Results 1 to 18 of 18

Thread: Start-up: drive spins up, down, all around . . .

  1. #1
    Join Date
    Mar 2002
    Posts
    9

    Default

    Two Problems:

    (1) My recent internally upgraded hard drive spins up for a few seconds, then spins down for around 10-20 seconds (grey screen), spins up again, flashing question mark appears X 3, frozen question mark for a few seconds, smiley face, then MacOS 8.6 appears as welcome as can be, then all is swell.

    But - gee whiz! Is this normal?

    Or is this a . . . Bad Thing?

    What's it all mean?

    The drive then seems to run just fine. If I then restart, the restart sequence goes normally without any of the above delays.

    Earlier drive back in the same spot starts up normally, with none of the delay or flashing question mark behaviors that the "new" drive exhibits.

    (2) Is this drive pokier than it ought to be?

    ATTO Pro-Tools 2.6, using a Max Transfer Rate of 8 MB, documents the following unblazing performance:

    Peak Read 3.58 / Sustained Read 3.57
    Peak Write 3.58 / Sustained Write 3.57

    I would have thought it a little perkier on the PowerMac 7200's internal 10 MB/sec SCSI-2 bus. But this drive, although old (circa 1995?), is new to me - I don't really know what transfer rates it should be achieving
    __________________________________________________ ______

    Troubleshooting information:

    (1) The Startup Disk Control Panel is highlighted to the "new" startup drive.

    (2) The Computer: PowrMac 7200 / 120 MHz / 256 cache / 256 MB RAM / 2 MB VRAM / MacOS 8.6. Note: a correspondent reports identical weird startup behavior with identical drives in his G3 upgraded 6115s - so it seems to be the drives themselves that are the problem, not bad SCSI ribbons or other individual computer defects. He reports that the drives seem to work fine so he just doesn't worry about it. Until I at least begin to understand what is going on, however, I will remain somewhat concerned.

    (3) The drive upgrade: I just replaced
    (a) the recent internal 1 GB IBM Ultrastar 2XP DPES-10380 50-pin (no adapters) single-ended SCSI-2 Fast 5400 RPM hard drive which always spun up normally, with
    (b) the "new" internal 4.55 GB IBM Ultrastar 2XP DCHS-04F 50-pin (no adapters) single-ended SCSI-2 Fast 7200 RPM hard drive which always spins up with this delay and flashing question mark.

    (4) The SCSI chain:
    (a) Internal: Motherboard(Terminated)*================= "new" hard drive(Terminated), ID 0. (I have temporarily taken the CD ROM at ID 3 out from the middle of the internal SCSI chain to simplify SCSI troubleshooting).
    (b) External: unplugged at present to simplify SCSI troubleshooting.

    (5) Drive jumper settings:
    (a) Jumpered: Autostart, Active Termination, Disable Unit Attention, Disable Sync Neotiations (also did a little experimenting - no discernible effects)
    (b) Not jumpered: Termination Power Enable (Thank you, Mrgregster!)

    (6) The drive is fully recognized and supported via Apple's Drive Setup 1.7.3. Drive testing runs fine, deep level format/zero all data runs fine, putting on the 8.1.2 Apple driver goes fine. Disk First Aid says all is well. Putting on a fresh MacOS 8.6 goes fine. But none of these actions change the weird delayed startup behavior.

    (7) The drive is fully supported by FWB Hard Disk Toolkit 3.0.2. AutoInitializes. Full overnight testing with all deep level options goes fine. The FWB 3.0.2 drivers run fine. Putting on a fresh MacOS 8.6 goes fine. But none of these actions change the weird delayed startup behavior.

    (8) The drive is reported as "Sector size 512 / User Defects 0" by Norton Utilities, which is good. All the following indicate that the drive is fine: DiskWarrior 2.1.1, Disk First Aid, Hard Drive Setup 1.7.3 (full testing and deep level format, drivers updated to v. 8.1.2, fresh MacOS 8.6), FWB Hard Disk Toolkit 3.0.2 (overnight testing, then Auto Initialize, Drivers updated, fresh MacOS 8.6), TechTool 2.5.5 (overnight full testing hardware and software, full expert mode), Norton Utilities 6.0.3.

    (9) PRAM has been zapped and desktop rebuilt, both via Tech Tool Lite 3.0.2. Also, 'cuda button held down for over ten seconds.

    (10) A second, identical drive replacing this balky drive does the same delay / flashing question mark dance on startup.

    I confess that I am almost beginning to run out of troubleshooting ideas. ;-)

    Your guidance would be most appreciated!

    Thank you,

    George2

    [This message has been edited by George2 (edited 27 June 2002).]

  2. #2
    Join Date
    Jan 2001
    Posts
    714

    Default

    IBM had problems with the Ultrastar line so you probably got a lemon drive. Find out if there is warranty left on the drive and contact tech support quickly for a replacement. Is it too late to get a refund? Best wishes.

  3. #3
    Join Date
    May 2001
    Location
    Lanham, MD 20706 USA
    Posts
    28

    Default

    George-
    I would remove the jumpers from Disable Unit Attention, Disable Sync Neotiations. (Internal single drives have very few jumpers installed).
    Keep jumpers on Active Termination and Autostart.
    I would also re-connect the CD-Rom for testing, because you want all the internal stuff to work together.
    Since the disc tests good and eventually starts up, I think the jumper settings are the best place to start.
    It is possible that the internal bus is terminated somewhere other than the hard drive, but this is not likely.
    Also, with a 7200 rpm drive, I think you should get better performance. But not sure. I think you should be getting around 6-7 MB/sec.
    This sure is fun, isn’t it?
    -Gregster

  4. #4
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Default

    You're 5 yr old "new" drive online specs:
    http://www.storage.ibm.com/hdd/support/dchs/dchstek.htm

    Check the link to the jumper settings as a start. There is an install guide etc. Also, why use HDT 3.02? Not a great driver even in its day it had trouble with hfs+ so I would grab the Drive Setup from 9.1 (v. 2.07) to use with any SCSI or IDE drive, even under an older system like 8.6. The most current FWB HDT does have some nice features and uses though and worth keeping around and upgrading - ESPECIALLY if there is a problem where the drive's mode pages are invalid or have corrupt data. I've avoided RMA's a number of times with using HDT 4.5.

    And while the Deskstar has had trouble (as have every ATA drive), the ultrastar line is not 'prone' to problems. Their current Ultrastar has excellent self-test, SMART, and you will hear it do a check every ten minutes of idle time (very reassuring once you know what and why).

    Gregory

  5. #5
    Join Date
    Mar 2002
    Posts
    9

    Default

    [Edit - two weird things now going on - the material I quoted from Angela's post is getting cut out, and much of my post is showing up in italics, even though I am not posting in italics. I don't even like italics. Strange. any ideas?]

    __________________________________________________ ______

    Angela wrote:

    <>

    Angela,

    Thank you for your thoughts and your kind wishes.

    This particular startup peculiarity is happening with four different identical IBM DCHS-04F hard drives - two in my possession (one Mac) and two in a correspondent's possession (two Macs) - so I think it is probably not a problem with an individual defective drive.

    Also - I think that the tremendous reliability problems we are hearing about lately with IBM drives are of relatively recent origin - and these are old drives, pre-current troubles. Probably many years out of warranty.

    Also - the drives seem to work fine once they finally spin up - if, perhaps, a little slow.

    Also - all the hard drive utilities I have used indicate that the drives are fine.

    I think that the problem may be something . . . something deeper . . .

    __________________________________________________ _______________

    mrgregster wrote:

    <<< Keep jumpers on Active Termination and Autostart. >>

    Will do.

    I jumpered Disable Unit Attention and Disable Sync Negotiations following the instructions on both the IBM and the LaCie sites to do so when using these drives in Macintoshes. Both sites are both quite clear on the matter. Of course - the same sites clearly recommended jumpering "Termination Power Enable," which is an apparent no-no on Macintoshes as per advice given repeatedly over the years on this board - so what do IBM and LaCie know?

    Disable Unit Attention and Disable Sync Negotiations are now unjumpered.

    Mac is turned off and turned on again . . . and . . . and . . . no change. :-(

    Same startup spin up/spin down/spin up/flashing question mark/smiley face/Mac OS 8.6.


    << I would also reconnect the CD-Rom for testing, because you want all the internal stuff to work together. >>

    OK - I have the CD ROM in my hand, even as we speak. It is an "Apple CD 12X speed CD ROM, 12-X SCSI, March 1997." As far as I know, it came with the machine and has not since been touched by human hands until I tore it from it's longtime happy home. I feel like a heartless landlord. There are seven jumper sites on the back of the CD ROM, which I never even knew were there - I will describe their current status. Oh - this CD ROM has always showed up as "ID 3."

    Parity - jumpered
    0 - jumpered
    1 - jumpered
    2 - unjumpered
    Unlabeled - unjumpered (odd - this one has only a single pin - maybe someone got hungry)
    Unlabled - unjumpered
    Term Power - jumpered

    Two issues:

    (1) I vaguely remember something about Mac SCSI not using parity, but if that were true, an Apple CD ROM shouldn't even have this jumper on it - should parity be jumpered or no?
    (2) As I understand it, Term Power should not be jumpered - but this seems to be how this came from the factory - unless someone just tossed this in here from another Mac and did not know what he was doing. Should this be jumpered, or no?

    The above two questions were why I disconnected the CD-ROM - I thought it might be messing up the SCSI chain, and didn't want to be concerned with the possible complications. But if you can tell me how to jumper it correctly, I will be happy to put the CD ROM back on the chain.


    << Since the disc tests good and eventually starts up, I think the jumper settings are the best place to start. >>

    Done.

    Looks like it's time for Plan B.


    << It is possible that the internal bus is terminated somewhere other than the hard drive, but this is not likely. >>

    Not sure what that means. Maybe you are referring to the way the internal CD ROM jumpers are set?

    With nothing else on the internal SCSI chain currently besides the motherboard on one end and the "new" terminated hard drive on the other end of the internal cable, I have done my best to simplify SCSI termination issues.

    But I will be happy to put the CD ROM back in the middle of the internal SCSI chain as soon as instructed in regard to the correct jumper settings.


    << Also, with a 7200 rpm drive, I think you should get better performance. But not sure. I think you should be getting around 6-7 MB/sec. >>

    This is my thinking, as well.

    It is a puzzlement!


    << This sure is fun, isn’t it? >>

    It's OK!

    When MacOS 10.x.x matures, I expect to get a G4 (shhh . . . don't tell my faithful PowerMac 7200). But in the meantime, I have become much bolder about experimenting with my near-end-of-life PM 7200: it has now become my SCSI learning box. My intention is to do everything on it wrong that one can do with SCSI, and then learn how to set it right - kinda like the surgeon in Gravity's Rainbow - bring it as close to death as possible, then bring it back from the grave. Leaping before I looked and buying cheap used obsolete IBM drives off of eBay seemed like a good beginning in the doing-everything-wrong direction . . . though I suppose I could have gone even farther and bought a pair of 80-pin drives . . .

    Finding MacGurus, reading most of the last couple of years of MacGuru SCSI bulletin board postings, learning to use ATTO Pro-Tools, learning of running Seagate Cheetah hard drives off a Miles or an ATTO card, learning about Granite cables and terminators - all are the parts of the learning curve that have come about since I began this little project, so it has probably not been a complete waste of time. And I don't really learn this stuff by just reading - I gotta get my hands dirty. I figure that I'll get 'em dirty cheap on my PM 7200, and I will then have the beginnings of knowledge to do 'most everything right with a new G4 - or, at least, know how to find the needed knowledge when things go wrong.

    So, yeah - it is kinda fun. No time pressure or desperation - this 7200 isn't going anywhere fast, and if something goes disastrous, I can just toss the old hard drive back in. Or if I actually destroy the thing - I buy the G4 a little sooner than planned.

    Everything that goes wrong with this box is something new to learn how to do right.

    Yeah - this is OK!


    So - next step?


    Thanks again,

    George2

    [This message has been edited by George2 (edited 27 June 2002).]

  6. #6
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Default

    Whatever said, narrow drives need to have a jumper on the pins for termination on the end drive. If there is a CDROM and a SCSI hdd, whichever is last.

  7. #7
    Join Date
    May 2001
    Location
    Lanham, MD 20706 USA
    Posts
    28

    Default

    George-
    Might be worth checking the jumper specs on the old HD to compare with how you are jumpering the new one. Very difficult for me to contradict the IBM and LaCie sites about recommended settings, but I am using two IBM drives in a 7600 and aside from ID settings, I think all I needed to do was enable autostart. (The hard drives would not wake up from sleep, now they do). [Can’t check it now, I’m not at home]
    I have no clue what to tell you about the CD drive jumpers, other than to say if it worked before, I would leave ‘em alone.
    Have you tried the recommended (IBM, LaCie) jumper settings: Term Power Enabled, et al?
    -Gregster

  8. #8
    Join Date
    Mar 2002
    Posts
    9

    Default

    Gregory wrote:

    << Whatever said, narrow drives need to have a jumper on the pins for termination on the end drive. If there is a CDROM and a SCSI hdd, whichever is last.Yeah - that's what I figure. >>

    The drive is currently the last and only device on the internal ribbon, and is terminated.


    << You're 5 yr old "new" drive online specs: http://www.storage.ibm.com/hdd/support/dchs/dchstek.htm
    Check the link to the jumper settings as a start. There is an install guide etc. >>

    Thank you, sir.

    That is one of the sources from which I obtained the information to Disable Unit Attention and Disable Sync Negotiations - and to Enable Term Power (whoops!) - and such.

    I haven't really looked into what they might have to say about the apparently slow transfer rates - I need to do that. But I have been focussing on the weird startup issue. First things first.


    << Also, why use HDT 3.02? >>

    An excellent question:

    (1) This may be part of (unintentionally) doing wrong things first. I was not yet a reader of MacGurus when I bought it. Plus, I have heard some good things about FWB HDT over the years. The bad things I have heard are mostly that it is both slow to upgrade and a bit gouging in regards to upgrades, in contrast to the free Apple Drive Setup or the pay once and upgrade forever Intech SpeedTools. But I got FWB cheap (obsoleted version - good only through Mac OS 8.6) and I don't intend to upgrade this box, so it seemed like a good idea at the time

    (2) HDT 3.02 offers "Full" support for the IBM hard drives I am using.

    (3) HDT 3.02 includes RAID 0 and RAID 1 software. If I can get these two identical "new" five-year-old 4.55 Gig IBM hard drives working right solo, the next step with this "learning box" is to learn about RAID 1 and RAID 0 by doing. I think that HDT will do the software RAID only when using the FWB drivers, but I am not certain about that.

    (4) FWB's HDT Test module with all the "Best" and deep level settings - overnight - was able to resurrect an old damaged hard drive (82 bad user blocks) that choked Apple Drive Setup 1.7.3, TechTool 2.5.5, and Norton Utilities 6.0.3. I am not using that hard drive except for fun - no vital data - but it is kinda neat to have a utility that can do things that others cannot do. One note - Apple Disk First Aid maintained that the damaged drive was fine throughout this episode, even while the other utilities were choking and dying left and right while noting the high number of bad disk blocks, some in vital areas.

    (5) The 300 page instruction manual has provided something of an education in regard to hard drives.

    << Not a great driver even in its day it had trouble with hfs+ >>

    Uh, oh . . . I am using HFS+ . . .

    Could you direct me to any more information on this possibly very important matter?


    << so I would grab the Drive Setup from 9.1 [v. 2.07) to use with any SCSI or IDE drive, even under an older system like 8.6.>>

    I thought that Drive Setup 1.7.3 was as far as one should go with MacOS 8.6. That Drive Setup 2.07 is for 9.x and such.

    Is it really OK to use more recent Drive Setups with Mac OS 8.6 on a PM 7200?

    Are there any advantages in using later drivers over Drive Setup 1.7.3 in this situation?


    <<< The most current FWB HDT does have some nice features and uses though and worth keeping around and upgrading - ESPECIALLY if there is a problem where the drive's mode pages are invalid or have corrupt data. I've avoided RMA's a number of times with using HDT 4.5. >>>

    I have delved into Drive Configure in FWB HDT 3.0.2. Tried setting it "optimize" the drive settings. But ATTO Pro-Tools 2.6 then graphed a widely sawtoothed jagged line on either reads or writes - it didn't seem to affect the numbers so much as the graph. I didn't know what to make of that, but it looked kinda like an EEG during a grand mal seizure, and that generally isn't considered good, so I went back to the original Auto-Initailized FWB 3.0.2 driver, and back to a smooth graph on ATTO's Pro-Tools 2.6. And all is calm and gentle once more.

    You use what you know.


    Gregory.

    Thank you for your help.


    All the best,

    George2

    [This message has been edited by George2 (edited 27 June 2002).]

  9. #9
    Join Date
    Mar 2002
    Posts
    9

    Default

    I have switched the drives around a bit and fiddled with jumpers and run some benchmarks. I think that the "old" 5400 RPM drive is running slightly faster than the "new" 7200 RPM drive. It seems to handle the drive buffer differently, as well.

    Hmmm . . . this may be a job for FWB Configure.

    But it is time to hit the sack - I'll try to report in tomorrow evening and we can try to figure out what to do next.

    "Old" drive jumpers - ID - 2, Enable Terminator, Disable Unit Attention. I just tried it on the end of the internal SCSI chain - everything spun up normally - no spin down, no flashing question mark.

    Gregster - if you have an internal SCSI CD ROM, could you let me know how the jumpers are set? I am particularly interested in regard to "Parity" and "Term Power." Also, anything on your IBM hard drives - particularly model numbers - and drive jumpers might prove helpful.

    Thanks again -

    George

  10. #10
    Join Date
    May 2001
    Location
    Lanham, MD 20706 USA
    Posts
    28

    Default

    George-
    Here are settings from a stock CD-Rom in a 7300:
    Parity: jumpered
    ID 0: jumpered
    ID 1: jumpered
    ID 2: OPEN
    (Next pin is a single pin): OPEN
    Next pin: OPEN
    Term Power: jumpered

    Did not crack open the 7600. Not a stock CD-Rom in there. Also, my IBM drives (circa 1999) are RAID 0, running off a Miles2. The CD is the only internal drive running on the Apple bus on that machine.

    BTW- Have you checked that (new) drive for bent pins? Just a thought.
    If you find one, you can carefully bend it back with a small screwdriver.

    In my previous post, I asked if you had tried the settings recommended by IBM and LaCie, including Enabling Term Power. I have to asssume you did that first. (?)

    I will be checking back, hopefully tomorrow, but surely by Sunday . . .
    Right now I have several Budweisers with my name on them.
    Good Luck

    Gregster
    ¬ó I thought I made a mistake once, but I was mistaken.

    [This message has been edited by mrgregster (edited 28 June 2002).]

  11. #11
    Join Date
    Mar 2002
    Posts
    9

    Default

    Gregster wrote:

    << Here are settings from a stock CD-Rom in a 7300:
    Parity: jumpered
    ID 0: jumpered
    ID 1: jumpered
    ID 2: OPEN
    (Next pin is a single pin): OPEN
    Next pin: OPEN
    Term Power: jumpered >>

    Exactly the same as the stock CD-ROM in my PM 7200. Looks like that's the way Apple meant it to be.

    Kinda goes against all I thought that I had learned about the motherboard providing term power, so don't provide term power on the rest of your SCSI chain.

    Am I wrong on that?

    << BTW- Have you checked that [new) drive for bent pins? Just a thought >>

    With the Big Guy - the 5" Bausch and Lomb magnifying glass - the one the mammographers use.

    The pins are straight as the day is long.

    Note also - what I am doing on one drive is more-or-less replicated on an identical drive - so it isn't a matter of an individually flawed drive. It must be something peculiar about this particular family of DCHS-04F IBM 4.55 GB 7200 RPM drives.

    A good thought.


    << In my previous post, I asked if you had tried the settings recommended by IBM and LaCie, including Enabling Term Power. I have to assume you did that first. [?) >>

    Yup. First things I did.

    Then experimented with some of these jumpers on and some off in various combinations. Also tried "delayed start," as well - saw an Internet post somewhere of someone's drive that would only work with delay start on, even if it shouldn't make any difference as SCSI ID 0 - go figure. Also tried jumpering "Term Power Enable."

    I could see no effect from any of these jumper settings, so between experimenting I am just leaving them as recommended by IBM (Autostart, Disable Sync Negotiations, Disable Unit Attention) and as MacGuru posts over the past couple of years repeatedly recommend (unjumpered: enable term power).

    You know -

    SCSI is beginning to make me think of women I have known. When everything is moving together, they are a beautiful sight to see. But when they are balky, there is just no figurin' -

    But ya gotta keep tryin'.


    << I will be checking back, hopefully tomorrow, but surely by Sunday . . .
    Right now I have several Budweisers with my name on them. >>

    Enjoy.


    I have a few more drive switching ideas and speed tests with different drivers to run before posting again and maybe giving you something more to work with.

    But this day is done.


    Thanks again,

    George


    [This message has been edited by George2 (edited 29 June 2002).]

  12. #12
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    George,

    Can you try something for me? Leave your CD-ROM jumpers as is for now. For the 2XP DCHS-04F (new drive) set the jumpers for Enable Autostart, Disable Sync Negotiations, Disable Unit Attention. Now partition the drive for the first partition not one bit or byte more than 2GB or 2048MB, but can be less. Second partition for the remainder available. Install your OS on the first partition of 2GB or less. See if that works.

    BTW, what IBM calls sync negotiation and in other parts of the docs calls the same thing T.I. sync. negotiation or target initiated synchronous negotiation is going to cause slower thruput. Asynch is much faster. So you do want to Disable Sync Negotiations. k

  13. #13
    Join Date
    Mar 2002
    Posts
    9

    Default

    Kaye wrote:

    << Can you try something for me? >

    Your wish is my unnegotiable demand.


    << Leave your CD-ROM jumpers as is for now.>>

    OK - the CD ROM is back in the computer, mid-SCSI ribbon, with stock jumpers.


    << For the 2XP DCHS-04F [new drive) set the jumpers for Enable Autostart, Disable Sync Negotiations, Disable Unit Attention. >>

    OK.

    One question before I begin:

    This drive will be on the end of the internal SCSI chain. Do you wish Active Termination Enable jumpered, as well?

    Just want to make sure:

    MotherboardT*=======CD ROM ============"New" hard drive (Terminated?)


    << Now partition the drive for the first partition not one bit or byte more than 2GB or 2048MB, but can be less. Second partition for the remainder available. Install your OS on the first partition of 2GB or less. See if that works. >>

    Is it OK if I do just one partition of 900 MB,then leave the rest of the drive unpartitioned as unused space?

    That way there will be no >2 MB partitions on the drive to cause any imaginable problems.

    Just trying to simplify.


    << BTW, what IBM calls sync negotiation and in other parts of the docs calls the same thing T.I. sync. negotiation or target initiated synchronous negotiation is going to cause slower thruput. Asynch is much faster. So you do want to Disable Sync Negotiations. >>

    That is helpful: this has been a matter of some confusion for me. It seems to be labeled "DIS TI SY" on this drive. That is what I have jumpered.

    _______________

    Gregory,

    As per your recommendation, I tried more recent Apple drivers. Apple Drive Setup 1.7.3 had updated the driver to v. 8.12. Apple Drive Setup 1.9.2 installed Driver v. 8.14 - I could see no obvious difference in the drive's behavior. Apple Drive Setup 2.03 claims to have updated the driver, but seems to have left it at v. 8.14. These updates may be doing lots of good things under the hood, but I am unable to see any obvious difference in my usage.


    Thanks, All.


    George

    [This message has been edited by George2 (edited 30 June 2002).]

  14. #14
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Yes, terminate "New" hard drive.

    Yes, OK if you do just one partition of 900 MB, then leave the rest of the drive unpartitioned as unused space.

    The sync neg/TI sync neg subject, this is what Mark James of SoftRAID told me when I asked him way back (I asked him about Tagged Command Queing as well):

    >Enable TI-SDTR
    >Installing a jumper enables Target Initiated Synchronous Data
    >Transfer Request Negotiation.

    Tagged command queuing is part of the SCSI spec, and should not require
    any jumpers. Synchronous communication will slow things down further.
    Asynch is much faster, and is the default on most all drives.

    k

  15. #15
    Join Date
    Mar 2002
    Posts
    9

    Default

    << The sync neg/TI sync neg subject, this is what Mark James of SoftRAID told me when I asked him way back [I asked him about Tagged Command Queing as well):

    >Enable TI-SDTR
    >Installing a jumper enables Target Initiated Synchronous Data
    >Transfer Request Negotiation.

    Tagged command queuing is part of the SCSI spec, and should not require
    any jumpers. Synchronous communication will slow things down further.
    Asynch is much faster, and is the default on most all drives. >>

    Kaye,

    Way, way over my head - for now, anyway - but perhaps one day I will understand it.

    Just a little more on the Disable Sync negotiations issue : "DIS TI SY" on the on-board "new" drive diagram (the IBM DCHS-04F 4.55 GB drive) is labeled "Disable Sync negotiations" on the more extensive jumper guide diagram for this particular drive on the IBM web site.

    The guide on the IBM support site for IBM drives running on the Macintosh operating system does state: << Ensure that the drive jumpers are set to disable sync negotiations. >>
    http://www.storage.ibm.com/hdd/support/rma/ckscsmac.htm

    ___________________

    << Leave your CD-ROM jumpers as is for now. For the 2XP DCHS-04F [new drive) set the jumpers for Enable Autostart, Disable Sync Negotiations, Disable Unit Attention. Now partition the drive for the first partition not one bit or byte more than 2GB or 2048MB, but can be less. Second partition for the remainder available. Install your OS on the first partition of 2GB or less. See if that works. >>

    Done.

    Please note that the setup you describe is more-or-less the configuration I started with except I have always terminated the drive on the end of the internal SCSI chain.

    The "new" 4.55 GB IBM DCHS-04F drive that has been exhibiting this odd spin up behavior (I actually have a handful of these "new" drives, so I may speak in the singular or in the plural at different times) has previously been in the computer with the CD ROM in the stock Apple CD ROM configuration with both the CD ROM and the hard drive jumpered as you have described (except for the hard drive being terminated). The hard drive was partitioned in seven equal 722 MB partitions (easy size for backups to a CD-RW, with a little breathing room), with the first volume as the MacOS 8.6 boot partition.

    But I am happy to try again from scratch - maybe something will come of it. Here goes:

    Step I - disconnected external SCSI chain, pulled PCI cards, disconnected printer, POTS and cable modems. Disconnected my 17" Samsung 700 IFT CRT monitor and hooked my old 14" Macintosh Performa Display into the back of the Mac. Down to a compulsively simple PM 7200 /120 MHz / 256 cache / 256 RAM / 2 MB VRAM / Mac OS 8.6 with nothing else going on to muddy the waters.

    Step II - reinstalled CD ROM with stock Apple jumpers as previously described. Reinstalled one "new" IBM DCHS 04F 4.55 MB drive onto the end of the internal SCSI ribbon with the jumpers set as follows: SCSI ID 0 (no jumper needed), Autostart, Active Termination Enable, Disable Sync Negotiations, Disable Unit Attention. Termination Power Enable is left unjumpered.

    MotherboardT*======CD ROM ID #3====="New" Hard drive ID #0 (Terminated)

    Step III - Started up from the MacOS 8.6 installation CD ROM via holding down "C" on startup. Using Apple Drive Setup 1.7.3 off the CD ROM, I updated the driver, restarted, and did a deep level format/ write all zeros with a single 900 MB partition. The rest of the drive was left as unused space. A fresh Mac OS 8.6 / Font Update was applied to the "New" drive. Set the "New" drive as the startup drive via the Startup Disk Control Panel. Shut down the computer completely.

    Step IV - started up the computer. Got the same few seconds spin up, then the same rapid spin down at about six seconds (almost as if someone was putting on the brakes), then about twenty seconds of quiet with a flashing question mark, then hard drive spin up, then smiling Mac, then the Mac OS 8.6 screen.

    Same as before - almost - but different. Before, I would get spin up, spin down, blank grey screen X 20 seconds, flashing question mark X 3, ten-second or so pause with frozen question mark, smiling Mac, MacOS 8.6.

    Now I get NO blank grey screen. The flashing question mark now begins much earlier, flashes more times (eleven?), and there is no frozen question mark between the flashing question mark and the smiling Mac.

    It is my perception that these changes may be due more to simplification and everything moving just a little quicker - only one partition on this "new" drive configuration vs. multiple partitions on the prior "new" drive configuration, much less on the nearly empty desktop now, no PCI cards currently installed - one or more of those sorts of things seems to be speeding things up.

    But the overall problem - the startup spin up/spin down/spin up/flashing question mark, smiling Mac, MacOS 8.6 - remains.

    On Restart, as before, all spins up normally, with none of these delays or spin downs or flashing question marks.
    ______________________________

    I have tried a few other things:

    (1) Replaced the "new" 4.55 Gig IBM DCHS-04F drive with the "Old" 1 Gig IBM DPES drive (jumpered for termination and Disable Unit Attention) at SCSI ID 0 - the "Old" drive started up the machine perfectly, as it always has.

    (2) Leaving the "Old" hard drive in the top bay at the end of the internal SCSI cable, I put the "New" drive back in in the lower drive bay unterminated with a SCSI ID # of "1."

    MotherboardT* ===== CD ROM ID #3 ======"New" Drive ID #1 ====="Old" drive ID #0 (Terminated)

    Set the "New" drive as the Startup drive in the Startup drive Control Panel. Shut down completely. Powered up - the "New" drive spun up, then down, then there was the some delay - not as much as before - no flashing question mark - the "Old" drive became the startup drive, despite not being chosen as such in the Startup Control Panel. Oddly - the "Old" drive was slowed down starting up the computer by perhaps 30 seconds while the "new" drive was spun down - so the "new" drive is delaying the other drive's startup on the same SCSI chain. On hitting "Restart," however, the "New" drive started up normally and became the startup drive.

    Another odd thing - when the "New" drive and the "Old" drive are both hooked up and I double click on the icon of the "Old" drive, the "Old" drive window opens but is empty of icons for about 20-30 seconds- cursor still moves via mouse direction, but clicking on things doesn't do anything - then the "old" drive icons show up and everything works normally. This does not occur, however, when starting up with extensions disabled - so there is some sort of extension conflict involved on this one - I haven't bothered to pin it down. Starting up from a turned off computer with extensions off, however, does not prevent the weird delays on startup with the "new" drive in any configuration - I have tried that multiple times.

    Another odd thing - when starting up from a CD ROM, the "New" drive goes through it's spinup/down/up delay thing - which is what these drives seem to do - but the weird thing is that it it also delays the CD ROM's spin up. Normally when you startup the computer while pressing down "C, " you can hear the CD ROM spinning up fairy rapidly - it vibrates the table just a little bit - with one of these "New" drives is on the SCSI chain, there is around a 30 second delay in CD ROM spinup.

    Another odd thing. When I put the "Old " drive in an external APS SCSI enclosure, I always turn the external SCSI on first before turning on the computer, and I can always hear the "Old" external drive spinning up as soon as I turn the external enclosure switch on. But when I put one of these "New" drives in the external enclosure and turn it on - it doesn't spin up at all. It just sits there. Silence. Until I turn the computer on and it goes through it's spin up/down/up song and dance. The fact that the "new" drives exhibit quirky behavior whether on the internal or the external bus indicates to me once again that this is a drive problem - not an internal SCSI cable problem.

    (3) Speed testing via ATTO ExpressPro Tools 2.6 gives further evidence that something is wrong with the "New" drive(s). In three out of four tests, the "Old" 1 GB 5600 RPM hard drive is faster than the "New" 4.55 GB 7200 RPM hard drive. Note that the PowerMac 7200's Internal SCSI bus is 10 MB/sec, the external SCSI bus is 5 MB/Sec.

    ATTO ExpressPro Tools 2.6 is set at Max Transfer Rate 8 MB, Expected Peak 10, Sample Size 1, System Disk Cache unchecked. I did not know if it would make any difference which drive I ran ExpressPro Tools off of, so I ran it off a 150 MB bootable RAM Disk running MacOS 8.6 wtih ATTO ExpressPro Tools 2.6 on the bootable RAMDisk. I include also the Bootable RAMDisk stats, just for kicks:

    -------------"Old" (Internal)--"New" (Internal)---"New" (External)--Bootable RAMDisk
    Peak Read---------3.73------------3.58---------------3.58 ----------------33.95
    Sustained Read---- -3.69------------3.57---------------3.57-----------------32.76

    Peak Write--------3.95------------3.57---------------3.57 ----------------14.05
    Sustained Write:----3.45------------3.57---------------3.57-----------------13.96

    Note that the "New" 7200 RPM 4.55 GB drives score poorer than the "old" 5600 RPM 1 GB drive in three out of four tests. That just ain't right.

    I get close to identical numbers on the "New" drives whether or not they are on the internal 10 MB/sec or the external 5 MB/sec SCSI chain, whether they are running the FWB 3.0.2 driver or the driver from Apple's Hard Drive Setup 1.7.3 (v.8.12) or 1.9.2 (v. 8.14), or whether or not the "new" drive is formatted as HFS or HFS+. I've tried 'em all.
    _________________

    An aside: the bootable RAMDisk.

    Seeing the above numbers, and noticing that the bootable RAMDisk seemed kinda perky, I threw a few more programs on the bootable RAMDisk - Norton and a Web Browser and such - I am writing you from the bootable RAMDisk now, and have spun down all the drives as I write - everything is spooky quiet in here, and the computer is running considerably faster than normal. A definite speedup compared to regular computing.

    It is thundering and lightening outside!

    I have a monster UPS with surge suppression/voltage regulation at my right foot ( APC Back-UPS Pro 1100), so I'm not worried about frying or disappearing disappearing the RAMDisk.

    Out of curiosity, I ran Norton's System Info "Hard Disk Performance" results on this PM 7200 running off a "New" hard drive vs. running off a RAMDisk:

    This computer running off a "New" IBM DCHS hard drive: 259 - Between a PM 9600/200 and a PowerCenter Pro 180 (hey - that's better than PM 7200 stock!)

    This computer running off a bootable RAM Disk: 1415 - Between a PM G4 400 and a PN G4 Cube 450 (zowie!)

    Hmmm . . . maybe I should get a newer computer, toss in lotsa RAM, and run off a bootable RAM disk?

    What are the fastest available computers that can run off of bootable RAMDisk?

    An 8600/300 and a 9600/350?

    Anybody ever run an ATTO ExpressPro Tools benchmark on a bootable RAMDisk on one of those machines?

    I just wish the Apple ROM would still support bootable RAMDisks.

    Ah well - I should look to the future, not to the past.

    Life only gets better. Ask anyone.

    End of Aside.

    ____________________

    I suppose I should focus on dealing with two matters:

    (1) The weird delay spinup/down/up on startup of these "new" drives.
    (2) Poor throughput on these "new" drives.

    But thoughts on bootable RAM Disks would also be welcome - if only to shoot the idea down.


    What next?


    Thank you,

    George

    [This message has been edited by George2 (edited 01 July 2002).]

  16. #16
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    George,

    Your aside first. I have tested a RAM disk (not bootable) at magician's request on my G4-800DP with ATTO EPT bench test:

    50MB RAM disk:
    ATTO 8MB, Sample Size 2, no System Disk Cache.
    PR-313.52MB/s, SR-163.87, PW-165.79, SW-159.06

    10MB RAM disk:
    PR-265.39MB/s, SR-177.90, PW-168.48, SW-166.27

    50MB RAM disk:
    Intech QuickBench¬ô 1.5, Extended Test Size: 10 MB
    Read: 172.918MB/sec, Write: 173.937MB/sec

    I did not test the 10MB RAM disk with QuickBench.

    I am going to need some time to digest your latest very thorough post. I've looked at it about a half dozen times and nothing sticks out. My pea brain is shriveling with age. I used to run a couple of Ultrastar 2XP drives in one of my NuBus Mac clones, the DCHS-39100 or DCHS09W, the 9GB wide model off of a NuBus JackHammer card. I was getting SR-9MB/s, SW-11+MB/s. Yes faster drive but NuBus machine. Your drives shouldn't be that much slower. I doubled checked my notes on the jumper settings I used. I don't see anything wrong.

    I think, if you want, send me one of the drives, USPS Priority Mail, which starts at $3.85, the 1-3 day service, and insured if you want. Email me your name and address and phone. I will test the drive in my PTP on both internal SCSI and a Bluenote. I'll check the jumpers, the SCSI cable will of course be different, different computer, will try different drivers, check Mode Page Parameters, etc., and return.

    Or we can continue here. What do you think? k

  17. #17
    Join Date
    Mar 2002
    Posts
    9

    Default

    << my G4-800DP >>

    I am not jealous. Not me. Not even the least little bit.

    << 50MB RAM disk:
    ATTO 8MB, Sample Size 2, no System Disk Cache.
    PR-313.52MB/s, SR-163.87, PW-165.79, SW-159.06 >>

    Neat!

    Stupid RAM Disk tricks: DiskWarrior 2.1.1 will recognize a RAM Disk and is happy to do it's thing on one - including optimizing the directory to a perfect pretty blue gradient, which would seem pretty useless since a RAM Disk is in random access memory, anyway, and shouldn't care what sequence the information on the RAM Disk is in. Norton Disk Doctor 6.0.3 will do it's thing, even a media check (what media?) - except Norton DD will not run the partition portion of the test (I guess you cannot partition a RAM disk). Norton Speed Disk 6.0.3 will Optimize a RAM Disk - and even erase blank space - again, probably utterly useless activities. TechTool Pro 2.5.5 and FWB HDT 3.0.2 pretty much ignore a RAM Disk. I believe Apple's Drive Setup 1.7.3 ignores them, too. There are some RAM Disk programs (RAMDisk+? AppDisk? I forget) which claim to be as much as 2X faster than Apple's stock RAM disk, but I have never benchmarked any of them.

    Is there any software out there that measures Seek Time and Access Time? It might be fun to see what they have to say about such benchmarks on a RAM Disk. I guess there shouldn't be much of any seek time at all . . .
    __________________________________________________ _______

    << I think, if you want, send me one of the drives, USPS Priority Mail, which starts at $3.85, the 1-3 day service, and insured if you want. Email me your name and address and phone.

    I will test the drive in my PTP on both internal SCSI and a Bluenote. I'll check the jumpers, the SCSI cable will of course be different, different computer, will try different drivers, check Mode Page Parameters, etc., and return.

    Or we can continue here. What do you think? k >>

    It is out of my hands.

    My drives are already quarreling over which one gets to make the trip.

    "I'm broker than he is!" Pick me!" "No, pick me!"

    See what you've done.


    Info sent via email.


    My thanks,

    George

    [This message has been edited by George2 (edited 01 July 2002).]

  18. #18
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    George,

    Returned your email with my address. Hate to hear quarreling.

    The bench test for Hard Disk Toolkit does tell Average Access Time and sometimes tells Average Seek Time. I use ramBunctious for 10MB cache for Exploder. Just ran the HDT test, it gives Average Access Time 0.0 ms. No result for Average Seek Time.

    I had an ATTO SiliconDisk II purchased off of eBay. It is a solid state drive but mine was only 64MB (and the RAM is proprietary and very expensive even now) and it is UW SCSI (40MB/s max). It had an Average Access Time 0.2 ms. Again no report of Average Seek Time. After playing with it for awhile, I sold it to someone on the forums here. k

Posting Permissions

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