LM334 constant current source "inverse" design?

[zdelarosa00]

Hello, I need some guidance for I am currently designing a constant current source for the following laser diode:

OPV300
https://www.datasheetarchive.com/pdf/download.php?id=b08e595083910be444c389371b5b3171d8738d&type=M&term=opv300

And since it is going to be a fairly on/off/brightness control, I want to make a simple constant current source.

I found some circuits online like this:
http://www.bristolwatch.com/ccs/LM334.jpg

And every other "lm334 constant current source circuit" has this configuration, since the LM334 pulls the current reference from V+ to V-, they all use PNP transistors (1st attachement), so, my question is, since every other example in the datasheet uses the same connection principle, can I "invert" the connection of this LM334 or would that mess stability/function?

I came from the top of my head with the following examples... There is two points to take into account here, I want to know if I could use a NPN transistor first, and second, can I, similar to this circuit (2nd attachement) use at the end a MOSFET and if yes, is this a usable current-voltage translation (I know there are complete ICs and OpAmp circuits for current-voltage converters, but I know if this simple setup would work, why yes or why not). The thing here is just have a reliable current source, not an instrumentation grade military specification circit.

Am I making sense here? Thanks in advance for your attention.


 

[Zero999]

The circuit you linked to is a bad idea. It depends on the Hfe of the transistor, which can very a lot, depending on the transistor and temperature. The same result could be achieve with an NPN transistor, by connecting the base to +V, via the LM334 current regulator and the laser diode between the collector and +V, with the emitter at 0V.

The one with the MOSFET is equally bad. It'll just act as a source follower.

Just use the standard LM338 regulator circuit, with no external transistors or an LM317L circuit which doesn't vary with temperature, like the LM338 does.

[zdelarosa00]

But how are the options where a PNP transistor a good idea if they are also dependen on hFE?...

And how the LM338 regulator with an NPN different than the LM334 current source and a PNP? That is why I certainly am not getting it.

[Zero999]

But how are the options where a PNP transistor a good idea if they are also dependen on hFE?...

As I said, the circuit using a PNP transistor is not a good idea, because it depends on the hFE.

And how the LM338 regulator with an NPN different than the LM334 current source and a PNP? That is why I certainly am not getting it.

Which circuit? All of the circuits posted in this thread are rubbish.

You're better off using a current source made with the LM317L. Refer to figure 30, section 8.2.15 on page 18 of the datasheet. Change R1 to 180R for 7mA. The load is the diode laser.
http://www.ti.com/lit/ds/symlink/lm317l-n.pdf

EDIT:
Schematic attached.
R1 = 1.25/7mA = 1.25/0.007 = 180R.

[zdelarosa00]

Ok. I would go to that LM317 circuit if needed. Thanks.
But I have another doubt...
What if I use an OpAmp to amplify the current source? Is that more stable?

And on another idea: Is it a good idea to put a transistor as active load at RSET in the LM317 adj resistor and change current trought there?

[Zero999]

Ok. I would go to that LM317 circuit if needed. Thanks.
But I have another doubt...
What if I use an OpAmp to amplify the current source? Is that more stable?

And on another idea: Is it a good idea to put a transistor as active load at RSET in the LM317 adj resistor and change current trought there?

No, you can't just replace resistors with transistors, like that.

If you want to use an op-amp, then there are plenty of op-amp constant current circuits.

What are you trying to do?

[zdelarosa00]

I really just want a simple to understand circuit for conrtolling a laser diode (OPV300)...
The thing is, I would like it to be as less integrated as possible, I mean, complete 1 voltage regultator 1 set resistor setups totally work but I would like to modulate the brightness fairly quick, and there is I2C digital potentiometers but are not nearly as fast.

So, that is why I was going for the current reference, set resistor, transistor and all components separately, for simplicity I thought of a MOSFET at the output since for me, givenis much more easy to produce a varying voltage to modulate brightness.

How fast? Maybe 400KHz tops or so.

At this point i am not really sure how to proceed.

[Zero999]

I really just want a simple to understand circuit for conrtolling a laser diode (OPV300)...
The thing is, I would like it to be as less integrated as possible, I mean, complete 1 voltage regultator 1 set resistor setups totally work but I would like to modulate the brightness fairly quick, and there is I2C digital potentiometers but are not nearly as fast.

So, that is why I was going for the current reference, set resistor, transistor and all components separately, for simplicity I thought of a MOSFET at the output since for me, givenis much more easy to produce a varying voltage to modulate brightness.

How fast? Maybe 400KHz tops or so.

At this point i am not really sure how to proceed.

More questions I'm afraid.

Do you want continuously variable control i.e. be able to set the current from any value between 0mA and 7mA?

Or do you just want to pulse the LED on and off at 400kHz?

How are you controlling this? Presumably a microcontroller.

What power supply voltages do you have available?

How critical is the current regulation?

If all you need is to pulse it at at 7mA, you have a nice stable supply voltage and current regulation is not important, then simply connecting the OPV300 to the output pin of the microcontroller via a suitable current limiting resistor will do the job.

If you need to continuously vary the current between 0mA and 7mA, with a decent level of precision. You need a DAC, clocked a double the maximum frequency of interest and a voltage to current converter. The voltage to current converter can be made with a fast enough op-amp, with sufficient output current capacity.
https://www.allaboutcircuits.com/textbook/semiconductors/chpt-8/voltage-to-current-signal-conversion/
https://www.allaboutcircuits.com/technical-articles/the-howland-current-pump/
http://www.ti.com/lit/an/snoa474a/snoa474a.pdf

[zdelarosa00]

OK I'm ready...

1) Continuous control would be useful but I'm almost certain I will just use pulsed modulation.

2)Yes a microcontroller is the source of the modulation, but for experimentation purposes a signal generator can be used.

3)I think power supply voltage is not a problem, +-32v 2A. And +5v 3A. Why is this pertinent?

4)Critical enough to not burn the laser.

5)That uC-DAC-OpAmp Translator sounds like a very good idea, I guess there is no transistor at the output here.

[Zero999]

I think you should try connecting the laser diode to the microcontroller output via a resistor.