Zilog Z80 bites the dust

Incidentally there's a book that might be of interest to people.

I can't say it's written like a bestseller, but the guy isn't a journalist or author, he created the Z80!. So I take back even the slightest criticism in that previous sentence : ) Plus, it's a first-hand account, so that's also a significant positive.

Incidentally there's a book that might be of interest to people.

I can't say it's written like a bestseller, but the guy isn't a journalist or author, he created the Z80!. So I take back even the slightest criticism in that previous sentence : ) Plus, it's a first-hand account, so that's also a significant positive.

Just read the excerpt you posted.  I've suffered a bit in PCB layout when it's a challenge to fit everything onto a tiny board... but yeah what Federico went through would have been a hundred times worse.

Hehe yes, it's quite interesting, having never done chip design, I would never have thought there could be such a simple problem, but it totally makes sense, that in the manual drafting days, once one dimension of the chip was established, you kind of just had to continue designing the chip to be as long as required to fit all the functionality : )

Yeah, I try and imagine it as having the constant fear of painting yourself in a corner.

When laying out with tape, you can't even see the corners, which are in different directions on different layers, until it is too late. Sometimes they are more like land mines hidden on different layers.... 

Land mine is a scarier analogy than paint.... and I'll have to read up on historic IC design to fully appreciate your post.  Sounds like you aren't talking about fitting concerns but accidentally connecting to the trace on a layer you can't see.  If so, then yeah, land mine sounds about right.

We used translucent tape (blue and red) for top and bottom.

I did the initial design with 1:1 scaling on 0.1" graph paper using blue and red pencils - the tape up was 2:1.

Never had to tape up anything worse than two layers before PCB CAD on PC turned up and saved the day !.

MK

Transistors necessarily have overlapping structures but the amount of overlap affects properties of the transistors including diffusion of impurities to dope the silicon. It was much more than just a 3D juggling act. 

Thanks Doug.  I have a clearer picture now.

I bought the book and have read it now. Very interesting stuff about the early development of logic chips. I hadn't realised just how manual the process on the first 8080 and Z80 was.

The book is somewhat spoiled by the random AI/philosophical stuff, of which there is far too much - you did warn us !

MK

I felt the same! Also I'm constantly surprised top science and tech journalists have missed out for decades on writing about these subjects for the general reader. Even a kind of 'Bill Bryson' level style for some of the flagship semiconductor stories would be interesting to read, at least anyone could get into that level of explanation, not too deep.

I liked the book Computer Architecture A Quantitative Approach although the price is crazy nowadays. A bit too much in-depth, it's not a light read, but that is offset slightly with real-world example chapters, where the author discusses how the theory applied to (say) Pentium, MIPS, etc.

Bob Johnstone is a New Scientist journalist, who writes extremely well, at a popular science level but for engineering topics, I've read several of his books. He seems to consistently do decent research for his books, going out and interviewing the people involved.