Why the strange pinouts?

--- Because you've got to bring all the inputs and outputs to the edges of the silicon to connect to the lead frame, and most most times (always?) you can't get all the I/O's to come out perfectly ordered just the way you want them to. It's like laying out a single-sided PCB. (At least it was back when I was doing that kind of stuff. I don't know how many "layers" are available now though...)

---

--- Same story, I suspect.

In truth, it doesn't matter much since you're going to have to do a PCB layout anyway, and even if all the I/O's came out of the chip ordered perfectly, the layout would vary from PCB to PCB.

-- John Fields

Ah, sweet mysteries of life.

I've always been annoyed by the fact that PIC microcontroller chips have their power pins in the wrong places. For the PIC12F675, for example, Vdd is pin 1, whereas Vss is pin 8. For an 8 pin PDIP package, ground is usually pin 4, and Vdd pin 8.

Some chips use pin placement in an attempt to deal with heat. For example, the TI SN754410 quad 1/2 H driver chip has 4 ground pins in the center, which double as heatsinks.

I've also heard that some of the AC chips have unique placement of power and ground in order to mitigate ground bounce problems, which can occur when all the outputs go up and down simultaneously. Since they are so fast, they can induce an oscillation on the ground plane that can affect other chips in the vicinity.

I'm sure most of the placement of pins for logic families is historical. I believe certain chips (some earlier cmos 4000 series) have fallen out of favor because of pin placement, and been replaced with chips that have more 'standard' placement.

I had problems with the f,g thing on a recent project. It would have been so nice to have a single sided PC...

I'm suprised there aren't CPLDs targeted towards prototyping/hobbyist use. Seems like using this technology, you could replace nearly any logic chip with a faster version that has somewhat more flexible pinouts. However, atmel and the other cpld mfgrs have proprietary programming protocols that forces one to buy expensive programmers. That makes it hard for people to do on the cheap in their garage.

You could have:

Cheers! Rich

It's because of the way the ins and outs landed on the chip when they designed it - you don't design a chip the way you design a circuit, and stringing a conductor all the way around the die to come out at a certain pin is a LOT different from doing it with a PCB trace. Rearranging inputs and outputs would have made the chips more expensive.

Lately, they've been doing it since designs are so complicated that one jumper or so here or there doesn't add that much to the cost.

Hope This Helps! Rich

The typical reason is for the convenience of the die layout. Another reason would be for economics, nobody would use a more expensive package that is also larger and therefore used more PCB real estate. There are also power dissipation reasons as well as electrical noise minimization. Much depends on the function of the IC.