R4 and R5?

A typo, yes.

 

the circuit, as given and even if you fixed the topology-related issue, will give output current that varies with the load resistance.

the follower in the original circuit exists precisely to counter that.

True in general.
But the values shown readily meet the the OP's requirements without the need for the added follower.
Even better accuracy (output impedance) can be achieved if R1 through R4 are changed to a higher value such as 100kΩ or 1megΩ.

 

But with it I cant use the OpAmps I have or even know which one to use for correct working.

that's the danger of simulation: opamp models don't model the amp's inability to deal with near-rail signals. your simulation may show that the opamp works fine 10mv or lower close to the rail / ground, when in fact most r2r opamps cannot even do that.

for simulation to be helpful, you have to know where its limitations come into play.

 

that's the danger of simulation: opamp models don't model the amp's inability to deal with near-rail signals. your simulation may show that the opamp works fine 10mv or lower close to the rail / ground, when in fact most r2r opamps cannot even do that.

for simulation to be helpful, you have to know where its limitations come into play.

So the circuit can't really be truly simulated? Even usig a r2r OpAmp the current is changing with the load.

 

So the circuit can't really be truly simulated?

I think dannyl has given you an unduly negative view of Spice simulators.
It can be simulated within the limitations of the op amp model.

You can test the model to see how well it matches the data sheet specifications for operation near the rails.
So you simulate the op amp by itself in a follower configuration and see how close it can go to the rails with various output load resistances, and then compare that to the data sheet spec.
I think you will find that most op amp models come reasonably close to simulating the actual values.

Even if the op amp can't go fully to the rail with you maximum desired load current it may still go far enough to meet your circuit requirements.

 

I would like to Update this topic just to say what happend next.

I mounted it in a Breadboard and everything went ok, so I designed the PCB, Only that R5 and R6 are 500 Ohm:



And here It is:



2.3 mA and a Load Resistor range from 0 to 1k Ohm.

Thanks to all.

 

That is a very pretty board. It is difficult to tell (a lot of scrolling up and down) but are the inputs to the unused opamps connected to anything? It is a good practice to connect them to something, even to each other, but it probably seldom matters, so this question is just out of curiosity.

 

That is a very pretty board. It is difficult to tell (a lot of scrolling up and down) but are the inputs to the unused opamps connected to anything? It is a good practice to connect them to something, even to each other, but it probably seldom matters, so this question is just out of curiosity.

No, they are not, but I will take that on count next time.