I'm just getting ready to do a small (20-board) run of 0.8mm QFP chips for a design, and so far with earlier runs (stencil+reflow) I've had a BANNED success rate with bridges (though 100% success rate after manual rework). Obviously not good, but it's doable.
But looking at future designs (microcontroller+bluetooth) the best part I can find is 0.5mm pitch QFN. Ouch--I thought just SMD was a barrier for hobbyist/semipro folks, but these fine pitches look painful. I'm not sure I could even place a part without some sort of assist. Are people doing these by hand with any degree of success, and if so what tools/techniques work best for you?
Any success with 0.5mm QFN by hand?
Moderators: adafruit_support_bill, adafruit
Please be positive and constructive with your questions and comments.
- alphatronique
- Posts: 231
- Joined: Fri Jun 25, 2010 8:30 am
Re: Any success with 0.5mm QFN by hand?
Hi
if you have right tool it not a problem
1-) have good tacky flux my favorite was http://www.zeph.com/zephlux.htm but Chip Quik SMD291-ND may a subtitle
2-) have good controller temp solder iron whit something like 0.5 to 1mm tip ..
3-) put flux not be shy ,put chip on it ,solder 2 opposed corner pin ,then remain of pin ...
4-) if chip have bottom pad ,put small dot of solder past ,then after solder heat the whole thing whit hot air rework station
for melt solder under ship ,pin whill hold chip for it not move..
the magic part was done by flux that avoid solder bridge ,whit bit of practice it become very easy to solder
http://www.youtube.com/watch?v=erb6-i54tbo
if you check on that video solder tip was special kind but most of the job was done by the good flux
and if your curious solder iron and guy use was metcal 500 or 5000 whit SMTC-1167 tip
that what it use here and do this way for production QTH .. to day my smallest part was 0.3mm pitch connector
if you have right tool it not a problem
1-) have good tacky flux my favorite was http://www.zeph.com/zephlux.htm but Chip Quik SMD291-ND may a subtitle
2-) have good controller temp solder iron whit something like 0.5 to 1mm tip ..
3-) put flux not be shy ,put chip on it ,solder 2 opposed corner pin ,then remain of pin ...
4-) if chip have bottom pad ,put small dot of solder past ,then after solder heat the whole thing whit hot air rework station
for melt solder under ship ,pin whill hold chip for it not move..
the magic part was done by flux that avoid solder bridge ,whit bit of practice it become very easy to solder
http://www.youtube.com/watch?v=erb6-i54tbo
if you check on that video solder tip was special kind but most of the job was done by the good flux
and if your curious solder iron and guy use was metcal 500 or 5000 whit SMTC-1167 tip
that what it use here and do this way for production QTH .. to day my smallest part was 0.3mm pitch connector
- vputz
- Posts: 79
- Joined: Mon May 03, 2010 5:43 am
Re: Any success with 0.5mm QFN by hand?
Hmm. I'll give it a try then and see how it goes. I'm actually using a Metcal for hand-soldering, but at least the first round on this will likely be paste and reflow, maybe with hot-air rework. Worth a try--nothing can possibly go wrong!
- alphatronique
- Posts: 231
- Joined: Fri Jun 25, 2010 8:30 am
Re: Any success with 0.5mm QFN by hand?
Hi
i find more easy to do 0.5mm pitch by hand whit flux + metcal that do it whit solder past
you need to inspect it by microscope anyway
if do whit solder past even whit good setup i still have some solder bridge , something like 1 on 200-300 pin
but i use lead free past so it more hard to work that leaded counterpart
hope it usefully for you ...
i find more easy to do 0.5mm pitch by hand whit flux + metcal that do it whit solder past
you need to inspect it by microscope anyway
if do whit solder past even whit good setup i still have some solder bridge , something like 1 on 200-300 pin
but i use lead free past so it more hard to work that leaded counterpart
hope it usefully for you ...
- juku
- Posts: 6
- Joined: Sat Sep 01, 2012 4:38 am
Re: Any success with 0.5mm QFN by hand?
As others have said, good flux is one of the keys. The other is a stereo microscope.
- vputz
- Posts: 79
- Joined: Mon May 03, 2010 5:43 am
Re: Any success with 0.5mm QFN by hand?
I have to admit, this is starting to sound like a poor candidate for even small-scale home production. Which would be a real pity, because chips like the Nordic bluetooth+M0 SOC are a lot of power in a small and cheap package. Well, I'll give it a go or three and see if I can make any headway.
- cstratton
- Posts: 294
- Joined: Wed Sep 29, 2010 3:52 pm
Re: Any success with 0.5mm QFN by hand?
Assuming access to a $100-class hot air station, I think small QFN chips may actually go faster than larger TQFP types, due to the ability to exploit self-centering behavior during the "float" stage.
I've also found that for prototypes, a stencil is not necessary - you can just tin the pads evenly with wire solder and an iron, then re-flux, place the chip approximately, and flow with the hot air.
A good hot plate works as well for decent FR4 boards - but you really need a good temperature control or you will just burn boards. Unfortunately it does not work on phenolic DIY board stock, which tends to boil and delaminate if heated through to soldering temperature.
A microscope is of course the nice tool to use, but for inspection only (ie, once it has cooled) a $7 10x loupe works - you just can't do any work under it as the working distance is only a centimeter or two.
I've also found that for prototypes, a stencil is not necessary - you can just tin the pads evenly with wire solder and an iron, then re-flux, place the chip approximately, and flow with the hot air.
A good hot plate works as well for decent FR4 boards - but you really need a good temperature control or you will just burn boards. Unfortunately it does not work on phenolic DIY board stock, which tends to boil and delaminate if heated through to soldering temperature.
A microscope is of course the nice tool to use, but for inspection only (ie, once it has cooled) a $7 10x loupe works - you just can't do any work under it as the working distance is only a centimeter or two.
-
- Posts: 164
- Joined: Fri Jun 11, 2010 9:21 pm
Re: Any success with 0.5mm QFN by hand?
Not at all hard AS LONG AS your footprint is designed such that the pads extend at least 0.5mm outside the package outline.
Either paste+reflow, or ( if it doesn't have a centre slug that needs soldering) smother it in flux, tack a pin then drag-solder with a large to medium-sized tip which has a small edge radius, so it gets right into the corner of chip and PCB.
The extended pads allow easy rework if you do get bridges - again plenty of flux and drag around the edge, braid if necessary to remove excess solder
Either paste+reflow, or ( if it doesn't have a centre slug that needs soldering) smother it in flux, tack a pin then drag-solder with a large to medium-sized tip which has a small edge radius, so it gets right into the corner of chip and PCB.
The extended pads allow easy rework if you do get bridges - again plenty of flux and drag around the edge, braid if necessary to remove excess solder
- mrxtreme
- Posts: 10
- Joined: Fri Aug 03, 2012 12:46 am
Re: Any success with 0.5mm QFN by hand?
Okay, I just placed 3, VEED-4 (3mm, 16 pin, 0.5mm pitch) parts on my board.
Others have suggested hand soldering, but there is no way I could get that to work. The pads are underneath and there isn't good way to get solder to reliably flow. All sorts of problems with opens. Plus there is no way to get to the center pad.
What I did was "real" SMT construction (solder paste and hot air reflow). But even that took some experimentation.
I tried hand applying solder paste with a needle, syringe, all sorts of tiny pointy things, using a microscope, but getting the right amount of paste was hit or miss. I ended up getting a stencil from OSHStencil.com for just a few bucks and oh man what a difference. It took a couple of tries to get the squeegee technique right but it put down just the right amount of paste.
Since the board is tiny I put it on a Mr Coffee mug warmer to get it a bit over 100 C and then I hit it with some 300 C hot air (using Pb-free paste). Within a minute, perfection! No bridges, no opens, no scorching, no problems. All three parts working beautifully. I reflowed the backside which has all the bypass caps, as well.
From now on, I plan to get a stencil whenever I order a new board. No more hand soldering for me. OSHStencils are tremendous. Their pricing is excellent. For my little board (front and back stencils) it was the best $11 spent. This was my first time using them and they had them done in less than 6 hours! Stupid USPS took several days to get them to me, however.
For QFN parts I found this reference very helpful when I designed the footprint.
http://www.hep.caltech.edu/~hitlin/CMS_ ... pps_04.pdf
I no longer have any trepidation about placing 0.5mm leadless parts by hand.
Later...
James
Others have suggested hand soldering, but there is no way I could get that to work. The pads are underneath and there isn't good way to get solder to reliably flow. All sorts of problems with opens. Plus there is no way to get to the center pad.
What I did was "real" SMT construction (solder paste and hot air reflow). But even that took some experimentation.
I tried hand applying solder paste with a needle, syringe, all sorts of tiny pointy things, using a microscope, but getting the right amount of paste was hit or miss. I ended up getting a stencil from OSHStencil.com for just a few bucks and oh man what a difference. It took a couple of tries to get the squeegee technique right but it put down just the right amount of paste.
Since the board is tiny I put it on a Mr Coffee mug warmer to get it a bit over 100 C and then I hit it with some 300 C hot air (using Pb-free paste). Within a minute, perfection! No bridges, no opens, no scorching, no problems. All three parts working beautifully. I reflowed the backside which has all the bypass caps, as well.
From now on, I plan to get a stencil whenever I order a new board. No more hand soldering for me. OSHStencils are tremendous. Their pricing is excellent. For my little board (front and back stencils) it was the best $11 spent. This was my first time using them and they had them done in less than 6 hours! Stupid USPS took several days to get them to me, however.
For QFN parts I found this reference very helpful when I designed the footprint.
http://www.hep.caltech.edu/~hitlin/CMS_ ... pps_04.pdf
I no longer have any trepidation about placing 0.5mm leadless parts by hand.
Later...
James
- alphatronique
- Posts: 231
- Joined: Fri Jun 25, 2010 8:30 am
Re: Any success with 0.5mm QFN by hand?
Hi
normally whit the right flux it not a issue to solder it by hand check for the 4300lf flux
http://www.solderandmore.com/servlet/th ... Categories
and for bottom pad i put tiny dot of past ,then solder by hand then add more flux and heat it whit hot hair
once ready it will self center on pad so no need of stencil ..
have some kind of package that pad not expose on the side often it was on Gyro and G-sensor
that one was bit more challenge ,so flip chip put flux then lite bit of solder on all pad
and flip back to PCB whit another good layer of flux (never to much flux)
and heat whit hot air ..
if what to progress need to master that since it mater of time that you have one to rework
same for drag soldering of fine pitch http://www.youtube.com/watch?v=5uiroWBkdFY look a 2:10
and for the last for all footprint refer to IPC-7351B that only reliable source of trouble free footprint
normally whit the right flux it not a issue to solder it by hand check for the 4300lf flux
http://www.solderandmore.com/servlet/th ... Categories
and for bottom pad i put tiny dot of past ,then solder by hand then add more flux and heat it whit hot hair
once ready it will self center on pad so no need of stencil ..
have some kind of package that pad not expose on the side often it was on Gyro and G-sensor
that one was bit more challenge ,so flip chip put flux then lite bit of solder on all pad
and flip back to PCB whit another good layer of flux (never to much flux)
and heat whit hot air ..
if what to progress need to master that since it mater of time that you have one to rework
same for drag soldering of fine pitch http://www.youtube.com/watch?v=5uiroWBkdFY look a 2:10
and for the last for all footprint refer to IPC-7351B that only reliable source of trouble free footprint
-
- Posts: 306
- Joined: Wed Jun 12, 2013 5:01 pm
Re: Any success with 0.5mm QFN by hand?
Bottom preheating really does help especially with a chip that has a 'center' pad.Since the board is tiny I put it on a Mr Coffee mug warmer to get it a bit over 100 C and then I hit it with some 300 C hot air (using Pb-free paste). Within a minute, perfection!
Good going and I like the idea of using the mug warmer.
- mrxtreme
- Posts: 10
- Joined: Fri Aug 03, 2012 12:46 am
Re: Any success with 0.5mm QFN by hand?
Marc,
Thanks for the tips. My parts are the type with no exposed pads which as you said are more challenging.
I will practice your technique for the time I need to do rework on an already populated board. I still contend that for a new proto, buying a stencil is still much easier and for my unskilled hands, faster than tinning all the pads, fluxing, and then reflowing. For me it was well worth $11 (which was including the first class shipping and tracking I paid extra for). Just swipe and plop the parts onto the board. I'm lazy
Next time I have to place a leaded part I'll try the drag technique with solder paste as shown in the video. That looks faster and easier than individually soldering each pin, which is what I've been doing. I also like the idea of using Kapton tape to hold the part. Never thought to do that. It always seemed like I needed a third hand to hold the fidgety part as I tacked the corners.
Later,
James
Thanks for the tips. My parts are the type with no exposed pads which as you said are more challenging.
I will practice your technique for the time I need to do rework on an already populated board. I still contend that for a new proto, buying a stencil is still much easier and for my unskilled hands, faster than tinning all the pads, fluxing, and then reflowing. For me it was well worth $11 (which was including the first class shipping and tracking I paid extra for). Just swipe and plop the parts onto the board. I'm lazy
Next time I have to place a leaded part I'll try the drag technique with solder paste as shown in the video. That looks faster and easier than individually soldering each pin, which is what I've been doing. I also like the idea of using Kapton tape to hold the part. Never thought to do that. It always seemed like I needed a third hand to hold the fidgety part as I tacked the corners.
Later,
James
- alphatronique
- Posts: 231
- Joined: Fri Jun 25, 2010 8:30 am
Re: Any success with 0.5mm QFN by hand?
Hi
it better to master rework before put 30K $ on a pick & place
since you will have more to rework for sure as board QTH increase
and stencil was nice but sometime on populated board it hard to handle
but again have good flux do miracle ,have assemble board whit that kind of accelerometer last week
and was able to touch-up some solder whit flux and litel bit of pad that extent was god for solder to flow under chip
and make good connection ,but agree that it not easy ..
and as trick make sure to inspect at every print that you not have hole closed by solder on QFN pad
you may end whit a pad whit near no solder so rework to done if you miss it . even whit automated optic inspection that may happen..
it better to master rework before put 30K $ on a pick & place
since you will have more to rework for sure as board QTH increase
and stencil was nice but sometime on populated board it hard to handle
but again have good flux do miracle ,have assemble board whit that kind of accelerometer last week
and was able to touch-up some solder whit flux and litel bit of pad that extent was god for solder to flow under chip
and make good connection ,but agree that it not easy ..
and as trick make sure to inspect at every print that you not have hole closed by solder on QFN pad
you may end whit a pad whit near no solder so rework to done if you miss it . even whit automated optic inspection that may happen..
Please be positive and constructive with your questions and comments.