I wanted to generate a laser pulse between 1 us and 50 us by creating a simple circuit with a source meter, as I didn't have a laser driver. At the beginning, the setup seemed to work as intended (I tested it with optical spectrum analyzer). However, after a few runs, the laser seemed to break and stopped emitting any light, which I confirmed by trying to view it on my laser viewing cards (https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=296). My circuit is shown below, any pointer is much appreciated!

Each run, the current source is on for 500 mA and then turned off, the signal generator is always on.

EDIT: Forgot to mention that the open-circuit voltage / voltage range is capped at 6V, also checked each line (wave gen, current waveform across resistor) with probes.

Circuit for driving laser diode

From the voltage waveform across the monitoring resistor, I didn't observe any spike in the voltage at the time that the power supply switches from CV to CC mode, which was one reason why I thought this would be ok. The sampling speed is 50 MS/s.

If the root cause was the current surge from the power supply, I believe this should be reflected from the voltage across the monitoring resistor? I initially was worried about the overshooting problem, but decided to adapt this circuit still because my signal generator cannot generate a very long >98% pulse.. I would think 50 MS/s would be enough to capture the transients..

I captured this for the entire run, but I didn't see overshoot in any of them (but I didn't plot them all). Essentially I agree with the problem with the power supply, but I have not observed anything in the recorded signal that would have suggested this, and was wondering if there can be alternative theories. I will try the different configurations suggested.

Voltage waveform across resistor

  • 1
    \$\begingroup\$ You might be running into the problem where if you open circuit a bench supply set to constant current mode, the output voltage goes up to the maximum allowable so you get a brief pulse of high voltage (and overcurrent) when you connect a load to it. \$\endgroup\$
    – vir
    Aug 1, 2023 at 22:37
  • \$\begingroup\$ Have you verified that the current waveform applied to the diode is correct? A clamp type of current probe would work great. \$\endgroup\$
    – qrk
    Aug 1, 2023 at 22:38
  • \$\begingroup\$ Sorry, forgot to mention that I capped the maximum voltage to 6V - also I undersetand the problem with the brief overshoot, but I didn't observe it across the resistor and was hoping that a sourcemeter would be more reliable than a bench supply in this regard.. Also yes, verified the current waveform by probing the two sides of the resistor. \$\endgroup\$ Aug 1, 2023 at 22:40
  • 2
    \$\begingroup\$ Asking the current source to play nice while switching from constant-current to constant-voltage seems risky. Might be better to shunt current to an alternate path, so that the power source is always in constant-current mode. \$\endgroup\$
    – glen_geek
    Aug 1, 2023 at 23:28
  • \$\begingroup\$ I agree with the others that it is unlikely the power supply is handling this gracefully. Scopes are not perfect, and don't forget there are 500 shots happening in 1 run, and you're not capturing all of them. \$\endgroup\$
    – Drew
    Aug 2, 2023 at 13:11

2 Answers 2


The power supply won't be able to switch between CV and CC fast enough for the laser diode. Use the power supply in CV mode and build a proper current source for the diode. Place the switching FET in parallel with the laser. When switching a current source you must create an alternate path for the current to flow. Something like this allows the charge to flow either through the laser or through the FET.:


simulate this circuit – Schematic created using CircuitLab

  • 1
    \$\begingroup\$ Thank you for your answer! However, I was confused on why this was not reflected from the voltage across the monitorning resistor, would it be possible to elaborate on that (I added more details to the original question)? \$\endgroup\$ Aug 2, 2023 at 11:35
  • \$\begingroup\$ The laser diode has a reverse breakdown voltage of only 2V. Switching constant current sources creates voltage spikes across all the stray inductances. Try putting a Schottky diode in reverse parallel to the laser diode. If I were doing this I would not use R1, and monitor the current produced by the photodiode. \$\endgroup\$
    – RussellH
    Aug 2, 2023 at 14:30

I'll propose a jankier, but simpler solution: Just use the power supply in voltage mode.

If this is just an experiment efficiency doesn't matter, so you can drive the laser with 12-24V using a beefy current limiting resistor. That should give you a relatively stable output current even as the Vf changes slightly with temperature.

Alternately, you could use a lower voltage, and just adjust it regularly to keep the output current in range.

The bypass configuration others have suggested is a good idea. It reduces the switching transients a lot. Generally speaking, Keep an eye on the parasitics. Add any long wires as inductors in your circuit. And assume there's a few pF of parasitic capacitance everywhere.

If you can afford to burn a few diodes, try some different configurations and gather more data. If not, experiment with the parasitics in LTSpice before trying again.

Either way, don't trust the power supply cv/cc switchover. That's just asking for trouble.

  • \$\begingroup\$ Thank you very much for your response! Essentially I agree with the problem with the power supply, but I have not observed anything in the recorded signal that would have suggested a current overshoot even at 50MS/s for the entire duration, so I was wondering if there can be other factors that could also contribute to this failure? I will try the different configurations suggested. I was also thinking of adding a zener diode or capacitor, would that help with the CV/CC switching? Or would you suggest that it should be avoided at all cost. Thanks! \$\endgroup\$ Aug 2, 2023 at 14:04
  • \$\begingroup\$ @user3602697 Well, using the power supply in this way is very likely to cause problems AND, your system doesn't work, so I would try eliminating that before you do anything else. If it does or doesn't burn out after reconfiguring, that will tell you something. \$\endgroup\$
    – Drew
    Aug 2, 2023 at 16:06
  • \$\begingroup\$ Something may be happening on a cycle the scope isn't capturing, or it's too fast for the scope. \$\endgroup\$
    – Drew
    Aug 2, 2023 at 16:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.