Re: reading from cassette

[mjmahon@aol.com (Michael J. Mahon)]

Don Bruder replied:

>In article <20030831204429.21879.00000286@mb-m04.aol.com>,
> mjmahon@aol.com (Michael J. Mahon) wrote:
>
>> Don Bruder replied:
>> 
>> >In article <hPr4b.240101$It4.113796@rwcrnsc51.ops.asp.att.net>,
>> > "desiv" <desiv@attbi.com> wrote:
>> >
>> >> So, why would the recording of the MP3 to the tape (intermediary) not
>have
>> >> the noise from the PC's ground?
>> >> 
>> >> (I am not an electronics guy, so that's a real question.  I'm not trying
>to
>> >> be funny..)
>> >> 
>> >> Is that a type of noise that the Apple would pick up but a recording
>> >usually
>> >> doesn't???
>> >
>> >No, that's a type of noise that won't be present (on any audio 
>> >equipment) if the hookup is done correctly. A correctly cabled audio 
>> >hookup from a computer to an audio device will have no noise such as 
>> >what I think you're probably referring to (the "hummmmmmmmm" that is 
>> >sometimes heard). If it does, then something about the wiring between 
>> >the computer and the audio device is wrong - more than likely, grounding 
>> >between the devices is faulty - and correcting that fault will eliminate 
>> >the hum completely.
>> >
>> >For lots more details (especially if, as it seems, you're having that 
>> >problem), check out the various audio newsgroups and look for 
>> >discussions about "Ground loops" and how to cure them.
>> 
>> These problems are indeed caused by "ground loops", or, more
>> precisely, current flowing in the ground wire(s), producing a
>> voltage drop across the resistance of the ground wire that is
>> added to the signal.
>> 
>> The problem is that computers are not "audio equipment" and
>> frequently do not have grounds as well isolated from noise on the
>> AC line (not just 60Hz noise, but often harmonics of 60Hz
>> caused by very non-linear loads, like switching power supplies.
>> 
>> As a result, a computer-to-computer connection which would
>> produce good results between audio equipment, may produce
>> excessive noise.
>
>As purely anecdotal evidence, I've cured *EVERY SINGLE* noise problem 
>I've ever encountered when doing audio work by the simple method of 
>making sure that both pieces of gear are solidly grounded *TO EACH 
>OTHER*. Whether that meant finding the ground plane and attaching a wire 
>to that on the computer, then figuring out a way to attach the other end 
>to the ground (such as it sometimes is, especialy in today's cheaply 
>made mostly-plastic junk...) of the amp/deck/whatever, or just using a 
>good set of cables with a dedicated ground wire running alongside, this 
>has fixed the problem for me every single time, without fail. 

This approach attempts to lower the impedance of the ground
connection enough so that the current flowing between the two
(or more) pieces of equipment does not create enough of a voltage
drop to interfere with the desired signal.

"Isolation of the ground from noise on the AC line" attempts to
reduce the currents flowing through the ground wires so that
the resulting voltage drop is not enough to interfere.

Using a #8 copper wire to bond the grounds together can be
regarded as a "brute force" method, while reducing the ground
current is more of a "finesse" method.  Both can work, but
folks who have difficulty identifying a good low-impedance
grounding point on each piece of equipment (which is most
folks) may encounter difficulty using the brute force approach,
and in the worst case, may wind up welding.  ;-)

Most audio gear have grounds that are pretty well isolated
from the AC line, so any remaining leakage or imbalance
will have a sufficiently high impedance that a simple shielded
cable will work fine.  It it doesn't then a lower-resistance
connection will usually do the job.  (This is not always the
right approach with equipment which is not so well isolated,
as noted below.)

>It's not a question of the ground being "isolated", per se, (A concept 
>which is practically the diametric opposite of what anything other than 
>a protective, or "earth" ground is supposed to be - ground in the 
>non-earth-ground sense is supposed to be a common-to-all-components 
>reference level for the entire device. That level might be 0 volts, or 
>it might be 50, or 90, or even 1500 volts. But it remains "ground" so 
>long as it is the "common" level for the device. As such, "isolating" it 
>defeats the entire purpose)

As I said, I was talking about isolation from the AC line.

> but a question of ground truly being 
>"ground" *FOR BOTH PIECES OF GEAR*. If you can measure a voltage 
>difference between the two items' ground planes, you're going to have 
>noise. No ifs, no ands, no  buts, you're going to have noise. The one 
>with the higher "ground" level is going to (unsuccessfully, usually) try 
>to draw the other item up to it's level, while the "other" item is 
>trying to drag the higher-level item's ground down to match its idea of 
>ground. If you "force" both items to agree to a single level by 
>electrically bonding their ground planes together, the noise goes away. 
>In my experience, that's a 100% certainty.

Again, you can either reduce the impedance/resistance or decrease
the ground current.  An isolation transformer is one means of isolating or
"floating" the ground of one piece of equipment at the ground potential
of another piece of equipment.  This also has the effect of reducing
the potential difference between the two equipment's grounds.

In general, a lot of ground current flow is a sign that something is
not working as it should.  Perhaps the insulation in a power
transformer is breaking down, or a line bypass capacitor is
leaking, or one of a pair is open.  Of course, some equipment is
notoriously _not_ isolated, with the equipment ground being
connected to one side of the AC line, but this is unusual except
on the line side of a switching power supply--for which the
switching-frequency transformer provides isolation.  However,
this isolation tends to be imperfect at line frequency harmonics
and switching frequency harmonics, because of unbalanced
capacitances and other deviations from ideal design.  The usual
way to soften the impact of these problems is to put a line filter
on the line input, which includes bypassing to the AC ground.
Again, inevitable imbalances in capacitors will result in incomplete
balancing of ground current and potential noise.

Good audio gear usually has a 60Hz power transformer for isolation,
with an electrostatic shield between the primary and the secondary
windings, to significantly reduce common-mode AC noise coupling.

It is _always_ better to design to minimize ground currents than to
have to patch things up later with brute force solutions.

For low-level audio signals, differential signalling is preferred in
professional equipment, precisely to eliminate the effect of
inevitable common mode (ground) noise.  This is a design
approach to dealing with the impossibility of perfect grounding.

I was proposing to the original poster that he use a piece of audio
equipment (a cassette recorder), which had reasonably good ground
isolation, as an intermediary device which could separately float
at the ground level of the PC when recording, and the Apple II when
playing back (or vice versa).  This is practically guaranteed to work
without any extra connections or knowledge of how to bond the two
computer's grounds together.

"Isolation" in this case means that the ground of the cassette
recorder will "float" with the ground potential of whatever it is
connected to, and only the difference between the signal lead
and the grond lead will register as "signal"--providing some of
the advantage of a differential connection.

The best isolation/float is obtained with a small battery-powered
cassette recorder, but most cassette recorders have small
transformer-isolated wall warts that provide sufficient isolation
from the AC mains.

>I've measured 40 volts of difference between the "grounds" of two 
>different pieces of audio (in this context, I'm using "audio" to refer 
>to *ANY* source or consumer of a desirable audio signal, be it computer, 
>TV, radio, turntable, or whatever, not neccesarily strictly "dedicated 
>to making noise" hardware) hardware in the past. The signal was so noisy 
>as to be almost unusable for any purpose, let alone approaching 
>"audiophile" quality. Placing a wire between the two items' "common" 
>ground (not earth/safety ground - that's a whole different kettle of 
>fish that may or may not be related to the "common" ground, depending on 
>the exact situation) killed the noise completely, presumably by forcing 
>both systems to agree that one level was ground. 

If the 40 volts you measured was present even in the presence of
a ground-to-ground connection, in the form of the shield of a cable
(which, for a six-foot cable would have only a few milliohms of
resistance), then that would cause enough current to flow to melt
the cable--which, of course, didn't happen (though it could in case
of a fault).

Instead, you measured 40 volts with a reasonably high-impedance
voltmeter.  The more interesting measurement is what _current_
flows between the two grounds, since that determines how serious
the isolation problem is and how much brute force would be required
to limit the ground-to-ground voltage drop to some maximum tolerable
value.  The problem is complicated by the fact that the current is
not a simple single-frequency sinusoid, but a complex, harmonic-laden
current.  Again, an ounce of prevention (in design) is worth a pound
of cure.

One design issue that is a property of the installation and not the
equipment is how the pieces of equipment are connected to AC
power.  If they are on the same AC circuit, or the same surge
protector, sharing the same ground, then that's about as good
as the installation could be.  If, on the other hand, they are plugged
into different AC circuits, or even on different sides of the 230VAC
service to the house, then very unpredicatable ground currents may
develop as other equipment in the house (or on the same transformer)
is operated.  This is clearly a situation to be avoided--and bonding
grounds together in such a case is potentially quite dangerous.

>Is it voodoo, or is it science? I really don't care... it works, or at 
>least, it has worked, repeatedly, for me. As long as "ground" in both 
>machines is at the same level, whether that level is 0VDC, +4VDC, or 
>even oscillating from -50 to +75 volts, as was once the case, the noise 
>"magically" ceases to exist.

Of course it's science, and of course, the whole idea is to get the
ground references of the two pieces of equipment to be the same.
But there are different ways to do that, and eliminating the problem
through proper design (not usually under the user's control) and
proper wiring is preferable to bolting everything to a quarter-inch
copper plate.  ;-)

If two poorly isolated pieces of equipment _must_ be interconnected,
then brute force may be the only way to go, but it's best to first try
to avoid ground currents if that is easily done (as in this case).

Most of what electrons do can be explained with considerable
precision.  Any residual belief in magic can be hazardous to your
health.  ;-)

-michael

Check out amazing quality sound for 8-bit Apples on my
Home page:  http://members.aol.com/MJMahon/