LTC Switcher CAD3 ?

Inductance =3D Al x turns^2 Henries

so the dimensions of Al are Henries. turns^-2.

If you three agreed on anything else, you got it wrong.

-- Bill Sloman, Nijmegen

the

Jamie gets it wrong - as usual.

-- Bill Sloman, Nijmegen

The only thing I did wrong was acknowledging your existence here.

Jamie

...

Bill: "per" means "divided by".

Tim

the

The usual failure of perception.

-- Bill Sloman, Nijmegen

Oops. Agreed.

-- Bill Sloman, Nijmegen

rote:

re - EPCOS

to 630nH per

of

squared.

eradian

tor

er

s.

fy

s at

That makes you a Holmesian, rather than a Sherlock fan. I've just re- read Julina Barnes' "Arthur and George"

which is essentially a portrait of Arthur Conan Doyle - a second rate author, who didn't recognise Holmes as his greatest creation.

David Lodge does the English author biography a whole lot better. "Author, Author" about Henry James, and "A Man of Parts" about H.G.Wells, are both substantially better, perhaps because the subjects had a bit more to offer.

-- Bill Sloman, Nijmegen

EPCOS

squared.

steradian

He claimed to have written one story during a break in a cricket game, and killed him off because he was bored with him.

I can't read either of them, for some reason. HJ is especially boring to me... I've never made it halway through any of his books. But I can, and do, reread the better Jane Austens and PG Wodehouses regularly.

I'm now reading Enigma, the Turing biography. It's amazing the ideas that had about computers in about 1950.

core - EPCOS

ir to 630nH per

er of

rn squared.

steradian

factor

ister

ysis.

who

ecify

nd

eers at

he

ul

the

...

I can't read his novels - nothing ever happens. He wrote a few science- fiction short stories in the sytle of H.G.Wells (who was a friend of his) which aren't too bad.

H.G.Wells was also a second rate novelist. His science fiction is interesting - the ideas are sufficiently interesting to transcend his defects as a novelist - if somewhat half-baked. His ideas about the shape of things to come weren't all that accurate, but his expectation that technological innovation would change the way we live was useful.

Jane Austen is well worth rereading. P G Wodehouse isn't - anything that Wodehouse could do, Terry Pratchett can do better. Terry Pratchett is a lot better educated than Wodehouse was, and it shows.

Read old copies of Astounding Science Fiction - around 1950 - for really bizarre ideas about computers.

-- Bill Sloman, Nijmegen

I think the Jeremy Brett series the best, especially those with Edward Hardwicke as Watson, although the portrayal of Mycroft is a bit over-the-top for my taste.

core - EPCOS

to 630nH per

squared.

steradian

factor

at

I have "IBM's Early Computers" and some stuff like that. The current concept of a base-2 synchronous-state-machine Von Neumann architecture seems so obvious now, but it sure wasn't in the early days. At least Turing appreciated that binary-decimal conversion could be done in software.

Maybe we should have stuck with the Harvard architecture. Microsoft has never learned the difference between code and data.

It would be interesting to consider what sort of computer could be built with early 1950s-sort of technology. Say, a few thousand dual triodes and core memory.

M14 core - EPCOS

pair to 630nH per

m,

umber of

turn squared.

d a steradian

ge factor

semister

nalysis.

ol, who

specify

as

, and

gineers at

o the

seful

ft the

on't

on

n/d...

the

id,

Most signal-processing chips seem to use it

Turing pointed out that there's no fundamental difference. I quite liked self-modifying code, and used it in my 1968 PDP-8 program.

Reliability would be an issue. Heated filaments don't last forever.Vacuum electronics can't hold a candle to solid-state devices when it comes to keeping on working.

-- Bill Sloman, Nijmegen

EPCOS

squared.

steradian

On a related note, Von Neumann berated his grad students for writing an assembler. Machine time was deemed too valuable for that purpose.

Dunno how you have a program loader for Harvard - all the Harvard machines I've seen just pointed at zero ( whatever that meant ) and ran.

Let's don't. Although the tiny JAN tubes might make that more fun.

-- Les Cargill

core - EPCOS

pair to 630nH per

number of

squared.

steradian

factor

semister

analysis.

who

specify

engineers at

the

Modern CPUs with MMUs can separate executable code from stacks and data; modern C compilers and OSs just don't bother much to use the facilities. That's why we have so many stack and buffer overflow trojans.

EPCOS

squared.

steradian

Check! And there are ways to extend the life in the region of an order of magnitude so that a tube hopper would not be sorely needed.

EPCOS

squared.

steradian

at

..then try using 6.3VAC on a semiconductor device - or B+ for that matter.. Tubes can resist B+ on any electrode, and due to current limiting on(small) supplies even the filament may resist B+.. Oh,do not forget RFI swamping out semiconductors - making them multi-legged resistors. Thresholds in tubes are about an order of magnitude better..

EPCOS

squared.

steradian

factor

at

..not to mention memory "leaks" which would be impossible with correct coding..

core - EPCOS

pair to 630nH per

number of

squared.

steradian

factor

semister

analysis.

who

specify

engineers at

the

Really? What a brilliant observation. What would s.e.d. do without you?

EPCOS

squared.

steradian

at

Tube computers were reliable enough to be useful. One trick was to run midnight "margin tests" (run test code with varying filament and B+ supplies) to weed out the weak tubes.

But, having read up a bit on early computer architectures, it would just be interesting to design a computer with just dual triodes. The HP9100 programmable calculator only had about 20 flipflops, discrete transistor potted SIP things, and a CRT display. Early PDP8's used discrete logic board, like a flipflop or a gate, with transistors and diodes. Those sorts of architectures could have been done easily with tubes. What was lacking in the 1950s was the concepts.