Upgrading early BIOS

M H Stein dm561 at torfree.net
Thu Aug 14 09:55:25 CDT 2008


Date: Wed, 13 Aug 2008 13:49:22 -0700
From: "Chuck Guzis" <cclist at sydex.com>
Subject: Re: Upgrading early BIOS

On 13 Aug 2008 at 15:36, M H Stein wrote:

>> AFAIK cable select has always been an option with Parallel ATA drives,
>> although it was rarely used in pre-UDMA days; the cable was obviously
>> different (40 conductors with standard connectors vs. 80 conductors and 
>> unique special connectors) and the order of the drives was usually reversed 
>> (i.e. the slave on the end of the cable).

>Nope--I've got a couple of very early Conner and Maxtor drives (less 
>than 100 MB)--CS isn't an option on the jumpers.

If you'd read more carefully, you'd have seen that I *did* say that "not all drives 
supported the CS option;" in fact, quite a few of the older drives did not.
I was responding to the posted statement that CS "had not been invented" in
the 486 days, when AFAIK, it was part of the early ATA specs (although
rarely implemented in practice).

> It's the motherboard that determines the cable type and mode from pin 34,
> which is grounded at the mobo end in a UDMA cable, so if the mobo isn't
> UDMA-aware I don't think the drives would be, and the cable wouldn't matter.
> Also, not all drives supported CS.

>Read the Wiki article you cited:

>"Pin 34:

>Pin 34 is connected to ground inside the blue connector of an 80 
>conductor cable but not attached to any conductor of the cable. It is 
>attached normally on the gray and black connectors".

>On UDMA-66 capable drives, it's the cable, not the mobo that dictates 
>speed.  Put an ordinary 40 conductor cable on a UDMA-capable drive 
>and mobo, and the configuration won't take advantage of it.  When the 
>changeover was in progress, we used to get support calls quite often 
>on the subject of UDMA cables and drives on non-UDMA mobos.

Well, I never knew that just replacing the cable in my 386 and 486 systems
would speed things up; live and learn. I always thought that if the mobo 
wasn't pin 34 aware and UDMA capable then the cable wouldn't have much
effect on the transfer mode. AFAIK in order to use UDMA all three items
have to be UDMA capable, the mobo, the cable and the drives, although
I'd say that, assuming it's UDMA aware, the mobo controls the speed 
depending on what kind of cable and drives it sees. Semantics?

You said that with a UDMA cable, any UDMA capable drive will operate in
UDMA mode (presumably regardless of the mobo); how does a drive know 
what kind of cable is connected? I always assumed that only the mobo can 
tell from pin 34.

BTW, a UDMA cable certainly works with non-UDMA drives and/or mobos, 
although you have to remember that it is also a CS cable, but I don't think
it'll be any faster.

>Similarly:

>"With the 40-wire cable it was very common to implement cable select 
>by simply cutting the pin 28 wire between the two device connectors; 
>putting the slave device at the end of the cable, and the master on 
>the "middle" connector. This arrangement eventually was standardized 
>in later versions. If there is just one device on the cable, this 
>results in an unused "stub" of cable, which is undesirable for 
>physical convenience and electrical reasons. The stub causes signal 
>reflections, particularly at higher transfer rates."

>Bottom line is that if you use a non-CS-capable (i.e. 40 conductor 
>"straight through") cable, it won't work. Depending on what the 
>mobo/controller end does with pin 28, you'll have either two masters 
>or two slaves on the same cable.

Agreed; I took it for granted, but with our detour into UDMA the most 
important point to be made relevant to the OP is that Cable Select needs 
a special cable and with pre-UDMA drives & mobos you're probably 
better off using Master/Slave and a standard cable.

>Cheers,
>Chuck

m




More information about the cctech mailing list