XTR111: EF pin with 0V input

Hi Jacob,

I have performed some bench testing in our lab and am observing similar behavior to what you've described. It does indeed appear as though when Vin is at GND, the error flag will not reflect a wire break. This makes sense - the fault detection is predicated on the idea that if there is a wire break, the current will be unable to flow, causing headroom issues at the current mirror that can then be detected. If there is no output current flowing, then the aforementioned condition will never arise, and thus we cannot detect a fault.

It seems the actual threshold at which the error flag does reflect a fault may be device dependent. The unit I tested saw the error flag go low when Vin exceeded about 2mV, but for Vin < ~2mV, the error flag would be high during an output open-circuit condition. This falls quite close to the expected offsets of the XTR111 device (and certainly would be a reasonable value for the offset of an internal comparator or similar circuity). Consider also that the device is not actually specified to operate linearly down to 0mA but rather down to 0.1mA, which in your configuration would be roughly Vin = 10mV. The inherently random nature of the component offsets would explain why you see the behavior on one unit but not on another unit.

Would it be possible for you to incorporate a deliberate offset into your application, so that for example your valid signal range was from 4mA-24mA or 4mA-20mA instead of 0-20mA? This configuration is common because not only does it allow other circuitry to be powered with that 4mA of offset, but in the case of a Vin of 0V, you will still be able to detect a wire break condition. All you need is enough offset current for the error flag to be consistently reliable regardless of Vin, so even 1mA or so of offset would likely work. You can use a pullup to VREG (as the XTR111 EVM does) to achieve this offset, allowing Vin to still go down to 0V.

Cheers,

Jon