EEVblog #434 - SMD Thermal Pad & Drag Soldering Tutorial

[EEVblog]

Dave shows another method for hand soldering a surface mount SMD chip with a thermal pad, using both liquid and gel flux.
A method using hot air is here:
Also, drag soldering TSSOP packages is demonstrated.
http://pcbzone.net/



Dave.

[Bloch]

That is the reason for the different positions ?

[AndyC_772]

It takes a real masochist (or sadist, perhaps, depending on who's going to be doing the assembly!) to design a part like that MLF into something that's going to be assembled by hand. Kudos... I think!

If anyone doubts that some devices are moisture sensitive, I saw an excellent example of just how bad the problem can be a few years ago. The chip in question was a microprocessor in a 1mm pitch BGA package, about 20mm square.

When the failed boards were X-rayed, it was clear that the device hadn't just suffered some tiny internal fracture - the entire BGA substrate had warped to the extent that the innermost solder balls were shorted together, and the outer ones weren't even making contact with the board.

It's normally a good rule of thumb that the larger a device is, the more moisture sensitive it's likely to be. It looked as though your MLF device had a stated shelf life of a year out of the bag... I'm a bit surprised they labelled it as moisture sensitive at all tbh.

[mikeselectricstuff]

This method works better with solder paste on the pad, then you're not  relying solder going through the vias, only heat. It also fills any gap between PCB and device pad which can be caused by bent pins lifting it slightly.

[Ketturi]

Nice tutorial. Soldering chips with chisel tip looks much faster than soldering one pin at time, but is fine pointy tip better for 2 pin parts? And I don't get how you can even solder something under camera, my hands are shaking much more even when sitting in good position.

[mikeselectricstuff]

I wonder if an angled half-silvered mirror might be a solution to the working under the camera problem. Or maybe just a small camera!

[Smatek]

The way with solder paste is the better one. In times with no rework station I solder such devices the same way like Dave.

With two mirrors you can see under the Board and solder the chip more easily.

<0>                                                      < your eyes
          _____                                         < PCB
          \                        /                        <  two mirrors
          ^                      ^
          ---------------------------------  < work bench

One tip for the flux. For years I am making my own flux. Simply dissolve in alcohol wood rosin. The consistency can be determined by you - from honey-like to super liquid. Rosin is very cheap to buy in musician business stores.
To apply, I use a dropper or a syringe with a blunt needle.
The advantage of rosin is that when it dries it develops a very good cohesion.

[Bored@Work]

That is the reason for the different positions ?

Polarity.

[IanJ]

but is fine pointy tip better for 2 pin parts?

Not necessarily because you want to transfer heat as well as solder.........and that's less easy with a small point. Ever noticed with a pointy tip you often end up laying over your iron at an angle to get the heat transfer.

Ian.

[Watermelon]

Looks like a poor soldermask to me. Soldermask i usually work with never damages when i'm drag soldering a chip.

On the second chip, Dave is stressing pin 1 mechanically to allign the chip correctly, i think its good practise to make sure the chip is on the board without mechanical stress. Could be done by just reheating the soldering joint.

[scarrier]

Concerning the layout, I have? heard that a ground plane like the one on the bottom is not very good at all because you have big crack in it and there may be a considerable voltage if you measure from each side of the crack. Can you demystify a bit the ground plane. heard thease crack in pcb acted like antenna also. What is truth about ground plane

[free_electron]

our soldering lady (we have a person in my lab that hand assembles small series, she's mil-spec soldering approved)  swears by the following approach :

flip the chip over , flux in the thermal pad. apply dot of solder on a wide tip and then apply solder on the heat pad
apply flux on the board.
put chip on the board
re-heat using hot air , or using the iron form the backside.

the advantages are :
solder does not have to 'wick' upward through the via's. you never know how many actually make contact. by applynig a dot of solder on the pad first you are sure the entire pad is covered in solder. this also goes for paste. there is no risk of 'voiding' or poor wetting. you have already wetted the entire surface.

over solder paste the advantage is that there is no risk of 'not enough solder' : meaning voiding and bad thermal contact. it also avoids having small droplets or balls of tin underneath the board

Now, in Dave's video  notice that he is using a solder mask defined pad. there is a square of copper and the solder mask defines how much copper is exposed. this is different from a copper defined soldering pad. here there is copper and an opening between the copper and the soldermask where the substrate is visible.

for packages such as BGA and MLF it is not advisable ot use solder mask defined pads. the problem is cracking of the solder mask. there is a risk that , under power , a remnant of lfux reacts with the solder and dendrites start growing.... eventually creating conductive pathways.

Always clean off flux !!!! i had it a couple of weeks ogo. a board that works fine.. and after a few days behaves weird i had 100k ohmson certain pins, ther pins were 30 ohms. flux remnants ! when i applied power to the board ( nothing was heating up , this is like  a 10mA conumption board ) i coudl actually see the fluxstart to bubble between component pins. after about half an hour there was resistance ! let it sit under power overnight and the flux had 'caramelised' itself to be about 100 ohm ...

[Pentium100]

OK, this may be a stupid question, but I have to ask it:

Since the chip has a separate ground (or whatever the thermal pad is connected to) pin, why not just use thermal grease between the chip and the thermal pad and solder just the pins? Nobody does it, so I guess there are some problems with this method, but I can't figure out what they are, does anyone know?

[BravoV]

why not just use thermal grease between the chip and the thermal pad and solder just the pins?

Solidified metal (solder) join beats the best thermal grease on earth out of the water when it comes to thermal conductivity.

[konfu]

Do I get it right that there is no TO-220 LT3080 anymore but a smaller MS8E (MSOP) based one? Nice "upgrade" ;-) I guess Dave also uses the transistor-based tracking pre-regulator design now instead of the i2c pot. Seems like the µSupply is having some progress. Can't wait for the next revision of the schematic to come!

[arko]

Hi Dave and All!

Happy to see that the µSupply is progressing!
This project really motivates me beyond any other, can't wait to see it complete (And any more progress videos along the way). 

Btw, seeing the chip fuming the flux out of itself, it reminded me I always wondered how well this kind of package resists the heat over time during soldering.

I am sometime really afraid to re-heat the chip one last time to perfect some joints. We usually have pretty nice reflow temperature profiles in the datasheets, but I usually find no info about short, high temperature limits. Any advice on hand soldering these ones? How much do they suffer from too long heat exposure? Is it a legitimate concern?

Arko

[robbag]

That Micro Supply has certainly evolved since we saw it last. I am very excited to see the final product.  Expected completion date?

[madhunm]

That Micro Supply has certainly evolved since we saw it last. I am very excited to see the final product.  Expected completion date?

Dave,

In the schematic you posted on the forum, you had used an ATMEGA324. However, in the video, you think I spotted an ATMEGA644. Any reason for the change?

Thanks,
Madhu.

[konfu]

More IO pins I guess (Mega328=32pins; Mega644=44 pins). So that's 12 more to be used for something good

[mswhin63]

Dave, good to see your video blogs are now being used to help Curtin university EE students how to solder. Just received my first week Lab learning requirement with 3 links to your website with a "learning requirement" before lab. 

[EEVblog]

Dave, good to see your video blogs are now being used to help Curtin university EE students how to solder. Just received my first week Lab learning requirement with 3 links to your website with a "learning requirement" before lab. 

Awesome! 

Dave.

[EEVblog]

More IO pins I guess (Mega328=32pins; Mega644=44 pins). So that's 12 more to be used for something good

Yes, I needed every pin I could get.
Have uploaded the latest version over the old one.

Dave.

[madhunm]

Have uploaded the latest version over the old one.

Dave,

Is it the uSupply USB Rev B Update (http://eevblog.com/files/uSupplyUSBrevB.pdf) ?

Thanks,
Madhu.

[konfu]

Hi Dave,

I can't find any updated schematic. Not sure whether I'm the one to blame or you ;-)

[deephaven]

The file linked to by madhunmn loaded ok for me.

I must say this seems to have just the right feature set that I would want