PDA

View Full Version : Noise beyond 60 Hz -SUMMARY



Mike Monroe
05-18-2000, 10:55 PM
ORIGINAL QUESTION:
> I am using a portable EMG system and utilizing 6 of 8
> available channels for data collection. Usually only 60hz is emitted by
> electrical devices, but I am finding out that computer monitors are
causing
> more than just 60Hz interference--they are causing the EMG system to pick

> up interference at approximately 60-62Hz, 120-123Hz and 180-184Hz during
a
> low exertion VDT task. I have tried various computer monitors
> and they all seem to cause this "excessive" interference. When I turn
the
> monitor off, only 60Hz remains (from other equipment). Has anyone run
into
> a problem like this? If so, what has been determined as the cause and
what
> could be done to remedy this situation?

My thanks to those who took the time to provide very helpful suggestions.
An LCD Display took care of my noise problem.

Mike Monroe




RESPONSES:

From:
Hartmut Witte
In a comparable situation we found that our force measuring chain (special
made
Kistler plates - amplifiers - ADC - Computer) owned a single peak in the
basic
noise level at 78 Hz (cf. attached file noise.doc). This turned out to be
the
frequency of the graphics card and the monitor. We minimized the
interferences
by testing several combinations of graphic interfaces, their position on
the
main board and monitors, but could not eliminate it, but reduce it to about
the
tenfold amplitude of the base noise level.
We remain as long as possible in the frequency domain, apply NO filtering
and
check the "last" data in the freq. domain for resonance effects. Up to now
(knock on wood!), there was no need for us to filter or smooth before
changing
into the time domain. In EMG-data, this for sure will be different. Perhaps
you
may use your two unused channels to control the noise level (freq.
dependant)?
If you find any graphic interface which does not use the main board as an
antenna, please tell me!



From:
Dave Grimshire

As you know computer monitors radiate electro-magnetic energy. The human
body is a very good antenna and collects the energy as you've seen on your
EMG. There are low radiation monitors made that should reduce the amount
of energy a monitor will direct to the front of the screen. Also small
monochrome monitors seem to have less radiation.

Dave Grimshire

From:
Jan Cabri jcabri@zatopek.fmh.utl.pt

Why don't you try a portable computer (laptop). If you do not
connect the laptop to the power line, I think it might resolve your
problem (I guess your electrodes are not shielded enough to prevent such
an interference). However, you will have to buy a pcmcia-a/d convertor
for this, which don't go cheap.
Best regards,

Jan Cabri, Ph.D.
Technical University Lisbon

From:
Gonzalez
It's normal to find components at integer multiples of the frequency
(60Hz) of the periodic signal. They are called harmonics and they will be
present as long as the periodic signal is not a perfect sinusoid.
What can be done? To turn off any equipment not involved in the
measurement, including lights. Also, apply some type of filtering after
acquisition of the signals, either analogue or digital filtering. Filters
used for supressing this type of interference are known as notch filters.
Tradeoff: The part of the signal (EMG) near 60Hz, 120Hz, 180Hz, etc will
be affected as well.

Regards,

Jose A.


From:
"Murray Maitland"

You might wish to look at your pre-amplifiers. Generally the
closer the preamps are to the signal source the better in terms
of increasing the signal to noise ratio.

You might look at using a notch filter to remove the 60 Hz
and harmonics. See The Scientist and Engineer's
Guide to Digital Signal Processing
by Steven W. Smith
California Technical Publishing
ISBN 0-9660176-3-3 (1997)
http://www.dspguide.com/pdfbook.htm

You might also look at your grounding since this can make noisy
loops. Although for surface EMG, it isn't usually a big problem.


From:
Gordon Chalmers
Here is a suggestion of various things to try. Have you tried a grounded
shield around the top, sides and back of the monitor? Try a tinfoil sheet
as a shield with aligator clip attached and run to ground contact on a
piece of equipment (and allow air space for ventilation). Also, can the
subject and subject keyboard equipment be physically and electrically
isolated from the data collection equipment? (e.g., not on same table top,
and/or large physical distance such as opposite ends of the room, perhaps
even try different electrical circuits in the room for research monitor &
equipment and the subject keyboard equipment, and the research subject and
his/her equipment could be placed on rubber mats to isolate).

Finally, since you know it is the monitor, and if nothing else works, you
could try a flat screen monitor (or perhaps this is one of the other
monitors you have tried), my guess is they would not emit strong electrical

fields as CRTs do.

From:
"Arnel Aguinaldo"
You are probably picking up the refresh rates of the monitors and/or the
processing cards of the PCs, all of which have the capacity to run
processes
of up to 180Hz, depending on the system. My experience has shown that
increasing sampling frequency and adding bandpass filters usually would do
the trick. Otherwise, I'd try other signal conditioning methods.



From:
"Peter Chupity"
Have you considered using an LCD display?


From:
Jason Harrison

You're measuring harmonics. The refresh frequency of the monitor is
probably 60Hz, then you have the horizontal frequency (much higher)
and the vertical sweep frequencies.

If you've already sheilded all your cables then you can try moving the
monitor or using a projection monitor to move the monitor as far away
as possible. Or use a notch filter to remove the frequencies from
your measurements.


From:
Matthew Pepper

Here are some ideas.

You are picking up harmonics of 60Hz. Probably generated by the power
supply.
This is likely if the supply uses rectification to transform the AC wave
form
to DC. The rectification process produces the harmonics. The transformer
will
also produce harmonics as it is not a linear system.

It may be that a switch mode power supply will help if the line
interference is
generated by your system. However if the monitor is the source its' power
supply will already be switch mode.

Otherwise the source can be other line powered equipment in the same lab as

well as any fluorescent lighting.

If the computer monitor is the problem you could try replacing it with a
LCD
display. If your system is flexible enough you could use a PCMCIA data
acquisition card [$400] and run the experiment on a laptop.

Regards

Matthew Pepper

From:
Einar Skavland Idsų
The signals at 60-120 Hz are the vertical screen updates (refresh rates).
On
current monitors a ray of electrons is fired at the monitor 60-120 times
per
second. Older monitors (and video cards) can only use the lower frequencies

(