From: "Tom Faloon"
Subject: Re: Automotive EMI
Date: Mon, 23 Sep 2002 23:57:09 +0100
NNTP-Posting-Date: Mon, 23 Sep 2002 22:55:18 +0000 (UTC)
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2615.200
X-Newsreader: Microsoft Outlook Express 5.00.2615.200
First you need to establish how the noise is getting onto the amplifier.
The ignition is certainly a possible source, but so is the car electrics.
You can expect a very 'dirty' supply on automobile electrics. (HF noise and
very large amplitude spikes.)
**** Find the entry path *****
Do some of the following to establish the noise entry path.
If easily accessible, disable the alternator by disconnecting its power
cable, or temporarily removing the drive belt. Run the vehicle and see if
the noise has reduced significantly. If it reduces, then you have conducted
interference, getting in via the car's electrics.
If you don't want to meddle with the alternator, then power the amplifier
circuit from a separate battery, keeping everything else, including sensors
and ground connections, as normal. If the noise is reduced significantly,
then you have conducted interference, getting in via the auto electrics.
Running the amp from it's normal car electrics, disconnect the sensors, and
their cables completely, and replace them with resistors having about the
same source resistance as the sensors, and fitted right at the amplifier
input. If the noise is reduced then you can assume it was being picked up by
the sensors / cables.
You may well find that some noise is getting in by both paths.
**** Reducing radiated interference *****
If radiated noise, from the ignition system, noise is entering via the
sensor leads, then do some of the following. (If not done already.)
Terminate the sensor cables in as low an resistance as the sensors will
allow. Preferably close to the sensor impedance.
Add some decoupling capacitors across the resistive terminations. (0.01uF
ceramic and / or smaller. - experiment !)
If that does not work, filter the inputs, but use passive filters, at least
at the front end, and use good RF construction practice. The repetitive
noise may be at a very low frequency, but it will have very high frequency
components, which will go straight through any active filter (or
differential amplifier) you care to build. (This stuff can cause serious
interference to an AM radio, and even to a TV, so it may have components up
to 100's of MHz. Once inside the amplifier it can get rectified, and
produce low frequency noise.)
**** Reducing conducted interference ******
If noise is entering via the car electrics, then filter/decouple the
amplifier well. (You will probably need an RC or LC filter to isolate the OP
Amp supply effectively, otherwise HF crud on the OP Amp power pins will go
straight through to the output. - It has no respect for the power supply
rejection ratio's shown in the OP Amp data sheet.
**** Other points ****
What exactly did you mean by 'when the load on the engine increases'
Are the sensors connected to ground at the point where they are physically
i.e. could they form earth loops.
Does the amplifier have a solid earth.
Is the amplifier in a screened box. (Not suggesting that is essential, just
wondering about the construction.)
Why 10 Hz cutoff. What decides that?
( The lower the better.)
EMI / EMC
Chris Matthaei wrote in message
I've been trying to do some low rate (50samples/sec) data acquisition of
on my engine (Ford 5.0). I noticed that as the load on the engine increases,
I see a
large increase in noise in the signal. I'm guessing this is due to the spark
increasing with load. Anyway, I'm wondering how I can clean this up. I was
of using a simple active low pass filter using a 741. 10Hz would probably be
cutoff. Does anyone think this will help? Any other suggestions?