Richard,
I asked this question last year and the general consensus was "no,
the magnetic tweezers will not damage an IC". However, this was not enough
to convince the skeptics in our company, so I actually had to "run the
numbers" and do some order of magnitude estimates. The following is the
reply I made. It is only necessary to read it if you don't believe my first
statement.
Gentlemen,
It has come to my attention that there are concerns with operators
wearing items containing magnetic material. Training asked me to
investigate the possible negative impacts of magnets on the production
floor. Furthermore, they have indicated that you may be interested in
knowing my findings. I can say that fears of magnetism are unwarranted and
the following outlines why.
Four possible failure modes of integrated circuits may be induced by
magnetism.
1. Induced Current
2. Induced Voltage (Hall effect)
3. Metal movement
4. Altering magnetic media (Floppy disk)
The greatest concern with magnets seems to be induced current, and induced
voltage. I think we all understand the concept behind induced current. The
question becomes; how big is the effect? To jog our memories a bit, lets
review Faraday's Law of Induction.
e = d?/dt where ? is magnetic flux ?=?B*A
B= magnetic strength
A= area bound by the circuit
A changing magnetic flux will induce a voltage and a current through
the loop. Looking at the equation, we can see three factors that come to
play in inducing a voltage:
1. the stronger the magnet, the greater the voltage,
2. the larger the circuit loop, the greater the
voltage,
3. the faster the magnetic flux changes (smaller dt),
the greater the voltage.
To get an idea of the magnitude of this phenomenon let's look at a
theoretical circuit. Assume a large circuit board has a particular loop
that twice circles the perimeter of a large board of 40 cm before connecting
to ground. In addition, assume that we will use a powerful magnet from a
gas powered camping generator having a magnetic strength of 5000 Gauss or
0.5 Tesla. For simplicity in the equation, we will assume that the magnet
is moved from a far enough distance that the magnetic strength changes from
zero to full strength. Furthermore, let's assume we are using GaAs
semiconductors and we want to make sure the induced voltage is less than
1.5V. What we want to know is how fast would we have to move the magnet to
get 1.5 volts in the circuit.
From the equation, we solve for the change in time dt.
dt = d?/e
= (B*NA)/ e where N = number of current loops
= (0.5 Wb/m2)(2 loops)(0.4 m)2 where 1 Tesla = 1
Webber/m2
1.5 V*s where emf is expressed in
volt seconds
= 0.106 seconds to move the magnet from far away to very close.
As we can see, the magnet must move very fast indeed. (Electric
generators work by increasing the number of windings, thus increasing the
area of the current loop). Furthermore, we can solve for the induced
current using Ohm's law. Assume 100? of resistance.
Imax = emax = 1.5V = 0.015 Amps
R 100?
Of course, this is a very unrealistic scenario. It does give an
idea of the magnitude of the problem. A magnet that is several orders of
magnitude less powerful will induce voltages and current several orders of
magnitude smaller. Furthermore, a proportionally smaller current loop will
induce a proportionally smaller voltage.
Interestingly, the more obscure Hall effect may actually have an
impact on the function of integrated circuits. The Hall effect requires the
circuit to be powered. For those who are unfamiliar with the Hall effect,
it is when opposite charges will separate to opposite sides of a current
conductor when exposed to a magnetic field.
To get a picture of this effect, imagine a flat sheet of copper the size of
a sheet of paper. Attach wires to two opposing sides of the copper sheet
and send current through the sheet of copper. Now place a large magnet
under the sheet of copper. The presence of the magnet will influence flow
of electrons and cause negative charge carriers to separate from positive
charge carriers. This is because the moving charge (electrons) are a moving
electric field; and by induction, experience a force perpendicular to their
direction of travel. The net effect is, if a voltmeter is connected to the
sides of the sheet of copper that are perpendicular to the direction of the
current flow, a voltage potential can be seen.
It is by this phenomenon that ABS brake wheel speed sensors, and crankshaft
position sensors operate in an automobile.
The formula by which this works is:
VH = IB where I = current and B = magnetic field strength
nqt where n = charge carrier density q = charge t =
thickness of charge carrier
Looking at the equation, we can see that the bigger the current or
the greater the magnetic strength, the greater the Hall voltage induced.
Inversely, we can see that the smaller the charge carrier or the thinner the
current carrying material, the larger the Hall voltage induced. Notice
also, width of the current carrier does not appear in the equation.
Again, the question is one of magnitude. Take for example, a 1mm
thick piece of silicon carrying 0.1 mA of current, subjected to a magnetic
field of 1.2T.
VH = (0.1 mA)(1.2 T)
(1 X 1020 electrons/m3)(1.6 X 10-19C)(0.1 X 10-2m)
= 7.5 mV
Again, this is not enough voltage to cause a latch up in logic
levels. Realistically, the magnetic flux seen from small magnets is orders
of magnitude smaller than 1.2 Tesla.
On the other hand, actual thickness of the doping (Germanium, Gallium
Arsenide etc.) is the actual current carrier on an IC and may be orders of
magnitude thinner than 1mm. In theory, if the sideways voltage potential
through a diode or transistor on the surface of the IC is made high enough,
it could change the logic level of an adjacent transistor. While this would
make for interesting research, I'm afraid we don't have the resources to
conduct it. For further investigation, the spacing of transistors, current
levels, charge carrier thickness and charge carrier densities value would be
needed.
Suffice it to say, in practical application, it is highly unlikely that a
change in logic levels will be seen. Furthermore, permanent damage due to
the Hall effect will not be seen.
I did ask other scientist and engineers around the world what they though
might happen to circuits exposed to magnets. Many responded back with just
how powerful of a magnetic field they have operated IC's without any
detrimental effects.
For metal movement, the concern might be breaking conduction bonds. Most
wires are encapsulated in plastic; restricting individual movement of
conductors. Crystal timers are not encapsulated because the piezoelectric
component of the crystal must be free to move. Therefore, crystal timers
are the most susceptible to mechanical damage due to magnets. With the
exception of the delicate gold wires, connections are made with solder and
are very robust. Magnetic fields would have to be so great that it could
apply a force high enough to exceed the tensile strength of solder. Even if
we dealt with superconductors, we would never see a magnetic field that high
in this plant. The delicate wires, on the other hand, are made out of gold
and are diamagnetic. It is in no form, ferrous. While gold can be made
magnetic, the magnetic flux strength required, or low temperatures required
are the topics of experimental physics. Therefore, the gold wires are not
susceptible to any magnets we might have.
Concerning magnetic media, I am not familiar with how it works. My limited
understanding is that somehow, variations in magnetic strength are induced
into the magnetic medium. These variations are then read via the Hall
effect. Since magnets are used to record and erase magnetic media, it is
theoretically possible that someone's magnet could change the magnetic
states of media brought in close proximity. Indeed, this has been reported
to be the case. This means that no one handling floppy disk should have a
magnet on his or her wrist. Furthermore, those of us who have a magnetic
paperclip holder on our desk should not allow floppy disks to come in close
proximity to it. Since the handling of magnetic media, on the production
floor, only occurs by technicians, it is not necessary to restrict magnets
from operators.
Thanks
Ryan Grant
Advanced Technology Engineer
MCMS
(208) 898-1145
[log in to unmask]
> -----Original Message-----
> From: Richard Hawkins [SMTP:[log in to unmask]]
> Sent: Thursday, March 01, 2001 10:16 AM
> To: [log in to unmask]
> Subject: [TN] IC Damage by magnetism?
>
> Do any of you know (definitely) if it is possible to damage an IC with a
> small magnetic field? Such as one on the end of a pair of tweezers?
>
> Thanks in advance for your help.
>
> Richard Hawkins
> Associate Designer
> [log in to unmask] <mailto:[log in to unmask]>
> Phone: (615) 382-3655
> Fax: (615) 384-4405
>
> CEI Co., Ltd.
> Plant 2
> 210 Evergreen Drive
> Springfield, TN 37172
>
> --------------------------------------------------------------------------
> -------
> Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
> To unsubscribe, send a message to [log in to unmask] with following text in
> the BODY (NOT the subject field): SIGNOFF Technet
> To temporarily halt delivery of Technet send the following message: SET
> Technet NOMAIL
> Search previous postings at: www.ipc.org > On-Line Resources & Databases >
> E-mail Archives
> Please visit IPC web site (http://www.ipc.org/html/forum.htm) for
> additional
> information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700
> ext.5315
> --------------------------------------------------------------------------
> -------
---------------------------------------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt delivery of Technet send the following message: SET Technet NOMAIL
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
---------------------------------------------------------------------------------
|
