New CCS Idea

[sijosae]

This is my first post here.

I like to introduce my new CCS.
This is just an idea. I did not test it yet.
Have you ever seen before?

I like to know your opinion or comment.
Thank you.

[janneman]

Sijosae,

VERY neat idea! I haven't seen it before.
I think it will clearly improve on the Vb stability over the left side diagram.
I have a few questions, though. With the high quality of the 317 as current source, most likely the current through the 317 is more stable than the current through the transistor. The current through the transistor is still depending on variations in Vbe, temperature and even voltage.
So, if your aim is a very constant, high impedance current source, the 317 alone is probably better then in combination with the transistor.
Just some thoughts.

Jan Didden

[sijosae]

I just tested this new CCS as follows.

V+ : mA
1.86 : 0.000
2.41 : 0.533
2.86 : 1.236
3.05 : 1.243
3.61 : 1.249
4.50 : 1.250
6.93 : 1.251
10.63 : 1.255
12.39 : 1.256
14.79 : 1.257

(2N2222 - 500 ohm, LM317L - 300 ohm)

IMHO, it shows very good results.

[sijosae]

Another application of this idea is a replacement of the zener diodes in the power supply. (voltage reference)

By using LM78xx ICs, we can get various and high voltage voltage reference simply.

e.g. LM7824 --> as a 24V zener diode.

[leadbelly]

Originally posted by sijosae
e.g. LM7824 --> as a 24V zener diode.

I'm hardly an expert, but isn't the LM7824 both more expensive and much noisier than a zener?

[janneman]

Originally posted by sijosae
I just tested this new CCS as follows.

V+ : mA
1.86 : 0.000
2.41 : 0.533
2.86 : 1.236
3.05 : 1.243
3.61 : 1.249
4.50 : 1.250
6.93 : 1.251
10.63 : 1.255
12.39 : 1.256
14.79 : 1.257

(2N2222 - 500 ohm, LM317L - 300 ohm)

IMHO, it shows very good results.

OK, lets do a quick calculation. Between 4.50 and 14.79 volts, the current varies 7uA, that's a dynamic impedance of about 1.5 megohms. Pretty good indeed, surely better than the LED version.

But what is the variation in Vb with the same steps? That will tell you what is better, just the 317 or the 317+transistor.

Jan Didden

[sijosae]

Thanks Jan Didden,

I completely agree with you.
Yes. I know the current through the LM317 is more stable.
However there is a lower limitation in the current. (datasheet of the LM317L)
In that case, current should be > 5 mA.

[janneman]

Originally posted by sijosae
Thanks Jan Didden,

I completely agree with you.
Yes. I know the current through the LM317 is more stable.
However there is a lower limitation in the current. (datasheet of the LM317L)
In that case, current should be > 5 mA.

Indeed, you're more flexible this way. Good thinking.

Jan Didden

[AKSA]

Interesting idea, Sijosae.

However, the voltage reference will always require a prescribed current from the collector source which will vary very slightly as the potential stress across the LM317 increases. Is this a good thing?

Another option might be to run the LM317 from another power rail (positive with respect the to the LED anode) and then run a resistor from this same rail to the emitter. Chosen correctly, this should compensate changes in Vbe with rising and falling voltage at the collector. This is an Elektor idea, incidentally, not my IP at all!

Cheers,

Hugh

[Ultima Thule]

Here is an accurate CCS with a voltage reference I have used several years with good results, and it don't get much better...

Cheers