Why we use 4-20mA?

Loops can be:
4÷20mA 2÷10V
2÷10mA 1÷5V
0÷20mA 0÷10V
0÷10mA 0÷5V

For example if we have a 4 ÷ 20mA loop, it will have a range of 20mA - 4mA = 16mA. This is domain.

By convention 4mA is the lower limit (LL = Low Limit) and 20mA upper limit (High Limit = HL), the rest of the field values are intermediate states (points).

How can we force the lower limit, quite simply, send 3 mA current in loop.
How can we force upper limit, as simply , send 20mA current in loop.

With domain can juggle depending on the application and the required measurement technique.

Examples - depending on requirements:

1. The simplest application, two states: ON / OFF. Assigned to each state, in a simple logic value, for example OFF = 4mA and ON = 20mA or in reverse logic OFF = 20mA and ON = 4mA.

2. Limit cutting at the low (cut off low limit) for example for signal value <4.05mA, ON state is achieved.
Limit cutting the upper limit (cut off high limit) for example for signal value >3.95mA, OFF state is achieved.

3. An application with three states:
Low Limit 0% 4mA
Middle Limit 50% 12mA
High Limit 100% 20mA

div. = (20 - 4) / 2 = 16/2 = 8 mA
8mA = 4mA + 12mA
12mA + 8mA = 20mA

4. An application with four states:
Low Limit 0 0.0% 4mA
Intermediate Limit 1 1/3 33.3% 9.33mA
Intermediate Limit 2 2/3 66.7% 14.67mA
High Limit 1 100.0% 20mA

div. = 16/3 = 5.33mA

5. An application by 5 states:
LL 0% 0 4mA
25% ¼ 8mA
ML 50% ½ 12mA
75% ¾ 16mA
HL 100% 1 20mA

div. = 16/4 = 4mA

6. An application with 17 states:
LL 1 0 4mA
2 1 5mA
3 2 6mA
4 3 7mA
5 4 8mA
6 5 9mA
7 6 10mA
8 7 11mA
ML 9 8 12mA
10 9 13mA
11 A 14mA
12 B 15mA
13 C 16mA
14 D 17mA
15 E 18mA
16 F 19mA
HL 17 G 20mA

div. = 16/16 = 1mA

7. From more points, increase the number "n" to improve the accuracy of the loop

div. = (20-4) / (n-1) = [16 / (n-1)]mA

On a single loop, can imagine orders for multiple devices eg 4:

LL LH
Device 1 4 8
Device 2 8 12
Device 3 12 16
Device 4 16 20

I'll just address the 250 versus 200 ohm dropping resistor. While 200 is a standard the 250 provides a standard 1-5 VDC signal. Also, or maybe primarily, since the 250 is non-standard that requires them to be specifically ordered therefore enforcing the likelihood of making it a high precision, temperature stable 0.1% resistor. Putting a 5% or 10% precision resistor in the loop kind of defeats the purpose.

Pneumatics aside, prior to the 4-20 mA standard evolving, it was 10-50 mA DC signals that were standard and a 500 ohm dropping resistor. The 10mA signal was sufficiently above the minimum power level required for the electro-mechanical components of the early transmitters. It has been too long ago for me to recall what the O&M said about the voice-coil circuit as used in Foxboro DP and pressure transmitters but I vaguely recall something like 6mA.

Another item that came with purely electronic transmitters and just general improvements in electro-mechanical and electro-pneumatic instruments was moving away from 48 VDC powered loops to the 24 VDC powered loop. Some advantages there were a safer circuit and one that was better suited for use in hazardous environments.

Bottom line is that as with all standards, something has to be decided upon at some point in time. That time may have certain limitations that disappear as technology improves and are the limitations are forgotten. A fair example might be why was 8 bit in computers an accepted standard for so long when 16 is so much better? Same reason Bill Gates selected "just" 256K for the memory of his first computers, he couldn't not see how anyone would need more than that!! Mainframe computers back in the mid 70s used magnetic memory core boards with just 4K on a single approximately 17" wide by 12" deep board!! I still remember when I saw my first 256K board in 1979!