Here are details about the issues my customer is having with the AD5421.
Attached is a screenshot of the current loop transmitter page.
"Our test setup includes (3) 9V batteries in series to make a 27V isolated loop supply.
The side of 27V connects to LOOP_P. The side goes to a 265 ohm resistive load
and the other end of the load attaches to LOOP_M.
We have 2 issues with the circuit.
1) If nothing is attached to the loop, we don't read anything out of the SPI interface -- the SDO
pin is always low. This seems odd since the part is supposed to detect certain faults in the
loop, and if the loop breaks such that there's no loop voltage, how would we find out? In the
circuit, D39 and R79 provide constant power to the IODVDD pin which I thought would avoid
this problem. Strangely, if I momentarily connect the loop and then disconnect it, we begin
getting data out of the SDO pin and this seems to go on forever, or as long as we test it.
My question here is, what's the proper way to use the part so it can be communicated with
when the loop isn't powered? (It's on a shared SPI bus, so the digital signals must behave.)
2) In debug of problem 1, we connected a diode from a 6V supply rail to the REGIN pin.
This "fixes" problem 1 so we can talk to SPI all the time, even when nothing's connected
to the loop. Then we started setting DAC values and observing the current through the load.
Here's a table of what we see:
DAC value mA (loop)
0x0000 4.0 (good)
0x2000 5.99 (good)
0x4000 7.97 (good)
0x6000 10.01 (good)
0x8000 12.78 (high)
0xC000 49.7 (very high)
0xFFFF 49.9 (very high)
As you can see, the low current values look correct but as the DAC value is increased, the
current eventually runs off the chart. I don't see any reference in the datasheet about how to
get 49 mA into the loop. I'm not sure that the 6V fix for problem 1 could cause such high current,
because I'm calculating around 13V on LOOP_P in the 49 mA case (so OR'ing in 6V shouldn't
make a difference.) Oh, our ALARM_CUR_DIR = 1 so we'd expect "full" current during alarms,
but this ought to be 24 mA not 49 mA.
If you can suggest which pins to check I can do that.
If you would like additional information, sketches or circuitry external to this page I can provide that.
But I thought we could start with the core circuit in case there's a problem right there."