So, I am working on a project using TO18 laser diode modules.
Using KISS context for design and control I am seeking some additional information.

The package is a LD/PD but I am only using the LD part of the package.
I did some looking around and found the basic current limiting control is the easiest way to stablize LD output.
As a referece, LM317 (https://www.allaboutcircuits.com/technical-articles/an-introduction-to-laser-diodes/)

However, I want to vary LD power and I have two ideas for that, but I am not 100% which is best for the LD (in terms of MTBF and actual ckt control).

1) from the reference link (figure 9 above), vary R1 accordingly
or
2) insert mosfet between the 317 and LD and control output total power via PWM on the fet, but run the 'on time' via full power (max current via R1 setting). Obviously I would need to choose PWM frequency carefully.

I could also do a combination of 1 & 2, hence vary the 'on-time' power (current) and then vary the intensity using PWM on fet, but I don't have a need for this from application perspective.

But, Fig 9 needs Adj fwd biased, so does using fet cause issue for 317 when PWM is 'off', thus floating fwd bias of Adj and causing issue? Would require ckt like this one to keep fwd bias at all times?

Suggestions?

 

I don't see any problem up to at least 500Hz. (not shown)

At least not with a 5-Watt standard Red LED.

The MOSFET is the first one in the list so I can't really vouch for that.

I didn't try placing the MOSFET between the 317 and the LED, but as a low side switch it seems to work ok.



Ok so I just tried it with a PMOS after the 317...it works but the waveform is just fugly.

Ok, so I was able to clean up the waveform on the PMOS , so either one should work for you.

 

Here is an alternative to PWM.

 

The frequency would be fairly low, maybe around 90-100Hz.

But, after more educating myself on LD packages, there's a thermal issue that affects LD light power. The hotter the package gets the lower the light power.

So, I am looking at this ROHM APC ckt. I assume PD is reverse biased, just enough to fwd bias that npn?

 

Well, I guess if your applications light output is so critical that prudent thermal management is not enough, then you will have to take other power control methods.

Maintaining a critical light output via feedback from a PD is out of my wheelhouse.

And there is certainly no way I could help you using a sim. (I know you didn't ask for that)

 

Well, I guess if your applications light output is so critical that prudent thermal management is not enough, then you will have to take other power control methods.

Maintaining a critical light output via feedback from a PD is out of my wheelhouse.

And there is certainly no way I could help you using a sim. (I know you didn't ask for that)

I can sim too. But no reply is a bad one. Thanks for helping.

My goal was just to use something not too complex under the KISS context. That ROHM ckt is the last one they list in their PDF, the other ckt's are more complex using opamp.

 

I assume PD is reverse biased, just enough to fwd bias that npn?

PD is photodiode with a reverse current proportional to the Laser diode output light.
When starting, there is no PD current, so the left NPN is off and the right NPN is fully on.
As the Laser turns on, the PD current starts to increase, increasing the left NPN current which reduces the right NPN current until equilibrium is reached at the set Laser current.

The set current is determined by the base-emitter voltage of the left NPN, which varies with ambient temperature.
For better temperature stable operation, you could use an inexpensive TL431 programmable reference in place of that transistor.

 

Thanks crutschow, that's about what I got from that ckt.
However, the PD npn is a bi-polar, so technically Im(PD) is a sum of base current + current across R3-->R4.
npn current is too small to be significant? Their math only uses R3+R4, etc.

Next question would be, although variable R3 gives the adjustable power swing for LD, if there's a nFet between right npn emitter and neg rail, and nFet is PWM 50% around 100Hz, I can save some on supply power (+ the power needed/wasted to run a PWM signal, like 555)?

Or, just run the LD DC mode (always on) and just vary it's current?

 

npn current is too small to be significant?

Its base current is typically about 1/100th of the collector current, so perhaps about 20-30µA.
To reduce that to zero, you could use a small N-MOSTET such as a 2N7000.

Next question would be, although variable R3 gives the adjustable power swing for LD, if there's a nFet between right npn emitter and neg rail, and nFet is PWM 50% around 100Hz, I can save some on supply power (+ the power needed/wasted to run a PWM signal, like 555)?

Sorry, I don't quite understand what you are proposing.
If you don't use a 555, how are you generating the PWM signal?
To save power you can use a CMOS 555.

 

Generally speaking, PWM can be used to vary power output, but not used to save power.

Average power output using PWM is basically the same as linear.

 

I would be adding an IC to make the PWM, and that IC uses power. I not sure the net savings is worth the extra complexity.

In realm of MTBF, having components not turn on/off frequently makes them last longer.