If anyone has an objection to the way I work, so be it.

I apologize if I was a bit cranky. I guess I've become a bit of a curmudgeon in my 82 years (my family sometimes complains).
It's just that one of my pet annoyances is posters who only answer one question and ignore others.
Just a short comment to say you are working on the unanswered questions would be good.
I do appreciate you moderating this site and your many learned comments.
Cheers,
Zapper

 

This sim shows the effect of changing load resistance from 250R to 1000R in 250 steps.
As you can see, other changes to the circuit would be needed to keep the 20mA pulse constant.

My sim didn't show any significant variation in current while stepping the load, as each current step is virtually on top of the other (below):
I did have to add capacitor C2 to prevent oscillations at the higher resistances.

So the difference is likely the LM317 model. The question is which model is closer to the real device(?).
I attached my LM317 files, if you care to experiment with them.

 

Hi Carl,
I was referring to my simplified circuit in post #16, or the very basic circuit in post #20.

Initially, I thought we were dealing with a specific 'one off' requirement rather than a range of applications

I have asked @Jake-S to expand his idea's on using an SSR.

E

 

I was referring to my simplified circuit in post #16, or the very basic circuit in post #20.

My simulation was in reference to yours in post #19.

 

hi C,
Got it.
I have run the RL change as a List, when it gets to 750R, the Load current in R2 drops below the required 20mA.
The 250R & 500R show OK, for the 750R, I(R2) is down to 16mA.

This problem is caused by the Zener diode, without the diode all OK.
E

 

This problem is caused by the Zener diode, without the diode all OK.

Okay, I didn't have that in my simulation.
But you didn't observe the oscillations I saw.
That may be due to LM317 model differences.

 

hi,
I did see those oscillations on the higher RL value, suspect the LM317 model.

BTW: the TS is concerned that the load voltage/current may exceed 5v/20mA, on an open load condition, so a clamp maybe required. I used a 5VZ on my basic circuit.

E

 

BTW: the TS is concerned that the load voltage/current may exceed 5v/20mA, on an open load condition, so a clamp maybe required. I used a 5VZ on my basic circuit.

Yes, that's a good solution.

Depending upon how accurate the model is, the capacitor I added to suppress the oscillations may also be required.

D4 is likely not needed as I doubt that the reverse leakage of the D1 small diode is sufficient to damage the LM317 circuitry.

 

Hi guys, apologies I moved onto another circuit for a while.

Below is the SSR idea I had - I could not find a DPDT relay for LTspice so used Cicuit Wizard - but it should show the idea.


I put a dummy load resistor in just incase the current source would come into any grief if periodically going open circuit and thus ramping up to its Vcc.

Obviously this is a more expensive circuit, but that doesnt really matter in this circumstance. I was hoping the above would give a more consistent 20mA, as the previous simulations show a sensitivity between the 4-20mA source and the clipped target 20mA under excitation of the LM317.

Hope the above makes sense, please share your thoughts.