17 Aug
2013
17 Aug
'13
11:38 a.m.
On Aug 16, 2013, at 1:31 PM, ard at p850ug1.demon.co.uk (Tony Duell) wrote:
SMD (and ESDI) drives have the data separator in the drive, so what's presented
on the interface is not the raw data to/from the head amplifiers. Unlike ST506 and
ST412 where the data separator was in the controller.
I think RK)5s do too, and maybe some of the other DEC drives. I am not
sure about the Philips X1215, etc, but I can easily look it up. IIRC the
HP7900 drive does not include the data separator, it's in the controller.
I had considered using an RK05 or an RK03 in place of the HP drive on an
HP9880 system, this is one thing I would have to sort out (it is not
difficult!).
Quite often the data separator was in the drive so you got seperate clock
and data signals from the drive to the controller when yu were reading,
but the data encoder was in the controller, so you sent the digital form
of the head signal to the drive. This encode, of course, had to use the
same encoding scheme and data rate as the date separator is designed to
decode
The
actual data to/from the drive is almost raw; there's just a
bitstream (at TTL level) from the controller to the drive to write,
and from the drive to the controller to read. So, other than
This sort of interface was quite common on hard disk units at that time.
SMD, for example has high-ish level seek, select, etc lines but raw data
to/from the head amplifiers. So do most DEC drives (RK05, RLs,
RK06/07,etc).
Having the data separator in the drive makes it a bit easier on the controller since
you get a clock from the drive to clock the data.
IIRC the Shugart SA4000 drives had a clock track on the disk and a fixed
hed to read it. The drive provided both the read _and write_ clocks to
the controller.
-tony