Alioth's Journal: D'oh

So I was laying out the breakout board for the W5100, the all in one ethernet and TCP/IP offload chip I talked about in my last JE.

It wasn't till I was three quarters done that the thought struck me that the footprint for the chip looked a bit too big. So I checked the drawing at the end of the datasheet against the dimensions I was looking at... sure enough, it was too big! There are, it turns out, at least 3 breeds of 80 pin LQFP package... and then I discovered the one that the W5100 actually is, doesn't have a footprint at all in PCB.

Fortunately, it's very fast and easy to create new components in PCB, it only took a few minutes to make the footprint. But this chip actually has even finer pitch pins than the FTDI FT245RL, which comes in an SSOP package. It's going to be an interesting challenge making the PCB - finer traces and closer spacing than I've done to date.

Re-doing the layout also meant I did a much nicer job this time around, too.

There's one rather contradictory thing in some of the W5100 documentation. I downloaded the layout guide. It had the usual admonishments about not routing the analogue signals near the crystal - and not routing power near it, either. But there's a 1.25v *analogue* power pin *right next* to one of the crystal in pins, which seems contradictory - because this forces you to route at least a short section of supply power really, really close (0.2mm, or 8 mil) from the 25MHz signal line from the crystal.

V small

[anominous]

That's kind of what I was getting at in my response to your last JE! I did a highly scientific analysis - opened the pic full screen, and moved around until my thumb was the same size of yours, then held various objects until I found one of appropriate scale - - - - I guesstimated the chip to be approx one quarter the area of a Rennie indigestion tablet ;) (and *really* didn't fancy hand soldering that!). Good luck - hope you can can do it!! rob

Re:

[Alioth]

For precise measurements, the plastic bit itself is 1cm x 1cm. Each row of 20 pins is 8mm long. I think it's quite doable, I'll just have to carefully check the PCB under a magnifying glass before etching to make sure there's no bits of the transfer paper still stuck in the gaps (which would act as etch resist).

For soldering I expect about the same difficulty as SSOP which isn't much bigger. It's more time consuming than hard - very carefully line the chip up on the PCB using a magnifying glass to see it be

Re:

[anominous]

Thanks for the info, the replies, and for putting up with me :) I'm assuming "drag soldering" is what it sounds like - - get it everywhere, and use wick to remove the unnecessary. I'm amazed at the density of chips one can work with at home - - am tempted to break out the soldering iron, RS Catalogue, and have a play :) Wishing you all the best, Rob

Re:

[Alioth]

There's quite a number of techniques for home soldering fine pitch SMD.

Drag soldering is pretty much what you thought - there seems to be two methods - the big blob of solder on the iron tip, which you drag down the pins, or the run-the-solder-wire in as you drag the tip down the pins (I've been using the latter). You get a lot of bridged pins with either, but the solder wick takes care of that amazingly well. The first time I did it with a fine pitch SSOP, I had no opens or shorts at all. I was amazed. The