search

particle-iot / core

Hardware design files for the Spark Core, a tiny Wi-Fi development kit.
https://www.spark.io/
Creative Commons Attribution Share Alike 4.0 International
367 stars 56 forks source link

Oval-shaped pads for the Core headers #22

Closed zsup closed 10 years ago

zsup commented 11 years ago

Suggestion from Bunnie: Make the header pads oval-shaped, to decrease the likelihood of jumpers.

technobly commented 11 years ago

What do you mean here by "jumpers"? Make the pads stronger so they don't break when desoldering so you don't have to add jumpers later?

zsup commented 11 years ago

Meaning solder jumpers, i.e. short circuits on the manufacturing line

technobly commented 11 years ago

Oh... hmm. Solder mask is suppose to prevent that.. but I guess the headers are hand soldered? If Bunnie is your CM contact.. I guess they know better.

zsup commented 11 years ago

Yep hand soldered

bunnie commented 11 years ago

Correct, because the headers are protruding from the top and have to be put on after SMT, and also due to the density of the top/bottom sides, it's not possible to do wave solder on the headers. Therefore, the headers will be hand-soldered. Technicians will likely be running knife-style solder blades that can heat multiple pads at once down the edges to quickly solder the units. Increasing the spacing between the pads can help reduce shorts.

Also, they requested white silkscreen fill in the pin header region. I've seen this practice occassionally to improve yield, I'd consider it optional but worth a try.

technobly commented 11 years ago

Ah... that makes a lot of sense, but to clarify isn't the solder mask white and the silkscreen black? http://i.imgur.com/bw2pqpE.jpg

zsup commented 10 years ago

Upon reviewing the board, it may not be possible to switch to oval-shaped pads, as many of the traces are tight against them with no room to move. @satishgn take a look and see if you think this is workable, but if not we might just have to skip this one.

bunnie commented 10 years ago

I think the intention is actually more to trim copper than to add copper to make the oval shape. i.e. what they want is to increase the gap between copper regions to avoid bridging.

I don't have the design files on-hand, so I don't know what your annular ring is currently. But, if you are escaping your pads toward the inside (I presume you are), and if the drill escapes slightly on the pad on the edges due to the thinned down ring, you should still be fine in terms of manufacturability and yield. The point at which you lose is when the drill goes through the board at the intersection of a trace and the pad.

Another thing you could try if you're really out of space is to bring the copper all the way to the outside edge (so it's being milled/scored during board singulation), this can also serve as a "solder sink" for excess solder; they are running the iron down the outside edge and i think having the copper available toward the outside will help the surface tension of the solder pull solder away from bridging between pads.

satishgn commented 10 years ago

Zach, this would be tough to do as many of stm32 signal traces are very close(with min 8mil clearance) to the header pins. How about reducing the drill diameter of the through-hole pads so as to just to accomodate the berge strips?

mohitbhoite commented 10 years ago

We concluded to go ahead with the current header pad layout with slight reduction in the pad and hole sizes. I have committed the changes made to the PCB under the dev branch and updated the spark library to have this part. @satishgn could you make sure if everything looks okay in there? (This would be first commit to the Core PCB!)

Thanks!