The other day, I came across an interesting device called a current limiting diode:

One producer is MCC (www.mccsemi.com). It makes a 20 mA version (CLD20), which was the highest current rating I could find. Other devices seems to be in the low, single-digit mA range. A review by Sze Chin (ECN, May 15, 2000, p.49-53) gave some interesting applications, including simple triangle and saw tooth wave generators based on charging a capacitor.

The only application the datasheet shows is its use for LED strings (http://61.222.192.61/mccsemi/up_pdf/CLD20(DO-214AC).pdf ).

Does anyone here have experience with them? Specifically, if one uses a CLD to create a triangle wave, how linear are the ramps near the top? I would order some and experiment, except they are a little hard to find in stock at DigiKey and Mouser.

John

 

They are basically JFETs with the gate tied to source, with a resistor in between to set the current.

 

That is how they are described. If one looks at the datasheet for the CLD 20, it has a limiting voltage of 4.5V. As I interpret that, it stops regulating current then the anode-cathode differential is <4.5V. Thus, the capacitor charge curve should become non-linear as the voltage differential passes that point. I was wondering how that effects the triangle (or other) waveforms. Imagine, you were using one of these in a 555 circuit to control triggering. I was looking for real-life experiences.

Are these things really just useful for lighting LED strings?

John

 

That is how they are described. If one looks at the datasheet for the CLD 20, it has a limiting voltage of 4.5V. As I interpret that, it stops regulating current then the anode-cathode differential is <4.5V. Thus, the capacitor charge curve should become non-linear as the voltage differential passes that point. I was wondering how that effects the triangle (or other) waveforms. Imagine, you were using one of these in a 555 circuit to control triggering. I was looking for real-life experiences.

Are these things really just useful for lighting LED strings?

John

Well, honestly, I don't see how you'd use them for LED strings. Sure, they would regulate current, but you could probably get away with a resistor, and handle a lot more current; plus, they are usually small devices, so can't dissipate much power. If you really needed current regulation, you probably wouldn't need super stable regulation, and you could use a simple NPN transistor current sink. If you want a sawtooth wave generator, then this circuit should do (a 555 timer sawtooth oscillator.) I've replaced the PNP current sink with a current source (which in reality would be the current diode, but the circuit simulator doesn't support it.) Note how voltage drops to zero as the discharge resistor may be very small (given that the current source limits the current that may flow into the discharge junction, protecting it against any damage.)

 

Here's a link to the the application note that I have:

http://www.centralsemi.com/PDFs/products/cld_application_notes.pdf

The triangle and sawtooth generators are on pages 143 and 144, respectively.

This link has some other information: http://www.centralsemi.com/product/cld/index.aspx

John

 

They have been used to implement 2 wire amplified quartz transducers for at least 40 years. In the late 60's, I worked for Walter P. Kistler of Kistler Instruments. They had/have transducers called "Piezotrons". PCB Piezotronics was started by a few Kistler deserters and they also have ICP constant current fed devices. At the time I was there, we used 4ma Motorola MCL1304 CC diodes. That was the highest current available then.

PCB has a good website that illustrates and explains how it all works. Kistler's is so-so.

http://www.pcb.com/techsupport/tech_signal.php

I left Kistler when I learned he was a devout racist belonging to and contributing to some far out organization which supported his view. I "walked across the street" to PCB and was offered a very good paying cherry job which for some stupid reason I declined. That was the 2nd incredibly stupid thing I did in my early life.

 

They used to be quite expensive little beasts - but worth it for some applications due to their simplicity. I've used 100V rated parts for dc/dc switchmodes to power the controller with wide range input without having to generate an auxiliary supply.

 

These devices have been around for a long time. I was always put off by the ±15% tolerance.

 

Yes the tolerance can be significant. The nominal 10% of the CR460 I used was ok, but thermal management can be needed if going to high voltage with largest current rating in range.