Quieting electronic noise

One of the most difficult things to diagnose in the automotive electronics world is electrical noise. Call it what you want, EMI, RFI, Spurious Induced Signal, whatever. Noise is always a signal element that is not wanted in a particular signal and usually is induced or injected into that signal. There are more ways to cause noise, and noise is often classified as different things: distortion, interference, jamming, etc. Just to make it easier to talk about, we will just call it all noise.

So, what is noise, where did it come from, how did it get into my system, how do I find it and most important of all, what do I do to fix it? What it is is a fairly simple thing to answer. Anytime there is energy in a signal that is not supposed to be part of that signal, it is noise.

Noise can come from any number of places. Electronic components such as resistors and

semi-conductors create a little bit of noise during normal operation, and often a lot of noise while malfunctioning. Noise can come from cosmic rays and particles from sun spot activity breaking through the atmosphere. It can be in our physical environment such as an electrical transmission line or a powerful radio transmitter. It can be a cross-connect between two signals. It also can be induced from one place to another.

A noise source can get its noise into our signal in several ways. It may be a part of the circuit itself, as in the case of a failing or failed part. It can be induced. Anytime there is a relative motion between a magnetic field and a conductor, a voltage is induced into the conductor (Figure 1). If an unwanted signal is AC, then the magnetic field it creates is constantly building up and collapsing. This building and collapsing is a motion in the magnetic field. Because the field itself is always moving, just having a conductor within the field will cause a voltage to be induced into the conductor.

Let me give you a common example of how a failed component causes noise. We all generally know how an alternator works. The field current is turned into a magnetic field, which is spun around a coil of wire inducing a current in the coil. What many are not aware of is that the electricity being made in the coils of a modern day alternator is actually AC. This AC power is run through the rectifier to be turned into DC. The rectifier is a group of diodes located in the rectifier assembly of the alternator. If one of these diodes is either completely open, or simply can no longer carry its full specification of current anymore due to a long life in an extremely hot engine compartment, then there is something called ripple on the alternator output.

Ripple is a DC signal with an AC signal “riding” on top of it. In other words, rather then the instantaneous voltage of the alternator output always being 13.5 volts, it will vary from say 11 to 14 volts. This easily can be seen looking at the alternator output with a scope (Figure 2). The frequency of the ripple varies with engine rpm. Ripple also can be detected by connecting an AC voltmeter to the alternator output. Anything more than a half of a volt (500mV) is unacceptable.

This ripple is present on all of the wires providing power to everything in the car. What effect it will have on various circuits will vary greatly. Many modern day computer modules, and almost all audio equipment, have filter capacitors to prevent the ripple from affecting that module.

Some do not have this added protection. Sensors that are connected directly to the power bus will have their output vary by the same amount as the ripple at the power wire. Older audio equipment used to exhibit “alternator whine” which is a sound in the speakers that varies in pitch with engine rpm.

The best repair for this problem is to replace the alternator. In some alternators, the rectifier assembly is replaceable.

A similar condition can occur on a 5-volt sensor power bus that comes from a computer. Inside the computer 12 volts is converted to 5 volts using a voltage regulator chip. If this chip is receiving ripple at its input wire, or it is being overloaded by a misbehaving sensor, ripple will be seen on the 5 volt power bus (Figure 3).

Another place for noise to be generated is a ground loop (Figure 4), which is caused by some resistance in the ground path. There is a voltage drop across that resistance. This voltage drop makes the voltage at ground different in different places. If a system is grounded in more than one place, and one of those places is on a different side of the bad connection, then there is a constant flow of current from the bad part of ground to the good part of ground. If this bad connection is completely static, it never changes its resistance, than there is not much trouble. This is rarely the case with a bad ground. It is normally all over the scope in its instantaneous resistance depending on the physical connection and the current running through it.

This causes the current in the ground loop to be constantly changing. So now a voltage level (ground) that all of the individual circuits inside a module depend upon is now actually a source of induced or injected noise. This is why it is so important for all grounds to be clean and tight.

It also is why checking, tracking and repairing bad grounds is always one of the first diagnostic or troubleshooting steps anyone should take.

Then there are the more exotic sources of noise. Induction sources like EMI (electro-magnetic interference) and RFI (Radio Frequency Interference). I recall helping someone on the ATSG tech line where the customer would bring the vehicle back two or three times a week with a code for an input speed sensor fault. This went on for weeks. The sensor, computer and even the o for that sensor were replaced. The technician finally did a better interview with the customer and found out that the problem always occurred at the same place on his way to work.

The technician arranged to have use of the vehicle over a weekend. He hooked up every scope, meter and scanner he had to anything that might be causing the problem and drove the vehicle on the same course that the customer

took to work. Low and behold the problem occurred at the same spot every time. The technician pulled over and recorded his various readings a few times, then he looked up to view the scenery and discovered that the road went right down the middle of a huge electrical generating windmill farm. He called the following Monday and I gave him a part number for some hard core MILSPEC microphone cable with 100 percent shielding to replace the speed sensor cable. Also made sure that he only grounded the shield in one place and made the leads to the sensor and computer as short as possible. Problem solved.

There are plenty of less exotic and unusual EMI issues. Often a spark plug wire or a wiring harness will not get properly mounted after some work and will get too close to each other. This can cause a small voltage spike in the signal wire every time that plug fires. Sometimes two signal wires will come into contact with each other injecting each other’s signals into the others.

Then there is RFI. This almost never happens in today’s modern electronics. This is when a radio signal gets induced into a signal wire. For this to happen the signal in the wire has to be extremely low (or unusually weak), and the radio signal must be very strong. The thing is that very is a relative term. Would it be possible for your laptops WiFi transmitter to induce a signal into say a speed sensor signal. Well, it is possible, just highly unlikely.

Another RFI scenario involves harmonics and resonance. The wavelength of a radio wave is the distance that the wave travels in one cycle. Mathematically it is the speed of light divided by the frequency (?=?/f). If the effective length of a wire just so happens to be the exact wavelength of the radio wave it makes it resonant and some crazy things can happen. This is the reason why use of radios is not permitted in blasting areas and singers with perfect pitch can shatter a crystal glass.

The defense against EMI and RFI is all about positioning and shielding, and shielding is all about grounding.

One last topic on fixing noise issues is filtering. Filtering is electronically removing the noise while leaving the original signal intact. A filter is a circuit that is added to the signal wire (Figures 5, 6, 7 and 8). There are an unlimited number of filters in the world. Some are specifically designed to solve a particular problem, and some are more general in nature.

The bottom line is if you are having generally wild and crazy things happening, use a scope or AC meter to check for noise. The usual causes for noise are failed or overloaded parts, or bad connections, especially grounds. I know I am not the first person that has harped on grounds as the most popular electrical problem in automotive electronics today. Here is yet another reason.