17 Apr
2002
17 Apr
'02
12:51 a.m.
Well ... if you're referring to MicroChip ...
Those onesies are what the stuff they sell you is good for. The third-party
stuff and freeware are both better in most cases.
How do you get their stuff to run under RT-11?
Dick
----- Original Message -----
From: "Tony Duell" <ard(a)p850ug1.demon.co.uk>
To: <classiccmp(a)classiccmp.org>
Sent: Tuesday, April 16, 2002 5:07 PM
Subject: Re: TTL computing
you
attempted
I hacked into the keypad matrix in order to use the memory in the calculator
to store the contents of 2"-wide traffic counter tapes for later processing.
Linux is supposed to be free.
Linux is free, which is one reason I run it (the fact that it's stable,
and easy to maintain is another...). But all the _free_ or low-cost
PLD tools I've seen do not run under Linux.
There's a fellow here in Denver who runs the Palasm software, a DOS
application, under DOSEMU. It doesn't work with the extended memory add-ons,
but then, neither does DOS.
no
If all you want to do is whittle, that's fine. If you want to make something
that's going to be of use, some planning is generally required. I've made
stuff ad-hoc, but normally got better results if I planned what I wanted
before simply cutting away anything that didn't look right.
BTW, I'm hot after a native-code line-by-line assembler/disassembler for the
805x. I've never seen a resident monitor that had one and I thoroughly
dislike being tethered to a PC just to assemble a few dozen lines of code.
finite.
> > Hmmm.. There's at least one IC
manufacturer who did support hobbyists
> > (by giving away samples, releasing the programming spec, etc) and then
> > discovered they did rather well when said hobbyists recomended those
> > chips for commercial applications. It was basically a form for
cheap/free
> > advertising.
> >
> Microcontroller makers seem to provide some simple means for programming
their
devices in all
cases. Those means are often quite limited, but generally
quite adequate for most purposes.
What do you mean by 'limited'? Provided I can program all the locations
of the internal 'ROM' then I'm happy. Maybe I can't do full
production-grade verification, but that doesn't worry me for making
one-off prototypes (if I was using said chips in production then I would
care about this, but equally I'd have no problem buying the right
commercial programmer).
> >
> > > > [1] I've seen many HP documents that were distributed to user
groups
> > > > under the NOMAS scheme. NOMAS = NOt MAnufacturer Supported. In other
> > > > words, use and enjoy the information, but we're not going to help
> any
> > > > more. Needless to say, such information (which included things like
ROM
> > >
source listings) is very valuable.
> > >
[...]
> That's not what I remember about HP's early calculators, though. I
> to find out what I could about the HP calculators
of the early '70's,
since we
had an
application for 'em, but all they'd tell us was that everything was
"proprietary" and they wouldn't tell us anything.
HP did claim that everything was 'proprietary' in the user manuals, sure.
And in fact some service-centre manuals contain little technical
information other than how to swap boards.
However, when the HP41 came out, HP were relatively kind to the user
groups. They sent out the 'byte table' (how user program instructions
were stored in memory) to the user group virtually as soon as the machine
was released. And a few years later when said users had figured out
enough of the bus to hook up a 'ROM emulator' (aka an MLDL box) and were
starting to crack the machine code instruction set, HP eventually
supplied the official instruction set and even the ROM sources.
We did get 'em to work, but bought the TI calculators instead.
Having used both HP and TI calculators, I would only ever consider the
former for real applications!.
What exactly were you trying to do? If it was some haardware modification
then you would have been on your own. If it was hooking something up to an
external connector (like the one on the back of a 9100, or the 9815, or
something) then there would have been an official interfacing manual. I
have what's almost an advertising flyer for the 9810/20/30 series which
describes the bus and even gives schematics of the 2 standard HP interface
modules.
> That's not how they presented themselves back in '72-73. I wrote them off
> then and have avoided them since. Since then I've used only one HP
product
Then I have to say you've missed out...
> > > > But as I've said enough times already, the computer and OS to run
said
> > > > software most certainly are not
free!
> > > >
> > > Tony, YOU've made the choice to avoid the software and hardware that
make
And the commercial ones all seem to need X. For all I'd prefer a
command-line system (so I could write a shell script round the program)
that would read in a text file at one end (containing the PLD equations,
Hardware Description Language program or whatever) and spit out a binary
file at the other end to program into the device.
Well, the most valuable part of the VHDL suite is the simulator. You can
refine your desriptions, simulations, and testbenches all you want. It sounds
like you're looking for another excuse not to tap into this very useful
technology, though.
> > Firstly, I am not critising them. I'll just choose not to use their
> > products. Now will you please stop flaming me for making that choice.
> >
> This isn't a flame, it is a caution that you may be isolating yourself
from
> the current technology, and that technology would
make things much easier
for
> you. With the programmable hardware you
wouldn't have to resolder wires
in
order to make
changes, and you wouldn't have to change connector pinouts
HAve you ever considered that I might actually prefer soldering wires to
typing in equations/programs on a computer? (I do, FWIW).
> because you changed the internals of a circuit. Maybe you should consider
> DOSEMU and WINE on some LINUX box. Then you could spend your time
> experimenting with those theoretical concepts as you want to, and generate wasted
hardware.
This sort of arguement comes up all over the place. I've seen it from the
model engineering crowd -- the people who like to design everything on a
mechanical CAD system and let a CNC tool make it .vs. those who like to
get a piece of metal from the scrap box, chuck it in the lathe and start
machining. Or the photographers who like doing digital manipulation of
their images .vs. those who only use traditional silver-based film. And
so on.
The only reasonable response is to agree that _both_ sides are enjoying
their hobby, and that's the only point of a bobby anyway.
> >
> > However, there's also a difference between 'not suporting' and
> > 'preventing others from supporting'. I don't expect Philips or
Intel to
> > sell me an 8051 cross-assembler for linux
> they won't sell it to you, but they'll give you what they have available,
and,
> with a little web-search, you can find useable
source code that can easily
be
adapted to
whatever compiler you have.
Agreed. THere's a perfectly good 8051 assembler available out there
(released under the GPL) which I use all the time.
The point is, though, there will never be similar tools for PLDs because
the manufacturers won't (for a variety of reasons, some more reasonable
than others) release the necessary information.
That doesn't seem to bother most folks much. I've heard complaints, but
it
doesn't stop most folks.
> > > > > you. If you specifically tell them they're not allowed to
build
your
> > > > > circuit, as you retain the
rights to it, but choose to share the
> > > information
> > > >
> > > > Odd, but when I publish a circuit, I expect people to want to build
it.
> I
> > > > am not doing this as some kind of 'how clever I am'
advertising!
> > > >
> >
> You just tell them that to protect yourself. If you suggest that they
build
So I tell a lie (Namely 'you may not build this circuit' when in fact I
want them to) in order to protect myself? %deity this world is screwed up!
It surely is!
> it, you could be considered liable for the solder splat on their lap, and
for
the burn on
their privates if they were doing it in their skivvies.
Thankfully I'm in the UK. I have been known to write 'Experimental
Design' on my circuits, though (a phrase which I was told implied that
the circuits do not necessarily comply with any regulations (EMC, etc)
and that it's up to the constuctor to check that).
> and it's generous of you to want to do that, but there are risks.
> >
> Using SSI/MSI parts makes for the greatest likelihood the end-user will be
> able to implement your design, but it doesn't increase the chances of
> completing the work, since the supplies of many of those old parts are
Are you sure that _none_ of the TTL (I am including all 74xxx families
here, HCT, ACT, etc) are still made? I am suprised things like the 8 bit
buffers and latches are unavailable, simply because putting that sort of
function into a PLD used up a lot of I/O pins for little benefit. I am
thinking of things like an address latch for an 8051 microcontroller or
similar.
I don't think I said that. I did, IIRC, say that not all of the old logic is
still available. I doubt that any of the "logic" is still being made, but the
datapath parts, along with muxes, etc, are still being made. There's no point
in making most of it any longer because the programmable stuff is faster,
cheaper, etc. The datapath parts are still cost effective, though.
> I have a couple of old circuits I published back in the '70's, for which I
> doubt anyone will find parts nowadays. If someone becomes accustomed to
using
Oh, agreed, some of the TTL parts are really hard to find now. But
actually, probably easier to find (or to substitute for) than some more
recent PALs/PLDs/etc which have also been discontinued.
Yes, the small PLD's, 28-pins and smaller, particularly in DIP packages, are
becoming scarce. Even PLCC parts of, say, 84 pins are become scarce.
> If, for whatever reasons, you decide you don't want to run those more or
less
> current OS' or the platforms that support
them, it's a choice you make. I
> routinely see completely functional hardware capable of running the
popular
> OS' in the surplus vendors' stores for
<$80. I bought a complete HP PC
unit
Not in the UK :-(.
And of course the purchase price of the computer is not the whole story.
There's the maintenance costs (which are likely to be significant if I
can't just pull the right spare chip out of my junk box). And there's
finding somewhere to put the darn thing!
Ah, yes ... the space problem. I'd say, if it breaks, chuck it, i.e. the
motherboard, and give the thing a heart transplant. The brain is still in the
hard disk, and that's the same. You can get the typical "last-generation"
motherboard for about $20 on eBay. It's really easy if you're not in a hurry,
so the thing to do is snag a spare of whatever's in the box you get, but
before you need it. That way, (1) the one you are using will never break, and
(2) if it does, in 20 minutes it's fixed.
> I'm sure you understand that IC vendors in business to make money by
selling
> their IC's in volume really can't afford
to spend more than 100ps per day
on
> hobbyists. With that short a time budget, they
certainly have to focus on
the
Of course. But as I said before, there's a big difference between 'not
supporting hobbyists' and 'preventing hobbyists from supporting
themselves'.
Well, every feedback I've gotten has treated hobbyists as a nuissance. It
seems they generate as many inquries for their $100 per year as the $100,000
per year customers.