Altair reliability myth - CP/M
This is a discussion on Altair reliability myth - CP/M ; Well, I have 50 days of uptime without a hardware failure. Using all
Rev. 0 boards..
I have to turn the thing off tonight. : ( Its sad to see it turned
off in such a long time...
Grant...
-
Altair reliability myth
Well, I have 50 days of uptime without a hardware failure. Using all
Rev. 0 boards..
I have to turn the thing off tonight. : ( Its sad to see it turned
off in such a long time...
Grant
-
Re: Altair reliability myth
"Grant Stockly" wrote in message
news:1170404886.648902.289050@s48g2000cws.googlegr oups.com...
> Well, I have 50 days of uptime without a hardware failure. Using all
> Rev. 0 boards..
>
> I have to turn the thing off tonight. : ( Its sad to see it turned
> off in such a long time...
What model do you have? 8800? 8800a? 8800b?
Was it factory assembled or built from a kit?
If built, did the builder add any filter caps or make
any other modifications?
Tom Lake
-
Re: Altair reliability myth
On 2 Feb 2007 00:28:06 -0800, "Grant Stockly"
wrote:
>Well, I have 50 days of uptime without a hardware failure. Using all
>Rev. 0 boards..
>
>I have to turn the thing off tonight. : ( Its sad to see it turned
>off in such a long time...
>
>Grant
It's possible. The later 8800A and especially the 8800B machines wer
far better in many small but additive ways. I achieved long uptimes
after modding my machine extensively.
However, try it with a 8800 (non-A) with 4 individual 4slot backple
segments (16 slots total), 4 88S4K 8k synchronous ram cards,
two PIOs, VDM1 and an 88ACR (cassette interface). As supplied
and unmodified the 8V bus in mine would be at 6.9V (undersized
transformer supplied), the +16V bus was near 14V and the bus ringing
was impossible to deal with.
Replacing the bus board with a 18slot doublesided one (wamco)
and rewinding the undersized transformers, upgrading the rectifier
system and replacing the filter caps with significantly larger values
made that an operation system.
However I had the problem of some cards with Tinplate edge
connectors and other with Gold plated and that resulted in a condition
that required remove and reinstall of every card to wipe the crud off
the connectors or the system would not run. Add to that that LI NY
is humid and salt air environment that tended to make the corrosion
grow faster.
All that power resulted in heat, not a trivial amount of it! So the
largest rotron fan was used (120cfm) and that was not enough to cool
it as the pattern of airflow is restricted in that box. The excess
heat also made the oneshots drift (CPU phase timing and frontpannel).
I have an stock 8800BT and the only problem it has is cooling and the
first owner solved that by sliding the top nack about 1.5" so the hot
air could escape out the top. Thats a loaded box (17 cards!) so it's
warm but the bus and power problems were solved in that model.
So I could insure it was covered I used to top off my 8800 that was
slotted along the front edge to improve airflow. It's seen uptimes of
2weeks without incident though for other reasons (noise) I rarely
run it for long periods.
Allison
-
Re: Altair reliability myth
"Grant Stockly" wrote in message
news:1170404886.648902.289050@s48g2000cws.googlegr oups.com...
> Well, I have 50 days of uptime without a hardware failure. Using all
> Rev. 0 boards..
>
> I have to turn the thing off tonight. : ( Its sad to see it turned
> off in such a long time...
>
> Grant
>
Let's not forget usability/maintainabilty issues - like those wobbly little
blue card guides. Always takes me 2 or 3 tries before I get cards in
straight!
-John
-
Re: Altair reliability myth
> What model do you have? 8800? 8800a? 8800b?
> Was it factory assembled or built from a kit?
> If built, did the builder add any filter caps or make
> any other modifications?
Brand spanking new 8800. : )
Grant
-
Re: Altair reliability myth
> I have an stock 8800BT and the only problem it has is cooling and the
> first owner solved that by sliding the top nack about 1.5" so the hot
> air could escape out the top. Thats a loaded box (17 cards!) so it's
> warm but the bus and power problems were solved in that model.
> So I could insure it was covered I used to top off my 8800 that was
> slotted along the front edge to improve airflow. It's seen uptimes of
> 2weeks without incident though for other reasons (noise) I rarely
> run it for long periods.
My vintage Altair 8800 has the upgraded 8800B transformer and big
caps, but still has the daisy chained motherboards. Proper grounding
is very important for the thing. I had at one time 6 8k seals cards,
the 2SIO, PMC, and disk card set. Never had the modem card in it
yet. I took that thing apart before ever trying it. I'm not sure it
would have worked.
Uptimes can be hard to establish because basic was sort of buggy. In
software development its easy to crash an emulator. ; )
Grant
-
Re: Altair reliability myth
"Grant Stockly" wrote in message
news:1170487480.011279.68780@p10g2000cwp.googlegro ups.com...
>> What model do you have? 8800? 8800a? 8800b?
>> Was it factory assembled or built from a kit?
>> If built, did the builder add any filter caps or make
>> any other modifications?
>
> Brand spanking new 8800. : )
How's the replica coming? I tried to email you from the
link on your Website but it was returned as undeliverable.
If you're shipping kits, I'd like to get one!
Tom Lake
-
Re: Altair reliability myth
I'm confused about what kind of "Altair" you are posting about, and
others had to ask you as well.
In one post you eventually say you have a "brand spanking new 8800" -
referring to your replica Altair 8800 which is built from new parts,
new boards, today, and sold by you. But in another post you say:
> My vintage Altair 8800 has the upgraded 8800B transformer and big
> caps, but still has the daisy chained motherboards....
If it's "vintage", that means it was built decades ago, from parts now
decades old, solder connections just as old, etc. "Uptime" for 30-year-
old parts have issues very different from uptime for new parts, new
solder joints, etc.
It's an achievement to build today a replica Altair which can run for
weeks. My post is no insult to you. But please be explicit as to when
you are posting about new, replica, copied Altairs - versus old,
vintage, original, made-by-MITS Altairs. Whatever you choose to call
either of them, be clear and not ambiguous, so you'll get appropriate
responses.
That said, it's a common problem when people post a query, but don't
provide enough information to respond except with more questions. I
get queries of that sort all the time: it's a lot of work to second-
guess the missing info, or a bit touchy to say "please provide more
information".
Herb Johnson
Herbert R. Johnson, New Jersey USA
web site
domain mirror
a>
my email address: hjohnson AAT retrotechnology DOTT com
if no reply, try in a few days: herbjohnson ATT comcast DOTT net
"Herb's Stuff": old Mac, SGI, 8-inch floppy drives
S-100 IMSAI Altair computers, docs, by "Dr. S-100"
-
Re: Altair reliability myth
On Feb 3, 2:10 am, "Tom Lake" wrote:
> "Grant Stockly" wrote in message
>
> news:1170487480.011279.68780@p10g2000cwp.googlegro ups.com...
>
> >> What model do you have? 8800? 8800a? 8800b?
> >> Was it factory assembled or built from a kit?
> >> If built, did the builder add any filter caps or make
> >> any other modifications?
>
> > Brand spanking new 8800. : )
>
> How's the replica coming? I tried to email you from the
> link on your Website but it was returned as undeliverable.
> If you're shipping kits, I'd like to get one!
Yes, try again! I have 2 left in the current batch of 20. What was
the undeliverable message? Try leaving a post on my forum or
something...
Grant
-
Re: Altair reliability myth
> If it's "vintage", that means it was built decades ago, from parts now
> decades old, solder connections just as old, etc. "Uptime" for 30-year-
> old parts have issues very different from uptime for new parts, new
> solder joints, etc.
>
> It's an achievement to build today a replica Altair which can run for
> weeks. My post is no insult to you. But please be explicit as to when
> you are posting about new, replica, copied Altairs - versus old,
> vintage, original, made-by-MITS Altairs. Whatever you choose to call
> either of them, be clear and not ambiguous, so you'll get appropriate
> responses.
>
I don't care about the relative uptime of new vs. old, the point is a
carbon copy Altair is reliable. The arguments in the past were that
Altairs were never reliable and there were much better computers out
there. If we went back in time, applied errata notes, and looked at
the 2 phase clock with a scope, we would have no problem with
reliability.
My other Altair is just as good, but I've never left it on for 50
days. Its noisy... I'm sure it would be the same. It was built very
well.
Grant
-
Re: Altair reliability myth
On 2 Feb 2007 23:28:49 -0800, "Grant Stockly"
wrote:
>
>> I have an stock 8800BT and the only problem it has is cooling and the
>> first owner solved that by sliding the top nack about 1.5" so the hot
>> air could escape out the top. Thats a loaded box (17 cards!) so it's
>> warm but the bus and power problems were solved in that model.
>> So I could insure it was covered I used to top off my 8800 that was
>> slotted along the front edge to improve airflow. It's seen uptimes of
>> 2weeks without incident though for other reasons (noise) I rarely
>> run it for long periods.
>
>My vintage Altair 8800 has the upgraded 8800B transformer and big
>caps, but still has the daisy chained motherboards. Proper grounding
>is very important for the thing. I had at one time 6 8k seals cards,
>the 2SIO, PMC, and disk card set. Never had the modem card in it
>yet. I took that thing apart before ever trying it. I'm not sure it
>would have worked.
Yes but, are you using a mix of aged (6mo to year old) cards with
edge connectors that are both tin plate (CPU, 88acr, 88pio and 88MCD)
or gold plate like the 88S4K rams with the nominal gold plate sockets
in the backplane?
As to grounding I had to add #12 busbar to the board grounds as the
voltage differential down the bus was bad with a full load of cards,
same for +8V and +16V.
With the backplane replaced the system was upgraded to 56k of Seals
8K static cards and all of the boards that had tin plate edges were
not reused (save for CPU). Guess what socket was flaky after 6months,
the cpu socket! Apparently the the gold edge connectors didn't like
the tin plate (or was it tin/lead??).
At least one 88S4K the gold edge connectors may have not had a flash
Nickle under plate wich results in creeping green from the copper
migration through the gold. The lack of Nickle under plating gold
is a bad thing.
There were inline improvents in the first months of production that
really helped. Mine was very early production and had every an all
bugs as shipped.
I was able to make the 88ACR (modem) work though it was fussy.
Changing from the modem board to a NZRI encoding with a
casette tape unit designed for saturation recording rather than
audio (Redactron) worked far better at 19200 baud.
>Uptimes can be hard to establish because basic was sort of buggy. In
>software development its easy to crash an emulator. ; )
Yes, very. However once running an establised application then you
could tell. Also I had after about a year morning sickness where
before starting you had to pull and reinsert the cards (edge connector
wipe) or it might not start or if it did it would crash running know
good code. I believe that was the tin gold edge connectors at fault.
The funky this was once up and past the first hour it could be counted
to run at lest 2weeks (longest run tried) only to fail the next power
up (if I didn't remember to push pull the boards).
I still have the machine (the 8800) but haven't used it since 1979 or
80. In 79 I built a Netronics 8085 Explorer as a hardware testbed
system and already had a working NS* Horizon that replaced
the MITS. That NS* is still in use and solid.
I also have a 8800BT that's CA1979/December (recieved jan 1980
accoding to the BOS) production and that machine is still running as
is and has a history that was 100% solid save for the 8" floppy drives
that MITS used.
Allison
-
Re: Altair reliability myth
> Yes but, are you using a mix of aged (6mo to year old) cards with
> edge connectors that are both tin plate (CPU, 88acr, 88pio and 88MCD)
> or gold plate like the 88S4K rams with the nominal gold plate sockets
> in the backplane?
My ancient Altair has the same card edge connectors as the new one,
made even by the same company. They are both gold plated.
My ancient Altair and current Altair both have tin plated card edge
fingers. However the current 20 kits all have gold fingers. Cost me
an extra $5 or so per board, so I hope it is worth it to the buyers!
-
Re: Altair reliability myth
On 3 Feb 2007 14:47:34 -0800, "Grant Stockly"
wrote:
>> Yes but, are you using a mix of aged (6mo to year old) cards with
>> edge connectors that are both tin plate (CPU, 88acr, 88pio and 88MCD)
>> or gold plate like the 88S4K rams with the nominal gold plate sockets
>> in the backplane?
>
>My ancient Altair has the same card edge connectors as the new one,
>made even by the same company. They are both gold plated.
However the very early Altairs (first 1000 or less) had some cards
gold plated and some tinplate edge connectors.
Also there were a bath of 88S4K (the later synchonous 4k card)
with bad gold plate that turns green.
>My ancient Altair and current Altair both have tin plated card edge
>fingers. However the current 20 kits all have gold fingers. Cost me
>an extra $5 or so per board, so I hope it is worth it to the buyers!
Only if the gold was done right by the plater. Copper needs a flash
of nickle to prevent copper migration through the gold. Nothing
conducts worse than green gold.
I never said the design couldn't be reliable only that the early
machines had severe teething pains.
Allison