelted) solder itself to gently push the component onto the hot pad. I hold the iron on the pad until the solder wicks up onto the component's contact. You may have to use the solder to hold the component down, as sometimes the solder's surface tension will tend to make the component stand up vertically once it does start to wick. Soldering the other side is generally very easy: just hold the iron on the end of the component and gently push a very small amount of solder onto the contact. |
|
![](templates/subSilver/images/spacer.gif) |
jecottrell
Joined: 16 Jan 2005 Posts: 559 Location: Tucson, AZ
|
|
Posted: Mon Oct 08, 2007 7:10 pm |
|
|
Libor: Yes, I have been looking for a better magnification system. I currently just use a lamp/magnifier. I need something more to inspect the quality of my work.
Newguy: I just got the board and tested the 0603 work. No problem, not much different than the 1206s. The glop and wick technique is what I normally end up doing. I'm going to try to make the pads a little thinner and see if it makes things easier. I find the wide pads are more likely to get shorted if the part isn't perfectly aligned. I think longer is definitely a help. It provides workspace away from the part and allows capillary action to do its thing.
Thanks for the ideas.
John |
|
![](templates/subSilver/images/spacer.gif) |
newguy
Joined: 24 Jun 2004 Posts: 1903
|
|
Posted: Mon Oct 08, 2007 7:31 pm |
|
|
You can also use the poor man's microscope - a handheld magnifying glass under the lamp/magnifier.
I actually do this all the time. My handheld magnifying glass has a tiny "bubble" for higher magnification. It's amazing how much that bubble and the lamp/magnifier enlarge your work. |
|
![](templates/subSilver/images/spacer.gif) |
Guest
|
|
Posted: Mon Oct 08, 2007 9:51 pm |
|
|
I agree with all the tips above as they all work - and I agree that sometimes the fastest way is just to solder the heck out of the thing and purposely bridge the pads then remove the solder with solder wick. This makes an amazingly good solder joint.
A good iron is required with a fine tip - the Metcal irons can't be beat on this. They deliver lots of heat when required - super for heavy, multi layer boards with a lot of ground planes.
Low cost Stereo microscopes can be bought on Ebay - this is a must for making sure the joints are nice and clean with no micro-bridges. Nothing like seeing! You only need about 20 power for the smallest component - usually I run at 7 to 10 power and only really crank it up when I need to see something really small.
Also don't throw out the idea buying or building one of the toaster oven reflow ovens that are available (search Google) - if you do this sort of work only once a month it will pay for itself.
Then lastly there are very reasonably priced proto assembly houses like Advanced Assembly in Colorado - they can solder anything and the cost is very reasonable.
HTH - My 2 cents worth. Steve H |
|
![](templates/subSilver/images/spacer.gif) |
libor
Joined: 14 Dec 2004 Posts: 288 Location: Hungary
|
|
Posted: Tue Oct 09, 2007 4:16 am |
|
|
newguy wrote: | ...then use the (unmelted) solder itself to gently push the component onto the hot pad. |
why not use very fine pitch tweezers to place and hold the component at this step ? ..you'd have a more firm grip on it ...all else I do the same. |
|
![](templates/subSilver/images/spacer.gif) |
jecottrell
Joined: 16 Jan 2005 Posts: 559 Location: Tucson, AZ
|
|
Posted: Tue Oct 09, 2007 4:33 am |
|
|
I do both. Tweezers are good for keeping parts from tombstoning, but solder works good for tweaking ICs and saves time switching back and forth between tweezers and solder. |
|
![](templates/subSilver/images/spacer.gif) |
newguy
Joined: 24 Jun 2004 Posts: 1903
|
|
Posted: Tue Oct 09, 2007 7:02 am |
|
|
libor wrote: | newguy wrote: | ...then use the (unmelted) solder itself to gently push the component onto the hot pad. |
why not use very fine pitch tweezers to place and hold the component at this step ? ..you'd have a more firm grip on it ...all else I do the same. |
As has been mentioned, it saves time. |
|
![](templates/subSilver/images/spacer.gif) |
rnielsen
Joined: 23 Sep 2003 Posts: 852 Location: Utah
|
|
Posted: Tue Oct 09, 2007 8:42 am |
|
|
I use a water soluable flux in my projects. This makes clean-up a snap instead of needing flux-off or other chemicals.
Flux! This is key. If needed, flow more flux if it has been burned off and you need to re-touch some pins. Extremely fine soldering tip. Lots of magnification and light. I use different magnifying glasses including ones that clip to my glasses and jewelers loupes (2X, 5X & 10X).
Fine pitch & surface mount components are fun to use but can be lost real fast in the carpet. Make sure you have it on right the first time. It's a real mess trying to remove it once it's on. You need something that will head ALL of the pins, at the same time, to remove the part or you'll end up destroying it to get it off.
Ronald |
|
![](templates/subSilver/images/spacer.gif) |
newguy
Joined: 24 Jun 2004 Posts: 1903
|
|
Posted: Tue Oct 09, 2007 8:49 am |
|
|
rnielsen wrote: | Make sure you have it on right the first time. It's a real mess trying to remove it once it's on. You need something that will head ALL of the pins, at the same time, to remove the part or you'll end up destroying it to get it off. |
Chipquik (pictured here: http://rocky.digikey.com/WebLib/Chip%20Quik/Web%20Photo/SMD1.jpg) works really well for removing parts, even relatively large ones like 28 pin SOIC packages. With only an ordinary soldering iron, too. |
|
![](templates/subSilver/images/spacer.gif) |
drolleman Guest
|
|
Posted: Tue Oct 09, 2007 11:46 am |
|
|
i do use qfn in prototyping, i now prefer qfn over all other packages. what i do is use an food saver for vacuum pump. the vacuum holds the qfn (or other serface mount devices) then i put the circut board on a foam to hold them together. then move the circut board in place under the chip. i use rosin flux under the chip and use an old curcut board and put a puddle of solder on it. this is done so only a small amount of solder will remain on the iron while soldering the qfn pads. if i put solder directly on the iron there will be too much on the iron and then solder will cross pads and then i need to clean it up. i can usually place / solder a 44 pin qfn in less than a minute. so the old through hole is history.
as for magnification i picked up a headband magnifier at a cheap store for 5 bucks. it has 3 magnifications. at level 2 qfn pads are huge and easy to see. |
|
![](templates/subSilver/images/spacer.gif) |
jecottrell
Joined: 16 Jan 2005 Posts: 559 Location: Tucson, AZ
|
|
Posted: Tue Oct 09, 2007 12:09 pm |
|
|
drolleman wrote: | the vacuum holds the qfn (or other serface mount devices) then i put the circut board on a foam to hold them together. then move the circut board in place under the chip. i use rosin flux under the chip and use an old curcut board and put a puddle of solder on it. this is done so only a small amount of solder will remain on the iron while soldering the qfn pads. if i put solder directly on the iron there will be too much on the iron and then solder will cross pads and then i need to clean it up. |
I think I understand. You essentially squish the QFN between the vacuum, the board, and the foam. To 'load' your iron with solder, you just tap it to another board that has a blob of solder on it. Then you touch each of the QFN pads?
I think I'll try one of those headband doofers. They look functional, and cheap. |
|
![](templates/subSilver/images/spacer.gif) |
Storic
Joined: 03 Dec 2005 Posts: 182 Location: Australia SA
|
|
Posted: Tue Oct 09, 2007 2:27 pm |
|
|
Tricky, I have done it and my hand soldering method is to use solder paste, as well with the foot print on the PCB, I had extended slightly the PCB pads so I can get to then,
Using a large magnify glass to view, place the QFN part and line up, having a fine tip soldering iron solder down a few pins and check if still line up, then I use a heat gun to solder the rest including the pad underneath.
It did take me a few go's to get a method that worked for me, to let you know, I did not use the heat gun at first, on one of my attempts, not all the pins got soldered by the iron, the heat gun fixed this.
solder paste and heat if you get it right, will line up the micro square onto the pads.
another method that you could consider (in large quantities) take the PCB and parts to a company that has the machine's to do this, I found the cost to be reasonable per board however high when you need to do lots. _________________ What has been learnt if you make the same mistake? ![Wink](images/smiles/icon_wink.gif) |
|
![](templates/subSilver/images/spacer.gif) |
|