21 Apr
2012
21 Apr
'12
12:16 p.m.
My expeirence is that more compelx ICs are noticably
less reliable than
simple ones.
That's much too simple analysis even if it were correct. Failure rates in
ICs are primarily down to processing, packaging and handling (and to a
lesser extent die area), not complexity.
I would think the 555 had a much better lifetime than any
microcontrolelr
therefore.
You would be wrong however. Today's ICs have a much better lifetime.
Statistically today's commercial grade ICs have a significantly better
lifetime than military grade ICs of the 555 era.
The origianl 555 (as opposed to the 7555, etc) is large-ish junciton
bipolar and thus less likely to suffer static damage than a MOS
microcotnroller. It also has a much wider supply votaeg range and is
less
likely to be damaged by supply problems.
That's not a definition of reliability - it's one of robustness. Within
their specifications the 555 would be less reliable (IMHO of course).
And of coruse if a microcontroller fails, unless you have the program
to
put into it, you're stuck. If a 555 fails, well, there's a tube of 50
in
my spares box. And I cna go to any number of suppliers anf buy one.
Agreed, but off-topic ;o)
-tony
Bob
21 Apr
21 Apr
1:35 p.m.
My expeirence is that more compelx ICs are
noticably less reliable than
simple ones.
That's much too simple analysis even if it were correct. Failure rates in
ICs are primarily down to processing, packaging and handling (and to a
lesser extent die area), not complexity.
I would think the 555 had a much better lifetime than any
microcontrolelr
therefore.
You would be wrong however. Today's ICs have a much better lifetime.
Statistically today's commercial grade ICs have a significantly better
lifetime than military grade ICs of the 555 era.
The origianl 555 (as opposed to the 7555, etc) is large-ish junciton
bipolar and thus less likely to suffer static damage than a MOS
microcotnroller. It also has a much wider supply votaeg range and is
less
likely to be damaged by supply problems.
That's not a definition of reliability - it's one of robustness. Within
their specifications the 555 would be less reliable (IMHO of course).
And of coruse if a microcontroller fails, unless you have the program
to
put into it, you're stuck. If a 555 fails, well, there's a tube of 50
in
my spares box. And I cna go to any number of suppliers anf buy one.
Agreed, but off-topic ;o) 
6:34 p.m.
On Apr 21, 2012, at 1:35 PM, Tony Duell wrote:
I hate to tell you this, but my comments are based on actual, practical,
observations. Not 'what the theory says'.
OK. How many 1,000,000's of parts have you looked at? What long term statistical
analysis have you done?
Extrapolating long tail effects from single samples does not yield accurate results.
Not based on the ICs I have to replace I'm not?
Small sample size.
Hmm... I am tlaking abvout the reliabiltiy of the whole system. In other
words, how often do I ahve to fix it? What has failed? It may be that the
IC has failed because it was essentially misused (supply glitches,
whatever), but a device more susceptable to those is likely to lead to
prodcut that fails more often.
All things being equal complexity will *tend* to result in a less reliable system.
However, holding all things constant is usually not possible. So making general
statements, is a bit of an overgeneralization.
I would argue that todays computer systems are *much* more reliable than
the systems that were in general use 10-20 (or more) years ago.
TTFN - Guy
My expeirence is that more compelx ICs are
noticably less reliable than
simple ones.
That's much too simple analysis even if it were correct. Failure rates in
ICs are primarily down to processing, packaging and handling (and to a
lesser extent die area), not complexity.
In other words, I'ev; had many thousands of things across the bench by
now. And I can assure you that complex ICs, modern ones, old ones,
whatever, are mcuh more likely to be dead than simpler ones.
Still a small sample set.
I would think the 555 had a much better lifetime than any
microcontrolelr
therefore.
You would be wrong however. Today's ICs have a much better lifetime.
Statistically today's commercial grade ICs have a significantly better
lifetime than military grade ICs of the 555 era.
The origianl 555 (as opposed to the 7555, etc) is large-ish junciton
bipolar and thus less likely to suffer static damage than a MOS
microcotnroller. It also has a much wider supply votaeg range and is
less
likely to be damaged by supply problems.
That's not a definition of reliability - it's one of robustness. Within
their specifications the 555 would be less reliable (IMHO of course). 
7:08 p.m.
Extrapolating long tail effects from single samples
does not yield
accurate results.
Neither does assuming that theory matches reality.
I would argue that todays computer systems are *much*
more reliable
than the systems that were in general use 10-20 (or more) years ago.
On what grounds?
It contradicts tony's experience. It contradicts mine as well. Do you
have experience pointing the other way? (I would tend to assume not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
If you want to argue that there are grounds for thinking they ought to
be more reliable, that's a relatively defensible stance. But reality
trumps theory, and so far it's looking as though experience is 2-0 in
the direction of your theory - or any theory that predicts higher
reliability out of today's machines - being an inaccurate description
of reality.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
7:34 p.m.
On Apr 21, 2012, at 7:08 PM, Mouse wrote:
Except that your reality is a *small* sample set. When you look at data
for 1,000,000's of components you'll see different results. Until you look at
truly large data sets and do the appropriate statistical analysis all you're
really talking about is anecdotal evidence.
How 'bout MTBF?
Extrapolating
long tail effects from single samples does not yield
accurate results.
Neither does assuming that theory matches reality. I would argue that todays computer systems are
*much* more reliable
than the systems that were in general use 10-20 (or more) years ago.
On what grounds?
It contradicts tony's experience. It contradicts mine as well. Do you
have experience pointing the other way? (I would tend to assume not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
Unfortunately I can't really point to it since it's proprietary to the companies
that I've work(ed) for. However, there are a number of papers available (I'll
have to dig up citations) on this topic.
If you want to argue that there are grounds for thinking they ought to
be more reliable, that's a relatively defensible stance. But reality
trumps theory, and so far it's looking as though experience is 2-0 in
the direction of your theory - or any theory that predicts higher
reliability out of today's machines - being an inaccurate description
of reality.
What I'm saying is that until you're looking at truly large populations
you can't really derive any significant fact based conclusions. You're
talking about anecdotal evidence and opinion.
TTFN - Guy
8:13 p.m.
On Apr 21, 2012, at 10:34 PM, Guy Sotomayor wrote:
How 'bout MTBF?
I think you're arguing about different things. Today's PCs are largely
junk and generally less reliable than those of years ago, but I'd say
that's more due to poor engineering on the part of the designers of the
actual systems. They take shortcuts, don't do proper thermal
engineering (which is a lot more important in today's PCs than it was
20 years ago), and generally do a poor enough job that you can't
expect an assembled system to last more than about 5 years with any
degree of confidence. BGAs, in particular, do a great job of pulling
their pins off the board when subjected to thermal cycles; if you don't
design your thermal relief well enough, it's one of the best ways to
hasten your system's doom.
Note that it's not generally the ICs failing in these systems, at least
not on their own. It's usually their containers that cause the various
failures.
Today's high-end systems, though, are much more robust and designed
with durability in mind (well, a lot of them, anyway). I'd say you're
about as likely to find an IBM Z-series mainframe alive in 30 years
as you are to find a VAX alive (not dead due to natural causes, in any
case) today.
I will cop to MOS being somewhat less resilient to actual abuse than
bipolar logic, and things with tinier gates do exacerbate that; beyond
that, I seem to recall that smaller geometries are a bit more
susceptible to electromigration, and that tends to get worse with
higher temperatures. Of course, the actual MTBF data doesn't always
bear that out, so I could have been told wrong.
- Dave
(amusing side note about MTBF: we had a customer who developed a
really awful board, but claimed it had an MTBF of over 100 years; it
turns out the way they "calculated" this was by adding the MTBFs of
the individual components)
I would argue that todays computer systems are
*much* more reliable
than the systems that were in general use 10-20 (or more) years ago.
On what grounds?
9:32 p.m.
On Apr 21, 2012, at 8:13 PM, David Riley wrote:
On Apr 21, 2012, at 10:34 PM, Guy Sotomayor wrote:
How 'bout MTBF?
I think you're arguing about different things. Today's PCs are largely
junk and generally less reliable than those of years ago, but I'd say
that's more due to poor engineering on the part of the designers of the
actual systems. They take shortcuts, don't do proper thermal
engineering (which is a lot more important in today's PCs than it was
20 years ago), and generally do a poor enough job that you can't
expect an assembled system to last more than about 5 years with any
degree of confidence. BGAs, in particular, do a great job of pulling
their pins off the board when subjected to thermal cycles; if you don't
design your thermal relief well enough, it's one of the best ways to
hasten your system's doom.
Yes, proper engineering is hard for some of this stuff. It's much harder due
to several facts including the "race to the bottom" in most of the PC space
and that tolerances are tighter which means the margins are less.
Thermals are *much* more important now because a) the devices themselves
are a bit more sensitive but mainly because b) most "stuff" is jammed into
smaller and smaller enclosures and c) cooling is more difficult because of
"point" sources which leads back to b) (i.e. ever smaller footprints leads to
ever smaller enclosures (wash, rinse, repeat ad nauseam).
I would argue that todays computer systems are
*much* more reliable
than the systems that were in general use 10-20 (or more) years ago.
On what grounds?
Note that it's not generally the ICs failing in these systems, at least
not on their own. It's usually their containers that cause the various
failures.
Yes, and there's always the great example of TI SSI & MSI packages
disintegrating after several years. So it's not new.
Today's high-end systems, though, are much more robust and designed
with durability in mind (well, a lot of them, anyway). I'd say you're
about as likely to find an IBM Z-series mainframe alive in 30 years
as you are to find a VAX alive (not dead due to natural causes, in any
case) today.
I was very careful to talk about the "system" and that includes all of the I/O
attached. I still argue that today's systems are much more reliable if only from
the standpoint of the I/O (specifically disks).
I will cop to MOS being somewhat less resilient to actual abuse than
bipolar logic, and things with tinier gates do exacerbate that; beyond
that, I seem to recall that smaller geometries are a bit more
susceptible to electromigration, and that tends to get worse with
higher temperatures. Of course, the actual MTBF data doesn't always
bear that out, so I could have been told wrong.
All which goes back to my point above which means that proper engineering
is required. It's no longer possible with today's systems to just "wire it
up".
Simulation (logic, thermal, electrical) is necessary to produce a working and
reliable system.
And in case anyone gets the idea that I'm against simplicity, I'm not. I feel
that there needs to be a balance. Proper design and engineering is required
for any solution.
And to finish off my part in this, classic computers and such are for me a
hobby. I do high tech, leading edge stuff as a job and in that role something
like the 555 doesn't really have a place (at least in the stuff that I'm doing).
While we may have some things that could be done by a 555-like device, there's
so much other stuff going on that it's better (from every dimension) to incorporate
the functionality into micro's or ASICs (FPGAs and CPLDs are only used for
prototypes?they're too costly when you're building x*1,000,000's/quarter).
- Dave
(amusing side note about MTBF: we had a customer who developed a
really awful board, but claimed it had an MTBF of over 100 years; it
turns out the way they "calculated" this was by adding the MTBFs of
the individual components)
Idiots! ;-)
TTFN - Guy
22 Apr
22 Apr
11:17 a.m.
On Apr 22, 2012, at 12:32 AM, Guy Sotomayor wrote:
And to finish off my part in this, classic computers
and such are for me a
hobby. I do high tech, leading edge stuff as a job and in that role something
like the 555 doesn't really have a place (at least in the stuff that I'm doing).
While we may have some things that could be done by a 555-like device, there's
so much other stuff going on that it's better (from every dimension) to incorporate
the functionality into micro's or ASICs (FPGAs and CPLDs are only used for
prototypes?they're too costly when you're building x*1,000,000's/quarter).
That is almost true in my world as well. Still, sometimes they're actually
the best solution; we had to provide a ~600KHz clock to a bunch of switching
power supply modules on one board to synchronize them (so they'd switch on
alternate phases, etc). A 555 (well, a modern CMOS version) was the true
obvious choice there, because it was a pretty foolproof part with few
externals in the right frequency range.
Project management at the customer (this is the same project that made us
put the 800 MHz PowerPC on board when a 50 MHz ARM would have done) was
furious at us for not dividing down the 100 MHz system clock on one of the
on-board FPGAs instead. Of course, this was a second spin, designed to fix
the nasty problem of the FPGAs not starting up reliably because the power
supplies were getting thrown off because they weren't in sync. :-)
I'm glad to say it was actually one of my co-workers who pushed for the 555.
Of course, he had background as a technician and would much rather have
something he could adjust himself in the lab without going through the
exercise of finding an FPGA compiler, finding the source code, modifying it,
waiting an hour for the damn thing to compile, then waiting the 15 minutes
(no, really) it took to program the flash.
Project management was less pissed off once they forgot about it a week
later, and it's never given us problems since (and they made thousands of
the board).
Long story short, it's all a matter of choosing the right part. Sometimes
a 555 *is* the right part, and sometimes it's not. For a power supply
frequency reference, I'll choose the analog-ish part over an FPGA/micro
any day, unless there's an astoundingly good reason to go with the other.
- Dave
12:47 p.m.
Long story short, it's all a matter of choosing
the right part. Sometime=
s
a 555 *is* the right part, and sometimes it's not. For a power supply
frequency reference, I'll choose the analog-ish part over an FPGA/micro
any day, unless there's an astoundingly good reason to go with the other.
_Precisely_
-tony
21 Apr
21 Apr
9:57 p.m.
Unfortunately I can't really point to it since
it's proprietary to the
companies
that I've work(ed) for.
Such data is highly guarded by companies, often because it can be
embarrassing. It certainly was when I did QA stuff at USR (you know, doing
real life manufacturing engineering, instead of sitting in my parent's
basement).
--
Will
22 Apr
22 Apr
12:10 p.m.
Such data is highly guarded by companies, often because it can be
embarrassing. It certainly was when I did QA stuff at USR (you know, doing
real life manufacturing engineering, instead of sitting in my parent's
basement).
Hmm.. I have a USR modem here. It's been fine for many years _after_ I
resoldered a lot of dry joints shortly after I got it (on the PCB and on
the secodnary of the transformer in the mains adapter). Oh, and
re-crimped the telephone plug.
So to be honest, I don't think much of USR prodcution or quality control.
-tony
2:02 p.m.
Hmm.. I have a USR modem here. It's been fine for
many years _after_ I
resoldered a lot of dry joints shortly after I got it (on the PCB and on
the secodnary of the transformer in the mains adapter). Oh, and
re-crimped the telephone plug.
So to be honest, I don't think much of USR prodcution or quality control.
I've sold **hundreds** of these modems here in Brazil and never had
returns like that you said :oP From the Sportster 28800 to the Courier
V-Everything-but-the-girl, and the older couriers (I had a 9600/14400 HST)
7:09 p.m.
I've sold **hundreds** of these modems here in Brazil and never had
returns
like that you said :oP
I was on the industrial side of USR (Total Control rack mount stuff). Very
high quality, and very reliable. The problem was that due to various
production problems, build quality was low, and yeild was terrible. There
were huge piles of scrap and rework, with so much stuff getting rejected at
final test, and tons of orders that could not be filled due to a lack of
passing units. I was working elsewhere in the company (System Integration -
essentially making complete dialup racks for both AOL and AT&T, but then
they found I was a good quality auditer. They (Big Lou, just under Casey)
gave me the power to squawk about anything that was out of spec. Fun times,
and I am sure I pissed quite a few people off. Various rules were set up,
so if I took a number of samples of a lot, and a flaw was found, the
*whole* lot was failed and sent back, wherever in the line it was pulled,
and reworked...and sometimes heads rolled.
I never did anything for the consumer side of USR, however.
--
Will
7:32 p.m.
On Apr 22, 2012, at 10:09 PM, William Donzelli wrote:
I never did anything for the consumer side of USR,
however.
I suppose that's where the issue is. USR actually made really good
consumer modems for a while, but around the time of 56k modems, I
seem to remember the quality really going down the tubes for the
consumer gear.
At least they weren't the only ones going down the tubes in terms
of quality then, because as I recall that's when the Internet was
really becoming mainstream; I imagine the requirements of producing
ever-larger quantities of stuff for the unwashed masses catches up
to everyone in the end. :-)
- Dave
10:03 p.m.
On 04/22/2012 10:09 PM, William Donzelli wrote:
We bought quite a few Total Control racks at Digex, a few hundred of
them. They were very nice boxes. I still have two of them; I nearly
broke my toe on one just this afternoon.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
I've
sold **hundreds** of these modems here in Brazil and never had
returns like that
you said :oP
I was on the industrial side of USR (Total Control rack mount stuff). Very
high quality, and very reliable. The problem was that due to various
production problems, build quality was low, and yeild was terrible. There
were huge piles of scrap and rework, with so much stuff getting rejected at
final test, and tons of orders that could not be filled due to a lack of
passing units. I was working elsewhere in the company (System Integration -
essentially making complete dialup racks for both AOL and AT&T, but then
they found I was a good quality auditer. They (Big Lou, just under Casey)
gave me the power to squawk about anything that was out of spec. Fun times,
and I am sure I pissed quite a few people off. Various rules were set up,
so if I took a number of samples of a lot, and a flaw was found, the
*whole* lot was failed and sent back, wherever in the line it was pulled,
and reworked...and sometimes heads rolled. 
23 Apr
23 Apr
7:51 a.m.
On 4/23/2012 1:03 AM, Dave McGuire wrote:
On 04/22/2012 10:09 PM, William Donzelli wrote:
>
> I was on the industrial side of USR (Total Control rack mount stuff). Very
> high quality, and very reliable. The problem was that due to various
> production problems, build quality was low, and yeild was terrible.
We bought quite a few Total Control racks at Digex,
a few hundred of
them. They were very nice boxes. I still have two of them; I nearly
broke my toe on one just this afternoon.
-Dave
Yes, I remember selling and supporting those total control racks. We
never sold many of them, and they were fairly high priced. There was a
serial connection for management in the back of them. Pretty
straight-forward to manage, IIRC.
USR was *the* modem company for a long time. I loved their products.
I still have a Courier dual-standard that supported both 14.4k AND
16.8k. Great modem. It's like a foot+ deep, flat, but I loved the
printed hayes command set reference printed on the bottom.
I think I still have a kick-ass USR modem card too. Maybe PCI, or ISA.
And I used a sportster for a long time. I will say that towards the end
of the modem-era, it seemed like power surges on the telephone line used
to fry the relay regularly. It never happened for years and years, and
then it happened perhaps twice in two years. I eventually ran it
through a UPS for surge protection, and that prevented future problems.
Keith
8:29 a.m.
On 04/23/2012 09:51 AM, Keith Monahan wrote:
USR was *the* modem company for a long time. I loved
their products.
I still have a Courier dual-standard that supported both 14.4k AND 16.8k.
Great modem. It's like a foot+ deep, flat, but I loved the printed hayes
command set reference printed on the bottom.
I liked the internal and older external USR ones - the cases on the later
externals made it look like they'd been given away inside boxes of
breakfast cereal, though.
Mind you, I like devices in big heavy square-ish boxes on the desktop -
small items just fall over easily, or get pulled off the desk by their own
cables, or still occupy space but with the downside that I can't stack
other items on top of them. The majority seem to think that miniaturization
is the holy grail though (and I suppose the smaller the case, the less
material used, so it appeals to the manufacturer too)
cheers
Jules
8:47 a.m.
On Apr 23, 2012, at 11:29 AM, Jules Richardson wrote:
Mind you, I like devices in big heavy square-ish boxes
on the desktop - small items just fall over easily, or get pulled off the desk by their
own cables, or still occupy space but with the downside that I can't stack other items
on top of them. The majority seem to think that miniaturization is the holy grail though
(and I suppose the smaller the case, the less material used, so it appeals to the
manufacturer too)
I don't mind miniaturization, as long as it's done intelligently. I
like my Netgear 8-port switch that's on my desktop an awful lot;
it's the size of about two packs of cards, and it has a nice, solid-
feeling metal case. It's also totally square, so I can stack other
things on top of it. The only irritant is that the power inlet from
the wall brick and the ports are on opposite sides, which would be
great if it were a rackmount switch, but not so great for a desktop
one.
On the other hand, I remember my 56k modem from the end of the dial-
up era; I think it was a Global Village one? In any case, it was
nice and small, but (as you said) so light that the cables tended to
pull it off the desk, and they made this "cute-looking" little bulge
at the top for the speaker, which meant you couldn't stack anything
atop it. Infuriating. I think I ended up just stashing it under
the desk.
- Dave
2:35 p.m.
On 04/23/2012 10:47 AM, David Riley wrote:
On Apr 23, 2012, at 11:29 AM, Jules Richardson wrote:
The older netgear stuff isn't too bad - I do remember it being boxy (and
possibly the cases being metal). Still a bit of the small side for my
'tastes' though; I think I find that if things have to be on the desktop at
all, they should have the footprint of a piece of paper, so that I can
stack real pieces of paper on top.
There's a Netgear router close to to where I'm sitting; rounded edges to
the case, vents on the top, and so lightweight that if anyone so much as
sneezes near it, it'll fall off the desk under the weight of the cables
connected to it. I should probably screw it to the underside of the desk or
something, but it's handy being able to get at the ports easily.
Mind you, I like devices in big heavy square-ish
boxes on the desktop
- small items just fall over easily, or get pulled off the desk by
their own cables, or still occupy space but with the downside that I
can't stack other items on top of them. The majority seem to think
that miniaturization is the holy grail though (and I suppose the
smaller the case, the less material used, so it appeals to the
manufacturer too)
I don't mind miniaturization, as long as it's done intelligently. I
like my Netgear 8-port switch that's on my desktop an awful lot; it's
the size of about two packs of cards, and it has a nice, solid- feeling
metal case. It's also totally square, so I can stack other things on
top of it. The only irritant is that the power inlet from the
wall
brick and the ports are on opposite sides, which would be great if it
were a rackmount switch, but not so great for a desktop one.
I don't think the Raspberry Pi's been given much of an airing here
(http://www.raspberrypi.org) but I've mentioned elsewhere that it's one of
the problems I see with that particular gadget: the connectors line all
four sides of the PCB, so you either live with a miniature system with
wires going off in all directions (ick!), or you put in a larger case with
trailing wires to ports on the rear (and/or front) - which of course pushes
up the cost of a final 'product' quite a bit.
On the other hand, I remember my 56k modem from the
end of the dial- up
era; I think it was a Global Village one? In any case, it was nice and
small, but (as you said) so light that the cables tended to pull it off
the desk, and they made this "cute-looking" little bulge at the top for
the speaker, which meant you couldn't stack anything atop it.
Infuriating. I think I ended up just stashing it under the desk.
I still much prefer big old CRTs to modern LCD displays, and it's rare that
there's not a stack of stuff parked on the top edge of the case. :-)
cheers
Jules
24 Apr
24 Apr
12:14 p.m.
I still much prefer big old CRTs to modern LCD
displays, and it's rare that
there's not a stack of stuff parked on the top edge of the case. :-)
One of the gret things abotu the HP9836 machines was that the monitor has
a large flat top and that there were no cooling vats on it. And being an
HP it was solid.
You could happly put a dot matrix printer or pen plotter on top of that
monitor when the system was in use.
-tony
23 Apr
23 Apr
10:06 a.m.
On 23 Apr 2012 at 10:51, Keith Monahan wrote:
USR was *the* modem company for a long time. I loved
their products.
Well, that's a distinction that was passed around for years. The
distinction used to belong to AT&T/WE modems, then Anderson-Jacobson,
Milgo, Racal-Vadic. By the time you hit the late 70's, there were
various modems-du-jour. Telebit, Motorola, Hayes... I used a Racal-
Vadic 3451 triple modem for quite some time. But modems of that time
were extremely dumb--so they tended to work forever if designed
correctly.
I think I have a floppy drive box somewhere that's powered from a
Racal power supply.
I still have a few of my BBS modems, all external and mostly USR--a
Courier HST, a Dual-Standard, a V.Everything... I think it's
remarkable that USR would sell you an upgrade kit (usually a PROM and
perhaps an SRAM chip) rather than demand that you buy a completely
new unit. Not something that would happen today.
I have some consumer-level modems that were used less than a week,
including the Zoom, Hayes "business modem", etc. I got rid of my
Sportster 56K modems within a couple of months of having purchased
them. Not very good, IMOHO.
--Chuck
10:33 a.m.
On 4/23/2012 1:06 PM, Chuck Guzis wrote:
On 23 Apr 2012 at 10:51, Keith Monahan wrote:
I forgot about the LCD-configured Motorola modems. They were built like
brick shithouses and while very expensive -- they were also reliable and
perfect for "dumb" devices that could not dial. They could do DTR-based
dialing etc. They were heavy and took up a lot of space, but I liked them.
I seem to remember the original Hayes modems being pretty good products
too.
Keith
USR was *the* modem company for a long time. I
loved their products.
Well, that's a distinction that was passed around for years. The
distinction used to belong to AT&T/WE modems, then Anderson-Jacobson,
Milgo, Racal-Vadic. By the time you hit the late 70's, there were
various modems-du-jour. Telebit, Motorola, Hayes... I used a Racal-
Vadic 3451 triple modem for quite some time. But modems of that time
were extremely dumb--so they tended to work forever if designed
correctly.
11:09 a.m.
New subject: Modems; was: Arduino vs 555 reliability
On 23 Apr 2012 at 13:33, Keith Monahan wrote:
I forgot about the LCD-configured Motorola modems.
They were built
like brick shithouses and while very expensive -- they were also
reliable and perfect for "dumb" devices that could not dial. They
could do DTR-based dialing etc. They were heavy and took up a lot of
space, but I liked them.
Wasn't Motorola the old Codex? The old leased-line modems were
marvels--ISTR that the Milgo version from the late 60s took up about
a 6U rack slot--for a single 2400 bps channel. It had all sorts of
equalization settings. Codex was a big competitor--and I think there
were some legal issues between the two companies.
Does anyone still have one of these beasts? I can't imagine why
anyone would hang onto one, but you never know.
--Chuck
P.S. And we can't forget Gandalf modems, either.
12:02 p.m.
New subject: Modems; was: Arduino vs 555 reliability
On 2012 Apr 23, at 11:09 AM, Chuck Guzis wrote:
Wasn't Motorola the old Codex? The old
leased-line modems were
marvels--ISTR that the Milgo version from the late 60s took up about
a 6U rack slot--for a single 2400 bps channel. It had all sorts of
equalization settings. Codex was a big competitor--and I think there
were some legal issues between the two companies.
Does anyone still have one of these beasts? I can't imagine why
anyone would hang onto one, but you never know.
I have a Bunker-Ramo modem from the 60's. Pic with two of the circuit
boards removed for viewing:
http://www3.telus.net/~bhilpert/tmp/BunkerRamo.jpg
It would be about 3U in a rack.
I think it's only 300 bps, but I should make some measurements.
Why do I keep it? Good question. Because no one else does? I'd rather
it be in a tech museum than my house.
1:26 p.m.
New subject: Modems; was: Arduino vs 555 reliability
Wasn't Motorola the old Codex? The old
leased-line modems were
marvels--ISTR that the Milgo version from the late 60s took up about
a 6U rack slot--for a single 2400 bps channel. It had all sorts of
Somewhere I have a slightly more recent 22400 baud leasd-line modem. It's
only 3U high. I rememebr one of the PCBs contains some bitslice ICs and a
parallele multiplier chip. I must investigate it...
P.S. And we can't forget Gandalf modems, either.
I rmeemebr -- heck, I still have -- the Gandalf LDS line drivers.
-tony
11:22 a.m.
On 4/23/12 12:06 PM, Chuck Guzis wrote:
I have some consumer-level modems that were used less
than a week,
including the Zoom, Hayes "business modem", etc. I got rid of my
Sportster 56K modems within a couple of months of having purchased
them. Not very good, IMOHO.
Weren't the Sportster 56K modems all WinModems?
(Now THERE'S a term that still fills me with rage, all these years later)
Doc
11:35 a.m.
On 4/23/2012 2:22 PM, Doc Shipley wrote:
On 4/23/12 12:06 PM, Chuck Guzis wrote:
Definitely not.
The WinModems were primarily off-brand modems, or they were cards
integrated into laptops or pci/isa cards.
I never actually saw a WinModem firsthand.
I used a Sportster 56k modem with my Amiga. There was no driver support. :)
Keith
I have some consumer-level modems that were used
less than a week,
including the Zoom, Hayes "business modem", etc. I got rid of my
Sportster 56K modems within a couple of months of having purchased
them. Not very good, IMOHO.
Weren't the Sportster 56K modems all WinModems?
(Now THERE'S a term that still fills me with rage, all these years later)
Doc
11:42 a.m.
The WinModems were primarily off-brand modems, or they
were cards
integrated into laptops or pci/isa cards.
No, there was a USR winmodem
http://www.56k.com/reports/winmodem.shtml
11:59 a.m.
On 4/23/2012 2:42 PM, Alexandre Souza - Listas wrote:
But not all 56k Sportsters were winmodems, which was the point.
Keith
The WinModems
were primarily off-brand modems, or they were cards
integrated into laptops or pci/isa cards.
No, there was a USR winmodem
http://www.56k.com/reports/winmodem.shtml
11:36 a.m.
Weren't the Sportster 56K modems all WinModems?
No, most of them (at least here in Brazil) were DSP ones. Some with
texas DSPs, some with proprietary (maybe the same texas with different
markings) numbers
(Now THERE'S a term that still fills me with
rage, all these years
later)
You're not the only one...
11:51 a.m.
On Mon, 23 Apr 2012, Keith Monahan wrote:
USR was *the* modem company for a long time. I loved
their products.
I still have a Courier dual-standard that supported both 14.4k AND
. . .
I think I still have a kick-ass USR modem card too. Maybe PCI, or ISA.
. . .
And I used a sportster for a long time. I will say that towards the end
It often seemed hard to believe that Couriers and Sportsters came from the
same company.
Couriers EARNED a reputation for being one of the best and most reliable,
and replaced Hayes.
Sportsters, . . . were often hard to give away.
Was the logo of USR the same as the logo for the USR in the Will Smith
"I, Robot" movie? (were they the same company?)
12:08 p.m.
It often seemed hard to believe that Couriers and
Sportsters came from the
same company.
Couriers EARNED a reputation for being one of the best and most reliable,
and replaced Hayes.
Sportsters, . . . were often hard to give away.
Dunno why. Sportsters were always reliable modems (here in Brazil). I
had a courier because I was filthy rich (hahahaha) and I wanted to have the
"ferrari of the modems" but it didn't perform better than my 33K sportster
How many of you used the BOCA 28800 modem? ;oD
9:01 a.m.
On Mon, 23 Apr 2012, Dave McGuire wrote:
We bought quite a few Total Control racks at Digex, a few hundred of
them. They were very nice boxes. I still have two of them; I nearly
broke my toe on one just this afternoon.
Those Total Control Net servers were awesome. I LOVED working with them
back in '97 when I was sysadmin for a local ISP. The only problem I had
was getting US Worst to get the damn PRI circuit orders right. Fun times.
*wistful sigh*
g.
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.diy-cockpits.org/coll - Go Collimated or Go Home.
Some people collect things for a hobby. Geeks collect hobbies.
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Buying desktop hardware and installing a server OS doesn't make a
server-class system any more than sitting in a puddle makes you a duck.
[Cipher in a.s.r]
9:43 a.m.
On 04/23/2012 12:01 PM, Gene Buckle wrote:
They really were fun times. We got so much done with that technology.
Frequently these days I'll find something, either hardware or software,
and think "wow, if I'd had this in the 90s, we could've taken over the
world!"
We eventually moved to Cisco AS5200s, which were also quite an experience.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
We bought
quite a few Total Control racks at Digex, a few hundred of
them. They were very nice boxes. I still have two of them; I nearly
broke my toe on one just this afternoon.
Those Total Control Net servers were awesome. I LOVED working with them
back in '97 when I was sysadmin for a local ISP. The only problem I had
was getting US Worst to get the damn PRI circuit orders right. Fun
times. *wistful sigh*
11:13 a.m.
Hmm.. I have a USR modem here. It's been
fine for many years _after_ I
resoldered a lot of dry joints shortly after I got it (on the PCB and on
the secodnary of the transformer in the mains adapter). Oh, and
re-crimped the telephone plug.
So to be honest, I don't think much of USR prodcution or quality control.
I've sold **hundreds** of these modems here in Brazil and never had
returns like that you said :oP From the Sportster 28800 to the Courier
V-Everything-but-the-girl, and the older couriers (I had a 9600/14400 HST)
2:14 p.m.
On Mon, 23 Apr 2012, Tony Duell wrote:
So, yes, its fine now. By the comments of some peopel
here, though it
shouldn't be in that reworkign the PCB should ahve reduced the
reliability. Maybe it did, if I'd got one that didn't need reworking, it
wowould go on even longer. But 20 years without problems, turned on and
used just about every day, doesn't seen unrelaible to me.
Competent "rework" fixed the other cold solder joints, etc.
Haphazard "just get it done" "rework" will only fix the one joint that
had
stopped making contact, possibly overheat adjacent pins and componenets,
add solder bridges, and in removal and reinstallation add physical
stresses to other bad joints.
So, "modern" incompetent "rework", as practiced by many companies,
could
indeed reduce reliability.
MY experience with Sportster(S) was similar, other than that I did NOT
make the needed repairs.
OTOH, I did not have any problems with their Courier(S).
We got rid of the unreliable Sportsters, and every WinModem that came
along. Deities! I hate those things!
Does anybody make a good reliable tiny portable USB modem?
The USR one that I got on eBay won't work with the supplied drivers nor
the ones from the website, but ALMOST works with some generic drivers.
In terms of "Customer Service" and "Tech Support", USR was
uninterested
and unwilling to provide any help with out-of-warranty Sportsters, but
provided free firmware update to an old Courier with unknown provenance.
Were they really the same company? (It must have been the Sportster
"division" that survived to become the one that Will Smith dealt with.)
--
Grumpy Ol' Fred cisin at xenosoft.com
3:18 p.m.
On 23 Apr 2012 at 14:14, Fred Cisin wrote:
Were they really the same company? (It must have
been the Sportster
"division" that survived to become the one that Will Smith dealt
with.)
When the V.Everything came out, I received word of it and that USR
was offering an upgrade program. I had a V.34 Courier, so I called
them and they charged my credit card for something like a $50 upgrade
fee. When I asked them if they wanted me to return the V.34, they
said, no--you keep it.
On the HST upgrades, they'd ship a new PROM and SRAM and make sure
that you sent the old ones back. If you didn't, they'd pester you to
do so. I never did figure out what use they had for old used
components. On the V.34 upgrades, you got to keep the old SRAM chip.
Nice people, but a strange company.
Anyone here remember the early USR (ca. 1977-79?) It was at a
Wescon, NCC or some other trade show that I think I remember them
showing an all-in-one computer (keyboard, display, disk drives and
printer in a large black case). If it wasn't USR, it was some outfit
with "Robotics" as part of their name.
Can anyone help out with a fading memory?
--Chuck
24 Apr
24 Apr
8:32 a.m.
On Mon, 23 Apr 2012, Fred Cisin wrote:
On a minorly related note...
http://www.simpits.org/~geneb/hayes.jpg
The static bags contain both versions of the S-100 DC Hayes 103 modem as
well as an Apple MicroCom ][.
I've been told there's one more non-modem gadget that I don't have, but I
can't recall what it is. (that was sold by Hayes in the Stack enclosure)
g.
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.diy-cockpits.org/coll - Go Collimated or Go Home.
Some people collect things for a hobby. Geeks collect hobbies.
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Buying desktop hardware and installing a server OS doesn't make a
server-class system any more than sitting in a puddle makes you a duck.
[Cipher in a.s.r]
9:01 a.m.
On Tue, Apr 24, 2012 at 11:32 AM, Gene Buckle <geneb at deltasoft.com> wrote:
On Mon, 23 Apr 2012, Fred Cisin wrote:
On a minorly related note...
http://www.simpits.org/~geneb/hayes.jpg
Ooh... a Chronograph. Very nice.
-ethan
9:59 a.m.
On Tue, 24 Apr 2012, Ethan Dicks wrote:
On Tue, Apr 24, 2012 at 11:32 AM, Gene Buckle
<geneb at deltasoft.com> wrote:
Yep. There's another variant of it that I want. The one I have has only
a single serial port on it. The original had two, one passed through so
you could connect a SmartModem to it as well.
g
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.diy-cockpits.org/coll - Go Collimated or Go Home.
Some people collect things for a hobby. Geeks collect hobbies.
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Buying desktop hardware and installing a server OS doesn't make a
server-class system any more than sitting in a puddle makes you a duck.
[Cipher in a.s.r]
On Mon, 23 Apr 2012, Fred Cisin wrote:
On a minorly related note...
http://www.simpits.org/~geneb/hayes.jpg
Ooh... a Chronograph. Very nice.
10:45 a.m.
On Tue, Apr 24, 2012 at 1:20 PM, Alexandre Souza - Listas
<pu1bzz.listas at gmail.com> wrote:
? Why?! To configure the clock?!
Presumably to not "consume" a (your only?) serial port just to set the
time... an external version of the "no slot clocks".
I've never seen that variant. Mine is the single serial port version.
-ethan
Ooh... a Chronograph. ?Very nice.
Yep. ?There's another variant of it that I want. ?The one I have has
only a single serial port on it. ?The original had two, one passed through so
you could connect a SmartModem to it as well.
10:57 a.m.
On Tue, 24 Apr 2012, Alexandre Souza - Listas wrote:
You could do that as well.
The idea being that the Chronograph would
respond to Chronograph commands and the modem to modem commands. That
would allow you to have both devices connected at the same time
in order to update a clock on the host computer, etc.
g.
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.diy-cockpits.org/coll - Go Collimated or Go Home.
Some people collect things for a hobby. Geeks collect hobbies.
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Buying desktop hardware and installing a server OS doesn't make a
server-class system any more than sitting in a puddle makes you a duck.
[Cipher in a.s.r]
Ooh... a Chronograph. Very nice.
> Yep. There's another variant of it that I want. The one I have has only
a single serial port on it. The original had two, one passed through so
you could connect a SmartModem to it as well.
9:02 a.m.
Hmmm...you remembered me I was crazy to find an old hayes or hayes-like
modem to build a hayes stack chronograph...But now I got a surprise, it is
WAY BIGGER than the other devices. I had never seen one besides its
"family". Nice stack, eh? :o)
---
Enviado do meu Motorola PT550
Meu site: http://www.tabalabs.com.br
----- Original Message -----
From: "Gene Buckle" <geneb at deltasoft.com>
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Sent: Tuesday, April 24, 2012 12:32 PM
Subject: Re: Arduino vs 555 reliability
On Mon, 23 Apr 2012, Fred Cisin wrote:
On a minorly related note...
http://www.simpits.org/~geneb/hayes.jpg
The static bags contain both versions of the S-100 DC Hayes 103 modem as
well as an Apple MicroCom ][.
I've been told there's one more non-modem gadget that I don't have, but I
can't recall what it is. (that was sold by Hayes in the Stack enclosure)
g.
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.diy-cockpits.org/coll - Go Collimated or Go Home.
Some people collect things for a hobby. Geeks collect hobbies.
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Buying desktop hardware and installing a server OS doesn't make a
server-class system any more than sitting in a puddle makes you a duck.
[Cipher in a.s.r]
11:52 a.m.
On Mon, 23 Apr 2012, Tony Duell wrote:
I've seen it claimed by some people here that any rework degrades the
rleiability of a board. While this may well be true, this modem board
which I reworkded has been turned on each day for about 20 years, used,
and then turned off again with no problems (even though turnign it on and
off may not be good for it either).
You could claim that if it had never been reworked -- that is if it had
not needed rework -- it would run ofro even longer. THat may be so. But
quite honestly, a consumer device that runs for 20 eyars is not
unreliable. I ahev no idea what the design lifetime of the unit is, but I
would be suprised if it was that long.
So I will stick to the comment that sensible reqork does not totally
degrade reliabiiltiym and that boards that have been reworked can
certainly be reliable enopugh to use.
So, yes, its fine now. By the comments of some
peopel here, though it
shouldn't be in that reworkign the PCB should ahve reduced the
reliability. Maybe it did, if I'd got one that didn't need reworking, it
wowould go on even longer. But 20 years without problems, turned on and
used just about every day, doesn't seen unrelaible to me.
Competent "rework" fixed the other cold solder joints, etc.
Haphazard "just get it done" "rework" will only fix the one joint
that had
stopped making contact, possibly overheat adjacent pins and componenets,
add solder bridges, and in removal and reinstallation add physical
stresses to other bad joints.
So, "modern" incompetent "rework", as practiced by many companies,
could
indeed reduce reliability.
Hmmm.... Having seen some of the soldring in commercial devices I do
wonder just how it was done... A red-hot poker, perhaps....
MY experience with Sportster(S) was similar, other than that I did NOT
make the needed repairs.
Well ,as their 'tech support' was abpit as useful as a gallium teaspoon,
I had to fix it myself.
OTOH, I did not have any problems with their
Courier(S).
I never had one, so I can't comment.
We got rid of the unreliable Sportsters, and every WinModem that came
along. Deities! I hate those things!
Err, yes... Wasn't the idea to use the host comnputer CPU to do some of the
DSP worK? No thanks... I can't believe the code exists for most of my
machines...
Does anybody make a good reliable tiny portable USB
modem?
Since the origianl use of RS232 was to link a terminal to a modem, that's
the one place where there should be no problems in using it. Note I said
'should'... At least it's one place wherne I will contiune to use RS232
ports to link to modems.
-tony
1:01 p.m.
[...] abpit as useful as a gallium teaspoon, [...]
That is a *delightful* comparsion!
Thank you for my afternoon's laugh.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
21 Apr
21 Apr
10:32 p.m.
Except that your reality is a *small* sample set.
Also, your reality does not figure the whole lifecycle of the part. You are
not considering the failures of parts that happened before whatever the
product is shipped off the factory floor.
Sometimes the failures are catastrophic, and other times they are latent
(ticking time bombs - these are the failures that really suck). Reworking a
board lowers the reliability of the parts on it - even innocent bystanders.
There is no getting around it. In my metal film resistor example, having
them on a board will statistically lower the reliability of surrounding
components, because some amount of those resistors will be either flawed
from the start or will be placed by the robots poorly.
They will require
rework, and will cause damage to some number of components that
are mounted
nearby. Those tiny resistors and caps can be real pains sometime, so every
manufacturing engineer is happy to see when they go away. Yield killers,
basically.
--
Will
22 Apr
22 Apr
12:14 p.m.
Sometimes the failures are catastrophic, and other times they are latent
(ticking time bombs - these are the failures that really suck). Reworking a
board lowers the reliability of the parts on it - even innocent bystanders.
This is something I have _never_ observed. I've reowkred many boards (OK,
small sample, but...). I've desoldered ICs and soldeed them back again.
And I have never noticed a larger failure rate for stuff I've repaired as
agaisnt stuff I haven/t
-tony
12:06 p.m.
On Apr 21, 2012, at 7:08 PM, Mouse wrote:
Except that your reality is a *small* sample set. When you look at data
for 1,000,000's of components you'll see different results. Until you look at
And you have _ersonally_ looked at millions of components? Sorry, I've
seen too much bogus data already...
Extrapolating long tail effects from single samples does not yield
accurate results.
Neither does assuming that theory matches reality. truly large data sets and do the appropriate
statistical analysis all you're
really talking about is anecdotal evidence.
How coome 2 of use here (at least) seem to have contradictory findings?
It contradicts
tony's experience. It contradicts mine as well. Do you
have experience pointing the other way? (I would tend to assume not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
Unfortunately I can't really point to it since it's proprietary to the companies
3:57 p.m.
On Apr 22, 2012, at 12:06 PM, Tony Duell wrote:
No, all I can do is look at the data and do analysis on it. Look, when the company
I work for is producing 20-30,000,000 "widgets" per quarter, the interval on the
line
is 3-6 per second (not in line with things like aluminum cans which are 1-3,000 per
minute) but quite high given the complexity of the devices we're talking about you
can't look at each device in detail.
Statistics (and good design practices) are the only way that high volume
manufacturing works. You look at trends in the test and reliability data. Trust me
when I tell you the attention gets pretty intense when failure rates go up by as little
as 0.01% and line shutdowns occur at 0.5% (then you don't want to be anywhere
around).
And of course then there's the warranty data. Which again is all statistically
derived
based upon life tests during the development (and spot checks during production).
Again we're talking ridiculously large sums of money at stake.
Because it's anecdotal. BTW here's the definition:
(of an account) not necessarily true or reliable, because based on
personal accounts rather than facts or research
Again, when you're dealing with large populations, looking at a few in isolation,
is I'm sorry to say just anecdotal.
Oh come now. And you expect us to believe comments based on data that's
not even available? Sorry, I'll stick with what my experience as taught
me, modifying it on the basis of more experience.
Fine. There's lots of things that I've started to say on this list to deal with
an
argument and couldn't because it would violate various NDA's that I have to
live under (I kinda like my job). Arguments on this (and other fora) while
fun aren't worth sacrificing my livelihood.
I'm sorry if you can't/won't understand how large companies work and the
restrictions that are put on employees and the products (though some have
alluded to here). Because if you did, I'm guessing that you wouldn't have
made the comments you did.
TTFN - Guy
On Apr 21, 2012, at 7:08 PM, Mouse wrote:
Except that your reality is a *small* sample set. When you look at data
for 1,000,000's of components you'll see different results. Until you look at
And you have _ersonally_ looked at millions of components? Sorry, I've
seen too much bogus data already?
Extrapolating long tail effects from single samples does not yield
accurate results.
Neither does assuming that theory matches reality. truly large data sets and do the appropriate
statistical analysis all you're
really talking about is anecdotal evidence.
How coome 2 of use here (at least) seem to have contradictory findings? It
contradicts tony's experience. It contradicts mine as well. Do you
have experience pointing the other way? (I would tend to assume not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
Unfortunately I can't really point to it since it's proprietary to the companies
4:27 p.m.
And of course then there's the warranty data.
How long is the warranty period for the widgets in question? In my
experience, warranty periods are ridiculously short these days (90
days, maybe as much as a year) reflecting a lack of manufacturer
confidence in device reliability.
Perhaps you're talking about something designed with more care. But I
don't think anyone has been claiming that it's impossible to build
reliable devices these days, only that most modern computers are crap
for reliability compared to most classic computers, and that complex
devices are, other things equal, less reliable than simple devices.
Anecdotes about one particular manufacturer aren't very relevant, not
unless it's one of the very few makers responsible for a significant
fraction of the mass-market computers made today. (And maybe not even
then, depending.)
Again we're talking ridiculously large sums of
money at stake.
So what? Companies not infrequently lose ridiculously large sums of
money for various reasons.
Again, when you're dealing with large populations,
looking at a few
in isolation, is I'm sorry to say just anecdotal.
So is your data...as far as we're concerned.
I'm sorry if you can't/won't understand
how large companies work and
the restrictions that are put on employees and the products (though
some have alluded to here). Because if you did, I'm guessing that
you wouldn't have made the comments you did.
I'm not tony. But I think I mostly understand those restrictions. I
don't expect you to violate NDA.
But neither am I convinced by "trust me, I can't show you the data,
just trust me". Especially when it runs directly counter to my own
experience. Even more so when your supporting anecdotes appear
irrelevant, as when you cite infant mortality rates (you can't check
anything but infant mortality in time to shut down an assembly line, as
you spoke of doing) as if it were relevant to with long-term
reliability of the assembled devices; this makes me suspect we're
actually talking about different things but, because we're using the
same words, mistaking them for the same thing.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
6:39 p.m.
On Sun, Apr 22, 2012 at 07:27:07PM -0400, Mouse wrote:
Well, not everywhere ;-)
In the EU for instance, the law mandates a warranty period of two years
against "defects (inherently) present at the time of sale"[0] and for
the first 6 months of that time, the onus is on the seller to _prove_
that the problem isn't a warranty case[2]. So, say, those super cheap caps
that will fail after 4 months of continuous operation would be _really_
expensive down the line.
Interesting side effect: a short time after that law went into effect,
a lot of devices that used to have only soft "off" switches (i.e. TVs,
VCRs, ... that you switch "on" and "off" via the remote (and which,
obviously, are under (at least limited) power 24/7)) all of a sudden
(re)gained hard (switching the mains) power switches. Yup, a noticeable
fraction[1] of the devices used to develop faults in the PSUs after a
couple of months due to penny pinching designs. Adding hard power switches
was downright dirt cheap compared to dealing with the sudden explosion in
warranty costs ;-)
Kind regards,
Alex.
[0] IANAL so thats my laymans explanation
[1] noticable to sellers & manufacturers, that is
[2] And it is good PR not to quibble to much in the next 18 months
of the warranty period
--
"Opportunity is missed by most people because it is dressed in overalls and
looks like work." -- Thomas A. Edison
And of course
then there's the warranty data.
How long is the warranty period for the widgets in question? In my
experience, warranty periods are ridiculously short these days (90
days, maybe as much as a year) reflecting a lack of manufacturer
confidence in device reliability.
10:11 p.m.
On 04/22/2012 06:57 PM, Guy Sotomayor wrote:
This is one thing that's quite backwards about our society. "I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
That is just plain fucked up.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
How coome 2 of
use here (at least) seem to have contradictory findings?
Because it's anecdotal. BTW here's the definition:
(of an account) not necessarily true or reliable, because based on
personal accounts rather than facts or research
23 Apr
23 Apr
1:17 a.m.
New subject: OT: I saw it, really I did / was Re: Arduino vs 555 reliability
On 2012 Apr 22, at 10:11 PM, Dave McGuire wrote:
On 04/22/2012 06:57 PM, Guy Sotomayor wrote:
This is one thing that's quite backwards about our society. "I
saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
That is just plain fucked up.
No, it's not. People lie, fabricate, embellish and misrepresent.
If you've been on one side of a legal or other matter you may have
found it frustrating not to be believed on your word, but if you've
been on the target end of other people's lies you'll very much
appreciate for their version (an anecdote) NOT being taken as
"necessarily true or reliable".
People - even if sincere - also confuse, conflate, misinterpret and
misunderstand. Even if you really, truly believe you saw it with your
own friggin' eyes, it still may not be an accurate or complete
version of an event.
Not only is this a basic tenet of modern justice systems, it's the
basis of scientific inquiry.
How coome
2 of use here (at least) seem to have contradictory
findings?
Because it's anecdotal. BTW here's the definition:
(of an account) not necessarily true or reliable, because based on
personal accounts rather than facts or research 
2:05 a.m.
New subject: OT: I saw it, really I did / was Re: Arduino vs 555 reliability
On Mon, Apr 23, 2012 at 10:17, Brent Hilpert <hilpert at cs.ubc.ca> wrote:
On 2012 Apr 22, at 10:11 PM, Dave McGuire wrote:
Totally agree. Humans are great at pattern recognition (to such a
degree that we have many false positives) and at self-deception. This
has become a quite popular subject with neuroscientists, and not just
(one hopes) so they can get magicians to perform at their conference
banquets.
http://www.smithsonianmag.com/arts-culture/Teller-Reveals-His-Secrets.html?…
http://www.nature.com/nrn/journal/v9/n11/full/nrn2473.html
Joe.
--
Joachim Thiemann :: http://jthiem.bitbucket.org
?This is one thing that's quite backwards
about our society. ?"I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
?That is just plain fucked up.
No, it's not. People lie, fabricate, embellish and misrepresent.
If you've been on one side of a legal or other matter you may have found it
frustrating not to be believed on your word, but if you've been on the
target end of other people's lies you'll very much appreciate for their
version (an anecdote) NOT being taken as "necessarily true or reliable".
People - even if sincere - also confuse, conflate, misinterpret and
misunderstand. Even if you really, truly believe you saw it with your own
friggin' eyes, it still may not be an accurate or complete version of an
event.
Not only is this a basic tenet of modern justice systems, it's the basis of
scientific inquiry.
12:14 p.m.
This is one thing that's quite backwards about
our society. "I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
Well, actually, people's evidence can be very unrliable, either
deliberately or innocently. So, yes, there is no reason why anyone should
believe anythign I say. But equally, I do believe what I observe
(probably incorrectly in some cases), at least until I've got further
relaiable evidence (not just 'I am telling you') to modify it.
-tony
2:12 p.m.
On 23/04/2012 20:14, Tony Duell wrote:
One thing that always puzzles me is that if Oracle Solaris boxes are so
reliable why does a years support on one cost so much more (typically 5
to 8 times more) than a years support on an similar spec IBM or HP Intel
or AMD box?
Dave
G4UGM
This is one
thing that's quite backwards about our society. "I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
Well, actually, people's evidence can be very unrliable,
either
deliberately or innocently. So, yes, there is no reason why anyone should
believe anythign I say. But equally, I do believe what I observe
(probably incorrectly in some cases), at least until I've got further
relaiable evidence (not just 'I am telling you') to modify it.
-tony
3:11 p.m.
On Apr 23, 2012, at 5:12 PM, Dave Wade wrote:
One thing that always puzzles me is that if Oracle
Solaris boxes are so reliable why does a years support on one cost so much more (typically
5 to 8 times more) than a years support on an similar spec IBM or HP Intel or AMD box?
The cynic in me wants to say, "Because the market will bear it."
- Dave
3:25 p.m.
On 23/04/2012 23:11, David Riley wrote:
On Apr 23, 2012, at 5:12 PM, Dave Wade wrote:
Perhaps more cynical is that they want you to think that if its that
expensive it must be good?
I seem to remember a few years ago they came and replaced every thing.
Recently I have had a duff PSU just sent by courier...
One thing that always puzzles me is that if
Oracle Solaris boxes are so reliable why does a years support on one cost so much more
(typically 5 to 8 times more) than a years support on an similar spec IBM or HP Intel or
AMD box?
The cynic in me wants to say, "Because the market will bear
it."
- Dave
3:45 p.m.
On Apr 23, 2012, at 6:25 PM, Dave Wade wrote:
On 23/04/2012 23:11, David Riley wrote:
I don't think that differs fundamentally from "the market will bear
it". Here we come back to P.T. Barnum and his d(suckers)/dT = 1/min
equation.
I mean, there are reasons for all manner of pricing on support
contracts, but they generally come down to "what will people pay for
this?" and "how can we amortize our development and support costs?".
- Dave
On Apr 23, 2012, at 5:12 PM, Dave Wade wrote:
Perhaps more cynical is that they want you to think that if its that expensive it
must be good? One thing that always puzzles me is that if
Oracle Solaris boxes are so reliable why does a years support on one cost so much more
(typically 5 to 8 times more) than a years support on an similar spec IBM or HP Intel or
AMD box?
The cynic in me wants to say, "Because the market will bear
it."
4:40 p.m.
On Mon, Apr 23, 2012 at 10:12:19PM +0100, Dave Wade wrote:
On 23/04/2012 20:14, Tony Duell wrote:
One thing that always puzzles me is that if Oracle Solaris boxes are
so reliable why does a years support on one cost so much more
(typically 5 to 8 times more) than a years support on an similar
spec IBM or HP Intel or AMD box?
Two effects at work:
- they're gonna charge as much as the market will bear and not a cent less
- the "it is expensive, it must be good" fallacy
Kind regards,
Alex.
--
"Opportunity is missed by most people because it is dressed in overalls and
looks like work." -- Thomas A. Edison
This is
one thing that's quite backwards about our society. "I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
Well, actually, people's evidence can be very unrliable, either
deliberately or innocently. So, yes, there is no reason why anyone should
believe anythign I say. But equally, I do believe what I observe
(probably incorrectly in some cases), at least until I've got further
relaiable evidence (not just 'I am telling you') to modify it.
-tony
6:26 p.m.
One thing that always puzzles me is that if Oracle
Solaris boxes are
so reliable why does a years support on one cost so much more
(typically 5 to 8 times more) than a years support on an similar spec
IBM or HP Intel or AMD box?
Others have mentioned two probable parts of this ("because it can" and
"if it's expensive it must be valuable").
I offer for consideration another possible part: the level of support
is different. While I am not familiar with modern support contracts of
either sort, I speculate that the Oracle support contract specifies
much tighter constraints on the supporter's performance (response time,
time to repair, etc), higher penalties if they aren't met, and a wider
net for what it covers. (As compared to the other contract, of
course.) I do know that, back when I did know a little bit about
support contracts, that such things had a a substantial effect on the
price of the contract.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
24 Apr
24 Apr
12:40 a.m.
On 24/04/2012 02:26, Mouse wrote:
Well at least in the UK IBM appears to provide its own
support perhaps
that's a contributing factor...
One thing that
always puzzles me is that if Oracle Solaris boxes are
so reliable why does a years support on one cost so much more
(typically 5 to 8 times more) than a years support on an similar spec
IBM or HP Intel or AMD box?
Others have mentioned two probable parts of this
("because it can" and
"if it's expensive it must be valuable").
I offer for consideration another possible part: the level of support
is different. While I am not familiar with modern support contracts of
either sort, I speculate that the Oracle support contract specifies
much tighter constraints on the supporter's performance (response time,
time to repair, etc), higher penalties if they aren't met, and a wider
net for what it covers. (As compared to the other contract, of
course.) I do know that, back when I did know a little bit about
support contracts, that such things had a a substantial effect on the
price of the contract. /~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
23 Apr
23 Apr
4:38 p.m.
On Mon, Apr 23, 2012 at 01:11:08AM -0400, Dave McGuire wrote:
On 04/22/2012 06:57 PM, Guy Sotomayor wrote:
This is one thing that's quite backwards about our society. "I saw it
with my own friggin' eyes" is discounted as "not necessarily true or
reliable".
That is just plain fucked up.
Unfortunately, it is not. IIRC there have been several studies into the
quality of various types of evidence. Turns out that "eyewitness accounts"
are among the most unreliable type of evidence around, complicated by the
fact that the margin of error for it is at best very hard to quantify
outside of a lab setup.
Even when people are _not_ attempting to lie, evade, shade the truth etc
and are in fact doing their level best to give correct accounts it isn't
worth much. And this is pretty much down to the fact that the human nervous
system and brain have _not_ evolved as a high grade recording system:
- data aquisition is incredibly lossy - the brain tends to filter out all
the "unimportant" stuff very early on, static signals will fade into
the background, the signals preprocessing done in our sensory organs
and lower parts of the brain have evolved to normally work basically as
high pass filters: fast changes (like a tiger jumping at you from the
underbrush) get noticed and processed at priority, very slow changes
(like someone _very_ slowly sneaking up on you) are normally not
noticed (yes, with training and continious attention one can notice
things like that, but it imposes a high cognitive load)
- data processing is optimized for pattern recognition and will in fact
detect patterns where there are none (Rorschach test, UFO sighting,
"ghosts" howling in a drafty attic, ...), it also favours "well
known"
patterns, i.e. is prone to get into a rut even if it doesn't really fit
- data processing ruthlessy discards any information that is currently
not deemed important (parade a few pretty girls in bikinis in front
of 20something boys and they won't notice anything else)
- memory storage is done in stages and and each stage has a pretty
aggressive garbage collector running (standard test: say, for the last
5 years you always drive the some route to work - if someone inserts
low grade (not immediately attenting grabbing) events into that route,
would you be able to recall them 3 years later? I bet you won't
- stress does horrible things to perception and memory retention
- the brain is very, very fond of delegating routine tasks to "lower
level automated processing": on that above commute, after five years, you
won't be paying careful attention to the route you drive, but mostly
do it on "autopilot" and unless something really extraordinary happens
(you wait at a traffic light and it suddenly rains red roses), you
will remember almost nothing at all of that commute a few days later
- but you've done it so often, your brain will happily synthesize the
memory from of bits and pieces of previous times you drove the route
- also: false memories - with the right recall stimuli, you'll swear up
and down that something happened in a specific way for a specific
event when in fact it didn't happen quite that way
Yes, the limits of the human data aquisition, processing and storage
system _can_ be overcome to a significant degree, but it requires training,
dedication and constant attention - which implies a significant cognitive
load.
Example: drive about 30-50 miles at night in fall. Assume that on random
stretches of the road, frogs will be crossing the road in large numbers.
Your task is to flatten as few frogs as possible, ideally none. And yes,
there are plenty of almost-looks-like-a-frog-from-50m-distance-in-headlights
leaves on the road.
Then a few days later, do a similiar drive but this time your task is only to
arrive safely. Compare how fatigued you are each time.
The human brain aggressively uses data pruning and routine task automation
because otherwise the cognitive load would be utterly exhausting.
The human brain is also very, very vulnerable to biasing of any kind.
The research into the human brain can be utterly fascinating and at
the same time be very depressing. To use the title of a talk at the
OSCON 2011 conference: "All your brains suck."
Kind regards,
Alex.
--
"Opportunity is missed by most people because it is dressed in overalls and
looks like work." -- Thomas A. Edison
How coome
2 of use here (at least) seem to have contradictory findings?
Because it's anecdotal. BTW here's the definition:
(of an account) not necessarily true or reliable, because based on
personal accounts rather than facts or research
11:33 a.m.
How coome 2 of
use here (at least) seem to have contradictory =
findings?
Because it's anecdotal. BTW here's the definition:
(of an account) not necessarily true or reliable, because based on
personal accounts rather than facts or research Fine. There's lots of things that I've
started to say on this list to =
deal with an
argument and couldn't because it would violate various NDA's that I have =
to
live under (I kinda like my job). Arguments on this (and other fora) =
while
fun aren't worth sacrificing my livelihood.
I'm sorry if you can't/won't understand how large companies work and the
restrictions that are put on employees and the products (though some =
have
I mosster certainly understnad that (having signed a fair few NDAs in my
time too, and yes, I do honour them). And that's the 'valid reasons' I
mentioend.
But to be fair, if I can't verify the data, why on earth should _I_
beleive it? You could be telling me anything.
-tony
12:45 a.m.
On 2012 Apr 22, at 12:06 PM, Tony Duell wrote:
You have contradictory findings because it's your limited and narrow
experience. Whatever or however many instances you have dealt with,
they come through a filter of circumstances before they get to you.
An assessment begins with a relative comparison of production
quantities, followed by an assessment of the selective process of how
and why the (very small) sample set happened to end up in your hands.
Modern systems are made by the tens and hundreds of millions, 1970s
minis were made by the thousands and tens-of-thousands.
I've received more digital cameras to repair lately than 1960's
discrete-component calculators. Does that mean modern digital cameras
are less reliable than 1960s discrete calculators?
Adding another data point, I will contradict the findings of the two
of you: I've seen more failures of [simple] SSI devices than
[complex] VLSI.
None of these 3 anecdotes (mine or those of the two of you) are
indicative of the comparitive reliability of these items.
Oh come now. And you expect us to believe comments based on data
that's
not even available? Sorry, I'll stick with what my experience as
taught
me, modifying it on the basis of more experience.
On Apr 21,
2012, at 7:08 PM, Mouse wrote:
Except that your reality is a *small* sample set. When you look
at data
for 1,000,000's of components you'll see different results. Until
you look at
And you have _ersonally_ looked at millions of components? Sorry, I've
seen too much bogus data already...
Extrapolating long tail effects from single samples does not yield
accurate results.
Neither does assuming that theory matches reality. truly large data sets and do the appropriate
statistical analysis
all you're
really talking about is anecdotal evidence.
How coome 2 of use here (at least) seem to have contradictory
findings? It contradicts tony's experience. It contradicts
mine as well.
Do you
have experience pointing the other way? (I would tend to assume
not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
Unfortunately I can't really point to it since it's proprietary to
the companies
12:56 a.m.
I've received more digital cameras to repair
lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
In case anyone still cares, in my case, the lesson from experience
still holds when corrected for such effects. Indeed, for several years
I ran more machine-hours of Suns than peecees and I _still_ had more
peecee hardware failures.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
1:32 a.m.
On 2012 Apr 23, at 12:56 AM, Mouse wrote:
Probably not, as it was a fairly ridiculous argument from the outset.
I've
received more digital cameras to repair lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
In case anyone still cares, in my case, the lesson from experience
still holds when corrected for such effects. Indeed, for several
years
I ran more machine-hours of Suns than peecees and I _still_ had more
peecee hardware failures.
What correction for what effect? Your [anecdote] is not particularly
surprising, nor do I see a contradiction.
2:07 a.m.
I've received more digital cameras to repair
lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
in my case, the lesson
from experience still holds when corrected
for such effects.
12:40 p.m.
On 23/04/2012 10:07, Mouse wrote:
What correction for what effect? "A is far commoner
than B, so seeing more instances of failing A than
of failing B does not allow any deductions about the relative failure
rates". I took that to be your point; even if it's not the point you
intended, it's a point I felt worth addressing.
I've personally seen far more failures of well-designed computers than
of experimental scramjets. Does that mean experimental scramjets are
the more reliable of the two? No; I just have never had anything to do
with scramjets, reliable or otherwise. My favourite is that as more people in the
UK are injured by tripping on
loose slippers than are injuried by chain saws then slippers are more
dangerous that chain saws. Fact is very few folks in the UK own chain
saws...
(Not sure how well slippers translates into West Pondian but lightweight
house shoes...)
I've
received more digital cameras to repair lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
in my case, the lesson
from experience still holds when corrected
for such effects. /~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
12:54 p.m.
[...] slippers [...]
(Not sure how well slippers translates into West Pondian but
lightweight house shoes...)
That's roughly what my (North American) eye read it as; I think that,
at least in that usage, it's one of those words that doesn't need
translation when it crosses the Atlantic.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
1:03 p.m.
On Apr 23, 2012, at 3:54 PM, Mouse wrote:
Yeah, we call them "slippers" here too in the Colonies. A lot of
clothing slang makes it across fairly intact, with the notable
exception of "pants" (which is what we call our trousers, much to
the confusion of our British friends).
- Dave
[...] slippers
[...]
(Not sure how well slippers translates into West Pondian but
lightweight house shoes...)
That's roughly what my (North American) eye read it as; I think that,
at least in that usage, it's one of those words that doesn't need
translation when it crosses the Atlantic.
24 Apr
24 Apr
6:20 a.m.
On 04/23/2012 03:03 PM, David Riley wrote:
On Apr 23, 2012, at 3:54 PM, Mouse wrote:
Yeah, we call them "slippers" here too in the Colonies. A lot of
clothing slang makes it across fairly intact, with the notable
exception of "pants" (which is what we call our trousers, much to
the confusion of our British friends).
It took me a while to get used to referring to sneakers rather than
trainers. Hearing people talking about guys wearing suspenders ('braces' in
UK English) takes much longer to adjust to :-)
I can't really think of any other examples - 'bath robe' vs. 'dressing
gown', I suppose - but you're right, there's a very high degree of
commonality. I found that US English is quite regionalized too, so even if
a different term (to UK English) exists in one part, it doesn't necessarily
mean that it's different across the whole country (the same is true of
pronunciation).
cheers
Jules
[...]
slippers [...]
(Not sure how well slippers translates into West Pondian but
lightweight house shoes...)
That's roughly what my (North American) eye read it as; I think that,
at least in that usage, it's one of those words that doesn't need
translation when it crosses the Atlantic.
23 Apr
23 Apr
1:04 p.m.
On 4/23/2012 3:40 PM, Dave Wade wrote:
(Not sure how well slippers translates into West
Pondian but lightweight
house shoes...)
It translates just fine. If I'm at home, I'm almost always wearing a pair.
Now a gentleman from the UK visited my company, and he was wearing
fabric straps over the shoulders to hold up his trousers.
And I said, "I'll have to pickup a pair of suspenders."
And the guy was laughing, saying, "why in the world would you do that?"
And I said, "Because they look nice."
Little did either of us know this:
http://en.wikipedia.org/wiki/Suspenders
Keith
24 Apr
24 Apr
11:37 p.m.
On 24/04/2012, at 5:40 AM, Dave Wade wrote:
My favourite is that as more people in the UK are
injured by tripping on loose slippers than are injuried by chain saws then slippers are
more dangerous that chain saws. Fact is very few folks in the UK own chain saws...
(Not sure how well slippers translates into West Pondian but lightweight house shoes?)
The stat I like best is that the world's most dangerous sport is Lawn Bowls, as more
people die playing it than any other sport? It has nothing to do with the sport's
demographic!
Huw Davies | e-mail: Huw.Davies at kerberos.davies.net.au
Melbourne | "If soccer was meant to be played in the
Australia | air, the sky would be painted green"
23 Apr
23 Apr
3:12 a.m.
It was thus said that the Great Mouse once stated:
I've had *many* more years of PC experience, and I've yet to see one fail
as bad as the SGI I used in college between '91-96 (let's see ... the SGI
required a new video card and the mother board replaced twice; also, the
video monitor slowly lost colors over time as well).
-spc (Although I've had to replace my work laptop once already ... )
I've
received more digital cameras to repair lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
In case anyone still cares, in my case, the lesson from experience
still holds when corrected for such effects. Indeed, for several years
I ran more machine-hours of Suns than peecees and I _still_ had more
peecee hardware failures.
8:41 a.m.
On 04/23/2012 05:12 AM, Sean Conner wrote:
It was thus said that the Great Mouse once stated:
I've had *many* more years of PC experience, and I've yet to see one fail
as bad as the SGI I used in college between '91-96 (let's see ... the SGI
required a new video card and the mother board replaced twice; also, the
video monitor slowly lost colors over time as well).
I suspect most of us can come up with isolated cases like that though (both
on the vintage and modern side of the coin).
I suspect the split's been roughly even for me, but (straying from the
topic a little) two points from my own personal observations:
1) I've had far more faults with modern hardware *that I can't fix* than I
have with older items,
2) I've had far more *software* glitches bring modern hardware to its knees
than I have with older kit.
Ultimately it's downtime and inconvenience that matters to me more than
failure rate, I think - and for me the older, simpler approaches win out
there (which isn't to say that it works out that way for everyone).
cheers
Jules
I've
received more digital cameras to repair lately than 1960's
discrete-component calculators. Does that mean modern digital
cameras are less reliable than 1960s discrete calculators?
In case anyone still cares, in my case, the lesson from experience
still holds when corrected for such effects. Indeed, for several years
I ran more machine-hours of Suns than peecees and I _still_ had more
peecee hardware failures.
1:27 p.m.
On Mon, 23 Apr 2012, Jules Richardson wrote:
1) I've had far more faults with modern hardware
*that I can't fix* than I
have with older items,
"I have always wished for my computer to be as easy to use as my
telephone; my wish has come true because I can no longer figure out how to
use my telephone." - Bjarne Stroustrup
2) I've had far more *software* glitches bring
modern hardware to its knees
than I have with older kit.
Ultimately it's downtime and inconvenience that matters to me more than
failure rate, I think - and for me the older, simpler approaches win out
there (which isn't to say that it works out that way for everyone).
And, therefore, 30 year old computer equipment is far more reliable than
CURRENT; it can't run Windoze!
12:34 p.m.
You have contradictory findings because it's your
limited and narrow
experience. Whatever or however many instances you have dealt with,
they come through a filter of circumstances before they get to you.
Sure, I don;'t, for example, see those devices whcih fialed in production
or initial testing. On the otehr hand, nobody else outside the factory
does either, so my expericne probably is relevant for what I do.
I've received more digital cameras to repair
lately than 1960's
discrete-component calculators. Does that mean modern digital cameras
are less reliable than 1960s discrete calculators?
OK, we can all be silly about this.
I can say without fear of contradiction that I've replaced more TTL than
microcontrollers in HP9800s for the simple reason that there are no
microcontrollers in HP9800s. And these machiens are actually a
counterexample to my comments in that the complex ICs -- the RAMs and
ROMs -- fail much less often than the TTL.
I've repaired more HP calculators than no-name 4-bangers. I've repaired
more Leica cameras than Kodak Instamtics. In bother cases that;s not
because HPs and Leicas are unrelaible, but rather because HPs and Leicas
are worth taking the time to repeir, 4-bangers and Instamatics often are not.
Adding another data point, I will contradict the
findings of the two
of you: I've seen more failures of [simple] SSI devices than
[complex] VLSI.
None of these 3 anecdotes (mine or those of the two of you) are
indicative of the comparitive reliability of these items.
Hmm... When I find that I rarely have to rplace a 555 (in somethign
commercal or something I've made) but that I do have to repalce VLSI
devices quite often, I know which I consider to be the more reliable.
Maybe incorrectly.
-tony
22 Apr
22 Apr
11:48 a.m.
> Extrapolating long tail effects from single
samples does not yield
> accurate results.
On Sat, 21 Apr 2012, Mouse wrote:
Neither does assuming that theory matches reality.
He's just
Extrapolating long tail effects from single samples
I would argue that todays computer systems are
*much* more reliable
than the systems that were in general use 10-20 (or more) years ago.
On what
grounds?
It contradicts tony's experience. It contradicts mine as well. Do you
have experience pointing the other way? (I would tend to assume not,
since if you had I would expect you to have mentioned it by now, but
assumptions don't make for good data either.)
21 Apr
21 Apr
11:49 p.m.
In all of this discussion, I'm surprised that no one's mentioned the
LMC555, a CMOS version of the original bipolar part. Much better
power characteristic (1mW typical) and good to 3MHz. Will operate
down to about 1.5V. Available in a variety of packages, including
micro SMD (about 1.4mm square).
--Chuck
22 Apr
22 Apr
12:16 p.m.
In all of this discussion, I'm surprised that no one's mentioned the
LMC555, a CMOS version of the original bipolar part. Much better
I am pretty sure I mentioend the 7555, which is another CMOS version.
It has many advantages (including the fact it doesn';t put an almighty
glitch onthe supply line when it switches), but it may be a little more
'fragile' than the bipoler one. Without checking, I am not sure of its
output current specification either, the bipolar 555 can drive quite
heavy load.
-tony
6:51 p.m.
On 22 Apr 2012 at 20:16, Tony Duell wrote:
It has many advantages (including the fact it
doesn';t put an almighty
glitch onthe supply line when it switches), but it may be a little
more 'fragile' than the bipoler one. Without checking, I am not sure
of its output current specification either, the bipolar 555 can drive
quite heavy load.
C555, IIRC, is rated at 100 ma. Good enough for a lot of things.
--Chuck
12:02 p.m.
OK. How many 1,000,000's of parts have you looked at? What long term st=
atistical analysis have you done?
And how many have you _personally_ looked at? I am not talking about
gettign data from elsewhere, which may or may not be reliable?
Extrapolating long tail effects from single samples does not yield accura=
te results.
True enough. But either I have been very unlucky, or there really is the
effect that complex ICs are less reliable than simple ones.I am not
talking about a few more microcontrolelrs failing than 555s. I am talking
aobut often having to repalce a complex IC and very rearle having to
repalce a simple one.
=20
In other words, I'ev; had many thousands of things across the bench by=20
now. And I can assure you that complex ICs, modern ones, old ones,=20
whatever, are mcuh more likely to be dead than simpler ones.=20
Still a small sample set. I would argue that todays computer systems are *much*
more reliable than
the systems that were in general use 10-20 (or more) years ago.
My first thoguht was 'have I jsut telepoted to a another uinverse' and
then I relaised you can't really compare like with like. Thew old
computers that I (and others here) run are oftne very expensive manchiens
(when new), as opposed to the <$500 stuff that most people have today. It
can't be fair to compare the build and reliabiltiy of those.
And I will agree that there are plentyo of (presuambly) modern systems on
the internet that are almost always avaialble (whether becuase each
machine is reliable or becuase there's redunancy or ... I don;t know), so
there must be plenty of reliable modern machines.
Comparing those to a home computer of 10 ro 20 eyars ago is unfair.
Compaing them to a minicomputer or similar is more reasoable, and I've
found such machiens (yes old ones) to be relaivble when properly used.
-tony
2:39 p.m.
[...] I relaised you can't really compare like
with like. Thew old
computers that I (and others here) run are oftne very expensive
manchiens (when new), as opposed to the <$500 stuff that most people
have today. It can't be fair to compare the build and reliabiltiy of
those.
I'd say, yes and no.
Take, for exmaple, one of my SS20s. A modern machine built with the
same kind of engineering care and attention that went into that SS20
quite possibly _would_ be the more reliable of the two - I have no real
basis for comparison there.
But when compared in terms of what's common - what my employer put on
employees' desks at work back 20, 30 years ago, versus what my employer
puts on employees' desks at work today (or, in my case, would if I
didn't bring in my own machine for the purpose)...then today's machines
are crap. Same if the comparison is based on looking at shelves in a
(then-)typical computer store.
Each basis for comparison has a place, and I get the feeling at least
part of the disagreement here is disagreeing over which one is more
appropriate.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
23 Apr
23 Apr
11:20 a.m.
[...] I relaised you can't really compare
like with like. Thew old
computers that I (and others here) run are oftne very expensive
manchiens (when new), as opposed to the <$500 stuff that most people
have today. It can't be fair to compare the build and reliabiltiy of
those.
I'd say, yes and no. from the local shop. I do not find that suprising.
Take, for exmaple, one of my SS20s. A modern machine built with the
same kind of engineering care and attention that went into that SS20
quite possibly _would_ be the more reliable of the two - I have no real
basis for comparison there.
My only conserni is things like lead-free solder. Tin whisker growth is a
problem, and I am not sure that it can be easily avoided. I do know that
it's a problem in all sorts of modern stuff...
But when compared in terms of what's common - what my employer put on
employees' desks at work back 20, 30 years ago, versus what my employer
puts on employees' desks at work today (or, in my case, would if I
Indeed... Go back 40 years and there was essentially only one desktop
computer (HP9830). It's built like a brick outhouse, and mine still runs
(albeit with some minor repairs over the years). On the otehr hand it
cost rater hmnore than a family car at the time.
Now you can get very cheap (relaitelve) computers. It's not suprinsing
that they are not wrll made. And alas too may people -- indeed too many
compaines -- buy on price alone.
I guess what really saddnes me is that you can't buy a PC built like an
HP9830 (metaphorically) at any price. Yes, there are relaible devies
made nwo, but jsut try to buy one.
-tony
3:47 p.m.
I guess what really saddnes me is that you can't
buy a PC built like
an HP9830 (metaphorically) at any price. [...]
You can't? Or you can't do so easily and conveniently?
[earlier in the same mail]
On the otehr hand [the HP9830] cost rater hmnore than
a family car at
the time.
I once saw a machine that, as far as I could tell, was architecturally
a peecee, but mechanically it was amazing. Honestly, it looked
designed for remote telemetry or some such, the sort of environment
where any kind of service call, instead of involving a truck roll and
half a day's time for a tech, involves chartering a small airplane and
at least two person-weeks of time.
Heck, if you have "rather more than [the price of] a family car" to
spend on it, you might even just spin up a small company and design and
build what you want.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
4406
days inactive
4410
days old
89 comments
21 participants
participants (21)
-
als＠thangorodrim.de -
ard＠p850ug1.demon.co.uk -
bob＠theadamsons.co.uk -
cclist＠sydex.com -
cisin＠xenosoft.com -
dave.g4ugm＠gmail.com -
doc＠vaxen.net -
ethan.dicks＠gmail.com -
fraveydank＠gmail.com -
geneb＠deltasoft.com -
ggs＠shiresoft.com -
hilpert＠cs.ubc.ca -
huw.davies＠mail.vsm.com.au -
joachim.thiemann＠gmail.com -
jules.richardson99＠gmail.com -
keithvz＠verizon.net -
mcguire＠neurotica.com -
mouse＠Rodents-Montreal.ORG -
pu1bzz.listas＠gmail.com -
spc＠conman.org -
wdonzelli＠gmail.com