5
\$\begingroup\$

I am not sure if this is a good question but I am curious. Consider the following PCB:

PCB

I have realized that although the lines leaving the IC in the middle start fairly close to each other, they are significantly separated through the edge of the PCB.

Now I am guessing that the width of and spacing between the leaving lines are determined according to a standard but I am not sure why they are decided to be this wide. Could you explain the reasoning behind?

\$\endgroup\$
11
\$\begingroup\$

What you're looking at is a PC expansion board using the ISA expansion bus. This uses the then-standard .100 inch contact spacing which was widely available. There are a number of reasons for this. First, the connectors were cheap and fairly reliable. Second, when the PC was introduced, DIP package ICs were the norm, rather than surface mount. This means that the large size of the connectors was not a problem, since any reasonable circuit board was also large by today's standards. Finally, the contact fingers had to be fairly large in order to make the contacts physically strong enough (by having enough glue area) to hold up to repeated insertions.

EDIT - It's worth keeping in mind that the ISA bus was introduced in the original IBM PC - in 1981. You can read about it here.

\$\endgroup\$
  • 1
    \$\begingroup\$ No. ISA is .100, not .156 \$\endgroup\$ – Chris Stratton Nov 7 '15 at 18:38
  • \$\begingroup\$ ISA indeed dates back to the age when sizes were still measured in (fractional) inches. The first metric pitch was PCIe (1 mm) \$\endgroup\$ – MSalters Nov 7 '15 at 19:35
  • \$\begingroup\$ @ChrisStratton - Oops. You are, of course, correct. I've edited to reflect this. Thanks. What was I thinking? \$\endgroup\$ – WhatRoughBeast Nov 7 '15 at 19:55
  • \$\begingroup\$ I had thought that the previous (1970s) S-100 bus also used a .100" pitch edge connector, but you know what? It's .125"! Miniaturization in action. \$\endgroup\$ – hobbs Nov 8 '15 at 6:27
5
\$\begingroup\$

Mechanical tolerances, mostly. Fine pitch (close spacing) on edge pins is highly prone to shorting out or even being misaligned by an entire pin space due to tolerances required for insertion, and in some cases even thermal expansion could come into play.

In addition, there are practical problems with making the other half of the connector (spring-pin contacts) at a very small size.

Essentially, robust and reliable (also, the bus spec tends to be in use for a long time, and cause great gnashing of teeth when altered, while the pitch for the on-board parts is more free to alter as manufacturing tolerances improve over time.)

\$\endgroup\$
  • \$\begingroup\$ Sorry for the edit - had a brain fart and edited your answer by mistake. \$\endgroup\$ – WhatRoughBeast Nov 7 '15 at 16:32
  • \$\begingroup\$ Considering that memory modules today use 0.6 mm spacing, which is about 4 times finer, it is more than this is a 16-bit extension of an 8-bit design from the early 1980's. Given that when introduced it was displacing things like the S100 bus, it was considered compact for that era. \$\endgroup\$ – Chris Stratton Nov 7 '15 at 18:40
5
\$\begingroup\$

Supplementing WhatRoughBeast's answer, it doesn't look so unreasonable if you look at a contemporary card from when the connector was designed. This is apparently a CGA graphics card. Note the connector pitch is the same as that of the IC pins. The connector takes up a small part of the huge card. enter image description here

\$\endgroup\$
0
\$\begingroup\$

0.100 was ubiquitous in through-hole, DIL-era electronics, and using a .100 edge connector also meant you could use a socket that connected to the mainboard with pins at a .100 spacing - that's what everyone had experience using, and it probably also was easiest to develop (and manufacture) PCBs with .100 parts using the prototyping and design tools available then. Also, digital electronics back then tended to take a more naive approach to the problem of having to deal with relatively high frequency signals: one tried to just not get lines too close to each other to control capacitive coupling and lowpass filtering...

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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