MicroZED

Thanks John.

I suppose the *with your finger* bit helps to break through any film of contamination. Like they used to say to use a pencil eraser on bare copper / veroboard before soldering.

For gold finish, do you think it is better to solder coat the pads first (give it some extra solder like the paste would) or just leave it alone (maximally flat surface).

Yes that makes it a bit harder for this sort of thing.

We do inspect the outer rows, but that's more a process feedback issue than individual board QC.

It can see down the rows pretty well, so it would see a clump of flux, or some stray solder balls. We use rosin flux, so we aren't concerned about leakage and corrosion, like you might be with water soluble fluxes.

We've found BGAs to be extremely reliable, both initially and in the field, way better than fine-pitch leaded parts. And there's no other way to get 780 or 1120 pins on a chip.

We don't. Rosin flux residue doesn't do any harm.

Don't you ever get the feeling that playing with tiny parts on circuit boards isn't, well, a very macho way to make a living? Shouldn't we be out there chopping down trees or running ships aground or something? But on cold rainy days, I appreciate electronics more.

As the parts get smaller, everybody needs optics.

Reconsider a Mantis. If you set the IPD knob to match your anatomy, they are comfortable and the view is stunning. And you don't have to hunch over like you do with a visor.

Would a thin layer of epoxy or whatever add any substantial strength to a thousand soldered balls? How would you get it in there?

We've had zero ball-joint failures in the field, must be many billion ball-hours so far.

BGA's fare a lot better in equipment that is mostly stationary. Less so where things move or vibrate a lot. But yeah, if you absolutely need many hundreds of connections there may be no choice.

I think that depends on the halides in there.

To compensate for that I just dusted off my old mountain bike and ride it after lunch. Those big tires feel very manly and make a manly noise on paved roads. The main reason is that I found that a brisk 30-45min walk every morning is not enough to keep the weight off. Seems the fuel efficiency of a human goes up with age.

I don't have space for a Mantis here, mostly because I sometimes have very large prototypes on the lab bench. There can't be anything big and heavy in the way. When I want to avoid the hunch I use a Veho 20x USB microscope. A Supereyes 500x if it gets really small but that one doesn't allow several inches of vertical work space. Those are cheap "toy scopes" but they work quite well. As I said before, the good old days are today.

We have had very few fail, not enough to count. OTOH, we have a boatload of programmers who think that all serial ports are the same. It's no issue for the assembly houses to do, just that none of the engineering groups I've been in has invested the time or money to do it. They wouldn't get enough practice, anyway, perhaps because they don't fail often enough. ;-)

The boards and the parts are worth too much, particularly the engineering hardware. Production has all that stuff but they're in Mexico (add customs to the mix).

At the PPoE, we never needed to. There were never any problems. Here, we have X-ray gear but really little need for it.

We use all without fears. Unless there is a good reason to use larger, our default size is 0402. We try not to use BGAs with ball pitches less than .8mm but not because they can't be placed but for cost and reliability reasons. Lower than .8mm and the board dimensions have to change or B&B vias become necessary. They're also tough with 2oz. copper. ;-)

In the wake of all the BGA issues many service providers have sprung up. I don't know this place, just an example:

I just wish soemthing like this would be available somewhat locally for whole prototype assembly. A perfect biz opportunity for techs that were laid off. But hardly anyone does it. Beats me why not, considering that the investment costs in equipment are very modest.

It took me a while to get the hang of TSSOP, but now those are pretty much standard. I do automated placement and mass reflow of those with only a little trouble. I've stuck to 0805 passives, as there is not that much advantage to me to go smaller. (For microwave or high speed stuff, there may be electrical advantages, but I'm not in that regime.) I had a real bad experience with a chip-scale comparator (AD CMP603) that was only available in that package, had 65 on each board, and massive number of shorts under the chips. Soured me on the whole concept.

Jon

In a former life, my boss had a CM company. He got laid off during one of the downturns and bought some placement machines at a penny on the dollar and went into business. Made a bunch and then sold out. He's done similar things a couple of times.

Underfill makes a huge difference, like a factor of two in chip size for a given reliability.

IBM uses lots of giant BGAs. They invented the ball grid array back around 1980, for the 3081 mainframe. The local name is C4, which has nothing to do with plastic explosives, but stands for Controlled Collapse Chip Connect. (I think the collapse that's in view is the solder balls coalescing with the solder on the board.)

Cheers

Phil Hobbs

I'm hoping to be able to get a Mitutoyo FS110 microscope, but I need a bit more bench space first.

Cheers

Phil Hobbs

For me it's not just bench space, I'd have to get more office space first. And I really don't want to move out from here because I thoroughly enjoy working from the house. Almost zero commute time, anytime I can take a break can be family time, on rainy days we can start a dog walk the millisecond it clears up, the younger Labrador checks on my work every 20 minutes, and so on.

Joerg wrote: For me it's not just bench space, I'd have to get more office space first. And I really don't want to move out from here because I thoroughly enjoy working from the house. Almost zero commute time, anytime I can take a break can be family time, on rainy days we can start a dog walk the millisecond it clears up, the younger Labrador checks on my work every 20 minutes, and so on.

Nope. Plus getting a building permit in this area has become next to impossible. It took our church more than five (!) years for an education building and a ballfield, despite the fact that we declared it open for use by the whole community.

My bench isn't very big either. My Mantis doesn't take much room. It tucks away when I'm not using it. And it can swing over/into pretty big rackmount boxes.

Exactly -- never heard of "repinning", while "reballing" is quite popular :->

Best regards, Piotr

That's too big for me. My main lab bench is a large angled bench where one wing has a big equipment rack and the other needs to remain free from three side. It is a very heavy duty bench that can (and sometimes has to) take loads where the one in your picture would probably get bent in the frame or where it would snap the wood.

something like this?

-Lasse