Slightly OT, but I trust this group more than the bulk of the WEB, sooo...

Considering modern fabrication technologies and techniques, are vias the bad boy they used to be (circa early 80's)? I spend hours, many hours, reducing the number of vias in my designs. However, I see some commercial products, like the Arduino, that have enough vias to give me fits. Are my fears unfounded?

_________________
Bill

www.aztecmcu.com

We have never, ever had a via go bad in our products. I've been designing boards with vias since the late 80's. The closest problem we had was a manufacturing problem at a particular board house maybe a dozen years ago, where a problem in their drilling operation was wiping epoxy over the exposed edges of pads on inner layers so the thru-plating couldn't connect to them. The boards were bad from the start though, not years later. Our vias on .062"-thick boards are .015" finished hole size, and on small .031"-thick boards, they're .008" finished hole size. I put .010" of pad around the hole, so for example a .015" hole has a .035" pad. If you don't really have a pad, ie, the trace just kind of falls into the hole, I suppose that constitutes asking for trouble.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources

Vias gave me fits on the neo geos, not for the manufacturing but if they had an acid spill of coca cola since the solder over the or the hole itself would rot, so I would need to run a wire through them and solder them to each side....

_________________
"My biggest dream in life? Building black plywood Habitrails"

I've never had a problem with vias. I use them as needed to maintain a neat board layout. My POC unit is full of them and works fine.

_________________
x86? We ain't got no x86. We don't need no stinking x86!

Okay, I'll man up about them then. It will probably save me a day a month and make my layouts a little easier.

I used to work for a teleprinter manufacturer back in the late 70's and early 80's. IIRC, about 30% of early failures we due to the vias thru-plating fracturing away from the trace. It was easy to fix, but expensive to get the machines in from and back out to the customers. We were paranoid about them.

_________________
Bill

www.aztecmcu.com

That must have been at the end of the problem era for them, because I bought several Heathkits in the late 70's and early 80's with double-sided boards and vias, and the assembly instructions specifically said you didn't have to solder the vias, yet to this day, none have failed that I know of. (I have to say "that I know of" because I gave away and lost contact with my $750 oscilloscope with a bazillion PCBs.) I do remember seeing ads in the trade magazines for PCB eyelets though, to fix bad vias I suppose (although the vias I design into boards now are much too small to put the eyelets in).

_________________
http://WilsonMinesCo.com/ lots of 6502 resources

I recall in the 1970s patching up PCBs with faulty vias by inserting bare wire through them and soldering. Tracking down the bad vias was time-consuming though.

_________________
x86? We ain't got no x86. We don't need no stinking x86!

We've come a long way, baby. Even 20 years ago I remember seeing boards with traces whose width was the same as their height (ie, thickness) with sharp corners and straight-up sides. How they can make a trace with a square cross section is beyond me-- but so is how they can make die-feature sizes something like 1/20th of a wavelength of visible indigo light!

_________________
http://WilsonMinesCo.com/ lots of 6502 resources