Even in the 1980s, PLCC sockets were available, and PLCC/SOJ was already being used in SIMMs.

  • They saved on the cost of a socket. Aug 21, 2022 at 3:02
  • Was PLCC sockets really that expensive.
    – Yuhong Bao
    Aug 21, 2022 at 3:02
  • 1
    Standard things a developer looked at: What was the price and availability of the different packages and sockets? How difficult is it to mount/dismount the chip in the field? Especially without special tools. (PLCC sockets are well known to crack in the corners if the chip is dismounted without the proper tool.)
    – UncleBod
    Aug 21, 2022 at 10:35
  • 3
    Hint: The 3rd dimension ;)
    – tofro
    Aug 21, 2022 at 10:38
  • 1
    Your question is unclear. Please edit to clarify what you're actually asking. ZIP, PLCC, and SOJ are types of packages, which are primarily used for chips, but can be used for other things. As written, your question is the equivalent of asking about a choice between apples and oranges. You're asking about using ZIP DRAM packages vs PLCC/SOJ sockets. It would be reasonable to ask about the choice of using ZIP vs PLCC/SOJ for the chip package, or the choice to use ZIP DRAM in a socket vs a PLCC/SOJ DRAM in a socket, but asking about a package vs a socket isn't really clear.
    – Makyen
    Aug 22, 2022 at 14:31

1 Answer 1


Most important is that ZIP offered higher density while still being thru-hole, thus compatible with existing manufacturing technology.

Sockets are generally avoided as sockets are a cost factor. They make only sense for

  1. (comparably) high cost components, or
  2. components with high failure rate, or
  3. components only populated as options

#1 is the main reason that sockets were avoided for average, mass produced devices.

#3 is also the reason why the 80287 was offered as DIP 40 instead of PGA - an FPU was considered optional way into the 1990s.

Likewise options were the only area where PLCC sockets were somewhat common, for example for RAM expansion with graphic cards during the mid 1990s.

While PGA sockets were more expensive than DIP, thy were still lower priced than PLCC sockets, due being essentially the same socket technology - just more pins. The same consideration is true for ZIP. Here as well it was only a different plastic carrier holding the same pin elements.

Equally important PLCC only makes sense, as a bridge technology between classic pinned and SOJ, if a SMD manufacturing process is used. But SMD was still in infant stage in 1980 - it wasn't until the 1990s that SMD became a common place - not at least due many components not being available as SMD.

Typical video cards of the mid 1990s are maybe the single best known example for PLCC socket use, like this No-Name S3 Trio64V2 of ca. 1995 shows:

enter image description here

Using sockets was a way to stay with a single board design, but still delivering variants with different memory population, here 512 or 1024 KiB. This wasn't so much meant for user side upgrade, but the ability to deliver two different memory sizes while producing only one board and add the second set of chips depending on incoming orders. A great idea during production ram up, as it would only need the minimum amount of RAM by default while keeping flexibility. This is especially underlined by RAM being, at that time, the most expensive part, thus minimizing initial investment has a high effect on ROI.

So everything comes down to a rather short window of opportunity in the mid 1990s for mass usage of PLCC sockets in large scale production.

Last but not least, While PLCC was first introduced ca.1977/78 by TI, it did only receive a JEDEC Standard in 1984, which marked the time other manufacturers started to adopt that packaging as well.

Bottom line:

ZIP offered higher packing density while still staying with existing manufacturing process.

  • I suspect 1984 is also when PLCC packages for DRAM was first thought of.
    – Yuhong Bao
    Aug 21, 2022 at 4:02
  • Wasn't those sockets known as SOJ IC sockets? I only heard about the 4-sided once being called PLCC IC sockets. I remember we used 4-sided PLCC IC sockets for pre-programmed ICs, but never for RAM of any sort.
    – UncleBod
    Aug 21, 2022 at 13:04
  • 1
    @UncleBod Those terms intersect. SOJ are SOIC with J type pins. SOIC was conceived as term for smaller versions of existing chips, mainly TTL, while PLCC used as for package descriptor for large devices (more then 40 pins). JEDEC defines PLCC with 2 and 4 sided pinout. RAMs being the most prominent of two sided, as two sided simplifies board design - like having multiple in a row for wider words (addresses and CS in parallel) or larger address space (data and addresses in parallel) or both :)
    – Raffzahn
    Aug 21, 2022 at 13:23
  • Sockets also hurt signal integrity, sometimes significantly, and tend to have worse density (pin count / PCB area) then the corresponding socketless design. (Note phrasing; there are of course high-density socket designs that have higher density than a low-density socketless design, but a socket generally requires additional space compared to a similar design without a socket.) Not too much of an issue in those days, but still yet more nails in the coffin.
    – TLW
    Aug 21, 2022 at 19:20
  • By the time of SDRAM TSOP was required anyway, but I am talking about before that.
    – Yuhong Bao
    Aug 22, 2022 at 1:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .