I'm studying loop powered transmitters (the 4-20mA thing), because I need to make one. I've found two different reference design which mostly use the same technique but with some slight difference.

First one is from TI, appnote TIDU299: TI Circuit

Ignoring the right side protection and filter this is 99% of every transmitter I've seens (except those using dedicated expensive ICs). The appnote explain the working quite exaustively. In short, the loop servoes the whole local ground to maintain the potential on R3 and R4 the same. They use a reference as a regulator instead of an LDO since it's cheaper.

The idea is that most of the current is pulled by Q1 and only a part by the DAC output, the resistor network is to map the output to the 4-20mA range.

It's a little disconcerting (at least for me) a circuit which moves it's own ground and convincing that the opamp is actually in common mode (the R3-R4 node is below ground but the virtual ground is at local ground and it works).

Sadly they don't talk too much about R6 but I think it's simply an emitter degeneration to have gain margin and keep the loop (the opamp, not the current!) happy and stable. As a variation I've seen simply a base resistor (shouldn't change too much)

Analog on the other hand (being Analog, master of horrible complications) proposes this thing, in appnote CN0179:

Analog circuit

The basic idea is the same but:

  • They gone MOSFET instead of bipolar. I think it's only a stylistic preference, right? the zener is for gate protection during transients and the BAS diode is simply for polarity protection or is there some other reason?

  • The feedback is done in another way, on the high side: from what I get they force the current on R2 to have the same potential as on R1 (a 10x current gain in this case). It's only a stylistic choice, it's to have the option to have a 3-wire transmitter or there are performance reasons? From an opamp view I'd say that working above ground has better linearity but they are using their (horribly expensive) rail-to-rail amps so I don't think that would be the reason.

  • They added yet another servo loop as a buffer from the DAC. In the global scheme it doesn't actually add gain apparently: at full range 5V output from the DAC gives 2mA on the central pole which are only 2V on R1. 2V mirrored on R2 gives 20mA, as desired.

Why have they did that? I think it's needed to have the DAC voltage (referred to ground) referred to the upper rail (since the final servo is working up there) otherwise the signal would be inverted or whatever.

All of these extra circuit complication is only to enable a 3-wire transmitter or is there some other (performance) reason? after all they present this as a 14-bit grade transmitter (with 0.1% components) while the TI circuit is by design the cheapest thing possible. One idea at first glance is that the DAC output is buffered and so maybe with less load the stage has more linearity?

Any other idea? (of course I will not use the Analog schematic but I think it's interesting to know why they did it that way)

  • \$\begingroup\$ You might also expand things a bit by looking at this answer and its question regarding a 4-20 mA sensor approach. \$\endgroup\$
    – jonk
    Commented Oct 22, 2021 at 19:16

1 Answer 1


The second circuit is only for a 3-wire transmitter so it's different.

As you have noted there is no feedback around the output current back to A1 in the 2nd circuit so if either transistor was a BJT there would be an hFE-dependent error. Mirroring off the positive rail allows the ground to be common with the power supply. I don't think it can be much simpler without compromising accuracy and there is nothing at all wrong with it (for example, it's very forgiving of supply voltage changes). There's no technical reason to use R-R op-amps, but it makes it slightly easier (avoids a couple of components) and allows more compliance.

The first circuit is a 2-wire transmitter- measuring its own power supply current- so it is quite different. I suspect R6 also prevents the circuit from latching up by collapsing its own supply voltage as well as limiting the maximum output current. The op-amp must be one that does not exhibit phase reversal or the circuit can latch up. It also can't draw too much current or you may not be able to get safely below the 4mA level.

I don't see any significant inherent difference between the two topologies in terms of accuracy- if you want high accuracy the resistors, reference and DAC are going to be relatively expensive.

  • \$\begingroup\$ So it's as I suspected it's more complicated just to allow a 3-wire (sourcing) connection. Thanks for the advice \$\endgroup\$ Commented Oct 22, 2021 at 13:30
  • 1
    \$\begingroup\$ @LorenzoMarcantonio You really have to watch the current draw with a 2-wire design- quite often, it's necessary to add galvanic isolation which is really not so easy with a 2-wire design. It's easy with a separate power supply. \$\endgroup\$ Commented Oct 22, 2021 at 13:52
  • \$\begingroup\$ I know the issues for isolation; however in my case I have both transmitter and receiver under control and I'm doing a look exactly to avoid a third wire \$\endgroup\$ Commented Oct 25, 2021 at 6:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.