Level Measurement DP type erratic behavior



We are using diaphragm type Differential Pressure transmitters for level measurement of different vessels. Each vessel consists of two transmitters one with PLC loop (HIMA with current isolator card CS4/420) {let name it A} and other with DCS loop {let name it B}. The transmitter A show erratic behavior while transmitter B is relatively stable.
• Both transmitters are of identical model and make and have exact same configuration & installed at same vessel and diaphragms are located at adjacent locations,
• The grounding connections were found ok.
• When transmitter A is forced to any value the trend behaves normal and sets to a fix value
• The transmitter A was replaced with new one but behavior didn’t changed.
what could have caused the erratic behavior?

thanks in advance


From where you are reading the value ? From DCS or From the instrument directly?


Both from DCS by getting trends and through instrument by reading current mA values, both are fluctuating. find attached image when transmitter is in operation. in images the green colored is tx B where as blue colored is tx A



in images the green colored is tx B where as blue colored is tx A


in images the green colored is tx B where as blue colored is tx A

  1. Out of curiosity, >From where you are reading? “Both (signals) from DCS”
    If signal Tx A goes to a PLC, how does the DCS ‘see’ the Tx A signal to be able to record it as shown in trend charts? Is there a Tx A signal going to the DCS as well?

  2. On the first trend chart (this forum does not number posts for identification), is the Tx A blue flat line at 09:15 onwards a known force condition ?

  3. How do you initiate a force condition? HART command to force 4-20mA output?

If Tx A reads OK during a force output condition for an extended period, (not just one minute because the Tx A noise appears to have about a one minute cycle time), then the loop electronics (PLC AI, Isolator, loop power supply, wiring; no ground loops) tests OK.

That leaves the pressure sensing. But a replacement xmtr has exactly the same problem? Exactly?

Then I have to ask

  1. By “diaphragm type DP transmitter” do you mean a DP with remote seals with capillary tubing? Dual (2) remote seals? symmetrical (same length capillaries)? Or a single low side capillary?

  2. Who installed the remote seal(s) on the transmitter? the xmtr manufacturer? a professional 3rd party seal installer? some guy in the shop?

  3. Is the DP high side at the tank bottom tap and DP low side on the top tap?

  4. How are two DP’s connected on the same tank?
    Are there two flanges at the bottom tap, one for xmtr A, the other for xmtr B?
    How are each of the DP high sides connected? Flanged? Piped from the same flange/tap ? Are blocking valves installed? One for both? One for each?

  5. What part of the world are you in? What’s the climate there now? hot desert? winter near freezing?

  6. Are the contents of this tank heated? Are there steam lines in the area of this tank, for heat tracing or process heating?

  7. What is the vapor blanket pressure at the top of the vessel?


Dear carl_e,
appologies for late replying

  1. As in trends both trends are made in DCS. The PLC (HIMA) loop contains signal splitter i.e. cs4/420, that then gives signal to DCS as well as current switching card F6250.

2)yes that constant line is force condition

  1. Yes HART is used to force 4-20 mA transmitter output.

  2. Yes it is Differential Pressure transmitter with two remote seals high and low, with capillary tubing

  3. The remote seals are installed by manufacturer

  4. The high side is at bottom

  5. These are flanged connections, two at top and two at bottom. There is no blocking valve. the flanges for each tx is different and is located adjacent to other.

  6. Its subtropical region. The temp in summer goes to forty+ and in winter falls to 10 deg celsius.

  7. Yes, and there is steam tracing installed around the flanges.

  8. the gases pressure is around 4kg/cm2

Please find attached datasheet


The flat, stable output signal during a HART output force means that the problem is not electrical, it’s a pressure measurement issue.

For a DP to read consistently low means either

  • high side (bottom) is reading too low, or
  • Low side (upper flange) is reading too high

What can cause erroneous pressure readings?

Air bubbles in the seal/capillary fill fluid will produce low readings. My experience is that factory filled capillaries are properly made, but I have seen shop fills that have left an air bubble in a capillary.

If the construction is not ‘all-welded’ (capillary welded at both ends to the seal and the DP process head) mechanics have been known to unscrew the capillary from the seal and then screw it back together. That always introduces an air bubble. DP’s with capillary seals should never have ‘vents’ in the process heads but I’ve seen it happen. Opening a vent will put an air bubble into the system and cause a low reading.

A leaking flange gasket will cause a low reading but a leak should be obvious with a blanket pressure of 4 bar.

Seal/capillary temperature differences between upper and lower seals will produce a measurement error. A higher temperature causes the fill fluid to exert a higher pressure. A lower temperature causes the fill fluid to exert a lower pressure.

If Heat trace were turned on only on the low side, that only could elevate the low side fill fluid temperature, drive the low side measurement up with the result that the DP (DP = high side minus low side) is low.

A kink (sharp bend) in the capillary can cause noise as the pressure jumps as fill fluid moves across the obstruction. But more often the kink passes no pressure. If the kink is in the high side, the level appears to be ‘stuck’ at one level. The level changes when there is a kink in the low side with an error based on the difference between the trapped pressure and actual pressure. All that said and your datasheet spells out armored sleeving on the capillary which helps prevent kinking.

There’s always the potential for a bad sensor in the transmitter. It isn’t common, but does happen.