UNIXPC RF shield for hard disk

[Howard E. Motteler]

In article <21302 at cup.portal.com> thad at cup.portal.com (Thad P Floryan) writes:

>That shield for the HD is there for a reason!  MOST of the "crap" from the
>CRT is directed downwards ... right onto the HD.  It also goes off to the
>left (so to speak).

Well, sort of...  But the machine meets FCC standards for radio noise
as it comes out of the box, as does the CRT.  All digital signals are
sealed inside steel or tin, and the HD shield is part of this.  The
flyback transformer, which is essentially the same as found in a B&W
TV, is not completely shielded, as the sawtooth waveform for the
horizontal sweep at ~ 15KHZ puts out a lot less radio energy (even
given the higher power levels) than a bunch of TTL level square waves
around 10 or 20 MHZ, as found on the motherboard and the HD digital
circuitry.  The TTL video signal to the monitor probably puts out more
RF noise than the flyback, and some care has been taken to shield the
video signals.

Inductive coupling of the flyback transformer to, say the HD or floppy
read heads might be more of a problem, but interestingly, the floppy
is not inside the machine's electrostatic shield, so the designer's
weren't worried about this.  Come to think of it, the switching power
supply probably puts out more RF noise and has stronger magnetic
fields than the flyback.  Switchers typically run anywhere from 10 to
50 KHZ or so, and all power for the machine is running through that
little transformer.  I wouldn't leave floppies sitting on either side
of the machine.  (If you set floppies on the right side of the
machine, you block the power supply air holes, anyway!)

>In the UNIXPC, operating the internal HD without the RF shield is asking for
>possible data corruption; the RF energy beneath the CRT is *VERY* strong; I
>discovered this by placing an AM/FM portable radio near the "base" of the CRT
>when I had the cover off and the system operating when testing some other
>things.

Well, an AM radio is quite sensitive as a switching noise detector,
and if the "hood" (the steel lid above the motherboard) is open at all
when the machine is running, it will put out even more RF noise.
The very first computer I ever owned, a Z80 system from "the digital
group" came with a demo that would buzz out the "star spangeled
banner" an a nearby AM radio, while drawing a flag on the screen...ok,
ok, it was 1976, the bicentennial and all...

Since the HD platter and heads are sealed in a metal can, its
vulnerability to noise should be restricted to the short path the head
signals travel from outside the can, to the drive electronics under
the can.  Essentially, if noise from its own TTL signals, stepper
motor pulses, etc., don't bother it, then noise or magnetic fields
from the switcher or flyback probably won't, either.  I would guess
that the intent of the HD shield is to bring the machine into FCC
specs for radio emmissions, more than to protect it from external
noise.  

Not that it necessarily proves anything, but I've been running without
the HD RF shields on my machines for a year and an half, and although
they may put out a bit more RF noise, I can run an AM shortwave radio
10 feet from the machine, when it's closed up, with no interference,
and I've never had any other problems.  Reasons for leaving the
shield off include: (1) On the big miniscribes (at least on mine) it
defeats the drives' shock mounts by pressing very tightly aginst the
top of the drive; (2) it restricts air circulation around the drive
electronics, (3) the "tin fingers" that press aginst it in the back
prevent the "hood" from propping up in the full open position; (I took
those off, too); (4) lazyness on the part of the person putting the
machine back together :-).

>If you can retrieve that shield, do so and re-install it!

If you can't, I can make you a great deal on a couple of barely used RF
shields :-)
-- 
Howard E. Motteler       |  Dept. of Computer Science
motteler at umbc3.umbc.edu  |  UMBC, Catonsville, MD 21228