Tips for building a grubDAC or similar SMD project.
Oct 3, 2012 at 3:09 PM Thread Starter Post #1 of 9

bwshockley

New Head-Fier
Joined
Sep 4, 2012
Posts
22
Likes
11
First, I'm not saying that my methods will work for everyone.  I'm just putting them down so that someone else may learn and they can try them for themselves and determine if it's right for their project.
 
That being said, I recently finished my first ever SMD project, the grubDAC.  I purchased a kit from beezar.com and it was an extremely thorough kit and well packaged.  You can see many photos of the build in my album shown in my signature.
 
Equipment:  I used an Aoyue 936 with a 0.8D Tip set at 325C {Edited - thanks cobaltmute} the whole time.  I also used the Aoyue brass sponge tip cleaner after each and every component placement.  I used .022 diameter 63/35/2 solder from Radio Shack.  I found it to be about perfect.  I also used a flux pen for wetting the board, but I'm not sure it was 100% necessary.
 
I held everything in a set of "helping hands" and actually used the magnifying glass to work.  I'd say 95% of the time I used the glass for the SMD parts.  I didn't use it for the through-hole components.
 
I used some self-clamping tweezers, with long handles, to help hold components as well.
 
I followed tomb's excellent write up, so I will only list out of few of the things he doesn't mention or didn't do.
 
 
ICs
 
Starting with the ICs, I wetting the board with flux and tinned the first corner contact.  I then used some blue painter's tape (doesn't leave residue) to hold the IC on the board while I used tweezers to press down and the soldering iron to melt the tinned contact.  This worked out great and kept the tape on there while I did a few more pins.  I then removed the tape and completed the job.  I didn't necessarily work on one side first, then the next and I didn't go in pin order.  I jumped around so that each edge/section of the chip had a little more cool time before doing the pin next to it.  I would do each pin my placing the iron tip on the PCB contact butted up against the IC pin.  I'd wait a couple seconds then try and apply the solder to the IC pin itself, down where it was butted against the iron tip.  This worked well for most of them and the solder quickly slowed down the pin and onto the pad. I'd then quickly remove the iron tip.  I'd check for a complete joint and reapply if needed.  I found it was much easier to apply more solder than to remove excess solder.  I did one pin at a time, again jumping round the IC so as to not heat one section of the chip too much at once.  Doing this method I was able to complete the main IC ship in a few short minutes with only one easily fixable bridge.  Between each and every solder, I cleaned the tip of the iron to remove added solder.  I think this was key to being consistent.
 
I proceeded to the Wolfson chip with the same method, starting with blue painter's tape and the single tinned pad.  I completed about 3-4 pins before removing the tape and continued like before.  No bridges on this chip.
 
SMD Components, including VDO Regulator and Oscillator
 
For the SMD parts, I started by following Tom's suggestion of tinning one pad, placing the component, then completing the other side.  However, I was finding this very difficult to do, as I'd sometimes end up with the component not seated down on the board, but the end on the pre-tinned side up a bit.  Maybe if I had different tweezers it would have worked better, but I could seem to get the leverage needed on the tiny part to drive it down onto the board.  If you do use this method of pre-tinning, put just a tiny bit of solder on the pad.  You can add more later, you just need enough to tack the component on the board.
 
The method I found most useful to me was using my large tweezers to clamp the component on the board without any pre-tinning.  Being very careful of what is under the board, I'd clamp the component down on the PCB board, used something to prop the other end of the tweezers so I didn't add a moment to the board.  The tweezers did not have a strong enough clamping force to worry that I'd crush a component, but they were enough to hold the component in place.  Id's then proceed to solder both ends of the component and carefully remove the tweezers.  If you follow this method, wrap a tiny bit of tap around the tips of your tweezers and mind the bottom of the board that you don't crush/scrap/jab or otherwise damage anything on the board.  This might have required more time per component, since I had to align it, then try and clamp without moving it, readjust as necessary, but felt it provided more precision and the soldering part was faster and easier and meant less time with high temps on the board and fewer retries.  I did this method for all SMDs, the VDO and oscillator.
 
Through-Hole Components
 
This is where most DIYers have experience and again use the method that works best for you.  For me, to get a clean close to the board mount, I would place the component on the board, with the leads through the holes as close as I could without bending the leads on the underside of the board.  I'd then tack one lead, then put the solder down and using pliers and the iron reheat the tack and pull/push the component into position, then remove heat.  Then I completely solder the other lead in place then go back to the first and complete that one.  For the larger capacitors on this build, I'd tack them, then using my thumb, push on the top of the capacitor while I reheated the tack until the cap was completely seated.  This works okay if you are fast, they can heat up quickly if you aren't careful.
 
Upon completion I double checked all my solder joints.  My advise is to start with what you think might be too little.  It is easier to add more solder than take it away.  Or course, having a solder wick or solder suction bulb is helpful.  Again, clean your tip after every component placement or if you feel you have additional solder on the tip before starting a new joint.
 
Oct 9, 2012 at 5:43 PM Post #4 of 9
OK, I appreciate that this thread is well-meaning and all, but it really overcomplicates the whole process of soldering SMT parts unnecessarily.
 
I have hand-soldered thousands of SMD parts with nothing more than regular multicore solder, a soldering iron, a wet sponge, a solder sucker and some solder wick.
 
You just tack the chip to the board accurately in 2 or 3 places, there may already be enough solder on the pads to permit this, and run a large blob of solder down an entire row of pins, feeding in some solder to ensure fluxing, dragging it with the tip of the soldering iron, to which it will cling due to surface tension. Then you pick up most of it with the iron and remove any remainder with the solder sucker or solder wick. Work on a side away from one that is tacked in 2 places or the chip may skew. Repeat for any remaining sides.
 
Don't worry about overheating the chip. Just work steadily and purposefully.
 
This is the right way to do it. Learn to do it this way, your boards will look professional and you will not look back
 
The solder wick is good for tidying up any pads if you have made a mess of things and had to remove the chip. A hot-air paint-stripper gun is sometimes useful for removing chips whole scale, or in the last resort, to remove a chip without risk of disturbing other parts, cut it off the board with the tip of a scalpel, severing the pins carefully, working very close to the chip so that the blade hits the chip instead of the board when you push down on it. Then collect the pins with a blob of solder on the iron. Better to lose a chip instead of the whole board.
 
2-terminal parts are placed by heating the solder on one pad and sliding the component into place. Solder the other end, then resolder the first end.
 
Flux is unnecessary, although it can be a help when working with BGA parts.
 
Clean the iron when necessary by wiping it on a wet sponge, i.e. when a lot of black flux residue builds up on it, but don't obsess about it.
 
I do use tweezers for moving parts around, but mostly I work on my knee. It's a skill thing that you can acquire.
 
w
 
Oh, sometimes it's a good idea to solder the biggest, most difficult chip first. This means that if you have to go to the hot-air gun to get it off, you can just turn the board upside down and heat it and shake it off, without worrying about anything else coming off, and there will be a good chance you can start again with nothing lost.
 
Oct 9, 2012 at 8:04 PM Post #5 of 9
hmm on your knee? think i'll stick to my tweezers =) I use drag soldering too, but I disagree about flux, I use quite a bit of it.
 
Oct 9, 2012 at 8:19 PM Post #6 of 9
I know lots of people like to have flux, but we just used to consider it a timewaster on the production line, and you can see lots of things here I've built without it. Yes, the girls who did BGA rework were allowed a squeeze bottle.
 
w
 
Oct 9, 2012 at 9:08 PM Post #7 of 9
no cold-war soldering here =P if I have a need to go quicker i'll build a reflow oven/controller

I just find I get shinier joints with added flux, you say potato, I say potata
 
Oct 10, 2012 at 9:58 AM Post #8 of 9
When building I flux larger parts (PCMs), but I don't flux smaller ones.

For the smaller parts such as TPS regulators, I actually do the tweezer slide in method.

On 2 terminal parts, I do the ground side first.

I do everything under magnification, even through-hole. You can watch the solder start to flow and that is good way to see what is going on.

For us amateurs, the most important thing is to check your work right away. I use a 8x jewelers loupe. Sometimes you can't come back to fix stuff, due to the tightness of the board. Yes it is slower, but there is a definite pride when your board works right, the first time, every time. :D
 
Oct 10, 2012 at 5:32 PM Post #9 of 9
+1 on the jewellers loupe.
 
The reflow oven:-
 
 

 
No controller required, there's a self timer or you can just peer in the window and turn it off when you see the solder paste change from grey to silver. It has a quartz-tube fire bar above and below. I don't worry about a preheat and cooldown profile. I don't worry very much in general...
 
I have the advantage of designing the PCBs so I get to make stencils too..
 
.
 

 
These are surprisingly uncritical to use in practice, a bit of bleeding from pad to pad rarely causes a problem that's very hard to fix. Mostly the components do the final alignment themselves as the surface tension and floating is sufficient to pull them around
 
I etch these using toner transfer and muriatic + peroxide. Tape over one side completely. Since I have inkjet printers at home, I do the prints at the local library. This means using the regular paper as the library won't let me use anything else, but I touch them up with a resist pen. You can see there's some pitting, but I'm not actually running a production line, so I just wash the stencils off with isopropyl 
regular_smile .gif
.

 
w
 

Users who are viewing this thread

Back
Top