Figure 1: Programming a HART instrument so that processes can be monitored from the control room. Source: SiemensIn our technical support department at Siemens, we talk to many customers who have done so and found that it does just work. This is especially true if customers have wired into a HART input or output card and only going a short distance. However, we have also found some people who do run into problems.
For HART communications to work effectively, you need:
- A minimum loop resistance of 250 ohms
- A resistance Capacitance (RC) value of the network of 65 μs or less
- To meet the HART specification on the power supply.
If you are using a HART input or output card, then the so-called HART resistor is built into the card. If your run is short (less than 100 meters) and you are using typical instrument-grade wire, then the RC value of the network will be under 65 μs. Most industrial power supplies meet the HART power supply specification.
Customers missing the HART resistor are our most frequent technical support calls. If the wire distance is less than 100 meters, then, typically, just adding a 250 ohm resistor will get HART communications going.
When the network is long or you are using multi-drop, then you really need to calculate the RC value of the network and verify that it is 65 μs or less. This forms an upper limit on the loop resistance. For longer runs, sometimes the value of the HART resistor has to be lowered for HART communications to work.
Using a power supply that does not meet the HART specification is rare but has happened. The requirements are:
- Maximum ripple (47 to 125 Hz) = 0.2 V p-p
- Maximum noise (500 Hz to 10 kHz) = 1.2 mV rms
- Maximum series impedance (500 Hz to 10 kHz) = 10 ohm
The impact of the wire comes in two ways. First, it's capacitive and resistance values will certainly affect the RC value. Second, it needs to have a good shield on it and it should be grounded at one end only. This will protect the communications from electromagnetic interference.
Myth 2: HART is so simple that you do not have to design the network.
Well, if going point-to-point and the distance from the HART input/output card to the device is less than 100 meters, then this is true.
As we discussed above, if you are (a) using a HART input/output card, (b) going under 100 meters, and (c) using good shield instrument-grade wire grounding at one end, then it will work. As it turns out, this is the most common situation. However, if the runs are long, and/or multi-drop topology is being used, then you will need to ‘design' the network.
Myth 3: You would never want to multi-drop HART.
The truth of this myth really depends on how fast the data is needed. HART multi-drop is very slow. To control a process, if you need to know that the value (level, temperature, flow, pressure, valve position) has changed in the last 0.5 second, then you would not want to use HART multi-drop. However, if you are simply monitoring slow-moving variables, then this can work and there can be a real cost savings in wiring and components by using HART multi-drop.
When HART is multi-dropped, the analog channel cannot be used so all communications must go through the digital channel. The digital channel runs at 1200 bits per second and has a throughput of around two to three messages per second. If two instruments are multi-dropped, the process variables are updated every second. If eight instruments are multi-dropped, the process variables are updated every four seconds.
One challenge in using multi-drop is in the design. Calculating the loop resistance and the RC value of the network is no longer a trivial calculation – and it is not optional. This is an application where the network has to be designed and components chosen carefully.
To say that you would never want to multi-drop HART, then, isn't necessarily the case, depending on these variables. If the application works, using HART multi-drop provides an advantage in capital cost savings on wires and components.
Myth 4: HART devices have fewer diagnostics than other protocols.