How do I solder a QSOP (16pin) package properly?

[Qacer]

A lady who used to work in Magnavox's soldering line told me that the best way to solder is to use a soldering iron that is wide and put some soldering flux on the pins. That works like a charm. I've soldered SOIC packages without problems.

However, I can't seem to properly solder OSOP packages. The solder sticks to the neighboring pin every time. Tsk.

Any tips?

 

[PottedMeat]

QSOP like this?

http://pdfserv.maxim-ic.com/package_dwgs/21-0055F.PDF

You can do it a couple ways ( i use a cheapo hakko pencil iron ) -

Tin the copper traces on the board and use braid to remove the excess. Then place the part with tweezers or your finger and apply the iron to the top of the pin - you should feel the pin sink into the solder. You can slide the iron across the pins fairly quickly once the component is fixed. There should be enough solder to make a decent connection. Use a magnifying glass to check the pin - trace connections.

Or you can brush flux on the traces, place the component, and tack it in place with some solder. Solder the remaining pins ( dont worry about excess ). Hold some braid up to the pins that have excess and suck some of that solder up.

Oh and if this is a homemade board watch the time you heat up those traces - it wont take long before the copper peels off.

 

[bobsmith1492]

Well, you could use a toaster oven to reflow it (saw that somewhere...)

The smallest things I soldered were SOT23; I can't remember if I did QSOP or not. What I usually do is poke some solder onto a corner pin, line the chip up, then melt the solder and slide the chip into place. Stick a dab of solder on the iron, hold the chip down, and tap the solder onto the opposite corner. Then it's rooted in place, and you can go from there. If the iron is too thick, wrap a thin piece of wire around the tip a few times and use the tip of the wire to solder with.

 

[MrDudeMan]

i did this every day for 8 months, so i have quite a bit of experience...you have to put flux down on the pads before the part goes down, and then put flux on top of the pins when it is in place. tack down opposite corners so it wont move, and then drag the iron across all of the pins with an extremely minimal amount of solder on the iron. i can solder a 150+ pin IC in about 15 seconds doing that without any pins sticking together. it takes practice, but it works like a charm. you definitely shouldnt need braid or a solder sucker because you shouldnt be using enough solder to warrant either of those. less is more.

the best way to do this is when angled tweezers. a fat iron tip is good, but not too fat. the small tips are a terrible idea because you dont get sufficient heat transfer and you end up heating the part up too long because you have to hold it on there forever. high heat - low time soldering is the way to go, and with a wide tip you can do that easily.


two pictures of parts i used to solder. the first one is just for "wow" because its so small, but the second one i soldered on a daily/weekly basis using the technique described above.

diode
dsp

 

[dkozloski]

I've seen several articles outlining a method for soldering surface mount devices using a toaster oven. It sounds rediculous but it's hard to argue with success.
example

 

[MrDudeMan]

Originally posted by: dkozloski
I've seen several articles outlining a method for soldering surface mount devices using a toaster oven. It sounds rediculous but it's hard to argue with success.
example

you could spend the time to learn how to do it and skip the entire process mentioned above. 25-30 seconds versus 5+ minutes, plus you never have to mess with solder braid because once you figure out the trick, you never get 2 pins stuck together. also, defaulting to this process is cumbersome at the least. what if you have heat sensitive parts on the board? doing it this way, you MUST solder down the controller first. its a neat trick, but a total waste of time IMO.

 

[Mark R]

Originally posted by: MrDudeMan
i did this every day for 8 months, so i have quite a bit of experience...you have to put flux down on the pads before the part goes down, and then put flux on top of the pins when it is in place. tack down opposite corners so it wont move, and then drag the iron across all of the pins with an extremely minimal amount of solder on the iron.

Do you reckon that this could be adapted to work with QFN chips?

 

[MrDudeMan]

Originally posted by: Mark R

Originally posted by: MrDudeMan
i did this every day for 8 months, so i have quite a bit of experience...you have to put flux down on the pads before the part goes down, and then put flux on top of the pins when it is in place. tack down opposite corners so it wont move, and then drag the iron across all of the pins with an extremely minimal amount of solder on the iron.

Do you reckon that this could be adapted to work with QFN chips?

i know someone was going to ask

i have successfully soldered a handful of ethernet controller chips in the QFN package with this method, but it wasnt easy. if you are designing the board, it can be done much easier if you extend the pad length .5-1mm beyond the standard size. it flows to the pins like it is QSOP or SOIC package with pad length extensions.

 

[Colt45]

tack a few pins,
add flux, run iron across the pins while adding minimal solder. If some of the pins stick together, run the iron across again, if still no good, wick off the excess, add flux, and run the iron across again

works for me on QFP and QFN and whatever else.

 

[MrDudeMan]

Originally posted by: Colt45
tack a few pins,
add flux, run iron across the pins while adding minimal solder. If some of the pins stick together, run the iron across again, if still no good, wick off the excess, add flux, and run the iron across again

works for me on QFP and QFN and whatever else.

the problem with messing up is heating up the part multiple times. maybe what you have been soldering didnt require as much precision, but if i had to heat up a part twice, it had to be replaced. there was no room for error because the circuit boards were part of actuator controllers for satellites. i challenged this procedure once and was proven very wrong by looking at the gain of an amplifier with parts soldered only once compared to an amplifier with 1 or 2 parts (out of 800+) soldered two or three times and it was very different. i guess it all depends on the project, but trying to solder QFN packages or anything similar without a pad extension or a different method isnt the best idea IMO.

 

[Qacer]

Originally posted by: MrDudeMan
... then drag the iron across all of the pins with an extremely minimal amount of solder on the iron....

Thanks, MrDudeMan!

I just need some clarification on the above statement. My soldering iron is about 3mm wide and 1 mm thick. Do I need to put the solder all across the 3mm width? When soldering SOIC, I've just been putting minimal solder on the side corner of the iron.

Also, how fast do you drag your iron? The current adapter that I have has very minimal (<0.5mm) pad clearance after I put the IC down.

Thanks again!

 

[bobsmith1492]

If any of the pads connect to the ground plane, you might have to focus the iron there longer to heat the plane up - just a heads-up.

 

[MrDudeMan]

Originally posted by: Qacer

Originally posted by: MrDudeMan
... then drag the iron across all of the pins with an extremely minimal amount of solder on the iron....

Thanks, MrDudeMan!

I just need some clarification on the above statement. My soldering iron is about 3mm wide and 1 mm thick. Do I need to put the solder all across the 3mm width? When soldering SOIC, I've just been putting minimal solder on the side corner of the iron.

Also, how fast do you drag your iron? The current adapter that I have has very minimal (<0.5mm) pad clearance after I put the IC down.

Thanks again!

put solder on the tip of the iron and when you drag it, it will jump from pin to pin. you will have to try it a few times but you will get the hang of it. dont go too slow but dont go too fast...its hard to explain, but if you do it a few times you will figure it out. dont push too hard either. you can bend the pins or snap them off. it may sound harder than it really is. it just takes practice.

Originally posted by: bobsmith1492
If any of the pads connect to the ground plane, you might have to focus the iron there longer to heat the plane up - just a heads-up.

it depends how the ground plane is laid out. i do my ground plane with air gaps between the pad and ground to avoid this problem. if the ground plane is just straight connected to the pads, i usually skip them on the drag because it messes it all up.

 

[Colt45]

Originally posted by: MrDudeMan

Originally posted by: Colt45
tack a few pins,
add flux, run iron across the pins while adding minimal solder. If some of the pins stick together, run the iron across again, if still no good, wick off the excess, add flux, and run the iron across again

works for me on QFP and QFN and whatever else.

the problem with messing up is heating up the part multiple times. maybe what you have been soldering didnt require as much precision, but if i had to heat up a part twice, it had to be replaced. there was no room for error because the circuit boards were part of actuator controllers for satellites. i challenged this procedure once and was proven very wrong by looking at the gain of an amplifier with parts soldered only once compared to an amplifier with 1 or 2 parts (out of 800+) soldered two or three times and it was very different. i guess it all depends on the project, but trying to solder QFN packages or anything similar without a pad extension or a different method isnt the best idea IMO.

I don't work on satellites

 

[MrDudeMan]

Originally posted by: Colt45

Originally posted by: MrDudeMan

Originally posted by: Colt45
tack a few pins,
add flux, run iron across the pins while adding minimal solder. If some of the pins stick together, run the iron across again, if still no good, wick off the excess, add flux, and run the iron across again

works for me on QFP and QFN and whatever else.

the problem with messing up is heating up the part multiple times. maybe what you have been soldering didnt require as much precision, but if i had to heat up a part twice, it had to be replaced. there was no room for error because the circuit boards were part of actuator controllers for satellites. i challenged this procedure once and was proven very wrong by looking at the gain of an amplifier with parts soldered only once compared to an amplifier with 1 or 2 parts (out of 800+) soldered two or three times and it was very different. i guess it all depends on the project, but trying to solder QFN packages or anything similar without a pad extension or a different method isnt the best idea IMO.

I don't work on satellites

in all fairness, neither do i. i was just the peon who had to solder everything.

 

[Qacer]

I was finally able to solder that darn chip. I tried MrDudeMan's technique, but I guess I need more practice. There are still pins that look like they are not connected, but I probed it and the connection seems solid.

Instead of soldering the IC right away I made sure that I had some freshly melted solder on the individual pads. Then, I just put a very, very small amount of solder on the iron and tapped the IC's pin while rested on the pads. That seemed to have worked.

 

[MrDudeMan]

Originally posted by: Qacer
I was finally able to solder that darn chip. I tried MrDudeMan's technique, but I guess I need more practice. There are still pins that look like they are not connected, but I probed it and the connection seems solid.

Instead of soldering the IC right away I made sure that I had some freshly melted solder on the individual pads. Then, I just put a very, very small amount of solder on the iron and tapped the IC's pin while rested on the pads. That seemed to have worked.

yeah that is fine. i know i keep parading the heat issue around but as long as you were pretty quick, im sure its not a problem. i thought id mention one other thing (since you pointed it out)...it is somewhat disconcerting to look at the pins after you to the drag trick because a lot of times they do look disconnected. i eventually just trusted myself and havent had a problem since the first few, but occasionally i see one that looks really disconnected. i usually probe it, just like you did, and it always checks out. as you said, it just takes practice.

edit: oh, one other thing. solder actually isnt extremely conductive, which is why you always want the pins as close to the pads as possible. the way you did it is fine, especially since you are learning, but as you get better, you need to put the pins down on the pads first with nothing in between to get the best connection possible.