dear Mark,

Ideal voltage or current controlled current sources can perform the function which you desire. Such sources are available in any SPICE simulator. This is if i understood you?. The constant current diodes are depletion junction FETS operating at VGS=o.

They can be controlled by their trans conductance parameters in the  SPICE.

In principle , all transistors can be operated as a controlled current sources.

This may be to start the discussion on the problem.

wish you success

The Following Reference to Circuit images of Current limiting Diodes from 1 Micro Amp to 1 Amp may be useful for Simulation.

1.  https://www.google.com/search?q=1+Micro+amp+to+1+amp+current+limiting+diode&tbm=isch&tbo=u&source=univ&sa=X&ei=p6oEVPefK9ihugTDjoKYCQ&ved=0CDUQsAQ&biw=1280&bih=587

P.S.

Dear Abdelhalim Zekry and Pisupati Sadasiva Subramanyam,

Thank you very much for your contributions.

Last summer I searched extensively for active current limiters and passive current limiting diodes online including via the visual method suggested by Pisupati. I noticed a curious phenomenon. The more a person failed to find a passive current limiter, the more words they seemed to write about it, and, apparently, the higher its Google page rank!

Since the best candidate so far is a single line of SPICE code, it occurred to me that there may be plenty of people who know how to code this, however, the solution is so obvious - TO THEM - that it is not worth making a blog entry about, so, leaving the solution with a low Google page rank! Because of the natural human tendency to write more about interesting unsolved problems more than boring solved ones, if the boring solved one is out there, I am unable to find it.

I also found that each time I changed to analysis parameters of NUTMEG, in some cases small instabilities would be introduced (or a few instances large ones) and I became convinced that the design of my current limiter was not stable enough to be blindly trusted. For the few fixed current level diode models I found at various manufacturer sites, I did not have this behavior, however, I could not make the needed adjustments to tune the circuit, which is an electrical analogue of a biological system, not an analog of a real circuit.

This is why I have turned to ResearchGate. Perhaps I should post where the SPICE modelers "hang out" but I have not found these places on the Internet yet.

I welcome your further suggestions.

For what it may be worth, I have included a drawing of the desired behavior.

Thank you,

Mark

Dear mark,

I think the problem became clearer. From my point of view, you want to synthesize the transfer characteristics shown in the figure above.

The given transfer characteristics can be best constructed using an ideal npn transistor  in the common emitter mode driven by a constant current at the base IB and the the collector is biased through a proper power supply voltage in seies with ad a proper collector bias resistance RC and a current sensing resistance RS of  one ohm.

To get unity current gain we have to make beta of the transistor = 1. We have to adjust the collector biasing circuit to get current saturation at ICsat of one ampere . Then assume VCC= 10 V, VCEsat =.3V, then (RC+ RS)= (VCC-VCEsat) /ICsat= 9.7/1= 9 .7 ohm. Then RC= 8.7 ohm.

The voltage sensed on the the one ohm sensing resistance= IC= Iout. The output is taken across the RS.

This circuit works as a current controlled source to produce the linear part whose current is limited by a passive current limiter formed by Vcc and RC.

A similar circuit can be built with common base transistor circuit.

I hope this solution will satisfy your requirements.

Ready to discuss any matter with you.

Wish you succes

Dear Abdelhalim Zekry,

Thank you for your kindness in answering my question.

I added a simplified schematic of the system I am modeling as an electrical analog in SPICE to my original question, alongside of the characteristic curve (which fails to conserve current above Imax, so, perhaps my curve itself contains bugs!) which I was also late to add to the original question.

The reason I want a passive device is that I do not want to repeat trouble I have had with SPICE as a beginner (I am picking it up again after 25 years) with active devices deployed to simulate passive devices.

There are nonlinearities around t=0 and v=0 I get with B supplies that I do not get with passive models such as those from Central Semiconductor:

http://www.centralsemi.com/content/engineering/spicemodels/spicecld.php and

Beige Bag Software:

http://www.beigebag.com/case_cld.htm

however, upon looking at the circuits, these models also contain ACTIVE elements!

So, before I write down the emitter-follower (common emitter) circuit you suggest, may I ask a few questions?

• Since I work in the 0-7V range, silicon and germanium junction drops are significant. What transistor parameter(s) can I adjust to bring these to zero, or, should I use a high-impedance device instead, for example a JFET?

• May I draw my own version of this elementary circuit (I remember it from college and from teaching myself a little electronics in high school) and code it in SPICE and send it to you to make certain I understand?

Sincerely,

Mark

I corrected my characteristic curve to end at Iin = Imax to eliminate the nonconservation of current problem.