Conference hydra::amiga_v1

    Hi Shaun,
    
    	I was going to say "RTFM" but I thought that I wouldnt.....
    
    1.	I hope that you meant HD17.device. Did you rename it
    harddisk.device?
    
    2.	are you sure that you followed step 8 in construction.doc TO THE
    LETTER?  i.e, did you ctrl<D> immediately the window occurred in the
    boot sequence???
    
    3.	Are you using ARP mount? I think that there was mention in the docs
    that I had trouble with the ARP mount. Use the DOS mount.
    
    Cheers,
    Glenn

    Hi Shawn
    
    Yeh what Glenn said I found that the ARP mount would not mount my disk.
    But I did have the same problems you had, that is the drive would
    format ok but not mount. I found that in my mountlist I had stuffed up
    the entry for the hard disk. Also make sure that the device driver is
    name WHAT YOU CALLED IT IN THE MOUNTLIST!! I even went to the extent of
    matching the case of the filename. Somebody said the the mountlist file
    has to be exactly what the device driver is called.
    One thing I did was make the mountlist entry automatically mount the
    driver, that way you can tell if it is working straight off by whating
    the disk drive light (it should flash as the driver gets loaded). If
    you are still having prob's I can send you my mountlist.
    I can be reached at
    Allin1 add : ASHLEY MARK
    or
    DAOVAX::ashley
    
    Have fun
    I love being a turtle. (groan)
    
    MARK.
    

    It's nice to hear from other Australian Amiga freaks - thanks for the
    response.
    
    I tried to mount it using the official Commodore Mount (Not ARP) - no
    luck - still the same handler. Then played with the mountlist (which I
    thought was correct!!) to get it EXACTLY as per the example. Typed
    format and voila - success - well partially.
    
    It format's but there are read errors all other the disk. I'll play
    with it more tonight when I get more time.
    
    Once again - thanks for your help.  Shaun.
    

    Great stuff Shaun!!
    Just thought that you might be interested:
    Did you know that there is a Spirit Technology product that does the
    same thing as this public domain project. The difference is, of course
    price, that the HDA-506 (that's what it's called) supports DMA so it's
    quicker, I talked to some salesman in Melbourne and he said that it
    transfers at about 3Mb/sec, which isn't bad!! This progect will do at
    best I think about 28 to 30 Kb/sec using a HD perf exer (public domain
    cann't remember the name sorry but I have somewhere). Good thing about
    it also is that it will support autobooting so you can chuck DF0 for
    good.
    The only thing holding me back is the green stuff that Mr (?) Keating
    takes out of your pay each month.
    
    Anyway enough of that.
    I would be interested in how you get on with formatting/revectoring
    your hard disk. I have tried two disks that had bad blocks not mapped
    and I found that the low level formatter would like the disk and
    revector the bad blocks for me but when I went to do the DOS format it
    would fail.
    What I was thinking of doing is formatting it on an IBM PC, and then
    trying to do the DOS format then. I think it would work, being and
    ST-506 type disk controller I would expect the format to be the same.
    Anybody want to shoot me down in flames?
    Let us know how you got on
                    
    Mark
    
                                   

    
    Please include your mountlist entry and directory of l: and devs:
    
    for RLL drives use hd27.device and rename it to harddisk.device
    and for MFM use hd17.device and rename it also
    
    put mount =1 in the mountlist of the partitions
    
    then mount dh0:
    and format it 
    
    	It should be Ok if the format worked out OK , How fast was the
    format/verify in the format.bas [ loaded question.. ]
    
    
    
    		.Pete. 
    ( Not by the amiga at present so I am doing this from memory )

    re: .4  "transfers at about 3Mb/sec, which isn't bad!!"
    
    I'd say more like UNREAL!  Even the HardFrame (one of the fastest)
    comes in under 1 Mb/sec.  I might believe 300Kb/sec.

    I Believe that you can achieve much greater transfer rates than
    28Kb/sec. Try an interleave of five or six. Five was the fastest but
    six was the most reliable. I achieved greater than 80Kb/sec with ST11-M
    and RD52. 
    
    
    
    
    
    

    I at work at the moment so I can't upload the mountlist but I do
    remember that it is now EXACTLY as is the the makehd.zoo that glenn had
    put together.
    
    As far as I am aware l: and devs: have everything that is required -
    harddrive.device, mountlist, fastfilesystem - and anything else that I
    shove in.
    
    Format.bas seem to work ok. I'm was trying out 3 different type of XT
    RLL controllers:
    
    	OMTI 5527 - Fast but I blew it up the other day and the data that
    		    it wrote was corrupt.
    
    	Adaptec - No quite as fast in the format but it works. Only one
    		  problem - if format.bas finds a bad block the program 
    		  locks up trying to read the controller.
    
    	Western Digital - Stuffed!!! Takes ages to format and fails at the
    			  end of it. Can't even get passed the format.bas
    			  stage.
    
    I will upload the mountlist later on tonight (hopefully). I'm not as
    much in a hurry to fix this now as I just put in an Order for a Amiga
    3000!!!
    
    Ta Shaun.
    

    
    RE .4
    
    Gidday Mark,
    
    Correct me if I'm wrong but does'sdnt format.bas write some special
    information on cylinder 0. This means that if you do your low level
    formating on a PC you won't get this information.
    
    A question - will this project accept two hard drives at the same time
    or is there an update to the software that I need to do to make this
    work??
    
    Thanks Shaun.

Re .6 - remember the 3Mb/sec figure came from a saledroid. It is almost a
given that that is MegaBITs/second, whereas the numbers you're quoting are
MegaBYTES/sec. Converting gives you around 375MegaBYTEs/second. That isn't
bad, but it isn't great, either.

	<mike

    Ohh Well,                                      
    
    I had to return the borrowed Hard Drive Controller Last night so I have
    no more chance to play with this project - Until I get the money to
    buy a Seagate Controller.
    
    With my 3000 turning up middle of next week I think that I will survive
    without a hard drive for now!!
    
    Thanks for everyone's help.
    
    Shaun.
    

Re: .10

I suspect that Mike is right about bits versus bytes.

>Converting gives you around 375MegaBYTEs/second. That isn't
>bad, but it isn't great, either.

375MegaBYTEs/second is a high end throughput for an IBM type controller.
Such controllers usually have significantly lower speed than the SCSI
controllers you find on an Amiga.

I just couldn't help taking a potshot at the clone world.

    Did I miss something?  Do you really mean 375 million _bytes_ per
    second?  That's nonsense.  ST-506 type drives top out at less than 100
    million BITS per second, and SCSI I transfers are spec'd to top out at
    1.5 million bytes per second for asynchronous transfers, 4 Mbytes/sec.
    synchronous.
    
    John

    He meant 375KilliBytes/Second - Basic slip of the mind to finger
    translation - I would wager.
    
    John

Yes, I meant 375KILObytes/sec as equivalent to 3MEGAbits/sec. So busy
fixing one translation problem that I introduced another.

375MEGABYTES/sec - now _that's_ a controller I'd respect!

	<mike

    Hey so I was wrong, my mistake!!
    I looked at the article again and it listed it at about 100Kb/sec. Ah
    well I cann't be perfect all the time.
    
    Shaun I know what your saying about the format, I had a drive that I
    think had a bad cylinder 0. It would format (low level) ok but when you
    went to format it in AmigaDOS the drive would do several Recalibrates
    and then the format would die with an arror something like invalid
    format.
    But I think the low level format is IBM campatible, maybe at the end of
    the format the progarm wrties out the info that AmigaDOS needs. Well
    that's what I think, that's the way I would do it I mean the contrlller
    has a ROM based formater built in so why not use it!
    

    My problem is not quite the same as this note, but I thought I would
    enter it here for the sake of keeping even remotely connected
    information in the same place.  Here goes.......
    
HEeeeeeeeeeeeelppp!!!!!

I'm in DEEP DEEP TROUBLE!!

I recently built the Super_CheapHD project. (for those who are not familiar
it is the MakeHD project with additional buffering of the bus). A few 
teething toubles were encountered along the way, but eventually I triumphed.
That was until recently when I began encountering requestors telling me 
that VOLUME...... HAS A READ WRITE ERROR!!!.

The problem was quite random and applied to all partitions of the hard disk.
It's frequency could be anything from hours to 10's of minutes. The HD,  an
RD53, is partitioned in 20Mb,20Mb,28Mb. The system allways boots (from 
Floppy transferring to HD to complete the boot) Currently the requestor 
will   appear after a few minutes. Once there, the system is unuseable, WB is 
locked up solid. The three fingered neck pinch will clear the problem but 
only until the next time.  If you are in CLI/Shell you know it's about to 
happen because you can hear the drive do a seek or reset? the next disk
access to any partition will bring up the requestor.

Running WB from a floppy you can mount the HD and work normally until
you get no response (No requstor appears).  INFO shows that all HD
partitions are no  longer mounted or recognised. 

FORMAT.BAS supplied with the MakeHD.zoo does a deep format with no reported 
errors. (not strictly true. In order to get Format.bas to work I have to 
tell it that RD53's have 1020 cylinders not 1024 otherwise all is OK)
AmigaDos Format doesn't report any errors either. I have tried using WB 1.3
and WB 1.3.2 Formatter with the FFS switch.

I have checked the interface for "dry" joints and cross connections. I have 
replaced the disk with and RD51 and another RD53 but the problem remains.
Since the Deep formatter works. I tend to think that the Hardware is OK.

I am using a Seagate ST-11M controller which has been checked on a 
Decstation 350 (Olivetti PC )with an RD53. It seemed to work OK during the 
couple of hours it was installed.

I've checked Mountlists till I'm blue in the face. Unless I'm blind to the 
obvious I can't see a problem. I have specified the stack at 4000 for each 
partition but removing this entry as per the mountlist in the project has
not improved matters.   

I have only 1 Meg of memory but a careful watch of memory remaining doesn't
    show that the machine and disks are starved of memory.  This notes
    conference warns of not running virusX with HD so I have removed it
    together with StarBlanker and now have a standard 1.3 WB

Before leaving home today I encountered the requestor again, and in trying 
to recover from it got a message that the bitmap for dh1: was screwed up.
Using FixDisk 1.2 on this partition, gives an error relating to Error #33 
............BadDriveType.   This may be a red herring but....... I rebooted 
the system and it now fails to boot because dh1:is screwed up.

I am also using the HD17.device from the makeHD.zoo as harddisk.device has 
anyone encountered difficulties using this handler.
 
As you can see I have checked so much I'm blinded as to what to do next. Has
anyone else had this problem with this project or with HD's in general?

Any help hints pointers will be gratefully recieved and acknowledged.


		Regards,
			David

                                                            

If the drive is external do you have it in a box with a FAN?
What you describe might be due to the drive heating up...

-steve

    Steve,
    	Thanks  for the prompt reply. RD53 is in an old PC380 box
    connection from the IBM controller is via 20 way and 34 way IDC cables
    about 4 feet long. (well within the 10ft allowed). Bus connection from
    AMI to the IBM controller is 2 inches. Probably 4 inches by the time
    you've allowed for PCB wiring. I have had it shorter, but it makes no
    difference.
    
    	I have also run the drive for some time before formatting it, to
    acclimatise the disk.
    I'll put my mountlist relating to HD in a subsequent reply....just in
    case.......
    
    		Thanks again 
    			David 
    

	Everything is working OK at the moment and has been for the last
    hour. In that time I have uploaded my mountlist as attached I am 
    reformatting dh1: at the moment using Amigados Format. I just noticed it
    very slow around 2.5 seconds per cylinder.
    
Just in case here the relevant portion of my mountlist.
    
DH0:
    Device = harddisk.device
    FileSystem = l:FastFileSystem
    Unit   = 0
    Flags  = 0
    Surfaces  = 8
    BlocksPerTrack = 17
    Reserved = 2
    Interleave = 0
    Priority = 5
    LowCyl = 0  ;  HighCyl = 300
    Buffers = 30
    GlobVec = -1
    BufMemType = 0
    Mount = 1
    DosType = 0x444F5301
    StackSize = 4000
#

DH1:
    Device = harddisk.device
    FileSystem = l:FastFileSystem
    Unit   = 0
    Flags  = 0                            
    Surfaces  = 8
    BlocksPerTrack = 17
    Reserved = 2
    Interleave = 0
    Priority = 5
    LowCyl = 301  ;  HighCyl = 600
    Buffers = 30
    GlobVec = -1
    BufMemType = 0
    Mount = 1
    DosType = 0x444F5301
    StackSize = 4000
#

DH2:
    Device = harddisk.device
    FileSystem = l:FastFileSystem
    Unit   = 0
    Flags  = 0
    Surfaces  = 8
    BlocksPerTrack = 17
    Reserved = 2
    Interleave = 0
    Priority = 5
    LowCyl = 601  ;  HighCyl = 1018
    Buffers = 30
    GlobVec = -1
    BufMemType = 0
    Mount = 1
    DosType = 0x444F5301
    StackSize = 4000
#

    Amigados format has finished and dh1:looks to have recovered. I retimed
    format and it took 4.5 seconds to format and verify each cylinder.
    
    This got me thinking. Do I have an interleave problem? Interleave was
    set to 6 during deep format. I also set no Precompensation and no
    reduced write current as well.
    
    As a test I have just copied 3MB of software from DH2: to DH1: no
    problems. It took 5.5 minutes (a complete WB 1.3  plus 3 fish disk
    archives and many assorted files). All arrived safely. This is
    definitely wierd.
    
    
    			Regards,
    				David
     

DH0:
    Device = harddisk.device
    FileSystem = l:FastFileSystem
    Unit   = 0
    Flags  = 0
    Surfaces  = 8
    BlocksPerTrack = 17
    Reserved = 2                         <--- if two tracks are reserved
    Interleave = 0
    Priority = 5
    LowCyl = 0  ;  HighCyl = 300         <--- shouldn't LowCyl = 2 ?
    Buffers = 30
    GlobVec = -1
    BufMemType = 0
    Mount = 1
    DosType = 0x444F5301
    StackSize = 4000
#

			Alan T.

    Alan,
    	According to "AmigaDos Inside Out", Reserved = 2, refers to the
    number of boot blocks
    
    	Thanks for your help
    			David
         

   David,
    
    	Might I suggest that you have a Hardware problem. I ran my unit for
    six months without the slightest hint of a glitch. I know of others who
    have had identical experiences.
    
    	The Deep format parameter of 6 seems about right, or at least it
    was right for my RD52. 
    
    	The problem as described does appear to be temperature related. I
    would, as a first step, replace the RD53.
    
    cheers,
    Glenn
        

    Glenn,
    	You may be right. I have at various times replaced the disk with
    another RD53 and an RD51. Both have the same problem. So if it is, it
    is in the interface. The I/f is not in a box at the moment so I will do
    a little bit of engineering and fix it up.  Last Friday  I had the
    problem 4 times in the first hour and after that it was OK. It managed
    to occur during an XModem transfer which has left as file that cannot
    be deleted. (in fact attempts to delete this file bring a visit from
    the GURU. I can copy it and delete it but not the original) However
    this is a minor unrelated problem.
        
    	Thanks for your help
                  		David.