SMT Soldering small devices

[TomWS]

BTW, is Chip Quik a good brand for T5 paste?  http://www.digikey.com/en/product-highlight/c/chip-quik/type5-solder-paste
Or maybe I should just ask: what is a good brand for T5 paste?

Chip Quik.

[WhiteHare]

@TomWS
I notice that the melting point on T5 paste is just 361°F.  That means I won't even need to defeat the temperature limit on a toaster oven in order to use it. 

Did you recommend it because the AB18x5 is especially temperature sensitive, or do you just prefer it generally?  Any downsides you've noticed with it compared to other solder pastes?

[TomWS]

@TomWS
I notice that the melting point on T5 paste is just 361°F.  That means I won't even need to defeat the temperature limit on a toaster oven in order to use it. 

Did you recommend it because the AB18x5 is especially temperature sensitive, or do you just prefer it generally?  Any downsides you've noticed with it compared to other solder pastes?

T5 is generally recommended when pad sizes are as small as the AB18X5.  The solder balls are smaller and more uniform than the commonly used T3 paste.  The reason I qualified the recommendation is that, if you have some large pads on the same board, you either need to shrink down the cream layer for these large pads or judiciously apply different pastes to the different regions of the board.  T5 is very 'runny'...

In your case, with a breakout that only has the AB18X5 and crystal (I'm assuming), then T5 with a stainless steel stencil should be a reliable setup.

FWIW,  my first attempt at using the AB18X5, I had to unsolder, manually paste, reflow, and then a couple of solder wicks later I had a working device.  I was using T3 with Polyimide stencil for that first board.  So, it can be done, but it's a slog...

You mean you haven't disabled the Toaster Oven thermostat?  Hmmmm, probably haven't done any large boards, I'm guessing...

Tom

[joelucid]

BTW a trick I've learned from Tom which works particularly well with T5 paste is this:

Take a very, very small brush and create a mix of flux and paste, much like a paint. Apply this paint across the exposed pads of your circuit, position components on top and reflow. This is a good method for reflowing components during rework where you can't apply a stencil. But I've found it works so well that for small circuits like my TH motes I don't even use stencils anymore.

I've also found that with this method no oven controller is necessary. Just put boards in, switch oven on and switch off / open door when the solder melts - done.

Joe

[WhiteHare]

Thanks for the tip!  Is it the Chipquick flux that you're mixing with the paste to make your paint, or something else?  Roughly what ratio?

Also, what brush are you using?  IIRC, in the past Tom mentioned using a toothpick as a paste applicator, so perhaps that would make for a kind of brush for this type of flux-paste paint.

[TomWS]

Thanks for the tip!  Is it the Chipquick flux that you're mixing with the paste to make your paint, or something else?  Roughly what ratio?

I use SRA No Clean Flux Pen: https://www.amazon.com/gp/product/B008OC0E5M
Mix 'til it runs like paint...

Also, what brush are you using?  IIRC, in the past Tom mentioned using a toothpick as a paste applicator, so perhaps that would make for a kind of brush for this type of flux-paste paint.

Go to any craft store and get a #0 round tip brush. 

Tom

[jra]

+1 the flux pen.  Amazon says I bought mine on March 31, 2014 and it still isn't empty.  I use a similar technique that involves a dedicated Waring SB30 burner https://www.amazon.com/gp/product/B000I14C7I and a dedicated 8" aluminum frying pan https://www.amazon.com/gp/product/B000HKJHVG.  No stencil, no timer, no thermometer (well, sometimes I use an IR thermometer https://www.amazon.com/gp/product/B007VAQ4QQ if things seem to be taking too long).  Apply flux/paste mix to pads, position parts, put board in pan, turn heat on high, remove pan when solder melts, remove board from pan with tweezers.

[WhiteHare]

I've ordered these brushes:  https://www.amazon.com/gp/product/B01HWCAGQQ/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

and I have the T5 paste sitting in my refrigerator now.

Not sure how the pen per se fits into this, other than as a flux dispenser, so I'm guessing you are probably creating a small puddle of flux with your pen on a pallet of some kind, then adding in some paste, then mixing the flux and paste together on the pallet, and then applying a thin amount of the flux-paste "paint" with the brush to your PCB pads.

[TomWS]

Gosh, with the 000 brush you'll be able to do 0.35mm pitch devices 

EDIT: With a very steady hand, of course...

[WhiteHare]

While waiting for my AB1815 breakout board to arrive, I thought I'd try the reflow oven system with the BQ25504, which has similar form factor.  On all three attempts, it "floated away".  So, now I see why you like the SRA 312: it's lower viscosity than the chipquik flux.  And it evaporates a lot faster.

Aligning the chip when placing it remains miserably hard.  I had been using tweezers, but now it looks like getting a vacuum chip holder is more or less mandatory.

BTW, Joe's method of simply switching on the toaster oven and later switching it off seems to work fine just fine.  Mine is a 1440 watt toaster oven (with convection fan) from Walmart.    Out of the box it can be set to go as high as  450F, and so I didn't even need to modify it.

[perky]

On all three attempts, it "floated away". 

I think there's something not quite right. Are you sure you don't have too much paste on the pads? Or do have you a fan creating a draught? The surface tension should automatically align it. The temperature profile might need looking at too (so uneven reflow?). Worth checking your pad sizes and solder mask against the datasheet for that package.

BTW do you pre-heat the board first? This is what a normal temperature reflow profile would do, I suspect this might actually be the cause in that the heat is not getting uniformally to the device because it's having to also heat up the PCB below it.
Mark.

[WhiteHare]

It sounds like Tom has a stencil, but instead of using it, he's painting it on.  Tom, what is the reason?

[TomWS]

I agree with Mark that the reason for the specific temperature profile is to give a mid temp plateau for 30-60 seconds or so to make sure the flux solvent is evaporated before it boils at the higher temperature and before the solder begins to melt.  It might be that your oven has too much power and may be overheating the flux too quickly and/or you're using too much solder.  The amount of solder needed for a 0.23x0.35 pad is pretty microscopic.  Also, the center pad needs very small dabs in the center of the 4 corners, not a big puddle in the middle.  There is also a learning curve associated with knowing how much solder is 'just right'.   A stencil helps a lot in this regard.

@Whitehare, I'm not sure what you're asking.  I use stencils on most of my boards, either polyimide or stainless steel depending on smallest pin pitch on the board.  I only use the paint method if I'm reworking a board or throwing a bunch of coarse pitched devices, eg SOT-23, 805 resistors, etc on a simple board.

Tom

[WhiteHare]

Thanks guys.  Sounds like with the AB1815 I'll need to do it by the book, which in this case means I'll need to use a stencil and install a thermocouple probe so as to adhere to the heating temperature profile.

[perky]

Yeah, the pre-heat plateau is actually very important as Tom says. I have used a heat gun to solder on larger LQFP type packages before, and I try to simulate that temperature profile rather than just bring the gun down on it. It's useful also to pre-heat the entire PCB (not to the plateau level, maybe 60 deg C or so) if you can when doing that so heat isn't immediately sucked away from the localized component.
Mark.