mlord@bwdls58.UUCP (Mark Lord) (08/20/90)
Like many netters, I've acquired so much new stuff from the net
that my hard drives have finally overflowed again. That's right,
I said drives.. plural.
My system has a 40Meg Seagate 251-1 (30ms, measured)
and a 40Meg MiniScribe 3650 (61ms, measured)
But, alas.. when TC++ 1.0 arrived, I discovered that I had to make
room for it somehow (10Meg.. wow!), and I don't like to delete things.
So I bought a new disk controller, you know.. the 2:1 thingie by PERSTOR.
The advertisements claim this thing will almost double the capacity of
most existing MFM hard drives. That's even better than RLL!
It cost me US$199 from Hard Drives International, who FedEx'd it to me
(here in Canada) for US$28 in just 4 days. Not bad for another country.
I got the 16-bit slot version: PS180-16FN. This card is BIG.
Full size AT form factor (as big as a card can get). This worried me.
Large cards are usually a sign of poor engineering. My old controller,
an Everex EV-346 (1:1 MFM) was about half the size, and very reliable.
Hardware installation was easy, even without reading the owner's manual.
Just unplug the old controller, and plug in the new one. Nothing fancy.
The PERSTOR card formats MFM or RLL drives equally, with 31 sectors per
track (MFM = 17 sectors, RLL = 26 sectors). This explains how one ends
up with nearly double the storage capacity. No special device drivers
are required, although the card DOES require use of it's onboard disk
BIOS rather than that built into the motherboard. This BIOS extension
takes up 8K of high memory space, either at C8000 or D8000. Since my
system has a 16-bit Paradise VGA card, it had to go at D8000 (both manuals
cover this point). In addition, special setup software comes with the card.
The drive can be low-level formatted ONLY with the supplied formatting
program, which is dumb but adequate. My Seagate drive formatted with
errors on only the tracks which were marked as BAD at the factory,
and my Miniscribe formatted completely without error (no factory bad
tracks.. an old drive from before the RLL era). Just to be safe, I ran
a couple of verify scans using SPEDSTOR, which found two additional bad
tracks on the Seagate drive. Repeat scans are recommended in the guidebook.
After that? Well.. DOS works just fine (3.3 and, ughh, 4.01).
That's right, I finally gave in and installed 4.01 (the idea of six separate
hard drive partitions did not appeal greatly to me.. so I finally gave in).
The drive is transparent to most software I have. The exceptions are mostly
due to software not coping well with DOS 4.01 and large partitions.
The only real differences are:
Norton 4.5 "FR" - the Format Recover option does not work.
- it will save the info, but does not recover it
properly. Possibly due to DOS 4.01.
PCKWIK Cache - Does not cope well with the tracks marked as "bad".
- When using the /T (buffer tracks) option, it tries
to read ahead, and when this results in reading a
bad track, it sits there doing retries for 10 seconds.
- It did not do this before, even though my Seagate
drive has always had bad tracks.
I simply used /T- with the cache to get around the retry nuisance.
The HYPERDSK shareware (comp.binaries.ibm.pc) does NOT have this problem,
Another difference I noticed, is that the Perstor controller does not do
track buffering, which makes low interleave values impossible for those of
us using 8Mhz ISA busses (my cpu is an 18Mhz 386sx). This means nearly
anybody who buys one is stuck with 3:1 interleave. This is not as bad as
it sounds, though. Consider this: there is almost twice as much data on
a track, so we are taking 3 revolutions to read double the data. This gives
an EFFECTIVE interleave of about 3:2 or 1.5:1. In practice, my personal
feeling is that it seems more like 2:1. (My old controller was true 1:1).
Conclusions:
1) It really works! I have 74Meg on my 40Meg Seagate 251-1
I have 73Meg on my 40Meg MiniScribe 3650
2) It is slower than 1:1, but faster than 3:1 interleave.
3) It is compatible with nearly everything I've tried (LOTS!).
but higher (DOS) level utilities DO work just fine.
Ie. most of NORTON, PCTOOLS, SPEDSTOR etc.. just not the very
low level stuff.
4) It is cheap! I have 147Meg (formatted) for le$$ than a big drive.
--
___Mark S. Lord__________________________________________
| ..uunet!bnrgate!bmerh724!mlord | Climb Free Or Die (NH) |
| Ottawa, Ontario. 613-763-7482 | Personal views only. |
|________________________________|________________________|malloy@nprdc.arpa (Sean Malloy) (08/21/90)
In article <3949@bwdls58.UUCP> mlord@bwdls58.UUCP (Mark Lord) writes: >The drive is transparent to most software I have. The exceptions are mostly >due to software not coping well with DOS 4.01 and large partitions. >The only real differences are: > > Norton 4.5 "FR" - the Format Recover option does not work. > > PCKWIK Cache - Does not cope well with the tracks marked as "bad". Most low-level disk utilities will not work correctly with the PerStor controllers, probably because they are not set up to handle 31-sector tracks. For example, IAU, a program that will re-interleave your hard disk for best performance with the data in place, will hum along thinking it's working, but _will_ trash the disk. Fortunately, when I did this, I had taken the precaution of making a brand new backup just before I started. >Another difference I noticed, is that the Perstor controller does not do >track buffering, which makes low interleave values impossible for those of >us using 8Mhz ISA busses (my cpu is an 18Mhz 386sx). This means nearly >anybody who buys one is stuck with 3:1 interleave. Equally bizarre, on my 25 MHz 80386, the formatting software would not work until I switched the machine from 25 MHz to 12 (I think that's the slow speed) -- it just gave one bad track after another. Slowing the machine down made the formatting run just fine. > 4) It is cheap! I have 147Meg (formatted) for le$$ than a big drive. It becomes cost effective when you have a) more than 1 drive, b) an MFM drive 40 Mb or larger, or c) an RLL drive of 80 Mb or larger. If you have two drives, it's always going to be cheaper than adding another drive. If you have one MFM drive smaller than 40 Mb, it's not really cost effective, nor is it for a 66 Mb RLL drive, but once you get past those breakpoints, it's all gravy. Sean Malloy | "The Crystal Wind is the Navy Personnel Research & Development Center | Storm, and the Storm is Data, San Diego, CA 92152-6800 | and the Data is Life." malloy@nprdc.navy.mil | -- _Emerald Eyes_, D.K. Moran
mlord@bwdls58.bnr.ca (Mark Lord) (08/22/90)
In article <9177@skinner.nprdc.arpa> malloy@nprdc.arpa (Sean Malloy) writes: > >Most low-level disk utilities will not work correctly with the PerStor >controllers, probably because they are not set up to handle 31-sector >tracks. For example, IAU, a program that will re-interleave your hard >disk for best performance with the data in place, will hum along >thinking it's working, but _will_ trash the disk. Fortunately, when I >did this, I had taken the precaution of making a brand new backup just >before I started. Ooops. I meant to talk about this in my original post, but it got lost. Yah.. even things like DM, SPEDSTOR, IAU etc.. ALL of which claim to handle any number of sectors.. none of these can format with this controller installed. The documentation even mentions this point. But it's not like one need to do a low level format daily :). -- ___Mark S. Lord__________________________________________ | ..uunet!bnrgate!bmerh724!mlord | Climb Free Or Die (NH) | | Ottawa, Ontario. 613-763-7482 | Personal views only. | |________________________________|________________________|