17
\$\begingroup\$

I'm searching around digi-key when I noticed one of the MEMS oscillators has these copper pieces on the edge of the package:

enter image description here

I'm not talking about the pads, I'm just talking about those little copper pieces on the side of the package. I've seen them before and I've always wondered but I never really thought to question it.
Now I'm curious.

What are those for? Are they the internal copper connections, and if so, why are they brought out to a visible location other than the pads you solder to?

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Part of the chip frame? \$\endgroup\$
    – Majenko
    Commented Jul 16, 2014 at 19:55
  • 1
    \$\begingroup\$ It would be interesting to know if they are exposed as a side effect of the way the package is made, or if they were part of a production test (or even laser-trim calibration?) interface. \$\endgroup\$
    – Chris Stratton
    Commented Jul 16, 2014 at 19:56
  • 1
    \$\begingroup\$ I'm quite sure they are usually exposed just because not exposing them would be a PITA... \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 16, 2014 at 19:58

2 Answers 2

Reset to default
24
\$\begingroup\$

For plastic overmoulded parts, the lead-frame (correct terminology) is held in place while the plastic is injected. For packages like PTH (Pin through hole) or j-leads etc. the surrounding support is cut away with shears and the individual chip is freed.

In the package you show, these pads cannot be sheared and must be moulded into the package. that means that the lead frame must exit the package while the moulding/injection is happening (for support - locating reasons). These are then sheared off the end of the package and the chip is then freed.

This applies to over-moulded packages only. Ceramic packages can be multilayered mini circuit boards equivalents.

That particular package is a VFQFN - Which is similar to a MLF (Micro Lead Frame) from Amkor. enter image description here

Here is a snip from their website

Here is a manufacturing drawing for a QFN ( a little more complicated than the package in the OP) On this I've drawn a red square which shows the boundary that the saw will cut on to define the package edge. Note: the N in QFN means "No Lead" the red circle shows the die attach stabilization structure as it is brought to the edge of the package.

enter image description here

And finally here is a picture showing the moulding process being modeled. I've put a red circle on this to show the die attach stabilization structure again. The external frame is not shown in this picture.

enter image description here

One remaining thing to note:

In the OP post the copper on the side of the package appears to be on a separate plane. Whereas the pads are exposed on both the sides and bottom, these guys are only exposed on the side. There are several ways to accomplish this, one way is to look at the very first drawing and note that the "Cu leadframe" or "exposed die paddle" have steps at the edges. The leads that do NOT need to contact on the bottom are etched back during manufacture.

\$\endgroup\$
7
  • 1
    \$\begingroup\$ That's great but I still don't get why we'd see two fins for one pin with the configuration depicted in your image... \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 17, 2014 at 7:40
  • \$\begingroup\$ @VladimirCravero What do you mean by "fin"? And why do you think there is a ratio of 2:1 ? In these MLF packages, because the lead/pad of the package is not trimmed it also does not contribute to any mechanical locating system. In the lead frames you show, the leads themselves support the die during moulding. Here they can't. It's likely that end bits on the end of the package connect to the individual leads to support them. \$\endgroup\$
    – placeholder
    Commented Jul 17, 2014 at 14:03
  • \$\begingroup\$ With fin I mean "tiny bit of metal that you can see on the side of the chip", I don't know if there's a word for that. What I'm saying is that your example depicts a VQFN, as you say, but OP image has two "fins" per pin that doesn't even seem to touch them. I understand that you can have a "fin" not touching the pin but I don't understand why they are two. To see another two "fins" example look at the first picture I posted, on the far right (PIN 8) you can see two "fins" for a single pin, but in a leadless QFP flavour I'd expect them to be fully connected, like your example.Hope that's clear \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 17, 2014 at 14:12
  • \$\begingroup\$ @VladimirCravero I'll see if I can did up a diagram from a lead frame vendor and post it later. In the mean time notice that in your PTH lead frame the 2 "fins are only on the corner pins. That is probably done for stability , to prevent rotation of that pin during plastic flow and mechanical strengthen the finished part. the other pins don't need that. \$\endgroup\$
    – placeholder
    Commented Jul 17, 2014 at 14:19
  • \$\begingroup\$ That would be awesome, I'm afraid I could not explain properly my doubts though... I'll wait for your drawings then, thanks. \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 17, 2014 at 14:21
21
\$\begingroup\$

preamble: this answer is getting quite a number of upvotes but please note that this other answer is great too and adresses smd parts.

Have a look at this image:

enter image description here

As you can see in the center there is the die, supported by a metal plate. All the pins are brought near the die with these metal fins. Now look at this:

enter image description here

All the fins are connected together before being like in the first image. You have this big frame with lots of "die supports" connected together. You place and glue the die, wire bond it, put some plastic then cut the chip from the frame. When you cut it these tiny copper/Al/whatever metal fins is all what remains from the former frame.

\$\endgroup\$
8
  • \$\begingroup\$ Oh interesting! I knew they connected the die to the package but I didn't know the package had a specific frame. So if I wanted to, I could buy just a bunch of PDIP 28 frames without a die in them, thats interesting. Would you happen to know why then that they have the end of the frame exposed on some but not others? I assume its just preference on the company but could there be any technical reasoning for it? \$\endgroup\$
    – Funkyguy
    Commented Jul 16, 2014 at 20:03
  • 1
    \$\begingroup\$ I'd hazard the guess that exposed frame is cheaper but the chip is less resistant. If you cut the frame before putting the die/frame in the mold you can have your nice chip with no fins, but that would require a more complicated mold I guess. And I think you can't but them, or better they sell them 1k at a time. \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 16, 2014 at 20:06
  • 1
    \$\begingroup\$ This imho doesn't actually address smd parts, like op posted. \$\endgroup\$
    – Passerby
    Commented Jul 16, 2014 at 22:29
  • 1
    \$\begingroup\$ @Passerby I agree and upvoted the other answer, I can't change the accepted though and I'd rather not delete this since it adds some informations about non smd components. I don't even think that integrating my answer with placeholder's is a good idea... I might add something though, what do you suggest? \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 17, 2014 at 7:36
  • 1
    \$\begingroup\$ @vaxquis the "die support", which is the big metal plate underneath the die, is usually both electrically and thermically connected to the die. The die itself is usually kept at the lowest potential in the chip, i.e. ground or Vss. \$\endgroup\$
    – Vladimir Cravero
    Commented Jul 17, 2014 at 7:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.