Taobao chips full of water, DIY baking doesn't fix the problem

[gperoni]

I spent two days on this already, it's time to ask for help

I have a batch of taobao chips (Digikey doesn't sell them). I could solder/boot some of them but I could never get all of the GPIOs to work. For the longest time I thought the problem was me soldering 0.5mm pitch BGAs the wrong way, or a bad land pattern, but I've realised yesterday the issue is most likely to be related to the small bubbles coming out of the chips.

I've bought a 33 Euro toaster oven and used it to bake the parts. I left the parts in at 125C for a day. It seems the situation is improved, but I still see bubbles coming out of the package.

I wonder if the chips might have been damaged by the Chinese seller handling them the wrong way, if I'm not baking them for enough time, if I should lower the temperature and leave it go for a week, if I need better baking equipment, etc.

The chip doesn't have a MSL rating that I'm aware of, my datasheet doesn't say and the manufacturer won't talk to me (hence the taobao sourcing).

So here is what happens to the chips:

Unbaked, top. This one was unexpected. https://gfycat.com/DimpledHotInganue

Unbaked, this side is what is usually expelling stuff, I once saw things coming out of it even without using flux to highlight it, unfortunately I didn't get that on video thanks to a windows bug. Also unfortunately for this video I moved the chip while applying the flux... https://gfycat.com/DisastrousCrazyKakapo

Baked, side. It's the same side, same position even. Less stuff is coming out, but still stuff is, and the chip is still not working. https://gfycat.com/NimbleIcyFoxhound

Oven picture:

[jmelson]

I have a batch of taobao chips (Digikey doesn't sell them). I could solder/boot some of them but I could never get all of the GPIOs to work. For the longest time I thought the problem was me soldering 0.5mm pitch BGAs the wrong way, or a bad land pattern, but I've realised yesterday the issue is most likely to be related to the small bubbles coming out of the chips.

WOW, NEVER, EVER saw anything like that!  Those chips are FILLED with water!  What you need to do is weigh them with a sensitive balance (as in 0.01 g resolution) and then bake for a day and reweigh.  Keep baking and weighing daily until the weight no longer decreases.  If you can figure a way to put the chips under vacuum while baking, it might draw the water out faster.

When I bake stuff, my recipe is this:  Heat to 50 C for one hour, then 70 C for one hour, then whatever temp is decided on for several hours.  Mostly, I bake PC boards that have been sitting, unpopulated, for a long time, to avoid blistering of the laminate.  That amount of bubbling indicates HUGE internal stresses in the package, and is destroying the die or wire bonds.  You need to ramp temperature up slowly to drive VAPOR out of the chip without too much stress.

I finished an old run of boards by getting some obsolete chips from China.  A lot of these chips failed.  When I manually soldered in replacements from the same stock, they always worked.  My theory was that the chips that failed during reflow were full of water, although not in the dramatic sense like your video, and that was damaging them.  Manual soldering just didn't heat them enough to cause problems.

Obviously, you are hitting these chips with too much heat too rapidly.  A slowly increasing temperature might solve the problem, but the AMOUNT of absorbed water seems, to me, to be incredible!  So, they might suffer internal corrosion or degradation of the silicon over time.

Jon

[chickenHeadKnob]

You didn't allow for enough time for moisture to leave the epoxy. I would low-temp bake at say 55 degrees C for a week or more.

[mrpackethead]

125 seems way to high to me.   we dry our components some times, and its at 70C for 72 hours.

[Ian.M]

You need to avoid boiling the moisture in any voids in the IC encapsulation. Don't exceed 95 deg C while baking until you are pretty sure nearly all the moisture has been driven out, which may take days, then follow a JEDEC standard high temperature bakeout profile suitable for the thickness of your IC package.  The RH inside the oven needs to be kept as low as possible - if the oven doesn't have reasonably good air circulation and venting you may need a tray of quicklime in there as well.  Quicklime is one of the few desiccants that's effective at high temperatures, that is reasonably readily available as you can make it yourself by calcining natural chalk or limestone.   Its nasty stuff and its highly caustic dust is hazardous by skin contact or inhalation.

[mrpackethead]

Maybe not so smart to buy from TaoBao. You have no idea

[gperoni]

As I said sourcing from Taobao was not my first choice.

So it looks like I need to find a way to seal the oven, get calcium oxide, bake for days and pray. Alternatives are a different taobao supplier for the same chip with a different datecode or using an actual baking oven from an assembly house.

Two questions: 1) Does it make sense to bake below boiling temperature in a high humidity environment in the meanwhile and 2) I probably exceeded 100C for a brief while with the batch of chips I'm baking, do I throw them out?

[mrpackethead]

Whats the IC?

[Ian.M]

*ALL* low temperature JEDEC bakeout profiles require controlled humidity, under a fairly low threshold, (IIRC, 5%).
The RH in a vented hot oven will be much lower than ambient, and if you know the ambient RH and the ambient and oven temperatures, the oven RH can be calculated so long as the chips don't contribute significantly to it (large oven, few chips).  If the ambient RH is fairly low (e.g. a dry room with a full sized dehumidifier running in it), you probably wont need quicklime in a vented oven, even for sub-100 deg C bakeout profiles, provided some air can flow in slowly at the bottom to let hot humid air escape from the vent.
Whether or not exceeding 100 dec C is damaging depends on the size of the voids in the IC package and how much you exceeded 100 deg C by.    You won't know till you try to use the chips whether or not you damaged them, or even if they were bad to start with which is quite probable if they were used parts desoldered over a charcoal brazier, washed in the river and dried on the pavement!   Depending on the relative value of the chip and assembled PCB, the difficulty of testing, and the difficulty of rework, it may or may not be worth trying to use them if the boards are for yourself, but if you are planning to sell them, you definitely need to find a better supplier.

[gperoni]

I just wanted to give you a huge thank you Ian, you absolutely helped me out here more than I could have ever hoped for.

The chips are $3 parts, they look original to me. I don't even think it would make sense to fake the substrate/packaging for that number (the bottom of the package is very unique and imho difficult to replicate). Unfortunately my aliwangwang is refusing my password or I would asked the supplier if the parts spent months underwater...