Circuit

Consider the following U-Shaped Micro Photoelectric Sensor Panasonic PM-T54P with two PNP Open-Collector Transistor outputs:

I chose the supply voltage to be 12V DC. In the place of the two loads I want to place two voltage meters (to be more precise the NI 9401), which can endure a maximum voltage of 5V DC.

After some calculations, I came up with the following first schematic for Output 1 (the one for output 2 is identical, so I leave it out):

^{simulate this circuit – Schematic created using CircuitLab}

Some explanations:

• The connection to ground at the pull-down resistor R_GND is supposed to prevent a floating input into the NI 9401.
  It should be possible to get a switch into the schematic between V_A and V_In, which physically opens or closes the connection, but I did not know how to do this.
• The resistance R_NI is supposed to limit the current into the NI 9401 to less than 250uA in case the circuit is compromised and the resistors R_A and R_GND are bridged, so that R_NI alone has to deal with the full 12V of V_A.

Questions

1. Do I really need the connection to ground at R_GND? Can one tell from the internal circuit on the I/O circuit diagram if e.g. Output 1 will float or will have 0V if the transistor Tr1 is non-conducting?

2. Let's assume that V_A rises due to a control error of the user up to 30V, but I still want exactly my 12V at V_In.
   Could I accomplish this task with a 12V Zener-Diode between V_A and V_In as shown in the circuit below?

3. Finally I want an additional protection for V_Out, which should be capped at exactly 4.7V independently of what happens before it.
   Could I accomplish this task with a 4.7V Zener-Diode between R_NI and V_Out as shown in the circuit below?

This is my draft for both question 2 and 3:

^{simulate this circuit}

And here is the result for a DC Sweep of V_In between V_In = 0V and V_In = 30V. It seems to be OK, right?

Update

Now it states here that the input impedance of the NI 9401 is 47kOhm, so I tried to create a new schematic accordingly, based on the remarks from Bruce Abbott and rioraxe. Would that be correct?

^{simulate this circuit}

Here's the result: