Instrumentation Troubleshooting Part 2
By Mark Williamson
Pulsed signals may seem relatively simple but there can be a whole lot more to it than one might expect.
There are a lot of instrument transmitters and sensors that send a stream of pulses to represent the readings and a couple of ways that these pulses are determined by the receiving equipment.
Besides the obvious bad connections, broken wires and again, operator error, there are other thing to consider.
I have seen several cases of crosstalk where wires in the same cable or bundle were affecting the signal.
In one case flow rate was counted by a powered magnetic pickup. It was showing flow when nothing was being displaced. Signal from the magnetic pickup was 5 vdc. My first thought was either the gap adjustment on the mag pickup or an intermittent connection.
The receiving equipment had a threshold of 3 vdc to count the pulses off of the drive line on the pump. Mind you, this was mobile equipment and power was supplied by a running engine. After all the usual suspects were found to be no problem, I make a quick breakout and used my portable oscilloscope to take a look at the signal. There was 7 vdc of noise and a quick check by increasing the speed of the engine saw it was alternator whine. In the build of the piece of equipment an extra power run had been placed it the wiring harness that wasn’t being used. The batteries had just been replaced most likely due to a failing alternator. A battery normally makes for a good capacitor but in this situation the connections at the battery were not clean and the extra power wire was stacked on the terminal next the alternator wire making an antenna for the alternator noise. De-terminated the unused power lead and had the mechanics change out the alternator and the problem was solved.
In another situation, 5 mag pickup signals on a 6 conductor shielded cable with a drain, readings were affecting each other as chemical pumps were engaged. There was a broken common ground and the signals were finding ground through the other signal wires. It was about a 200 foot run of cable.
Finding a short circuit on a long run of cable is not that hard. If you have visually inspected the condition of the cable from one end to the other and all looks fine, using your ohmmeter, measure the shorted wires on one end and then the other. On one end you will read the resistance of the wire and the other will be near zero ohms.
A TDR meter is nice to have to find the distance to an open in a wire on a long run of cable.
Some receiving equipment have a threshold setting for the incoming signal. If the threshold is set too high, you will not see any counts and if it is set too low you will be counting noise.
Some receiving equipment use a charge pump circuit, converting the pulsed signal to a voltage while other systems actually counts the number of the pulses.
Some transmitters source the pulses while others have open collector outputs where the signal out has a pullup resistor tied to power. A cracked, broken or loose resistor will cause a problem.
On one job, I got to work 8:00 AM and the boss said I had a 10:00 o’clock flight to catch and it was a 45-minute drive to the airport. They had a nuclear incident, the NRC had already been notified and the site had been evacuated. I was fairly new on the job so with schematics and data in hand I flew out to Bakersfield. On the flight out I was studying the data and realized the difference in most of the count readings was evenly divisible by about five hundred and twelve and I felt sure that the radiation levels could never be as high as what they were saying.
I called them from the flight and told them the call the NRC and report that is was only instrumentation failure and that no radiological event had occurred and after studying the schematics told them to have resistor ready when I got there. It was a cracked resistor on the main data bus and the vibration of the engine in the mobile van was being counted. When I got there, with rad meter in hand, I went straight to the problem and had them back in service. If the boss hadn’t been so quick to rush me out the door, I could have fixed it a lot quicker with a phone call to a local tech.
There are many types of flow meters. Some or the common ones are electro-magnetic, corilis and impellers or turbines. You can’t use electro-magnetic like a Pro-Mag 55 if the fluid is nonconductive.
Using the Pro-Mag 55 as an example, there is a compensation probe in conjunction with3 other probes. Static buildup can cause problems and I have seen the extremes where the compensator probe was covered over. One was with cement where reading started going low and another where the flow rate was slow enough that sand had covered the probe causing the reading to go higher. The biggest problem I have seen is loose connections in the terminal strips, either from temperature changes or vibration, loose connections causing problems. Any buildup on the probes will cause bad readings.
Looking at corilis meters you can have many problems stemming from something affecting the proper operation. A blockage or buildup in the tubes is one issue. Suspended particulates settling in the bent tube is another problem. The lack of dampening in the mounting can cause problems due to vibration and mounting units too close to each other can be another problem. These units work using a vibration and how much gain it takes for the tubes to gain sympathetic vibration. They are also used sometimes for measuring density so that mass calculations can be made.
Proper mounting is critical. Shock mounting and position is very important. Where the curved tube is mounted on the bottom there is a chance of not being cleared, freezing and rupturing the tube or build up. Mounting the curved tube on the top, at slow rates, entrained gasses can collect causing bad readings.
Using turbine blade flow meters, you can have problems of objects blocking the blades and problems with the axis being fouled plus if you are running an abrasive, wear on the impeller affecting the reading.
I found a lot of things in turbine flow meters like rags, a cell phone and a radio and broken axis pins.
One of my first encounters with a flow rate problem was with a mag pickup clocking bolts on a drive line. After getting the grating removed discovered only 3 out of 8 bolts still had nuts on them and 4 were missing. It was right below where the operator stood.
Setting up the instrumentation it is important to understand the math involved for the proper indication to the customer or user. Whither it be units per pulse or pulses per unit, one being the inverse of the other. In calculating pulses from a mag pickup, number of teeth, turns ratio, displacement of a pump and efficiency.
A lot of this is nothing more than common sense but common sense seems to be a rare commodity these days.
Comments