Page 4 of 5 FirstFirst 1 2 3 4 5 LastLast
Results 61 to 80 of 97

Thread: Really Worried about heat issues.....

  1. #61
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default Thanks Costa

    Costa

    No need to apologize. I too have been extremely busy, and only now was getting around to re-installing the replacement Giga CPU........ until I found out last night that someone forgot to include the little screws that hold the fan in place on the Giga card.

    While I'm waiting for Giga's distributor to mail the screws to me, I can get busy with testing my memory sticks...... this will be good, because I can test the memory sticks with the stock 867 Mhz and with the Giga Dual CPU replacement as well.

    I will be sure to let you know how the tests come out.

    Many thanks

    Boyd 9

  2. #62
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default Initial Memtest Results

    Here are the initial test results, as suggested by Costa. All tests were run in single-user mode using memtest 4.05M.

    Test 1)
    J21 = Hynix stick
    J22 = 'Good' Micron stick
    J23 = 'Faulty' Micron stick

    There were massive test failures during the 2 test sequences I ran. Tests that failed included: Compare Sub; Solid Bits; Block Sequential; Checkerboard; Bit Spread; Bit Flip; and Walking Ones.

    Very interestingly, on the first sequence both Walking Ones and Walking Zeroes passed. On the 2nd test sequence, Walking Ones failed...... it does indeed pay to follow the instructions and run the tests multiple times!

    After the test was over, it gave me an 'error code 4', but I have not been to the memtest website to see exactly what that means.

    Also, after the test was over and I typed 'exit' and hit return, the computer was not able to fully boot up into Mac OS X (it froze/hung up on me)...... I had to shut the computer down by holding down the power button.

    Test 2)
    J21 = 'Faulty' Micron stick
    J22 = 'Good' Micron stick
    J23 = Hynix stick

    In single user mode, memtest allocated 1442 MB for the test, and then the computer hung/froze (i.e., memtest never got to the next part where it says that 'memory has been locked successfully'). I waited for at least 15 minutes for something to happen...... nothing did. I shut down the computer.

    Test 3)
    J21 = Hynix stick
    J22 = 'Good' Micron stick
    J23 = Empty

    I did this as a control test. 2 sequences in memtest. Both passed OK.

    Test 4)
    J21 = 'Good' Micron stick
    J22 = Hynix stick
    J23 = Empty

    I did this as yet another control test. 2 sequences in memtest. Both passed OK.

    I will make yet another attempt to run the configuration in Test 2, but as with previous troubleshooting, it already looks like a pattern is emerging.

    Best regards

    Boyd 9

    P.S. Just realized there's a newer version of memtest. Will try that with scenario #2 and see if anything happens.....

  3. #63
    Join Date
    Sep 2002
    Location
    Milan - Italy
    Posts
    525

    Default

    O.k. Boyd 9,

    thanks for the testing and follow up.

    Obviously the Micron "Faulty" module is the culprit. Junk it or see if you can still get a replacement for it (if you bought it through crucial.com, they should provide you with an RMA number to use).

    As for the Hynix module, as I wasn't able to check it's data sheet, we have to assume it has NOT been tested for refresh sequences around 6 ns (143 MHz). Therefore, you want to use that in J21 slot, so that your Memory Controller won't be pushing that frequency if and when needed.

    Remember, your internal MC will get it's timing parameters during the booting sequence by reading what it finds in the SPD chip found on the module fitted in the FIRST RAM slot. I am thus assuming the Hynix bears slightly "less" specifications than the other MICRON stick you have.
    This way, you will be forcing your internal MC to run at the Hynix best timing parameters, while your other MICRON stick (bearing better specs), will handle such timing effortlessy.

    If you need the full 1.5 GB of RAM for your computing needs and if you need to buy a new module, make sure you get the following minmum specs:
    • PC-133 tested.
    • CAS Latency (or just "CL") of 3 (or 3-3-3) at 133MHz.
    • Refresh sequence within 7.5 ns minimum
    • Max DRAM devices density 256 Megabits.


    So, ... you've experienced that even a BIG NAME like Micron can provide a "faulty" SDRAM module (provided it was assembled by MICRON/CRUCIAL) every "now" and "then". It happens... Just never be afraid and do your testing like you have done. Pinpointing the problem is like having resolved 50% of it.

    All the best.

    Costa

  4. #64
    Join Date
    Jun 2003
    Location
    UK
    Posts
    1,317

    Default

    Ciao Costa,

    good to see you back, we've missed you.

    Grazie mille.

    biggles.
    "illegitimis non carborundum"

  5. #65
    Join Date
    Jan 2002
    Location
    NW Montana
    Posts
    8,197

    Cool

    Costa Dude, your amazing. Thanks for all your help on the forum and the things we all learn from it. I understand what you say ... but no clue how you get there. - Randy

    Just added... Darn you guys talk funny over there. Just kidding.

  6. #66
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default Utterly flabbergasted.... NOW what?

    Thanks, everyone (especially Costa!) for all the great information. This has been quite an education for me too!

    Having said that, I really don't know what to make of the following test results...... any takers?

    I downloaded the latest version of Memtest (4.1) from their website, and according to Costa, tried this configuration....

    J21 = 'Faulty' Micron stick
    J22 = 'Good' Micron stick
    J23 = Hynix stick

    As you'll recall, this was test #2 from yesterday, that I had TRIED to run but couldn't get it to work. This time, in single-user mode, memtest DID launch successfully. I ran three sequences of tests (took all day, just finished 15 minutes ago).

    ALL TESTS PASSED OK

    HUH????

    I'm so confused, I can hardly see straight. What, exactly does this mean?

    1) That the portion of the 'faulty' Micron chip that truly IS faulty wasn't available for Memtest to check, and so everything checked out OK?

    2) That I have some other, as yet untraced problem?

    Of course, I don't trust the 'faulty' Micron chip anymore...... and why do I always get test failures with memtest when it's in any slot but J21?

    Anyway, the screws finally arrived for the fan assembly for the Giga card, so I can go ahead and install the CPU upgrade, but, in the meantime, what do I make of the information I found out today? Absolute LAST thing on earth I expected to see!

    Boyd 9

    P.S.: FWIW, I did see significantly LESS crashes and kernel panics with the 'faulty' Micron stick removed, so I guess its back to 1 GB, I think. Unfortunately did NOT purchase from Crucial (which I now regret).

  7. #67
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Default

    and why do I always get test failures with memtest when it's in any slot but J21?
    Costa mentioned that J21 "sets" the basic timing that the memory controller will use. If that chip is slower, it 'sets' the controller to use that as the slowest common denominator. Or, the faulty area is in the first 80MB which memtest can't test.

    In any event, get an RMA for it. RAM has never been cheaper even though that will not be a concern. (at least for DDR DRAM it is pretty much below cost)

  8. #68
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default

    Thanks TZ,

    That sounds like a pretty good game plan..... and I feel like I've done enough troubleshooting as it is..... best to return it and get a new one.

    Again, many thanks to everyone. Hopefully, the Giga Dual CPU updgrade will go without a hitch!

    Boyd 9

  9. #69
    Join Date
    Sep 2002
    Location
    Milan - Italy
    Posts
    525

    Default

    Ciao Boyd 9 and hi everybody else,

    ... it's good to see you too... makes me feel still "alive", somehow...

    Quote Originally Posted by Boyd 9
    Having said that, I really don't know what to make of the following test results...... any takers?
    Yes, ... that shows you that quality products are hard to die...
    TZ got it right: Memtest skips the first 80MB of available RAM addresses on the DRAM devices during it's "thing"; your issue may be on those very 80MB.
    DON'T TRUST the module: go ahed and replace it. I am confident that even a Micron official would advise you the same.
    Where I work, if during our final quality testing over our production we have doubts over the standards one of our products is bearing, we wouldn't ship it until we would be 110% sure it meets all the specifications required. If further testing shows it does not or still leaves us with doubts, we simply don't ship it at all.
    IOW, play it safe and go for reliability.

    RWM,
    thanks for your kind words. As to understand how I got there, it's all my fault...

    TZ,
    I would love to go on completing that "Upgrading Guide" we have started together, but I am afraid my office duties will still keep me from "being regular" here on the fora's for some time.

    So, to everybody, be prepared to see me dissapear for another "while".

    Biggles,
    If you ever fly to Milan, drop me a note before hand: we may get to drink something together while you stop in town for your next morning's flight...?
    Rick has my coordinates... (I have your mail, already... )

    Take care and "ciao" for now.

    Costa

  10. #70
    Join Date
    Jun 2003
    Location
    UK
    Posts
    1,317

    Default

    Costa,

    how long does it take to drive from Torino to Milano?

    I sometimes fly to TRN but I'm only there for 30 minutes or so, just long enough to turn the airplane around back to Luton. But the idea of renting a car at TRN and making a long weekend of things has it's attractions. And I couldn't possibly do that without making a date for dinner with you and yours. I have a week off coming up at the beginning of July, is this something we could put together? I know you're really busy, and I haven't even mentioned the idea to Da Boss and the Under Boss (16 year-old daughter), both of whom I am sure could be convinced of the merits of a shopping spree in Milano! And then, of course, there's the small matter of food and wine, and how should one go about that if not with the expert guidance of a Milanese?

    I don't have your e-mail address, but mine's to be had in my profile. Drop me a line if you think this could happen.

    Ciao.

    biggles.
    "illegitimis non carborundum"

  11. #71
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default

    Hey! I'd love to go to dinner in Milano too! Swing by and pick me up, will ya?

    In the meantime, have installed the Giga Dual CPU card..... completely and utterly unstable at 1.46 GHz, so kicked it down to 1.33 GHz.... so far so good..... no kernel panics or general 'weirdness' at 1.33 as of yet (it sure is great of Giga to make user adjustable speed settings, isn't it?). Although I must admit my machine doesn't *feel* a whole lot faster, I can, indeed, multi-task like a son-of-a-gun! Wakes from sleep just fine, too!

    I've got an email in to Giga tech support with the hopes that maybe a voltage tweak will make it stable at 1.46 or 1.4 GHz, but in the meantime, the faulty Micron memory stick is another story....

    Mac HQ in St. Louis (where I bought the stick) won't take it back OR exchange it, because a) it's been longer than 30 days; and b) because they switched memory vendors just days after I purchased my stick (I wonder why!)....... I looked at the Micron website, and they say to take memory issues to Crucial..... I called Crucial, but they won't take it back because it doesn't have a 'Crucial' stamp on the stick...... looks like I'm stuck with a $100 memory stick and so much for trying to give business to your local, independent Mac dealer...... (sigh)...... On the bright side, maybe I can make a Christmas tree ornament or a key-ring out of it.....

    I think maybe I could impact my speed issues with the upgrade by cloning my boot volume - - currently residing on an old IBM 5400rpm 66 ATA HDD - - and switching to one of the other internal drives (7200 rpm 100 ATA Seagates). Anybody care to recommend their favorite cloning software? Has anybody tried the "Drive Genius" reviewed in latest issue of Mac Addict?

    A Million Thanks again to all you Gurus for guiding me through this process!!!!!

    Boyd 9

  12. #72
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Lightbulb

    When I buy from Crucial, I go to Crucial, get what they recommend, as that is the only way you know it will be mac-compatible programmed and lifetime, If it didn't have Crucial sticker, it was sold because crucial didn't want it or like it - how can they sell you memory and say it is crucial's anyway? I'd bet it is more likely 'generic' SDRAM.

    Having all your memory of same spec and speed does make a noticeable difference.

    Definitely get your system off that older drive, you want 50% free space on boot volume, 8MB cache buffer, I'd go with getting another drive soon.

    Best cloning method has to be Disk Utility -> Restore. (There are other programs but why? TriBackup (4.03) makes sure that the ownership flag on the volume is off (which would prevent it from being bootable). Get Info on Target volume, make sure "Ignore Ownership on this volume" is off, not checked) and you are ready to proceed.

    There are a number of backup programs and tips in the Troubleshooting Guide.

    You might just want to do a clean install and then, later, import or synchronize your user folder. Personally, I'd leave or put /Users on a 2nd disk drive instead, your system will run better, and that keeps the OS separate in case you need to restore. I would also clone your clean system once you have it with just the basics, to another volume for backup or as an emergency volume.

  13. #73
    Join Date
    Jan 2002
    Location
    NW Montana
    Posts
    8,197

    Default

    I think maybe I could impact my speed issues with the upgrade by cloning my boot volume - - currently residing on an old IBM 5400rpm 66 ATA HDD - - and switching to one of the other internal drives (7200 rpm 100 ATA Seagates).
    Absolutely swap the 5400 RPM drive out or use it as backup and emergency boot/repair drive. You will see a nice bump 2-3 times as fast.

    Read about different drives and performance at Storage Review. - http://www.storagereview.com/comparison.html

  14. #74
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default

    Yeah, now that one of my memory sticks has bit the dust, I'm definitely thinking of all new memory sticks.... either from MacGurus or Crucial.... some place that guarantees for life, at the very least! But I can limp along with 1 GB until prices drop some more (and to think that my original Mac Plus shipped with 1 MB.... how ungrateful I must seem )

    I think it's also time I looked seriously at a SATA drive, now that I've got a nice dual Giga under the hood. (I guess I'll be cruising the SATA/PATA forum next to start gathering info).

    The reason I asked about cloning software is because the last time I tried it, it didn 't go well..... but now I'm thinking it was probably because of the memory issues.

    Oh, BTW, when I originally posted I was worried about heat issues...... I replaced the HDD ATA ribbon cables with the rounded variety...... this has significantly impacted internal air flow for the better (the ribbon cables were partly blocking one of the fans, I'm pretty sure)..... I also yanked the ZIP drive out, and moved one of the 3.5" HDD into there instead..... maximum temperature I've seen so far with the new Giga is around 37 to 38 C (this was while burning a DVD in the superdrive), and most times it stays around a comfortable temp. of around 29-33 C. SWEET!

    Again, thanks for the help/info

    Boyd 9

  15. #75
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Default

    Memory prices and unlikely to drop, they are in some cases losing money and below price. There is even a glut of DRAM on the market.

    The zip bay doesn't get much air, if you use the optical cable, it is not rated for ATA-4 or ATA/66. I'd use FW if you must use the drive.

  16. #76
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default

    TZ,

    That's nice to know (and good incentive!) about memory prices..... seeing as how I saved all that money installing the Giga myself, why, I can save even more money and buy all new memory chips too (the corresponding increase to my productivity will surely be worth it, right?)

    Lest you think I've rounded the bend, putting the 80 GB HDD in the Zip bay wasn't about airflow..... I had other ideas in mind..... Since the Tempo ATA 133 Card sits in PCI slot 4, it was no problem to run a rounded cable from there to the back of the HDD in the Zip bay, leaving the superdrive as the only device on the optical cable.

    My other line of thinking was that the Zip bay sits relatively close to the new Giga card..... the other HDDs sit on the bottom of the case (where, in theory anyway, cold air comes in). My idea was to monitor the temperature of the HDD in the Zip bay to give me an idea of just exactly how much heat the Giga card is generating.... Of course, this assumes a lot. I don't, for instance, know if the SMART temperature sensor is on the top, front, side or back of the 80 GB HDD.

    Whatever the case, I'm happy, because things seem to be going well, the inside of my QS case is less cluttered, and if that has impacted the air flow for the better (which I think is the case), then all's well that ends well.

    Besides, with one HDD in the Zip Bay, I still have room (if the PSU will cooperate) for two more HDDs (hopefully SATA drives, with a brand new SATA card from Mac Gurus).

    Boyd 9

  17. #77
    Join Date
    Jan 2002
    Location
    NW Montana
    Posts
    8,197

    Default

    Damn, boyd 9, I wish my QS and Giga was that cool. (it's the older hotter drives not the Giga) I would be ing! My slot fan bearings just died on me last week or more .. and we finally got some hot weather. I have 3 hard drives and with the fan they run around....

    Bottom Rear - the stock Seagate 40 gig - ~37
    Top Rear - IBM 60GXP 60 gig - ~41
    Center/Middle - new Seagate SATA 120 gig - ~38

    Without the slot fan...

    Bottom Rear - the stock Seagate 40 gig - ~38
    Top Rear - IBM 60GXP 60 gig - ~45 as high as 50 ouch!
    Center/Middle - new Seagate SATA 120 gig - ~41

    I need one new 120GB ATA drive and just pull my 40 and 60 gig drives out.

    Glad things are working good for you
    Randy

  18. #78
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Default

    Randy,

    I think I'm just pretty lucky w/ the heat issues..... that and the fact that my office is in the basement..... ambient year round temp of 68 F.

    There was tons of good advice here on different threads about heat issues that I followed also (also at xlr8yourmac)...... I pulled my QS away from the wall for better circulation in back of the case, cleaned out the dust bunnies from underneath, installed a PCI slot fan, plus the cables, I'm sure made a difference.

    Previously, with the ATA ribbon cables, I know they were partly blocking the inlet for the case exhaust fan. Now with one HDD in the Zip bay, and the 2 remaining HDDs in the bottom of the case (side by side), the new rounded ATA cables tuck neatly up underneath the exhaust fan when you close the case..... no blockage of the fan port surely means better airflow inside. Just doddling around today, the hottest temp registered 35 C..... I couldn't be more pleased!

    Anyway, I may be finished with adding internal HDDs in my QS..... I was drooling today over the Hot Swap SATA drives and enclosures..... maybe Santa will come early this year???

    While I still have not investigated 1.4 GHz on the Giga card, it seems ABSOLUTELY rock solid at 1.33 GHz....... and I couldn't have done it w/o all the help from all you gurus!

    Good luck with cooling your rig!

    Boyd 9

  19. #79
    Join Date
    May 2005
    Location
    Stinking Desert National Memorial
    Posts
    394

    Question What's Up with this?

    While testing out my new Dual Giga at 1.4 GHz, it seemed somewhat unstable at that speed, so thought I'd run a comparison at X-bench between the processor set at 1.33 and at 1.4, with everthing else identical, just to see what I'd find.

    The final scores were mostly identical (144.19 vs 142.10). For most things, obviously, the 1.4 was a bit to significantly faster. But for the part of the test that looks at the HDD, the 1.33 scored better.

    Out of 10 Hard Disk Tests, the 1.33GHz scored better than the 1.4GHz on 9 of them (that would be the Uncached Read of 4K Blocks).

    Is this odd in any way?

    Everything else about the tests was identical, and yes, I have installed the software that Giga provides for correct reporting of the 1.4 GHz speed.

    Any opinions or thoughts much appreciated.

    Boyd 9

    P.S./BTW - - Following the Gurus advice (esp. TZ), have opted for three new sticks of RAM and I've decided to go SATA! This all comes after a successful clone of my system from the 40 GB 5400 rpm IBM to the 80 GB 7200 rpm Seagate. Just as an example to say how much difference the better HD makes, my X bench scores went from 95 (stock 867 MHz QS) to 125 after the Giga install and from 125 to 142 after switching to a faster HD!

  20. #80
    Join Date
    Jan 2002
    Location
    NW Montana
    Posts
    8,197

    Default Hummm

    Processor cache stays the same right?
    Your disk drives were the same?

    boyd9, I'd just be guessing.

    Try Quickbench from this Google search and see if it scores the 1.33 higher than the 1.4 GHz.

    Randy

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •