Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!ucsd!nprdc!malloy
From: malloy@nprdc.arpa (Sean Malloy)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Perstore and Maxtor
Keywords: controllers
Message-ID: <1477@skinner.nprdc.arpa>
Date: 14 Feb 89 19:22:50 GMT
References: <16761@srcsip.UUCP>
Reply-To: malloy@nprdc.arpa (Sean Malloy)
Organization: Navy Personnel R&D Center, San Diego
Lines: 41

In article <16761@srcsip.UUCP> rogers@orion.UUCP (Brynn Rogers) writes:
>What is the interleave of a perstore controller (1:1, I guess)?

The manual I got with my controller recommends the following
interleave values:

	4 for a PS180 with XT ROM on any machine (the manual refers
		to 4.77 MHz XTs and 6 MHz ATs)
	8 for a PS180 with AT ROM on AT machine (the manual refers
		to 6 MHz ATs)

(The ROM is one of the socketed chips on the controller. There will be
a label on it indicating ROM type and version number. Mine says XT9-1.03)
The Perstor controllers put more sectors on a track than either MFM or
RLL controllers do: 31 for the PS180, 34 for the PS200. The additional
sectors radically change the best interleave from what the 'intuitive'
value is based on the drive specs.

The Perstor controller's data transfer rate is _very_ dependent on the
correct interleave. I discovered this the hard way. I have a 10 MHz XT
clone, so I figured I would be able to decrease the interleave, and
set it to 3:1 when I installed a 40Mb drive. A couple of days later, I
downloaded a file with some information about using Perstor
controllers from SIMTEL20, and found the following suggestion: Get a
copy of CORETEST (available on SIMTEL20), then format the disk with
various interleaves -- the format only needs to run for about 90
seconds for CORETEST to find enough formatted space to work with --
and compare the data transfer rates for various interleave values,
picking the interleave just before the one that _reduces_ the rate
when you increase the interleave.

When I ran CORETEST on my 3:1 interleaved disk, it gave me a rate of
29.56 Kb/sec. Reformatting the disk to the recommended 4:1 gave me a
rate of 241.78 Kb/sec (at 5:1, it dropped to 181.36, so 4:1 was the
best).


	Sean Malloy
	Navy Personnel Research & Development Center
	San Diego, CA 92152-6800
	malloy@nprdc.arpa