What is HART? advantages and disadvantages of HART

 

Go to
http://www.hartcomm.org/protocol/about/aboutprotocol_what.html

Click ont the "about the HART protocol" button at the top left to see "what is HART?" and "benefits of HART".

 

In my personal opinion, a major advantage of HART is that it is an open standard supported by many device, software, interface, and signal conditioner manufacturers so you can buy the parts from all kinds of manufacturers. HART has helped displace the proprietary protocols used with 4-20 mA devices in the past. From this perspective, the importance of HART cannot be overstated.

The useful thing about HART is that it enables remote monitoring of device health, remote setup and re-ranging, as well as remote diagnostics. This is what is called: Intelligent Device Management:
http://www.eddl.org/DeviceManagement/Pages/default.aspx

In my personal opinion, There aren't really any disadvantages of HART per-se. However, HART does not do "control" (the subject of your post). HART is superimposed on the 4-20 mA analog signal. That analog signal is what is used for control. Many so-called "digital control systems" in fact have analog control loops relying on 4-20 mA. There are no "HART control systems". The control system has pass-through of the HART signal into the intelligent device management software, this is really where HART is put to good use.

There are several latent disadvantages of 4-20 mA. Two-wire 4-20 mA devices are power constrained to 4 mA which limits the diagnostics and other functionality they are capable of. For instance, some valve performance diagnostics may only be on demand. Thermocouple failure diagnostics alerts after the fact, not before.

Systems that use 4-20 mA for control also use on/off signals for logic which has similar drawbacks as there is no intelligence (remote setup and diagnostics) in discrete devices like on/off valves and intelligence in electric actuators (MOV) etc. cannot be accessed (because all these devices don't support HART).

There is only one 4-20 mA device per pair of wires, and some devices require multiple pair of wires, for valve position feedback and auxiliary measurements etc. This means a lot of wiring, engineering, installation, and many I/O cards etc. The number of connections (cut, strip, crimp, label, and screw down) counting intermediate field junction boxes is huge.

The 4-20 mA signal is just one real-time variable per wire. This is not sufficient to support complex/advanced/sophisticated devices like electric actuators (MOV), multi channel temperature transmitters, and gas chromatographs etc.
http://www.ceasiamag.com/article-6631-thedigitaldrive-LogisticsAsia.html

Since 4-20 mA only handles transmitters and positioners, the system require many other types of I/O cards too to handle the other signals such as DI (NPN, PNP, and AC etc.), DO (NPN, PNP, Triac, dry contact etc.), temperature, and pulse/frequency etc.

Similarly, many kinds of safety barriers are required.

All the 4-20 mA (and other) signals have to be marshaled to the right terminals or the right I/O cards with intermediate connections in field junction boxes. There is lots of engineering work to get this right, and the installation has to be loop checked. We are used to doing all of this, but it doesn't mean it should have to be that way forever.

Adding 4-20 mA devices late in the project require lots of engineering changes because the limits are hard: for example 8 inputs per card or 12 signals per multi-core cable. This means more I/O cards, more cables, more junction boxes, etc. affecting many documents and drawings.

Similarly, adding 4-20 mA signals to existing devices (such as feedback from valve positioners) late in the project require lots of engineering changes; more I/O cards, more cables, more junction boxes, etc. affecting many documents and drawings.
Change device type such as from on/off valve to 4-20 mA control valve or electric actuator requires I/O assignment to be changed and even I/O cards to be added, possibly more cables, more junction boxes, etc. affecting many documents and drawings.

4-20 mA loops are not time synchronized between input scan, control, and output scan. Therefore the control response period is an aggregate of multiple scan times (i.e. longer than the controller cycle time).

There is a risk that during maintenance the 4-20 mA range in a replacement transmitter is set wrongly resulting in skew due to range mismatch between the transmitter and the system. This affects the controls and alarm functions.

Similarly, additional skew may result from differences in the 4-20 mA current calibration in the transmitter output and the system AI card.

Distortion of a 4-20 mA signal such as 18 mA reduced to 16 mA due to high loop resistance and insufficient voltage swing or 16 mA become 18 mA due to ground loop current leakage is not possible to detect because anything between 4 mA and 20 mA looks valid.

The 4-20 mA signal does not use the transmitter over its full sensor limit. The 4-20 mA narrows the measurement down to a portion of the sensor capability used in normal situation, but in abnormal situations it may be helpful to know what the real value is, beyond the normal 4-20 mA range.

Due to the range mismatch skew and current calibration skew, a 4-20 mA five point loop test is required when transmitters and positioners are commissioned. This is a time consuming process. We are used to it, but it doesn't mean it should have to be that way forever.

The 4-20 mA is set below 4 mA or above 20 mA on sensor failure. This mimics an extreme process value, thus triggering alarms and driving the loops to shut down the process - even though it is the sensor which has a problem, not the process. This kind of spurious trips reduces plant availability and productivity.

Because 4-20 mA wiring with AI cards is so labor intensive and expensive, plus the cost of the position transmitter, valve position feedback is rarely provided for all but the most critical valves. However, this is useful information for operators, and can even be used to improve control strategies with true bumpless transfer even on valve hand operation.

4-20 mA only supports a single signal, so multi-channel devices are not possible.

Firmware upgrade on devices using 4-20 mA signaling requires circuit board replacement since there is no way to download the firmware.

Cheers,
Jonas

 

Further to my earlier post;

In view of the disadvantages of 4-20 mA (and on/off signals) I figured it would be interesting to review what advantages a totally digital solution has.

FOUNDATION fieldbus (FF) communication is 25 times faster (31.25 kbit/s) thus enabling real-time digital closed loop control, there is no more analog 4-20 mA.

FF can also take the place of discrete signals too.

FF provides faster diagnostics report by exception, delivering diagnostics within seconds of detection in the device

Since FF has greater bandwidth, loading configuration and diagnostics screens in the intelligent device management software is faster

Bus-powered (two-wire) FF devices are not limited to 4 mA, they can draw more power for more powerful microprocessor and sophisticated software. This enables more advanced diagnostics such as thermocouple degradation detection before failure, and continuous valve performance monitoring. http://www.ceasiamag.com/article-4987-turninguptheheat-LogisticsAsia.html In the future, bus-powered (two-wire) magnetic flowmeters and Coriolis flowmeters may use the additional power to support larger pipe diameters and lower conductivity, etc. More powerful microprocessors will enable configuration of DP flowmeters without special software. Backlit display on two-wire devices could be another possibility. Radar level transmitters with stronger radar signal for greater sensitivity and higher accuracy is already a reality. Gas chromatographs and tank gauging systems are already based on FF and are glimpses into what the future holds.

FF supports intelligence in discrete devices such as on/off valves and electric actuators etc. These account for half of the devices in the plant, so if not digitally integrated, only half of the plant is intelligent.

FF supports multidrop networking of 10-16 devices per pair of wire reducing wiring, engineering, installation, I/O cards, and connections.

FF supports many real-time signals per wire/device, including sophisticated devices like electric actuators, multi channel temperature transmitters, and gas chromatographs etc., drastically reducing the number of wires required.

With FF, only a single types of "I/O card" (interface card) is required, because analog and discrete as well as input and output devices all have the same FF interface and can use the same card.

Similarly, since all FF device signals are the same, only one type of safety barrier is required.

Since FF is digital, the signal marshalling, binding each device to its tag in the system, is done in software which is very easy and flexible.

Adding a device with FF is easy since the number of devices per bus (within reason) is not a hard limit. Additional wiring may not be required.

Using additional signals in FF devices is very easy since this is a simple matter of software configuration.

Changing a FF device to another type of device is easy since the FF signal is the same for all devices, just a matter of software configuration. There is no need to change any I/O cards or do other re-engineering.

FF loops using control-in-the-field (CIF) are time synchronized from transmitter to positioner. Therefore the control response period is simply the bus macrocycle, no more. This is very fast and precisely periodic, ideal for PID control.

Since FF devices transmit the PV in real-number engineering units, there is no range required and therefore no range mismatch and associated problems.

Similarly, since FF has no 4-20 mA, there can be no current calibration drift.

FF communication includes error checking that would detect and reject distorted communications

Since FF communication does not use range, transmitters are used over their full sensor measurement limit, which could provide valuable in abnormal situations.

Since FF has no range mismatch, calibration skew, or signal distortion, traditional five point loop check is not required. A simple plausibility check will do.

For FF devices, the process value has an associated status (device failure distinguished from process problem) which tells the control system and operator if there is a sensor failure (or other problem) without masquerading as an extreme process value, thus not causing a spurious trip, thereby reducing downtime and improving productivity.

Valve position feedback is built into FF positioners and "piggyback" on the bus. There is no need for separate position transmitter, wiring, or input cards. Position feedback is provided for all valves at no additional cost.

Even more wiring can be saved using FF multi-channel temperature transmitters. One multi-channel transmitter connects up to 8 sensors, replacing as many as transmitters, associated wiring, and I/O cards.

FF supports firmware upgrade by download over the bus, without replacing the circuit board.

For now, I would personally use HART for the instruments on my Safety Instrumented System (SIS) but use Foundation fieldbus (FF) for the control system.

To learn more about fieldbus and stuff you did not know about HART take a look at the yellow book "Fieldbuses for Process Control: Engineering, Operation, and Maintenance" buy online:

http://www.isa.org/fieldbuses

Cheers,
Jonas

 

Thanks Jonas. That was a very well written and helpful post.