Richard M. Barrett
Senior Applications Engineer
I. IntroductionAccelerometer based monitoring systems can be tested to verify proper installation and operation. Testing ensures data integrity and will pinpoint most problems. The trouble shooting techniques presented in this paper are very simple and can be performed by most monitoring systems and data collectors.
Most installation and sensor problems can be detected by measuring the bias voltage of the sensor. The bias voltage will indicate bad cable routes and failed sensors. Many on-line systems trend the sensor bias voltage. Other problems can be detected by analyzing the time waveform and FFT spectrum.
The following section will explain sensor operation and how it relates to the bias voltage, time waveform and FFT response. The next section is separated into different fault indications. Finally a trouble shooting chart is given.
II. Accelerometer Operation and ResponseMost accelerometer faults can be diagnosed by measuring the bias voltage of the sensor amplifier. If the bias voltage is within correct limits the sensor is almost always operating properly. Most cabling faults can also be isolated by measuring the bias. After the bias is checked the time waveform and FFT spectrum will verify fault diagnosis or proper operation.
What is Bias Voltage?
The majority of accelerometers, piezovelocity transducers, and many pressure sensors have a biased output. The bias voltage is also referred to as the bias output voltage (BOV) or rest voltage. Biased outputs are characteristic of two-wire sensors that measure dynamic (AC) signals. Vibration and pressure are dynamic signals that vary with frequency. The bias voltage can be explained as follows:
Imagine that the supply voltage that powers the sensor is on the ceiling of a room. Further imagine the common lead or ground is the floor of the room. The vibration signal can be thought a bouncing ball on a bungee cord. Obviously the room confines the ball and will not allow it to bounce above the ceiling or below the floor. This is analogous to the vibration signal. The voltage output can bounce up and down but can not move above the power supply voltage or below ground.
Since the sensor is a two wire device, the vibration signal must be carried by either the supply voltage wire or the ground wire. In other words, the bungee cord must be attached to the ceiling or the floor. If the signal is carried at the supply voltage it can only swing negative. Conversely, if the signal is carried at ground it can only swing positive. Biasing the output provides a platform from which the signal can travel in both the positive and negative direction. This is analogous to bringing a table into the room and attaching the bungee cord to it.
From the table the ball could travel positive toward the ceiling or negative toward the floor. The height of the table provides a rest or neutral point. If the ceiling height is 8 feet and the table is four feet tall the ball could swing equal amounts up or down. However, if the table is increased to six feet tall the ball can only travel upward two feet and will hit the ceiling long before it hits the floor.
Most portable data collectors supply 20 volts of power to the sensor, whereas most on-line systems supply 24 volts. The bias voltage should be set at approximately one half of the supply voltage to maximize the amount of swing in the positive and negative directions. Most two-wire sensors produce an 8 - 14 volt bias. Figure 1 shows a typical schematic of a biased output sensor. Figure 2 shows the change in amplitude range as the supply voltage is decreased.
When the signal amplitude runs into the supply voltage or ground clipping occurs. Clipping the vibration signal distorts the waveform. In other words, it no longer matches the true vibration it is attempting to measure.
Figure 1. Typical Schematic of Biased Output Sensor
Figure 2. Change in Amplitude Range with Decreased Supply Voltage
Measuring the Bias Voltage
One may ask: How can the supply voltage line be lowered to the bias voltage and still provide power to the sensor? This requires an understanding of the current regulator built into the power supply. This device is usually called the constant current diode. It provides a constant current to the sensor regardless of the supply voltage or vibration voltage from the sensor. Most data collectors supply 2 milliamps of current to the sensor; most on-line systems supply 4-6 mA.
If the current limited power supply is probed with a voltage meter the supply voltage will be measured on the side of the constant current diode connected to the power supply. The bias voltage will be measured on the side of the constant current diode connected to the sensor. Figure 3 shows a schematic of a sensor power supply containing a 2-10 mA CCD (constant current diode).
Figure 3. Sensor Power Supply Containing a 2-10 mA CCD
The bias voltage should be measured periodically to check sensor operation. The best measurement device is a voltmeter. However, most data collectors can measure the bias if the sensor is powered from a different source (other than the data collector). When using the data collector as a voltmeter the DC voltage input is used. Oscilloscopes can also measure the bias by selecting the DC coupled input.
The bias voltage can be trended with many on-line systems. Trending the bias provides a record of sensor operation. If the sensor is disconnected or slowly develops a fault the bias trend will show when the event occurred. Since the bias changes with temperature the bias trend can show when machines are started, shut down, and crudely detect bearing faults and other heat generating problems. Figure 4 shows a bias voltage trend from an on-line vibration measurement system.
Figure 4. Bias Voltage Trend from an On-line Vibration Measurement System
The bias voltage will also indicate the condition of the cabling and connectors. If the bias level measurement is equal to the supply voltage, the sensor is disconnected. A measurement of zero volts indicates a short in the system. An unstable bias voltage can indicate poor connections, but can also be caused by a clipped signal or severe electro-magnetic interference.
Time Waveform and FFT Spectrum Fault Analysis
Time waveform can be measured with an oscilloscope, most data collectors and on-line systems. The time waveform will give an immediate indication of a clipped signal. Usually the signal looks truncated or flattened on one side and normal on the other. Severely clipped signals will cause the waveform to look jumpy. Contact noise from poor connections can cause a similar jumpy reading.
The FFT spectrum will give another quick indication of signal quality. The one times operating speed vibration is usually present and a good indication of proper operation. Presence of a large ski slop usually indicates distortion from sensor overload. However a noisy sensor that has been integrated will also produce ski-slope.
Cable routing faults can also be detected by analyzing the FFT. Multiples of line frequency usually indicate improper shielding or grounding. If the time and frequency measurements read zero the sensor is not operating or disconnected.
III. Fault IndicationsOpen Bias Fault: Supply Voltage (18 - 30 V)
When the bias equals the supply voltage, the sensor amplifier is disconnected. In most cases the problem is in the connector or cabling. First check the cable termination at the junction box, data collector or monitoring system. If the cable is connected to a terminal block make sure the wires are secure and tightened down - and in the proper terminal!
If a connector is used it can be disassembled or replaced. However, avoid disassembling or removing the connector until all other fault sources have been checked.
Next check the sensor side. Many times the sensor must be disconnected for maintenance - sometimes no one reconnected the sensor! If each end appears good check all other terminations, splices and connectors. Also ensure that the cable is not crushed or cut.
If the cable route and connections appear good, further test the connectors. Cable continuity can be tested by shorting the signal leads to the shield wire and measuring the opposite end with an ohmmeter. Depending on the cable length, several ohms to several hundred ohms should be measured from each wire to the shield. If the cable and all connections are in proper working order the fault is in the sensor. However, open faults within the sensor are very rare.
Short Bias Fault: 0 Volts
When the bias measures zero volts, power failure or a system short is usually the problem. First ensure that power is turned on and connected. If the power supply is on then there is usually a short in the cabling. Like the open fault, it is very rare to have a shorted sensor.
The most obvious fault location is in the terminations. Check to make sure that a frayed shield is not shorting across the signal leads. Many times a crushed cable can produce a short. Use an ohmmeter to check electrical isolation between the leads. Disconnect the cable from all other devices and measure between all signal leads and shields. When measuring between the signal out and common leads, the ohmmeter should measure infinite or at least above 50 megohm.
Low Bias Fault: 1.5 Volts
Bias measurement of approximately 1.5 V(sometimes 0.6V) usually indicates that the wires are crossed and powering reversed. Sensors with reverse power protection use zener diodes to shunt the reversed power. Thus when the power is reversed the zener voltage is measured. Sensors without reverse power protection do not usually survive mis-powering and are permanently damaged.
Damaged Sensor: Low bias, high bias
Bias readings other than those above usually indicate sensor damage. Common sources of sensor damage are exposure to excessive temperature, excessive shock, mis-powering, and electrostatic discharge. Excessive temperature is the most common cause of sensor failure. Sensors caught in a fire are usually destroyed and can show various bias readings depending on the failure mode within the sensor. Long term temperature failures are marked by a slowly declining bias voltage. In many cases bias returns to normal as the temperature decreases and the sensor can still be used in lower temperature applications. Figure 5 shows the bias trend of a sensor failing from long term temperature degradation in a paper machine dryer section. Note how the bias voltage increases when the machine cools during machine shut down.
Figure 5. Bias Trend of a Sensor Failing in a Paper Machine Dryer Section
Excessive shock, mis-powering and electrostatic discharge can permanently damage the amplifier of unprotected sensors. Bias readings of 3.5 to 5 volts usually indicate a "blown" amplifier. Industrial sensors should contain protection devices to prevent failure.
Erratic Bias and Time Waveform
The bias voltage should remain stable and unchanging. Shifting bias indicates a very low frequency signal that is not filtered out by the bias meter. In rare cases this indicates a real low frequency signal, however in most cases this indicates a fault. Primary causes of erratic bias are thermal transients, poor connections, ground loops, and signal overload. Each of these faults will also be visible in the time waveform as an erratic jumping or spiking of the signal.
Thermal transient signals cause thermal expansion of the sensor housing materials. This is detected by the sensor as a low frequency signal. The problem is most evident when using low frequency sensors. One way to reduce the effects of thermal transients is to cover the sensor with a thermal boot. Thermal boots are available from most sensor manufacturers.
Poor or contaminated connections can also cause low frequency bias and contact noise. Look for corroded, dirty or loose connections. Repair or replace the connection as necessary. Non-conducting silicone grease should always be applied to connectors to reduce contamination.
Ground Loops are developed when the cable shield is grounded at two points of differing potential. Always ground the shield at one end only! An easy test for ground loops is to disconnect the shield at one end of the cable. If the problem disappears it was probably a ground loop fault. Figures 6 and 7 show a connection susceptible to ground loops and a correct installation where the shield is tied at one end only.
Figure 6. Example 1, Connection Susceptible to Ground Loops
Sometimes spurious spikes from fast thermal shifts, lightning strikes, and shocks can overload the sensor and cause a momentary shift in the bias voltage. The shift in bias can trigger alarms and protection system shutdown devices. To prevent triggering alarms and shutdown a longer delay can usually be programmed or hardwired into the monitoring system. The delay prevents the system from taking action until the sensor has settled.
High frequency, high amplitude vibration signals can also overload the sensor and in severe cases cause bias shift and erratic time waveform. However overload problem are usually detected by observing truncated waveforms and large ski-slope spectrums.
Figure 7. Example 2, Connection Susceptible to Ground Loops
Truncated Time Waveform: Sensor Overload
Truncated (flattened) time waveforms indicate that the signal is clipping into the supply voltage or ground. The clipping causes the amplifier to saturate and become overloaded. Some common causes of overload are severe pump cavitation, steam release, impacts from loose or reciprocating parts and even gear mesh.
One way to reduce clipping is to use a higher power supply voltage and ensure that the bias is centered between supply and ground. However the bias voltage and supply are rarely adjustable. For example, if you are using a 15 volt power supply and a 12 volt bias, clipping will occur sooner.
Long cables can also reduce the amplitude swing at high frequency and may be a problem in some applications. The easiest solution is use a lower sensitivity sensor. A sensor with 10 mV/g sensitivity will have a hundred times more high amplitude range than a similar 1V/g sensor.
Sensor overload may also produce a ski-slope spectrum. If the amplifier saturates, intermodulation distortion occurs. This causes low frequency noise also referred to as washover distortion. Figure 8b and 9a/b show distortion due to pump cavitation and gear mesh overload.
Figure 8a. Normal Pump Signal
Figure 8b. Ski-slope Spectrum
Figure 9a. No Gear Mesh Overload
Figure 9b. Gear Mesh Overload
Sometimes the ski-slope response can be due to integration noise and near machine noise. In these cases the ski-slope levels are much lower. Figure 10 shows integration noise on an accelerometer. Figure 11 shows ski-slope due to ambient machinery noise at very low frequency.
Figure 10. Integration Noise on an Accelerometer
Figure 11. Ski-slope Due to Ambient Machinery Noise
Mounting Resonance Spectrum
Mounting resonance can give false indication of high frequency machinery faults such as gear mesh and bearing problems. The problem is most evident when using probe tips and magnets. However mounting the sensor on thin plates such as machine guards can also lower the mounting resonance. Figure 12 shows the resonance of several common mounting techniques.
Figure 12a. Mounting Resonance Plots for Probe Tip
Figure 12b. Mounting Resonance Plots for Magnet
Figure 12c. Mounting Resonance Plots for QuickLINK
Figure 12d. Mounting Resonance Plots for Stud Mounted Configurations
Re-measuring with higher resolution will usually discriminate machine vibration from mounting resonance. However amplitudes of machine signals at sensor mounting resonance will be greatly increased. In some cases the mounting resonance can cause sensor overload if it is excited by a machinery fault.
After the mounting resonance signals can be severely attenuated. Figure 13 shows the attenuated response of bearing noise measured with a tip coupled probe.
Figure 13: Response Plot Using Probe Tip Shows Mounting Resonance Excited by Bearing Fault
Line Frequency Harmonics in Spectrum
Harmonics of line frequency usually indicates interference from motors, power lines and other emissive equipment. First ensure that the sensor shield is grounded (at one end only!). If the shielding is good, check the cable routing. Never run the cable along side high voltage power lines and only cross at right angles. For example, if a power cable is 440 Volts and 60 Hz and the signals from the sensor are at the milli and microvolt levels any cross talk can severely corrupt the data.
IV. Trouble Shooting ChartBelow is a trouble shooting chart for sensors with a 12 volt bias. For sensors with other bias voltages the same concepts apply - only the stable bias range will be different.
BOV Spectrum Time Waveform Fault Condition Action 0 No signal No signal No power or cable/connector short
- Test/turn on power
- Test cable isolation
- Repair/replace cable
0.6 - 1.5V No signal No signal Reversed powering
- Reverse leads
2.5 - 5V No signal No signal Damaged amplifier
- Replace sensor
10 - 14V Stable High low frequency "ski slope" High amplitude high frequency noise High frequency overload (steam release, air leak, cavitation, etc.)
- Repair steam leak/dump
- Use less sensitive sensor
- Place rubber pad under sensor
10 - 14V Stable Very high low frequency "ski slope" no high frequency signal Choppy Damaged amplifier
- Replace sensor
10 - 14V Stable Good signal strong 60Hz 60 Hz Inadequate shielding
- Connect/ground cable
10 - 14V Stable High low frequency noise High frequency spikes ESD Arcing impacts
- Reroute cable
- Use less sensitive sensor
- Place rubber pad under sensor
10 - 14V Stable High low frequency noise Jumpy/Choppy Intermittent connection
- Repair connection
18 - 30V No signal/weak 60 Hz No signal Open cable connections
- Repair connection