I need a CW laser diode, with an integrated photodiode, with wavelength around 800nm and power 1-1.5mW. Although I didn't find anything that exactly matching my specifications, I did find this from Thorlabs and this from Roithner Lasertechnik. Unlike some of the other laser diodes I browsed through, these only specify a maximum power (of 10mW and 5mW, respectively), rather than a typical power, so this seems to imply that I can adjust the output power (in this case, to 1-1.5mW). Am I correct here? If so, then what are the consequences of doing such a thing (is any aspect of the laser diode and/or beam quality degraded in any way by this)?

  • \$\begingroup\$ I've definitely seen more detailed data sheets for laser diodes. \$\endgroup\$
    – Andy aka
    Jul 16, 2021 at 11:38
  • \$\begingroup\$ @Andyaka Hmm, well at least they're reputable / relatively trustworthy companies. I'm sure I can ask them for more information if necessary. \$\endgroup\$ Jul 16, 2021 at 11:41
  • 2
    \$\begingroup\$ mouser's laser diode catalogue for 780-820 nm laser diodes suggests you can get diodes with better datasheets. I'd put "have good datasheet" over "have heard their name before". \$\endgroup\$ Jul 16, 2021 at 11:42
  • 3
    \$\begingroup\$ also, since this is invisible and these diodes can produce much more than 1.5mW: I know you're probably way more familiar with laser safety than I am, but in case anyone else should be reading this in the future: Please do not look into the laser with the remaining eye; always operate lasers in enclosures with interlocks, and wear appropriate eye protection. \$\endgroup\$ Jul 16, 2021 at 11:46
  • \$\begingroup\$ @MarcusMüller Haha, I doubt that. But yes, I have calculated the (ocular) MPE limits (as per ANSI/etc. laser safety standards), which is why I'm looking for 1-1.5mW. \$\endgroup\$ Jul 16, 2021 at 11:49

2 Answers 2


Laser diodes have what is called a threshold current which is the current through the diode at which it starts to lase.

With your thorlabs diode, the typical threshold is 25mA (The catalogue uses this term rather than the datasheet. https://www.thorlabs.com/catalogpages/V21/1234.PDF)

With your diode running at 2V 25mA you'd be delivering 50mW to the diode and it would start to lase, after that a parameter called slope efficiency comes into effect which in your sheet is typically 0.5mW/mA. To get 1.5mW out you'd need to drive around 28mA into the diode.

Thorlabs don't seem to indicate what level of optical power output they define as lasing though so you might be well over 1.5mW by the time you reach that point.

Since you've got a bit of spread with regards to threshold current and slope efficiency you might have a hard time locking to 1.5mW without an external power meter and calibration of the diode driver.

Edit: It looks like the diode has an integrated photodiode which you could use as part of a closed control loop for controlling power. Which was probably your plan.

This isn't the best datasheet, to be honest, I would expect to see some LIV curves here that describe the diode a bit better.

I'm not an optical engineer but I don't expect an impact on beam quality based on how much power you supply the diode.

Additonally, note that this diode can put you into Class 3B territory which is why you need to make sure that if you use this diode or a diode capable of driving >5mW that you build your design so it is not capable of driving enough current to exceed the safe limit, attenuate the laser, or use a laser that is impossible to drive beyond the limit.

  • \$\begingroup\$ Damn, I actually wanted to lock to the power without using the PD; the PD is for something else. \$\endgroup\$ Jul 16, 2021 at 12:12
  • 1
    \$\begingroup\$ Not saying you cant do both, you could buffer the PD output and then send it to your driver for the control loop and also monitor it/do whatever you want to do with it? \$\endgroup\$
    – VBwhatnow
    Jul 16, 2021 at 12:37
  • \$\begingroup\$ This en.wikipedia.org/wiki/Self-mixing_laser_interferometry is what I wanted to do. I know you said you're not an optical engineer (I'm a novice to everything, so you know more than me), but what do you think? See here youtu.be/MUdro-6u2Zg \$\endgroup\$ Jul 16, 2021 at 12:38
  • \$\begingroup\$ The risk with running the laser at too low a power is that it might have poor SMSR (side mode suppression ratio), meaning it could run multi-mode. It might give better results to run it near its nominal operating point, and using an optical attenuator to reduce the optical power hitting the target. (Of course this attenuator affects your back-reflected signal twice as much which might reduce the fringe contrast in your interferometry scheme) \$\endgroup\$
    – The Photon
    Jul 16, 2021 at 16:33
  • \$\begingroup\$ @ThePhoton Ahh, yes, that's an excellent point; the reflection back into the built-in photodiode is critical for me. Could you recommend anything? I'm willing to go into the visible 650-700nm range, but I still can't manage to find anything with 1mW power and built-in photodiode. \$\endgroup\$ Jul 16, 2021 at 16:49

I suspect you're going about this the wrong way.

As VBwhatnow has answered, for the Thorlabs part, trying to run at 1.5 mW will probably be right down near the threshold current. Also note that you really can't tell exactly how you'll need to drive your laser: at the very least, for the nominal test power the photodiode current can vary over a 20:1 range and still be in spec. This doesn't mean that it will vary in operation, it means that from unit to unit the sensitivity can vary wildly.

I'd suggest that what you need to do is run the laser somewhere near maximum power. Note that, unless you have a power meter you will run the risk of overdriving the laser, so be careful. Once you have the laser running at a stable power, put an attenuator in the output. Thorlabs will provide you exactly what you need (for a price). Consider running the laser near 5 mW, and putting a 3x attenuator to get it near 1.5 mW.

I've mentioned a power meter. You don't say what your tolerance is for your desired power, but be aware that you'll need something to measure the power. The specs on the diode are simply not tight enough to closely predict what you'll get. Notice, for instance, that at 10 mW the Thorlabs laser will typically require 50 mA, but no more than 70 - and maybe as little as 25. So, if you drive 25 mA through it and it lases (which will be difficult to tell, since you can't see the output) you might be at the maximum 10 mW. Or you might be down around 1-2 mW. If your target is 10 mW, so you increase the current to 50 or 70, a sensitive laser will be grossly overdriven and may not survive.

You could roll your own using a commercially-available photodiode with a large(ish) active area and a well-defined sensitivity, and get decent results. Or you can try eBay, as long as you keep caveat emptor firmly in view. As a start, a lot of lab-grade laser power meters being offered do not include the (required) sensor head. Or you can get a new one from someone like Thorlabs. They sell good stuff, and charge accordingly.

  • \$\begingroup\$ Damn, it sounds like the better option in every way is to just find a LD that is rated/calibrated for 1-1.5mW typical power output. But, as I said, finding such a thing, with an integrated PD, was the problem in the first place. \$\endgroup\$ Jul 16, 2021 at 13:00
  • 1
    \$\begingroup\$ Heh. Well, I suspect finding one in that range is going to be a bit of an effort. Generally speaking, when it comes to mW lasers, more is better, so there isn't much of a market for 1 mW LDs when 10 mW units are available. Like I say, I suggest you go with a higher-power unit and put a known attenuator in the path. \$\endgroup\$ Jul 16, 2021 at 13:17
  • \$\begingroup\$ I just found this digikey.com.au/product-detail/en/… ttelectronics.com/TTElectronics/media/ProductFiles/… What do you think? \$\endgroup\$ Jul 16, 2021 at 13:17
  • \$\begingroup\$ This en.wikipedia.org/wiki/Self-mixing_laser_interferometry is what I want to do. See here youtu.be/MUdro-6u2Zg \$\endgroup\$ Jul 16, 2021 at 13:18
  • \$\begingroup\$ That's a good video, so you have a good starting point. However, using a VCSEL LD is not a good idea. its' linewidth is so broad that I doubt you'll see any interference. Also, why are you fixated on 1 mW? What is wrong with 10? If you're worried about eye safety, get a cheap pair of laser safety glasses. Also, why are you working at 800 nm? Using a visible LD will make adjusting your setup much easier, since you'll be able to actually see the beam. \$\endgroup\$ Jul 16, 2021 at 14:07

Your Answer

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

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