[Coco] [Color Computer] [coco} IBM SCSI hard disk question..

[Gene Heskett]

George's Coco Address wrote:
> 
> ----- Original Message ----- From: "Frank Pittel" <fwp at deepthought.com>
> To: "CoCoList for Color Computer Enthusiasts" <coco at maltedmedia.com>
> Sent: Saturday, June 10, 2006 3:42 PM
> Subject: Re: [Coco] [Color Computer] [coco} IBM SCSI hard disk question..
> 
> 
>> On Sat, Jun 10, 2006 at 01:45:28PM -0500, George's Coco Address wrote:
>>>
>>> ----- Original Message ----- From: "George Ramsower" 
>>> <yahoo at dvdplayersonly.com>
>>>
>>>
>>> >A few years ago, I obtained a drive cabinet with three IBM DNES-309170
>>> >E182115T drives in it. P/N 25L1910
>>> >
>>> >Since these drives are considerably newer(NOV-99),  than the current
>>> >drives
>>> >I'm using, I'm curious if they can be used on a Ken-Ton interface.
>>> >They have the tiny, 68 pin connector for data exchange vs. the 
>>> larger 50
>>> >pin connector I'm used to.
>>> >They are 9gig drives.
>>> >
>>> >If they can be used, how would I go about doing this using OS-9 L2? My
>>> >system is almost totally stock OS-9 L2.
>>> >
>>> >George
>>>
>>>
>>> Today, I did some research and found there are adapters to convert 
>>> the 68
>>> pin to the older 50 pin connectors. What I've found is that the newer
>>> drives will "retro" to the older scheme to accommodate older systems. 
>>> They
>>> are used a lot on older Mac systems and some PC stuff.
>>>
>>> Still, before I invest in one of these adapters ($15-$25), I would 
>>> like to
>>> know if it will work with OS-9 LII and the KenTon interface.
>>
>> While I don't know how big of a drive os9 can access the scsi standard 
>> has been
>> around for a long time and is in my experience well followed.
>>
>> Frank
> 
> This too, is the way I understand it. However, I don't know how well the 
> driver in OS-9 does this. I would think it is okay.

Let me see if I can remember it all, or enough to make sense at least.

1. The code to do that, by doing a cluster size of more than 1 sector 
was in the original os9 level 2.  Kevin's Christmas present rbf.mn 
removed all that code, so at that point os9 was saddled with a maximum 
drive size of about 130 megabytes maximum.

2. However, when I first tried to re-transplant it into rbf.mn sometime 
around ed 34 as I was doing the nitros9 conversion, one vital bit was 
missing, a definition that took me about a months reading and 
cross-checking Kevins book before I felt confident enough to try it, and 
it worked.

So if you have the latest version of rbf.mn, about ed 36 I think, Boisy 
correct me please, then os9/nitros9 can handle almost any drive size one 
might consider using on an os9/nitros9 system just by increasing the 
sectors per cluster value in the disks LSN0.  I think this is DD.BIT, 
but my manuals are 1000 miles SE of me ATM, sorry.

I believe that there has now been some bits redefined in some HD 
descriptors to facilitate handling this at formatting time, but this 
would have to be confirmed by the vendors as to how to use it on a case 
by case basis.

I believe that Boisy and friends have also fixed the os9/nitros9 format 
command to handle this transparently in the current releases. Back when 
I was doing all that, we did not have such a capable format command, so 
I had to first format the test disk, and then adjust it using dEd to 
actually use and test the multiple sector cluster code.  As that was 
about an hours messing for each cluster size I tried, I quit at 8 
sectors per cluster in my testing, but I'd have no reason to think that 
16-32-64, even 128 might work equally well.

A cluster size of 2 will do a 256 meg (decimal megs), 4 will be needed 
for a 512 meg drive, and 8 for a 1GB drive.  It only can be incremented 
in powers of 2.

A 1 GB drive would need to use a cluster of 8, an 8GB drive would need 
16 to get to 2GB, 32=4GB, and 64=8GB.  This is assuming the driver hands 
os9 the first half of a 512 byte sector if the seek addresses LSBit is 
0, and the second half of that sector if the LSBit is a 1.  If not, then 
the drive size can double because half of it won't be used.

Using that logic, the os9/nitros9 maximum drive size would be reached 
when that byte in the LSN0=$FF, or 32GB.  For os9/nitros9, thats one 
gargauntuan drive considering the whole archive will probably fit on one 
700 meg cd.

> 
> These drives which I have are a mystery as to whether or not they even 
> work. The cabling inside the cabinet they were in are the 50 pin variety 
> and the drives were not connected to anything except the cabinet. I 
> expect they don't work, but I wouldn't know until I determine how to 
> effectively interface them to OS-9 L2 and try to format and use them.
> Once I learn exactly how to do this first stage, then I can discover if 
> they are actually usable.

Most of the quantum 6 and 8GB drives have died, often a violent death, 
plowing grooves in the disks from crashed heads that reach halfway 
through the disk in the ones I've looked at, so those labels should be 
considered suspect at best.  We at one time had an array of 8 of those 
in a commercial server at the tv station, but I don't recall that all 8 
of them ever worked at the same time, anytime.  They were pretty bad...

-- 
Cheers, Gene