steve@hite386.UUCP (Steve Hite) (02/01/90)
About 6 weeks ago I purchased a new SE/30 1/40 from my university.
It was a great deal...a new ImageWriter II was included for a total of
$3550. This was part of an Apple holiday promotional with colleges to
encourage students to buy Macs. With memory prices falling, I thought
it was a good time to go ahead and fill it up to 8 megs. Hey, I've got
FullWrite, MPW C and SADE...what can I say. :-)
Fearing for my warranty (read my signature), I got Komputer Kingdom
(read Dvorak's Feb '90 MacUser column :-)) to install it during a
routine logic board removal (read my signature). They gave me back the
1 meg of 256Ks that came with the machine. I was very surprised to find
that they were 120ns chips...nah, I was upset! My question is this...
Why does Apple put low speed memory chips in their high speed SE/30s?
Are they too cheap to use 80ns?
Needless to say, I'm not too smart about hardware. But it seems to me
that if I had bought only 4 megs of 80ns memory (upgrading the SE/30 to
5 megs total) then I would have to replace the 1 meg of 120's with 80's or
my fastest memory access speed would still only be 120ns. Hmm...that stinks,
Mr. Pepsi!!
BTW, I think the Apple EEs did an EXCELLENT job to keep a 16MHz machine
from choking on 120's with no parity errors...can you do it with a IIci :-)?
-----------------------------------------------------------------------------
Steve Hite -=* How to extend your Apple 90 day warranty *=-
...uflorida!unf7!hite386!steve Get a new SE/30, have the first 2 system
boards flake out on you, 6 weeks later,
the 3rd one finally works...voila!..Instant
42 day warrany extension..."free"...thanks
for the memories Apple! <pun intended>
clarson@ux.acs.umn.edu (Chaz Larson) (02/02/90)
In article <15@hite386.UUCP> steve@hite386.UUCP (Steve Hite) writes: > > Why does Apple put low speed memory chips in their high speed SE/30s? > Are they too cheap to use 80ns? > > Needless to say, I'm not too smart about hardware. But it seems to me >that if I had bought only 4 megs of 80ns memory (upgrading the SE/30 to >5 megs total) then I would have to replace the 1 meg of 120's with 80's or >my fastest memory access speed would still only be 120ns. Hmm...that stinks, >Mr. Pepsi!! > I am under the impression that the SE/30-II-IIx-IIcx machines are designed for 120ns RAM, and can't take advantage of faster SIMMs, so that your 8 megs of 80ns RAM are only being addressed by the OS at 120ns anyway. <chaz> -- ---------------------------------------------------------------------- Chaz Larson - clarson@ux.acs.umn.edu "Hey, I'm no Jack Kennedy..." - Flaming Carrot ----------------------------------------------------------------------
jcocon%hazel.cs.clemson.edu@hubcap.clemson.edu (James C. O'Connor III, 2841) (02/02/90)
From article <15@hite386.UUCP>, by steve@hite386.UUCP (Steve Hite): > Needless to say, I'm not too smart about hardware. But it seems to me > that if I had bought only 4 megs of 80ns memory (upgrading the SE/30 to > 5 megs total) then I would have to replace the 1 meg of 120's with 80's or > my fastest memory access speed would still only be 120ns. Hmm...that stinks, > Mr. Pepsi!! > BTW, I think the Apple EEs did an EXCELLENT job to keep a 16MHz machine > from choking on 120's with no parity errors...can you do it with a IIci :-)? All 16MHz Macs will run as fast with 120ns chips as with 70ns chips. The wait states are hardwired, regardless of memory speed. Apple is not dumping. 80ns chips are really a waste in anything less than a IIci, unless you get an accelerator card. Apple deserves flaming for somethings, but this isn't one of them. Jim
rieman@boulder.Colorado.EDU (John Rieman) (02/02/90)
In article <15@hite386.UUCP> steve@hite386.UUCP (Steve Hite) writes: > > About 6 weeks ago I purchased a new SE/30 1/40 from my university. > ... >1 meg of 256Ks that came with the machine. I was very surprised to find >that they were 120ns chips...nah, I was upset! My question is this... > > Why does Apple put low speed memory chips in their high speed SE/30s? > Are they too cheap to use 80ns? > Too smart would be a better explanation. Why should they put in more expensive chips that give the same performance? The SE/30, like every other machine I'm familiar with, accesses memory at a fixed speed. 120ns or faster chips will support this speed -- but faster chips don't give you any better response. Think of it this way: The processor puts a request out to memory. It waits 120ns. Then it gets the data it requested off the bus. As long as memory can respond in 120ns OR LESS, everything is fine. If memory responds in 80ns, the data just sits there for another 40ns, until the processor grabs it. (The processor isn't actually just sitting dead for 120ns -- it's preparing to grab the data. And yes, the story is more complex than this, I'm sure. But you get the idea.) -john rieman@boulder.colorado U. of Colo.
tbutler@wpi.wpi.edu (Tim Butler) (02/02/90)
In article <15@hite386.UUCP> steve@hite386.UUCP (Steve Hite) writes: > Why does Apple put low speed memory chips in their high speed SE/30s? > Are they too cheap to use 80ns? If they used more expensive ones the price would jump $1000. Seriously though, faster RAM wouldn't speed anything up. The system would still allow 120ns for memory access. BUT, at least you can take faster chips with you when you upgrade in 2-3 years. (So save those 120ns 256K chips) tim -- Tim Butler tbutler@wpi.wpi.edu Department of Mechanical Engineering Worcester Polytechnic Institute
steve@hite386.UUCP (Steve Hite) (02/02/90)
In article <15@hite386.UUCP>, steve@hite386.UUCP (Me) writes: > > Needless to say, I'm not too smart about hardware. But it seems to me > that if I had bought only 4 megs of 80ns memory (upgrading the SE/30 to > 5 megs total) then I would have to replace the 1 meg of 120's with 80's or > my fastest memory access speed would still only be 120ns. Hmm...that stinks, > Mr. Pepsi!! > > BTW, I think the Apple EEs did an EXCELLENT job to keep a 16MHz machine > from choking on 120's with no parity errors...can you do it with a IIci :-)? Thanks to all of the people who quickly pointed out how wrong I am about my SE/30 hardware assumptions and to those e-mail letters pouring in that will echo the first day's batch :-)...I ASSUMED that there were no wait states put in and that it was designed to access RAM faster than 120ns. I also have a 16MHz 386 AT Clone with no wait states and it demands faster memory to run properly so my fleeting past experience with RAM speeds definitely skewed my thinking (thanks Paul M. for being polite about it). I am not alone in my ignorance, though. The kind folks in our local Mac Users Group are also under the false assumption that if you put faster SIMMS in your Mac (i.e. replace them all with 80's) that your Mac will run faster. Bruce Cantie <cantie@cs.buffalo.edu> pointed out to me that there is a write-up in the current MACWORLD discussing this issue. I promise to watch some old TV Kung Fu episodes to regain a "humble as the dust" outlook on things that I do not fully understand. Thanks. ------------------------------ Steve Hite ...uflorida!unf7!hite386!steve
siegel@endor.harvard.edu (Rich Siegel) (02/02/90)
In article <2827@draken.nada.kth.se> d88-jwa@nada.kth.se (Jon W{tte) writes: > >Nope ! > >A 16 MHz machine will not run at all if the SIMMS are 120 ns, since >the data won't be there when the CPU wants it. Nope! yourself. Your calculator may tell you one thing, that the 120ns rise time isn't fast enough to respond to the 16MHz bandwidth of the processor, but the hardware designers and implementors at Apple would beg to differ; the memory is accessed at one wait state, and accesses are additionally bottlenecked by little things like bus arbitration. > >It _is_ bad of apple to put 120 ns SIMMS in a 16 MHz machine. Period. > It would be if said 16MHz machine with 120ns memory didn't work, but since thousands of Mac II's, IIx's, IIcx's, and SE/30's do work, your assertion is empirically disproved. If you were correct, then my 40MHz Mac II which has 4x100ns and 4x120ns memory would not even start up. R. ~~~~~~~~~~~~~~~ Rich Siegel Staff Software Developer Symantec Corporation, Language Products Group Internet: siegel@endor.harvard.edu UUCP: ..harvard!endor!siegel "When someone who makes four hundred and fifty dollars an hour wants to tell you something for free, it's a good idea to listen." ~~~~~~~~~~~~~~~
d88-jwa@nada.kth.se (Jon Watte) (02/02/90)
In article <7852@hubcap.clemson.edu> jcocon%hazel.cs.clemson.edu@hubcap.clemson.edu writes: >All 16MHz Macs will run as fast with 120ns chips as with 70ns chips. The >wait states are hardwired, regardless of memory speed. >Apple is not dumping. 80ns chips are really a waste in anything less than >a IIci, unless you get an accelerator card. >Apple deserves flaming for somethings, but this isn't one of them. Nope ! A 16 MHz machine will not run at all if the SIMMS are 120 ns, since the data won't be there when the CPU wants it. "Luckily" for customers with 120 ns SIMMS in their SE/30's (I have 100 ns, I checked) the classification temds to be conservative and allow for temperature changes and aging. It _is_ bad of apple to put 120 ns SIMMS in a 16 MHz machine. Period. h+ -- --- Stay alert ! - Trust no one ! - Keep your laser handy ! --- h+@nada.kth.se == h+@proxxi.se == Jon Watte longer .sig available on request
george@swbatl.UUCP (6544) (02/02/90)
In article <2827@draken.nada.kth.se> d88-jwa@nada.kth.se (Jon W{tte) writes: >In article <7852@hubcap.clemson.edu> jcocon%hazel.cs.clemson.edu@hubcap.clemson.edu writes: > >>All 16MHz Macs will run as fast with 120ns chips as with 70ns chips. The >>wait states are hardwired, regardless of memory speed. > >>Apple is not dumping. 80ns chips are really a waste in anything less than >>a IIci, unless you get an accelerator card. > >>Apple deserves flaming for somethings, but this isn't one of them. > >Nope ! > >A 16 MHz machine will not run at all if the SIMMS are 120 ns, since >the data won't be there when the CPU wants it. > >"Luckily" for customers with 120 ns SIMMS in their SE/30's (I have >100 ns, I checked) the classification temds to be conservative and >allow for temperature changes and aging. > >It _is_ bad of apple to put 120 ns SIMMS in a 16 MHz machine. Period. > Double Nope! The first posting is correct. Only the IIci needs the fast chips. According to the Macintosh Family Hardware Reference, the Mac II (no mention of IIx, IIcx, IIci, or SE/30 in my book) requires 120ns chips or faster. Of course, this is a 16MHz machine (approximately). According to TN 176, the IIci requires 80ns or faster. According to TN 230... "Although the 68030 is capable of a burst mode to more efficiently access contiguous blocks of memory, this feature is not enabled on the Macintosh SE/30. Enabling this feature would require significantly more complex control logic and faster (read ''more expensive'') RAM." The "more expensive" is theirs, not mine. Anyway, I don't see any mention of speeds in my documentation other than what is noted above. Unless someone with more complete resources can show otherwise, I would have to say that 120ns RAM is fine for everything but the IIci (Ignoring the portable here, doesn't it use SRAM?) -- / George D. Nincehelser \ uunet!swbatl!george \ / / Southwestern Bell Telephone \ Phone: (314) 235-6544 \ / / / Advanced Technology Laboratory \ Fax: (314) 235-5797 \ / / / /\ 1010 Pine, St. Louis, MO 63101 \ de asini umbra disceptare \
d88-jwa@nada.kth.se (Jon Watte) (02/03/90)
In article <1463@husc6.harvard.edu> siegel@endor.UUCP (Rich Siegel) writes: And many others with him. >In article <2827@draken.nada.kth.se> d88-jwa@nada.kth.se (Jon W{tte) writes: >>A 16 MHz machine will not run at all if the SIMMS are 120 ns, since >>the data won't be there when the CPU wants it. > Nope! yourself. Your calculator may tell you one thing, that the >beg to differ; the memory is accessed at one wait state, and accesses are >additionally bottlenecked by little things like bus arbitration. Yeah, right, wait states. I stand corrected. In fact, two cycles on a 16MHz machine uses ~125 ns. Plenty of time there... Though, with lots of memory and poor ventilation in a cramped SE/30, I still feel better with a little margin... Yes, I'm paranoid, I even back my HD up, as often as every one or two weeks ! > If you were correct, then my 40MHz Mac II which has 4x100ns and >4x120ns memory would not even start up. That's, uh.... many wait states. :-) Wonder why that's not mentioned in the ads ? h+ -- --- Stay alert ! - Trust no one ! - Keep your laser handy ! --- h+@nada.kth.se == h+@proxxi.se == Jon Watte longer .sig available on request
jimvons@ashtate (Jim von Schmacht) (02/03/90)
How can that High Speed SE/30 work with those Dastardly Low Speed 120ns Simms?
Easy - look at the following table and be enlightened:
680x0 Clock Rate Required Memory Speed
8mhz 150ns
16mhz 120ns <-- HINT HINT
20mhz 100ns
25mhz 80ns
33mhz 60ns
The SE/30, having a 16mhz 68030 requires 120ns simms (the Mac II has had 'em
for years). You weren't dumped on, you got what you paid for. Faster memory
is USELESS unless the processor is fast enough to require that speed. Oh,
and by the way, they WERE'NT cheap chips. Remember when apple bought memory
a while back thinking prices wouldn't come down and took a bath. Don't flame
Apple for doing business in a profitable manner (i.e. if it don't need wings
on the sucker, why put 'em there) - or do you believe in the universal
altruism of mankind?
--
Jim von Schmacht Senior Member, Project Test Staff Ashton Tate Corporation
Disclaimer: Standard Issue
--------------------------------------------------------------------------------
"It isn't the years - it's the mileage" -Indiana Jonesfozzard@boulder.Colorado.EDU (Richard Fozzard) (02/03/90)
In article <2829@draken.nada.kth.se> d88-jwa@nada.kth.se (Jon W{tte) writes: > >Yeah, right, wait states. I stand corrected. > >In fact, two cycles on a 16MHz machine uses ~125 ns. Plenty of time >there... Though, with lots of memory and poor ventilation in a cramped >SE/30, I still feel better with a little margin... Here's an interesting note: when I bought my se/30, i tranferred in to it 2 megs from an old SE, marked "-15", which i presume means 150ns. (I also put in 2 meg from a mac II marked "-12"). I have been using this w/o any problems for about 9 months, and i routinely fill up memory with many different programs. Dont ask me why it works, but it does. rich ======================================================================== Richard Fozzard "Serendipity empowers" University of Colorado/CIRES/NOAA/ERL/FSL fozzard@boulder.colorado.edu (303)497-6011 or 444-3168
d6maca@dtek.chalmers.se (02/04/90)
In article <15@hite386.UUCP> steve@hite386.UUCP (Steve Hite) writes: >BTW, I think the Apple EEs did an EXCELLENT job to keep a 16MHz machine >from choking on 120's with no parity errors...can you do it with a IIci :-)? Apple don't use parity checking, except on the IIci. BUT, can someone tell me why I can use 150ns memory in my MacII? (for over two years and no problems) - Martin Carlberg - Chalmers University of Technology, Gothenburg, Sweden
philip@Kermit.Stanford.EDU (Philip Machanick) (02/04/90)
In article <1990Feb3.161313.7380@mathrt0.math.chalmers.se>, d6maca@dtek.chalmers.se writes: > > Apple don't use parity checking, except on the IIci. BUT, can someone tell > me why I can use 150ns memory in my MacII? (for over two years and no problems) > > - Martin Carlberg > - Chalmers University of Technology, Gothenburg, Sweden The speed rating of the RAM is an indication of expected reliability; there's no widget inside the chip that says, "Hey, I'm being clocked too fast", and turns on a fire alarm. Running 150ns RAM in a Mac II (or SE/30) is like using a single density diskette as double density. There's nothing to stop you doing it, but fear of pushing something beyond/close to its maker's design tolerances. Philip Machanick philip@pescadero.stanford.edu
jdevoto@Apple.COM (Jeanne A. E. DeVoto) (02/04/90)
In article <1990Feb3.161313.7380@mathrt0.math.chalmers.se> d6maca@dtek.chalmers.se (Martin Carlberg) writes: > [...] BUT, can someone tell >me why I can use 150ns memory in my MacII? (for over two years and no problems) When a chip is marked 150 ns, it means the chip is certified for at least that speed. Some parts may actually be faster; fast enough to work at 120 ns. HOWEVER, it's not a good idea to use them at less than their rated speed, unless you feel lucky. A chip may work at 120ns in favorable conditions early in its lifetime, only to fail at the higher speeds later on. With RAM prices so low, I'd suggest anyone who is using low-speed parts should seriously consider replacing them. Even if you haven't had trouble yet, why take a chance? -- ====== jeanne a. e. devoto ======================================== jdevoto@apple.com | You may not distribute this article under a jdevoto@well.UUCP | compilation copyright without my permission. ___________________________________________________________________ Apple Computer and I are not authorized | CI$: 72411,165 to speak for each other. | AppleLink: SQA.TEST