sensor terminology

Good reply!

Well why is it called a 'loop'?

Because the signal travels to and fro down the twisted pair wires (or whatever wiring is used) So the signal is in a loop to and from the reciever or transmitter.

John.

Every electrical source / load configuration builds a loop because the current flows (very simply spoken for DC) from the source to the load via wire#1 and back from the load to the source via wire#2.

But you don`t only want to power your sensor, you want a result back.
In this case the source is your loop supply which may be even an unregulated DC voltage of, lets say, 12 to 24 V. And your load is the sensor which "decides", determined by your measured value, how many current it draws from your supply. So, by only observing your supply, you get your signal. The advantage of this concept is that you need only 2 wires between supply and sensor - conventionally powered sensors need (at least) 3 wires: supply, signal(back) and ground.

The sensor`s power is taken from the lower 4 mA that will ever flow, independant of your measurement. If you have a temperature sensor with 0 ... 100°C range 4mA will flow at 0°C, 12mA at 50°C and 20mA at 100°C - so the whole range is represented by 16mA per 100°C.

Another advantage is that a broken wire will show 0mA which is an error (because it`s < 4mA). With other systems 0...20mA (instead of 4...20mA) you would not know whether your wire is broken or the temperature is 0°C, compared with the above example. Because of that function, 4...20mA systems are also called "live zero" systems.

An additional advantage is that your loop is low resistant and differential which makes it very stable against EMI.
Note that this is an analog interface, not digital. But sometimes a small amount of digital signals are added (without corrupting the analog information) in order to transmit calibration data to the sensor (also see HART-protocol).

A good introduction of this technic is shown at analog devices or texas instruments (formerly burr brown chips).
Regards Uwe

Now that is an explanation and a half.

Thank you for submitting your reply, very interesting and educational.

One question though from the view point of the layman, What if you power more than one sensor with this loop, could it still work and how will the system tell the difference between the multiple sensor power takes?

Also would like to know the price comparison to lets say a conventional sensor system with PLC via control relays?

Thank you in advance.

Case491

<One question though from the view point of the layman, What if you power more than one sensor with this loop, could it still work and how will the system tell the difference between the multiple sensor power takes?>

It doesn't work like that. On a control loop there is only scope for one signal transmitter. It is the transmitter that generates the 4-20mA that other devices in the loop use to determine and present the position in the measuring range that the transmitter is trying to send. If there is a need for several analog sensors, then there is a need for a similar number of control loops.

<Also would like to know the price comparison to lets say a conventional sensor system with PLC via control relays?>

However the final control is arranged, there is always a need for a measurement. And if the measurement is over a range, as compared to a there/not switching function, then an analog transmitter or its modern digital cousin is always required. If one has a suitable transmitter for a task kicking about in a drawer somewhere then discussion of costs is somewhat academic.

So there is no way you could hook this up to a multiple loop and assign a dedicated transmit and response time per unit in the loop? A bit how they do it with a bus around a machine and each item has an ident code or digital signature, similar to the various hardware items in old computers with scusi.

An adressable multiplexer (relais or solid state) with common output to one load resistor would solve that problem. An adress counter incremented by timer brings you the cyclic analog signals. This is often done in large fields of temperature sensors where the sensors response time is much larger than the cyclic signal acquisition time - so you have no loss of information.
Regards Uwe

Unfortunately it`s not possible to drive more than one sensor in that way and get different signals. But as often there are two exceptions for special application:

1. parallel connected sensors sum their current in the load`s current knot; so if you have n sensors the total amount of current divided by n gives you the average of all,

2. serial connection makes the sensor with the lowest value limit the current in the loop; so you`ll get a minimum function fo all n sensors.

The question of costs will be answered best with a concrete example because there are too many determinants. As a rough thumb rule a loop makes sense when you have cable lengths of more than 50-100m (depending on your environment) because the difference of cable costs eats the difference of loop device costs.

And sometimes loops are really good - think of digital display meters of analog 4-20mA-loops. They need to be powered, and they use the "lower" 4mA via a DC-DC-converter to power the LCD.

Post#10 had the question of resistance and differential signal configuration:
Lets assume you have a supply voltage of 24V DC and a current of 20mA max. Lets further assume your current source has a minimum voltage of 2V. Then the allowed loop resistance is (24V - 2V)/20mA = 1100 Ohm which is rather low compared with the high impedance of con-conductive EMI-sources. The EMI-source and the loop resistance build a voltage divider greatly reducing the disturbances with the appropriate factor - one of the advantages of such loops. The loop resistance is the sum of wire#1 (source-load), the loadimpedance, wire#2 (load-source back), source impedance. The lower source and load impedance is selected the more wire length is allowed.

The signal may be called differential because you have only two wires, exactly parallal, maybe twisted (which is the best) and with the exact amount of current flowing, only with different sign. So theoretically the radiated field strength is zero and the influence by external field disturbances is zero, too. That makes wiring and EMI protection very easy.
Regards Uwe

great answer!

uweka,

What do u exactly mean, by 'the loop is low resistant and differential'?

In addition to the information in other replies above, a feedback control loop consists of:

• A setpoint. "I want to do 40mph now as I've left a built-up area (set point change)".
• The measurement. "I'm only doing 30mph".
• A comparison between the setpoint and the measurement. The result of this subtraction is called the 'error'. "I'm 10mph too slow".
• Some form of corrective action algorithm intended to reduce and eliminate the 'error', incorporated into a thing called the 'controller'. "Push harder on the gas pedal/accelerator pedal by 'this much'. I've driven this van before and it is no match for my Ferrari - wouldn't dare push that hard in that!".
• A final control element. "The engine throttle opens wider".
• There will be disturbances to the 'process'. Going uphill, the van slows down. Going downhill the van speeds up. The hill is the disturbance. "Speed is dropping. Squeeze harder to keep speed up".
• Then there's the 'process' itself. "This van has a 0-50mph time of 22.7 sec fully loaded on the flat. My Ferrari will do the same in 4.1 sec."
• Then there's the behaviour of the measuring element to consider. "The speedometer wobbles about a bit and is a bit slow in responding".
• Loop up to the top and start again.

There will be other control loops running in parallel with this one. The simple example is the one that keeps the van on the road.

• I want to steer down the centre of this lane and turn right at the next intersection (set-point change).
• I'm a bit over to the left.
• Turn the wheel slightly to the right.
• The van steers over to the right, having lurched a bit.
• Oh, now it's going too far to the right and the camber on the road surface has changed.
• I need to allow for that soft tyre/tire on the back. Mustn't turn the wheel too suddenly.
• Loop up to the top and start again.

Get the idea?

The human brain is an excellent multi-loop controller. Tuning all the control terms properly for driving a road vehicle comes with experience, and the result of getting it right might be called skill or competence.

DC 4>20 ma looped (even 1 wire) around what ?to generate what?Power to feed other Sensors?

Heard nothing more sacrilegious

"Loop" not only in Sensor terminology but throughout the "Control" is used which means:

A signal is fed-back to the control-element to regulate any changes in out-put.

In speed, Voltage, Current etc etc controls is used.

Not only in Electrical & Electronics but also every-where like Mechanical, Hydraulics, Pneumatics etc etc Even in Automobiles "Loops" are involved.