AN90: Current Sources for Laser Diodes

Hi,

Thank you for your question. I am reviewing the information you provided and may come back with follow up questions. I appreciate your patience as I work on your query. 

Regards,
Gilbeys

Thanks, GLADION .

I'd like to add the following:

There are cathode-grounded ("CG") laser diodes, i.e., laser diodes where the cathode needs to be connected to ground. I understand that this requires the use of a design as shown in Figure 5. However, I have a laser diode that can be connected either way, "floating" as in Figure 2 or "CG" as in Figure 5.

Without having much knowledge in this field, my guess is that the simple circuit of Figure 2 may be less noisy just because it has less components that introduce additional noise. I hope that makes sense.

Best regards,
Dan

Hello Dan-o-tronic ,

Sorry for the late response. Regarding on having a diode on figure 2, It depends on your input voltage surpassing the Vbe of Q1 + Laser diode's ON voltage + the drop on the 1ohm resistor. At Appendix A of AN90, it stated that laser diodes have typically 1.2V to 2.5V junction voltage.

So for example at max input current (250mA), Vbe + On voltage (laser diode) + Vdrop on 1mohm = 0.6V + 1.2V(minimum laser junction voltage) + 250mV = 2.05V. If your input is at maximum of 2.5V, the Input will be 0.45V higher (2.5V - 2.05V) than Q1's base which is still not enough to forward bias the diode.

I think adding the diode will not affect the overall circuit but if you want to minimize the noise, I suggest not to put it as it will contribute additional noise on the system.


Best Regards,
Kenneth

Hello Kenneth11 ,

Thank you for the explanation and the calculation of Q1's base voltage. That was helpful to understand when this diode is required: I simulated this circuit. After I had changed the input voltage divider to accommodate an input voltage range of 0 - 5 V, the laser diode current suddenly ramped up when input voltage was high enough to forward bias the "enable" diode. Only in this case the "isolation" diode (as added with red markup) was required. Great. On puzzle solved. 

From you comment regarding noise I conclude that in order to reduce noise it is generally a good idea to remove circuit elements that are not required. That leads me the comparison of circuits in Figure 2 and Figure 5 regarding noise. As I wrote earlier, I think that the simple circuit of Figure 2 should be less noisy than circuit of Figure 5 just because it has less components (i.e., instrumentation amplifier A2) that introduce additional noise. So if my laser diode does not require ground connection, I should go with circuit in Figure 2 for achieving low noise. Does that make sense?

Thanks.
Dan

Hi Dan-o-tronic ,

Yes. Unless you need a grounded cathode application, I would go with the simpler circuit for lesser noise.

Best Regards,
Kenneth

Hello Kenneth11 ,

that's great, will do so.

Now one issue came to my mind, the selection of the pass transistor Q1. AN90 uses FZT-849. This transistor is rated for a continuous current of 7 Amps. My application only requires 100 mA through the laser diode (at 2 V forward voltage). I assume that there should be a transistor with lower noise figure if it is rated for only, say, 200 mA continuously.

Could you please suggest a transistor with less noise? That being said, I'm unable to extract any figure of merit regarding noise from the FZT-849 datasheet.

Thanks.
Best regards,
Dan

Hi Dan-o-tronic ,

2N3904, also an NPN transistor which is used for Q2 can be your replacement. It has a continuous current of 200mA. Switching speed is also comparable with FZT-949 ( 65ns Ton and 35ns Toff ). It also has a noise figure of 5dB typical.

Best Regards,
Kenneth

Hi Dan-o-tronic ,

2N3904, also an NPN transistor which is used for Q2 can be your replacement. It has a continuous current of 200mA. Switching speed is also comparable with FZT-949 ( 65ns Ton and 35ns Toff ). It also has a noise figure of 5dB typical.

Best Regards,
Kenneth

Dear Kenneth11 ,

I was offline for two weeks so I could not reply earlier. Thank you for your suggestion. However, the 2N3904 is (according to my online search) a through hole component, whereas the FZT-849 is surface mounted. I'm looking for a SMS alternative. Thanks.

Best regards,
Dan

Hello Dan-o-tronic ,

Just checked some NPN transistors that has an LT spice model and found this BC847 with 200mA continuous collector current and in a SOT package.
Noise figure is also indicated on the DS for your noise calculations.
Here's the datasheet for you to check.

45 V, 100 mA NPN general-purpose transistors (nexperia.com)

Best Regards,
Kenneth