Level transmitter — art not science

There is much more to setting up a transmitter than a mere equation. Design and control errors can certainly compromise performance. Follow these simple tricks to get it right.

By Dirk Willard

“That, was unexpected,” I told myself. It was a rainy night in the Philippines and the separator was overflowing concentrated pickle liquor. “Thank God this is not in the States,” I muttered. My request for additional alarms had been over-ruled. Later, we added a warning: “Overflow at 75%” to the faceplate. The problem was due to a differential pressure (dP) transmitter. These transmitters are relatively simple and so don’t always get the attention they deserve.

This Philippines mishap points out the first rule about spanning a dP transmitter: choose the specific gravity to be the most critical (or dangerous) liquid (which wasn’t done for that separator). The basic equations are: zero = -h (Sfill) {4 mA (0%)}; calibrated span = zero + H(Sp) {20 mA (100%)}. Sfill is the specific gravity for the diaphragm fill; Sp is the specific gravity of the process liquid; H is the height of the liquid in the tank at 100%; and h is the distance between the “dry leg” at the top of the tank and the “wet leg” at the bottom of the tank. The most common error in the calculation is a sign change.

However, there is much more to setting up a dP transmitter than a mere equation. Design errors and control errors can compromise performance. The positioning of the nozzles for the wet leg and dry leg is a common design error. The dry leg should be installed at, or near, the level of the overflow. If it’s installed below the overflow, the operator will be flying blind while the tank overflows. Erring on the high side is better. The lower nozzle should be sited high enough from the bottom of the tank to allow easy maintenance but not so low as to become a stress concentration point or to pose the danger that accumulated solids in the tank could block the nozzle. Fractures in plastic and fiberglass tanks are more of a problem than for metal but thin-wall metallic tanks can suffer damage at stress concentration points from long-shaft agitators typical in fermenters.

One of the most foolish installations I ever witnessed was locating the wet leg at 21% of the straight-side height. The location was justified because of a jacket. Although the operators had learned to live with this design error, it raised an unnecessary risk of pump cavitation. Rather than programming an off-set, the optimum solution would be to install the wet leg in a tee below the pump nozzle. This works best if the tee is oversized so that the velocity in the tee is fairly low.

Another common design error is the valve selection or arrangement for servicing the diaphragms. Often a valve is installed in front of the transmitter so its diaphragm can be removed. Unless a tight shutoff is required for safety reasons, select a butterfly valve. Tight shutoff cannot be assured for all ball valves unless the valve is designed for applications where there is positive pressure only on one side. In addition, ball valves require more torque to open than butterfly valves; avoid ball valves larger than 3-in. Sometimes a diaphragm valve is used in place of a butterfly valve. Occasionally the wafer or diaphragm in the valve may interfere with the transmitter diaphragm. It is a good idea to install a second flange with a short spool to avoid this problem.

Control problems are much easier to fix than design errors. In his book, “Reliability, Maintainability and Risk: Practical Methods for Engineers,” David Smith said that a $100 mistake in design translates to a $10,000 mistake in start-up. In my experience, this is an underestimate.

A common mistake in controls is not testing the high and low alarms. Overflow may occur before 100%. You can estimate the low alarm from statics but the best way is to monitor the pump for cavitation. Set the low alarm at a comfortable level above the cavitation point, typically, 15%. Sometimes, it is desirable to trip the pump at the low level alarm. The behavior of operators has an impact here.

As with all things involving operators, apply “Poka-Yoke.” This is a method invented by the Japanese to mistake-proof an activity. The heart of the method is: find out how an operator can take unauthorized short-cuts and prevent them. As Trevor Kletz points out in “What Went Wrong?” operators may let the trip worry about the level. As long as the instrument(s) works, the equipment and process are safe. This problem also exists for high cut-off. Redundant level switches at the high or low position, whichever is most crucial, may be desirable. However, monitoring operator response is a better solution. Installed properly, the dP level transmitter is highly reliable. It is a small wonder that it has become a mainstay in the chemical industry.

Dirk Willard, senior editor

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.


No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments