At 07:40 pm 25/08/2001 +0100, you wrote:
>> >If 'internally' covers electronic desgin, then you've clearly never
>> >worked on PDP11/45s, PERQs, HP9100 calculators, or a dozen other older
>> >machines I could name!
>>
>> Nope, afraid not. You could say I missed the "golden age" of computing
>
>I didn't work on them when they were current. My first computer was an
>MK14, in about 1977-1978 (I still have it, but the regulator shorted and
>took out most of the expensive chips so it's currently dead...). My
>second computer, about a year later, was a TRS-80 Model 1.
Well, they were pretty current then, weren't they? The follow-up to the
Mk14 didn't arrive until 1980, and although I'm not sure about the TRS-80
1, I don't suppose it was *that* old at the time.
Incidentally, if I ever get good at this stuff, I'll probably take a ROM
dump of my TRS-80 Model 102, and figure out how to update it into the 21st
century (i.e. replace a 19 with a 20), and give it a 100-year new lease of
life :)
>> (although I'd say I didn't); having got into it in the early '80s (about
>> '81 I think). Bear in mind I was at primary school at the time, so I have a
>> good excuse.... :)
>
>I think what happened to me was that in about 1986 I managed to buy an
>old P850 minicomputer for \pounds 25.00 (it was sold as scrap). The front
>panel (toggle switches and filament lamps) is beautiful, so I deceided to
>have a go at repairing it. I learnt about how processors worked from that
>machine (even though it's very unconventional!), and decided to grab any
>other machines that I could and learn about those too.
More or less what's happening with me now.... Except i don't have a
mainframe to play with, or indeed the space to keep a mainframe in....
>> However useful the Wintel box, it's (to use a technical term) dull as f**k.
>
>Which is why I don't (and won't) have one... I've got a PC, but (a) it
>runs linux and (b) the hardware is 'much hacked' :-))
Unfortunately, my day job requires Wintel, so I have several (3) here;
although one only runs DOS, one runs Linux, and this one runs W2K.
>> That sounds about right. I also wonder how many tricks have been lost &
>> rediscovered several times....not necessarily in the field of computing.
>
>Far too many... Which is why I read almost all the old books on
>engineering, radio, electronics, etc that I can. A lot of the really
>ancient stuff (radio/radar stuff from 50 years ago, for example) is still
>_very_ applicable today...
Somehow, I'm not in the least bit surprised...
I used to have a Sinclair Cambridge Programmable calculator (I've no idea
where the calculator itself is these days, but I *should* still have it
somewhere). That came with four books c/w example programs, etc. Some of
these cover extremely complex maths & electronics - so I really ought to
write an emulator or something. Fourier analysis in 35 program steps? Try
*that* in VB :)
>Heck, I noticed a very nice mechanical edge-triggered system (basically a
>pair of level trigers with mechanical hysteresis) the other day. In a
>clock dating from 1360 or so. Same trick works in electronics....
I had to go look "hysteresis" up... Still not 100% sure where/why that
would be useful...
[GPIB]
>I'm pretty sure some HP 9000-series Unix boxes have HPIB disks, BTW...
Blimey. I wonder if they'd work on a PET...?
[Maplin]
>
>Odd... I never found Maplin to have the range of either RS or Farnell
>(and this was in the mid 1980s). Their quality was lower as well (I had a
>few dead TTL parts from Maplin, never had a single dead one from RS or
>Farnell).
It could be the cost thing - AFAIR RS were always more expensive than
Maplin, so (naturally) the school had Maplin catalogues and not RS (or
Farnell). My first introduction to these was back in the days when Dad
built PCBs of various types; most of the components came from RS or Farnell.
>Yes. Maplin now have a very poor range of components. I use them for
>things like DB25 connectors, cable, and so on because they're convenient
>(there are a couple of shops near enough to me that I can easily go and
>pick up the bits I need). But for ICs, etc I go to RS now.
Pretty much the same here, although I have to go all the way out to Marble
Arch... I could go to Stratford, which is probably nearer, but invloves a
drive; not something I like to do when the shops are open (too damn busy).
>> >The things that are not easy to find are the
>> >screw-down posts with M3.5 therads used on the chassis mount connector
>> >for the jackscrews on the cable connector to screw into. The quickest way
>> >for me to get those is to make them from scratch. More normal people, I
>> >guess, ask a friendly local model engineer...
>>
>> Should I add a lathe to my list of requirements? :)
>
>I did :-)...
I nearly got one for free when I rented my shed up in Liverpool, but the
owner (who was moving into another section of the shed) wanted it back,
unfortunately (he had a good excuse - he runs a metalworking shop). We got
to keep his pillar drill, but it's hardly useful for electronics, being 8ft
tall and taking drill bits up to 1.5" diameter.... Mind you, I've also got
a CNC drilling machine, ideal for electronics. Hopefully we'll get that up
and running soon, so PCG manufacture will actually be very possible.
>Seriously, a small metalwork lathe (small meaning the size of workpiece
>it'll handle -- the machine itself is anything but...)
Don't I know it - I remember the lathes from school being mosters, and the
one that we briefly acquired at Liverpool was pretty damn big.
>is a nice thing to
>have. But it's not that useful for classic computing unless you need to
>make mechanical parts, or spacers, or things like that. I use mine a lot,
>because it's there. If I didn't have it, then I'd find another way, I
>guess...
>
>A lathe becomes useful when you have other, related, hobbies like model
>engineering.
:) I'm deliberately trying to keep my number of hobbies (especially
expensive ones) down; motor racing takes up a *lot* of space, time & money;
and computing - although smaller - also takes time & money... I'd need to
inherit a big bag of cash before I can make enough time for anything else...
[Connectors]
>> That makes sense. TBH, any kind of in-line connector would do - except
>> maybe a DIN type (for the soldering reasons given elsewhere)...
>
>I try to be somewhat sane in my use of connectors, so I'd not use a 1/4"
>jack plug/socket for a monitor power connection (for all it would handle
>the current). I'd use something commonly associated with 'power'
>connections like the Molex or AMP series I mentioned. It makes life a
>little easier when you're trying to work out what a particular connector
>is carrying...
True... I've used plug-block extensively in the past, largely because it
doesn't involve solder. But then this was mostly on car electrics, where it
doesn't really matter how big it is, so long as it's robust enough to
survive the vibrations.
[Oscilloscopes]
>> got two other offers - so one way or the other I should end up OK. Not sure
>> if it'll be a Tek though.
>
>Well, Tektronix are (IMHO) about the best, but there are plenty of other
>brands that will be OK for the moment...
I bow to your superior knowledge: If the old man comes through, it'll
probably be one of the HP's I used way back when; I'm not sure of the brand
of the other two.
>FWIW, Stewarts of Reading are advertising various Tektronix 400-series in
>the latest Elektor magazine. And a few cheaper 'scopes that might do
>(there's a Gould for under \pounds 100.00 which looks possible!).
Hopefully, my local newsagent will be procuring a regular copy of Elektor
soon. I seem to remember avoiding Elektor as a young lad getting interested
in electronics, mainly because it seemed too complex. But then, I did have
learning difficulties[1]...
[1] I was too lazy to learn anything :)
[PIC]
>
>If you want to make a processor from TTL chips, or from discrete
>transistors, or even from valves, well, that's a reasonable thing to want
>to do. It's going to be slower than many microprocessors you can buy now
>[1], but that's probably not the point. You can't fail to learn a lot
>about procesor design by making one.
:) Can you *really* see me soldering hundreds of transistors onto a PCB? I
guess a valve processor would be useful in the winter months - it'd keep
the place nice and warm :)
Then again, as a learning project, I expect it would be extremely useful.
ISTR it being possible to build a latch from two transistors, or is it three?
>If you want to build a traditional microprocessor system (say a Z80 +
>EPROM + RAM + a few I/O chips), well, that's fine too. It's a different
>type of project, but it's equally interesting.
That was where I was thinking of starting. I'm not sure what I'll build to
begin with, whethere it'll be something useful, or just something to play
with.
>If you want to make an interface to link your PC to your pocket
>calculator (as I often do...) and it turns out that you need a little
>more than you can easily do with TTL, then you use a microcontroller
>(such as a PIC). Again, it's a different type of project, but that
>doesn't make it 'cheating' or 'invalid' or 'a waste of time' or anything
>like that.
Fair enough - it was just that looking through the recent "computerised"
projects in EPE, they all seem to use one of the PIC chips. Another problem
seems to be that they're evolving so rapidly, that models keep going out of
production (which leaves the hobbyist with something of a problem if (s)he
can't re-design the system based on another chip). OTOH, Z80s have been in
production for at least 21 years, and are still available now...
>If you want to learn about the internals of processors, or how to link
>things to an address/data bus type of structure then a PIC is probably
>not what you should be looking at.
Indeed.
>> Eek - self modifying code? Dangerous....
>
>You mean you _don't_ write self-modifying code? Odd....
It was drilled into me from a very early age that self-modifying code was A
Bad Thing.
Besides, it's impossible to do in Visual Basic - which is (currently) my
mainstay.
>You might find a PIC comes in handy on a more conventional microprocessor
>system, e.g. for encoding a keyboard or similar. So don't ignore them
>completely...
Good idea; I'd not got around to thinking about polling keyboards, etc.
There's also a lump of stuff to do with driving LCD matrix displays in EPE
(I don't actually *have* any copies of EPE - not recent ones, anyway, but
their website gives a description about what's been in them). It struck me
that using cheap LCD matrix displays would be a lot easier than creating a
CRT driver, and also means that project "X" could be put inside of one box
(maybe two, with a separate keyboard).
Cheers!
Ade.
--
B-Racing: B where it's at :-)
http://www.b-racing.co.uk
Received on Sun Aug 26 2001 - 07:15:35 BST