[comp.sys.mac] Is an

truesdel@ics.uci.edu (Scott Truesdell) (10/06/89)

alex@grian.cps.altadena.ca.us (Alex Pournelle) writes:

> I tend toward external drives on Macs (PCs too, occasionally).

Well, for dozens of reasons, external drives make sense. Ease of
swap-out in case of repair, sparing the inadequat power supplies
of most microcomputers, et. al.

But EMOTIONALLY and for convenience and neatness, the reasons for
internal H.D.s are more compelling for the majority of microcomputer
users. I've got a Mac SE/30 with the internal Quantum 80 and ethernet
card and it's just so CONVENIENT to tote it all over the place.  Of
course, when I took it to Boston in August, the Quantum became stuck
several times, including in Logan Airport when the security people
wanted be to turn it on to assure them it wasn't a bomb. (I had a
bootable floppy!). I guess humidity is the big culprit with this
Quantum stiction problem. Absolute reliability in dry ol' So.  Cal.

> Apple supplies are a 
>constant source of terror and error, have been since the Apple II.  Price 
>pressure (even Apple feels it) means they cut corners, and ask Sony to 
>do so on the supplies.  This is an invisible headache cause which many 
>people never notice.

The sad part is it's not that expensive to build 'em a little better.
From the OEM's reference, it'd add $1 or $2 to build 'em robust, no?

> The $65 Taiwan supply the cause (no overvoltage protection,
> lousy filtering, worse regulation).  But its day, in high-end 
> 386/486 computers, is on the wane.

Praise whatever dieties are currently in vogue for this favor!

>Moral: Power is more a problem than you think.

I believe it! Your description of this as an invisible headache is so so
true. re: the computer boots so the PS must be ok... what to look at next?
Unless the PS actually shuts itself down, it never comes under suspicion.

>[ref to Apple's power supplies], but the ones I've had apart don't make me
>very reassured.  The best indicator, alas, is time.
                                      ^^^^
                                      ackkk!!! there's that word again!
 
Seriously, though, even power supplies from "quality" vendors are
subject to this pricing pressure. Case in point: One of our favorite
OEMs, MicroNet, uses minimum PS technology in the external units. The
only RMAs I've had with MicroNet have been PS related and it's
(unfortunately) a higher average than I would like.

>Hmmm.  If you have to have 3-1/2" form-factor, I'd go with ST-157N.

Really? I tend to shy away from Seagates when given an alternative.
Knowing the nature of your work, I value your opinion and will 
reevaluate my policy on these drives.

>I am not encouraged by the Apple/Quantum "cure" pronouncement.

Me neither.

>  A ROM change doesn't
>make up for the fact that *Quantum isn't plating the media well enough*.
>And then they don't lube them well enough, either.  No amount of
>post-manufacturing shenanigans is going to cure stiction--they can't add
>torque to the motor.  If they run it harder at start-up they risk
>burning the coils--almost invariably fatal.

I thought about that. Throw more juice at it and watch it fry. As far
as remedial fixes go, it's a hack. I can only hope this hack isn't going
to the production line but rather they implemented a cure at the source;
better media, lube, and motor.

>Priams don't stick; CDCs don't stick.  Maxtors don't seem to (not big
>personal knowledge pool there).  Seagates, MiniScribes, Plus HardCards
>and Quanta do.  The first two are easy for us to rescue, the last
>two--not so easy.

Hmmm Plus HardCards are made by Quantum, right? Or at least they are made
and engineered by the same mother company. So this is an interesting 
coincidense.

>Call me prejudiced against Quanta; I've seen entirely more than my share
>come through the door.  And their tech. support hasn't been very
>helpful, I'm afraid (love to change that!).  Parts are essentially
>unavailable.

They're selling all they can make, I'll bet.

>> When the CDC Swifts start shipping [...]  

>Erm, didn't know CDC was STILL in the teething stage on those. 

Charles McConnethy told me they have been working with Imprimis on 
these teething problems. I forget the precise nature of the problem[s]
though I remember that they were slight and almost solved.

>Even Apple might use 'em.

Hurray!  They use the Wren V 173 megger for their 160MB drive. I love
that drive for the size range. Fast, quiet, and relatively cool.

>I might suggest Conners but my experience with the Compaq Conners has
>been--uninspiring.

Is that with the special ultra-power-miser version of the Conners?  I
haven't used Conners personally, but one thing that stuck in my mind
when first reading about them in Computer Design several years ago was
(a) the depth of experience behind their engineering dept. (starting
with old man Conners himself)  and (b) their method of designing in
reliability and low cost by making up for RELATIVELY low tolerance
parts compensated with intelligent firmware. In a production environment,
the relaxation of a few ten thousandths of an inch translates into 
big long-term cost savings.

>>good for economy, and, as an aside, they offer the best performance.

>'Sides, the Mitsubishis and Fujis of the world haven't got into SCSIs in
>a major way.

   :^)

>Have to see your lab[s: 4 to 6 depending how you look at 'em]
>sometime.  Sounds fun....

Sure. Any time. A lot of rather vanilla implementations with the odd
mighty Mac (8MB/173MB/24-bit color/19" screen/video I/O, etc) work-
station thrown in.

  --scott  714/856-5697
 
--
Scott Truesdell

truesdel@ics.uci.edu (Scott Truesdell) (10/07/89)

> The Mac has it all over the PC for true
>dragability--size and weight too.  But I get a mite worried about the
>power supplies with all that in there.

  <Ulp!>  Hear! Hear!

> If your 5 and 12 Volts don't
>vary (usual first sign of power supply error in "classic" [?!] Macs is
>the screen jitters) much (1/2%?  1%?) with all that plugged in--go for it.  

I don't know about the Pluses, but the SEs may be adjusted for output
under load: a good idea after even such a small modification as a
memory upgrade.

>All you CMS users out there, beware: CMS was, for over 8 months, using a
>35-watt power supply in their external 80meg (Seagate ST-277N) drives,
>which take about 75 Watts at startup.  The result is predictable:
>KA-BLOOIE! drives.  Fried beyond repair.
>They kept it pretty quiet, no recalls or anything.
>If you have one of 'em, see if it's got a Skynet supply board.  If so,
>you should contact CMS.

See, everybody? This is just one more horror story  in a [still]
emerging consumer market. Until there gets to be a Good Computing Seal
of Approval or some kind of watchdog agency to declare products at
least reasonably engineered, we'll always experience goofs like this.
<flame> <flame>

In the Golden era of computing, or even today, still, in the mainframe
and technical and business oriented machines (you know, the ones
costing over 6 digits?) the mfgs wouldn't dream of letting their
reputation be smirched by something so trivial to deal with. From an
engineering standpoint, THESE ISSUES ARE TRIVIAL.

>(backpedal backpedal, sounds of coughing and shuffling) I have re-
>re-considered this.  The Seagates are,
>despite my constant grotching about commodity drives, not bad.  

Case in point: all those cheap ST-225s out there in PC clones that
basically just keep on tickin'.

> [various comments about experience and where the state of the art
> lies in small drives...] So I'm stuck for a recommendation within my
> experience.  So I mentioned Seagate.

McConnathy repeated states to me and others that, when not DOA, the
Seagates remain the most reliable in their price range. Their solution
at MicroNet: burn the daylights out of them on the test bench before
they ship 'em. Results: very few returns (says Charles).

>[Conners has] some brilliant engineers.  I'm just not sure they let them
>make engineering decisions.

Hear! Hear! Engineers DO NEED Marketing to tell them when it's soup.
The demands of the marketplace in ANY industry will always carry out
this tug of war between Engineering and Marketing. For engineers, the
journey is the reward, and what they learn in the process of bringing a
product to life usually instantly obsoletes the current project. They
understand this and look foreward to the next project where they can
bring their newfound knowledge and experience to bear.

Marketing, on the other hand, no matter what the industry, tends to
lean towards the used car salesman mentality. That's why when "we" go
into computer stores (or wherever) we tend to shun places where the
salespeople decend on us like greasy pirranah and instead gravitate
towards the back where the tech is busy burning his or her fingertips
on a soldering iron. :^)

>Is anyone interested in this stuff, or should Scott and I take it to
>mail?

  Yea!  <babble> <babble>

Anyway, I would guess we've about thrashed it to death anyway.

  --scott

--
Scott Truesdell