So…I developed a nanosecond pulse driver for LED (the 3W Nichia “star” variety that were brand new at the time) back in 2006-7 to observe ink droplet formation in high speed printing systems. LED because it was safe for operators and didn’t require a special curtain. I built the driver with a totem output. The top side connected to 60V and the bottom to -5V on the theory that removing any charge on the LED die was essential to eliminating tail current and speeding up the output. This arrangement was successful. My employer applied for the patent but another inventor had just filed before me. Such is life!
Fantastic, this is really cool to see. I work for Lightware Lidar, and our bread and butter is pulsing laser diodes really hard for miniscule periods of time (for the purposes of range finding), so this hits really close to home for me. I also love that long exposure of the beam at different distances, you can really make out its profile.
Something like that was part of my masters work 40 years ago. I used coax cable as load lines. This gives an almost perfect rectangular waveform of the diode current. The length of the load lines defines the pulse width.
Ah the famous coaxial line pulse forming network. Try making it out of foil sheet and roll it up like a jelly roll capacitor. Add some extra turns of foil in the center and it transforms the current wave to a very high voltage quickly. It doubles the input voltage with every quarter turn. ❤
I believe range-finding technology was used for military technology and weaponry before it was ever applied to the automotive industry. For instance: Collision avoidance technology.
You've done a very nice design, assembly and experimental testing project with overdriving visible laser diodes. This video brings back memories of when the Laser Sensors Team that I worked on at the U.S. Army Night Vision and Electronic Sensors Directorate (NVESD) did some testing of overdriving inexpensive commercially available continuous wave (cw) near infrared (NIR) diode lasers to produce nanosecond scale pulses for rangefinding, lidar, and free space optical communications applications back in the late 1980s. It was fun in the initial experiments cranking up the peak current until we blew up each diode laser to determine the overdriving limits. We would always say when we reached the limit for each diode "It blowed up real good!" That was a phrase that two farmer characters, played by John Candy and Joe Flaherty on SCTV in the '70s, would say in the "Farm Film Report," in which the two farmers discuss the latest movies and film theory - a routine parodying Siskal and Ebert. Supposedly the best part of a film is when things get "blowed up real good." They even had some celebrity guests (other cast members dressed up as Dustin Hoffman, Bernadette Peters, etc) who were "blowed up real good" during special "Celebrity Blow Up" segments.
Hi, I'm a test engineer and I find your experimental setup and your entire experiment incredibly useful! I have been trying to puzzle out for months how to measure pulsed laser diodes and what equipment I can use there (I haven't been able to find very short rise time transimpedance amplifiers), so the average power method and back-calculation of peak power seems very insightful, thank you! However, for the love of God (and I know I'm just a random person on the internet but I'm also an electronics engineer with a passion for signal integrity and proper layout techniques), do NOT split the ground planes and place just a single pour ground plane over your entire board. It will improve your noise a lot. I don't know who in the literature suggested to put these R7-R11 resistors there and for what reason, but the split plane underneath the resistors is gonna create a very strong EMI that will radiate everywhere. Moreover it will add inductance to your dI/dT loop which in your case is from point C to H1.2 to the laser diode, then back through H1.1 all the way to the GND of Q1. This loop should ideally have a *uniform* ground plane underneath and be kept as short as possible to minimise EMI (both radiation and pickup).
Excellent, you are welcome! I am not a test engineer, so good advice like this is always welcome! The bottom plane under the PFN likely contributes significant capacitance, slowing things down as well I imagine, diode leads could be shorter, and the FMMT415 could be used in place of the ZTX415.
@@LesLaboratory the bottom plane doesn't add as much capacitance as you'd think because you have a two layer board so the ground plane is really far away from the signal layer (since this is a two layer board and the gnd is in the bottom layer which is usually 1.2-1.6mm thick) and because the ground plane split that you have is very minimal. That said, from an electrical engineering perspective the most important thing for a pulse forming network is not necessarily low impedance or inductance, but impedance matching of the transmission line and the source impedance (and ideally the load impedance as well). So my point is, you can put a solid ground plane underneath and then calculate via your EDA software the trace width that you would need to achieve a specific impedance (say 50Ω or sth). Having a signal pass over a split plane is a big no no in PCB design because that spot there is acting as an antenna and it both receives and emits radiation (which is harmful both to other devices nearby and to your device if it picks up stray signals). Moreover the extra zero ohm resisors are adding inductance to the loop and that both increases the rise/fall time and it can create a gnd bounce (where your "ground" is not really at 0V during the transient but at a slightly higher voltage than the actual gnd plane exactly due to the zero ohm resistors which have some small resistance and additional inductance). Also, important thing, I wouldn't use zero ohm resistors in the PFN even if it's just for stitching (which imho is a bad idea to begin with as I said), but I would rather use some specific resistance value like 0.5Ohm or whatever because then you have a controlled value, zero ohms are just some random low value with no effort in controlling the value. In any case, I would suggest at a future iteration to just ditch all of these parallel low side resistors at the very least. Second step would be to try to calculate a target impedance to hit for the trace that delivers the pulse.
@@rolmie I'm sure there are a bunch of other failure modes as well, such as dendrite growth, and photon damage to the facets, but so far so good! I should just let it run at high current for a few days or weeks and see what happens! The results in the literature seem good though in terms of lifetime.
@@rolmie @BreakingTaps might even have a use for it. With such short duration pulses, thermal effects will be absent, but the number of photons from each pulse, might, possibly, be able to micro machine on the order of atoms or molecules at a time!
@@LesLaboratory ...another fan of Jim Williams here...actually spoke to him briefly about "flyback switching regulators"...must have been back in the 80's...brilliant guy!!
Essentially the design of the pulse circuit is to keep the length of the current path as short as possible, I made a pulse laser with bonded chips that gave the same result at 70V as with discrete components and PCB at 350V . I changed from avalanche Transistor to FET. It is still produces today with no improvement.
As a former PCB layout professional, I can say that your HV traces need much more clearance away from ground. Even a small amount of moisture in the air or using it at a bit higher altitude, (I live at 6,000 feet,) could cause arc-over of the HV. A spacing between the conductors and ground of nearly .15 inch would be the minimum spacing for generic PCB material (without any sort of conformal coating) that I would use under any circumstances with just over 230 V applied, while .2 inch would be much better since your PCB layout clearly has the room for it. With the possibility of an unknown higher voltage being applied for experiments I would design the bd with a .25 inch gap between the HV and Ground wherever I possibly could, (including a wider footprint for the transistor and the LASER connector,) with a painted-on HV dielectric where the traces cannot be given that clearance in order to ensure reliability of operation. Any sort of contamination between the traces, from moisture to dust reduces the dielectric insulation between the traces, increasing to probability of arc-over. Circuits used near the saltiness of the oceans also require extra clearance. Even altitude has a pronounced effect, so with all of those unknowns it's better to just provide as much clearance as you can afford to in the creation of a HV bd design made available to your users.
Don't think this is of concern here with avalanche breakthrough from the transistor of a little more than 100V. Inductance on the other hand plays a major role at these frequencies...
For sure, the design is not optimal, nor hardened against the elements, it's just experimental. IPC-2221 suggests 1.25mm minimum clearance up to 300v between conductors on an un-coated board (0-10007ft), but it is generic and probably superseded by more rigorous standards. The pin pitch on the ZTX415 is really quite small, and so creepage could occur on the package itself as well. Far better to use the FMMT415 and ensure sufficient clearance on the rest of the board. For this I just wanted proof of concept. For another project I have been working on I have taken much more care, since I have up to 1kV on the PCB!
I think creepage distance is much less of an issue than you imagine. First creepage distance is much less of a risk on the parts of the circuit with nanosecond pulses since there will not be enough time to get a breakdown propagating unless the field strength is very high. On the DC part it can be an issue but not as big as if it was 230 mains voltage. When designing for 230 V mains the nominal peak voltage is 325 V and then you need to be able to handle voltage spikes of a few kilovolts lasting for tens of microseconds from switching transients and distant lightning. In a small DC-supply like this there will be much less transients. More clearance is better but not if it causes more inductance in the pulls path. The loop area needs to be as small as possible. To spread out the legs of e.g. the laser diode will reduce the peak power since the inductance increases. It is better to improve the insulation if needed by coating, poting or oil immersion.
Another thing to consider is the threshold current of the diodes. They typically won't lase until they are at a couple of hundred ma of current. You might want to think about being able to apply a dc ofset to get to threshold, then modulate on top of that. Of course now you will have to improve the heat sinking of the diode. I would also be curious what the diode is doing thermally.
I have seen biasing like before, and this was even done in earlier efforts, but didn't seem to be required at these extreme currents. It would still be an interesting thing to try out, now that all the other issues are ironed out. Someone suggested there is a paper where they were getting picosecond pulses out of an LD that was biased and fed with RF pulses. There is way more work to be done here!
This is _very_ cool! At that pulse duration and peak power it might be starting into the range of micromachining (when coupled to a microscope objective for a small spot). Especially in something like silicon, the absorption depth will be very small, around 1um iirc? Really love that long-exposure shot too, neat to see the individual pulses!
Thanks! Somewhere in that region, it is very shallow on silicon. This should work with 405nm diodes as well. There are even 375nm diodes in production now but they are $$$!
Thank they are superb. Process tracking, fast delivery, and I always seem to rack up coupons for SMD assembly as well, which is nice for when I need it.
When I worked at a National Lab, I designed and built a laser system that would gain-switch a seed laser diode using rf, resulting in a 50ps FWHM pulse at various rep rates. This would be sent to a 5W fiber laser amplifier. The average power out of this system would remain 5 watts, but the peak power was dependent on the rf duty cycle. I reached peak powers of 1.6kW in this system.
@@JohnHansknecht That sounds awesome! I have read about these, very very nice! I'm going to have to aim for picoseconds with these diodes, things could get interesting.
@@LesLaboratory You won't get power out of seed lasers that are gained switched. Just about any laser diode can be gain-switched. No special diode needed, but you need to bias it near threshold through an impedance block (resistor and sometimes inductor) and capacitively couple an rf energy of about a watt. A typical 10mW seed diode will put out barely a milliwatt when gain-switched, but if it is fed into an amplifier, this is where the results can get interesting. Takes a very fast (many Ghz) scope with photodiode input to setup the parameters for proper switching. This is usually out of reach of most hobbyists.
I did wonder if some of the microwave PTFE laminates and GaN microwave devices might give faster rise times, but how about a machined coaxial cavity? That might give a really serious and near-constant discharge current when compared with coaxial cables. If the RF impedance is a good match, then it shouldn't have any overshoot or oscillation. You'd still need a hyperfast pulse generator to drive the microwave device of course. Once you get past 30 GHz, you are really into chip-bonding territory, but there you are well into the picosecond domain. Not much point as the laser packages are hugely suboptimal and Stone Age - unless you can hack them up and get within a millimetre or two of the diode chip. Very cool work, fascinating as ever.
You are actually correct about the cavity method. This was actually used already for high-Q filters, and for Piston-tuners. Use of any kind of high-Q cavity structure, with lowest losses, creates the most compact bandspread (aka, sharpest waveshape) of the rebound signal. Now actually being used for quantum physics, as "vacancy devices" to create optical manipulations between goezinta and goezouta photonic driving impulses. Works great, and is highly efficient as opposed to other methods.
Interesting, this is similar with a few other comments here. Microwave engineering is not my forte, but I will have a read around and see if there is anything suitable. Picoseconds is a really tempting domain.
I bought some cheap laser glasses from amazon and I wanted to make sure they actually worked. To test them I dawned an oculus quest headset. This essentially air gapped my eyes from the laser so that there was 100% no way I could get the laser light back into my eyes. Then I projected the laser onto a target where I could see it, and then held the glasses I had purchased in front of the target.
Thanks! Oh totally. 445nm is cheap 405 not so much, but now its been well tested, its time to try I think. There are UV laser diodes now as well, so things are getting real interesting!
Yep! There is some interesting tech out there now for this particular field. I'm also interested in playing with SiC stuff for high voltage work as well.
Cool that you used 2 diodes from Roithern Laser, its an Austrian company but they seem to have the widest range of diodes - worldwide! We are also using them ...
I don't quite understand the "long exposure" photo at 12:55. What is the beam reflecting off in mid-air? The earlier shot with the flapping paper is essentially using this laser as a glorified strobe lamp. The size and spacing of the spots would be the paper's speed times the pulse width and pulse repetition period. The pulses themselves would be spatially a couple of feet long (using the rough 1ns/foot speed of light). You could observe this using smoke and a picosecond-ish camera shutter.
I was also wondering how a long exposure shot resulted in discrete laser dots like that - I don't understand... I guess there has to be a "moving piece of paper" or something in that photo too, but you just can't see it in the exposure, and you only see the dots
This is probably explained aliasing of the camera internal sampling rate vs the function generator frequency used to trigger the laser driver. This could explain the spacing between the spots and the length of those spots. Normally, those spots should be 4 or 5 feet long because the optical power metered was about 6ns.
Ah! I thought about videoing how this shot was done, but didn't think anyone would find it very interesting! Maybe I should have! I tried doing a long exposure, whilst quickly wafting a piece of paper along the beam length, and it works, as on camera, but not particularly well. Right behind the Laser you can see a pair of rails. This was a Linear motorised rail from a printer. I mounted an arm to it with a white target on the end, and that sits in the path of the Laser beam. The camera shutter is opened and the target moved from one end to the other, quite quickly. This was driven using an Arduino.
To answer your initial question, I would believe you. Some IR laser diodes worked like this. It was in an oldish book, "more electronic projects for the evil genius" chapter 4 (I had to look it up) where I saw this. Uses a thyristor to dump a capacitor through an IR laser diode. It is on the Internet Archive now, but I still have the physical book somewhere.
I had a thought of a cool technique for lidar, but I couldn't wrap my head around how to do it with a pulse: Use Radar ideas: Analoguely: modulate the beam's intensity with a frequency ramp, say 100-1KHz and back down. Mix the PD return with the outgoing beam (some lasers used to have PDs in them), and get a beat frequency out of it; the frequency of the beat will tell you the range to the object without needing to solve ns timing for TOF sensing. If you want to get even more accurate, you can start figuring out the phase difference (some guys did that with a Red Pitaya). I'm sure you could figure out how do that with say a 100-1KHz pulse interval and figure that out, but the math is making my head hurt... and "phase" is...?
Interesting idea but I don't see how the beat frequency will remain constant for a non- linear frequency sweep; nonlinear because the sweep goes up and comes down. Due to this, the difference between the outgoing and incoming frequencies will not remain constant for a given distance.
@@akarshchaturvedi2803 Good point. I'd probably make it be a constant speed sweep from 1KHz to 20KHz and then reset back to 1KHz. You should have a nice constant split except at the end and that would be an out-of-bounds step versus a ramp to zero and a ramp back up to steady.
This vid came just in the right time! I've been working on a small project to make dToF LiDAR. So far the digital part (tapped delay line on FPGA) is straight forward, but I'm looking for a fast LD driver that capable of 10ns pulse width.
Great! You can extend the pulse width by attaching a length of un-terminated coax in parallel with the capacitor bank (or even in place of it!) to widen the pulse width (see application note 47). There is a nice video on this by Dave Jones @EEVblog here: th-cam.com/video/uBYMePUFinQ/w-d-xo.html
Don't forget the old Tunnel Diodes. We built our first sub-nanosecond pulser, using these, along with some IBM digital transistors, capable of 1GHz bandwidths. The problems were horrible. Temperature unstable, moving or squeezing the coax changed the speeds and shapes, device failure due to avalanche breakdowns, absolutely NOT-repeatable. Other than that, it worked great.
Oh man do I have to comment on this. You mentioned pumping dye lasers with diodes, but in my opinion there would be an even more prominent example that might be of interest, Ti:Sapphire. With the power of blue/green LDs rising and getting ever closer to the absorption peak at 488 I think that could make for a wicked laser, especially which the prospect of modelocking in mind. Another way of generating short pulses from a diode that came to mind was having them DC biased and then modulated with RF. I would have to look that paper up again to a| how short those pulses got. Great video as always!
@@jobidi99 oh for sure, Ti:Sapphire is the answer. I'm hoping some company in China starts producing the optics at a low enough price for everyone to have fun! Interesting, if you find the link please share it. One notable observation though, we might be at the limit for pulse width out of these things, certainly at this power level. The input pulse is only 2 nanoseconds, but the output is closer to 7 nanoseconds. I'm not entirely sure why. Perhaps enough energy is dumped in the die that it just continues to glow and Lase, or perhaps the capacitance of the die is storing energy.
@@LesLaboratory Sharing links on yt is always a bit difficult, but the paper i was thinking of is "Nonresonant self-injection seeding of a gain-switched diode laser" and available on reserchgate. This apparently manages to reach the picosend regime. Another paper I can't find anymore this technique was used to seed a more powerful solid state laser
What I wonder immediately: Would a SiC-based semiconductor as switching device as well as an extremely low-rise-time driver bevore the transistor significantly reduce the pulse-width and thus the "overclocking" capability? A suitable inductor between R13 and the collector could also help with that.
nanosecond gain switching of LDs is *all* about the driver. and yes... $2500 seems to be about the bottom of the price point. the upside, as always, is control and "turnkeyability".... these drivers have buttons, displays, standard connectors, etc.
When you took the average power measurement, is the power meter you're using a photodiode or a thermal sensor? If these are high peak power pulses, the photodiode will saturate and result in a false average power reading.... something to consider. Or consider using the pyroelectric sensor you made in a previous video since it more equipped for peak power
The power meter is a silicon sensor, but, the integration time is about 2 seconds, so the pulses are ignored and we get a fairly accurate average power. Unfortunately because the output is in the nano joule range, my Pyroelectric sensors wont even detect it, which is a shame. Like you I would rather see a direct measurement than a derived one.
I'm considering studying optics & lasers next year (MSEE) and videos like this make me very very excited. Amazing content as always Les! I was wondering if you could cover pico and femto as well in the future? Best, TT
Thanks! I would love to get down to pico and pemto. It has already been remarked that the Raman Laser build probably resulted in fs pulses, but at the moment I have not way to detect them...yet. As this channel grows, I'm sure I will find a way to do it inexpensively :-)
I was watching a TH-cam channel FUEL TECH .Luis was demonstrating his company's ignition systems (internal combustion engines) . Spark plugs haven't changed in100years. Then I saw you light a match in a laser beam,and I got the idea of, laser beam spark plugs you should call Luis seldom at Fuel Tech. You guys could make millions.
I'm not sure how likely it is for the laser diode to fail, but it may be a good idea to mount a screw terminal where the laser diode normally gets soldered, that way it's easier to swap the diode. It looks like the 2.54mm terminals would fit as-is on this PCB.
GREAT VIDEO...again! The image starting at about timestamp 12;52 absolutely stunning!! If I understand things correctly (that the pulses re 7ns wide at 5KHz bursts, then each 'segment' of those flashes should contain about 30,000 bursts of light? As for the retroreflector, I have built some from three plane mirrors and found that the beam returns "PARALLEL" , but not directly to the source if struck off-center...is this true in your manufactured reflector? THANKS MUCH!!!!!! ps: Are you still planning to do a video on an improved series-injection pulse transformer, as in the 1,000,000 Watt laser video??
I buit a similar curcuit with the first "high power" laser diode i played with. It was a 30W 915nm pulse laser diode used for a military rangefinder. Despite the high rating it only made pulses of 4 to 8nS at the 30W and the beam was truly invisible. One other use of avalache transistor arrays is for driving pockels cells used for laser beam q switching.❤
Nice! In a future video, I will be doing an avalanche Pockels cell driver. I have already designed and tested a PCB, I just need a suitable Pockels cell to test it. Well, I mean I have a commercial one, but I want to be able to build one!
@LesLaboratory you can get the laser and all out of a surplus Tornado aircraft laser rangefinder and target seeker.They swapped them out for newer diode and erbium based "eye safe" versions in 2018 Lots of other fun bits to play with in those too. ❤️
Yes. Just behind it a metal rail is visible. This is the linear carriage from an old printer, driven by a stepper motor connected to an Arduino. The carriage had a white plastic target mounted to it. Opened the shutter on the camera, and turned on the actuator. A real neat effect.
Have you looked into using GaN FETs? They are capable of extremely short pulses, and the ones I've looked at from EPC have packages that are designed for absolute minimum inductance.
Hmm, I will look into that. I was thinking for the next board using an FMMT415, I only used through hold because it is easier to re-work the prototype. Not thought about GaN FET's so thanks or the tip, I will go look those up!
*Nanosecond Pulsed Visible Laser Diode Driver: Pushing the Limits* * *0:24** Powerful Visible Laser Diodes:* The video focuses on the NUBM44 and NUBM47 blue laser diodes, the most powerful visible single emitter diodes available, capable of 6 watts (and potentially up to 7-8 watts with reduced lifespan) of optical output power. * *0:54** Pulsed vs. Continuous Wave:* While high power infrared pulsed laser diodes are common in LiDAR applications, visible laser diodes are traditionally operated in continuous wave mode with current limiting to prevent damage. * *1:40** Scientific Literature:* Research indicates that pulsed operation of high power visible laser diodes can achieve peak powers of several watts, with pulse durations under 10 ns and currents of 50-60 amps. [From bussi7859's comment] Optimal performance requires minimizing the length of the current path. * *4:33** Achieving High Peak Power:* The video highlights experiments demonstrating peak output powers approaching 30 watts from commercially available visible laser diodes using a simple avalanche pulsed driver circuit. The 445nm laser diodes performed exceptionally well, reaching over 200 nanojoules and 25 watts peak power. * *6:12** Avalanche Driver Circuit:* The video explores using a Jim Williams avalanche driver, known for its nanosecond rise times, as a simpler and potentially more cost-effective alternative to commercial MOSFET-based pulsed laser drivers. [From laserdan's comment] GaN FETs might be an interesting alternative due to their short pulse capabilities and low inductance packaging. * *6:50**-**7:00** Driver Design and PCB Layout:* The video showcases a custom-designed PCB for the avalanche driver, emphasizing a large ground plane and thick traces to minimize inductance, drawing inspiration from commercial designs. * *11:28** High Voltage Power Supply:* The driver requires a high voltage power supply (around 188 volts in the demonstrated setup). * *12:35** Nanosecond Pulses:* The driver successfully generates nanosecond pulses, visible as individual flashes when a piece of paper is moved rapidly through the beam path. * *13:13** Power Measurement:* Measurements reveal a pulse width of 6.97 ns at a repetition rate of 5 kHz, with an average output power of 2.34 mW. * *14:47** Peak Power Calculation:* Based on the measurements, the calculated peak power is an impressive 67.1 watts at a drive current exceeding 80 amps. * *15:18** Laser Rangefinder Application:* The pulsed laser is used to demonstrate a simple rangefinder using a beam splitter, a corner cube retroreflector, and a fast photodiode. The time delay between the transmitted and received pulses accurately measures the distance to a target. [From bussi7859's comment] Rangefinders operating at 10kHz have demonstrated ranges up to 6km. * *17:00** Fluorescent Dyes and Future Applications:* The pulsed laser's potential for fluorescence analysis and use in confocal microscopy setups are discussed. Future work includes experimenting with different avalanche transistors, laser diodes, and pulse forming network parameters to further optimize performance. [From MathewMoss-fp9ju's comment] There's interest in exploring quantum entanglement experiments with the pulsed laser and crystals like GAGG. [From arktik75's comment] There's also interest in exploring the possibility of pumping dye lasers with this pulsed laser. I used gemini-1.5-pro-exp-0801 on rocketrecap dot com to summarize the transcript. Cost (if I didn't use the free tier): $0.08 Input tokens: 21048 Output tokens: 756
Thank you for the video! I assume you made sure the oscilloscope actually has 50 ohms input impedance for your measurements? As you may be aware, most of them have a high input impedance by default. Some can be switched into 50 ohms input mode.
Thanks! Yes, its a TDS784 and has a switchable impedance of 50ohms. For fast measurements like this, it's the only way. On my little 100MHz scope I use a 50 Ohm terminator right before the input.
Do you think it's possible to create a pulsing circuit that alternates between 2 voltages? This way you can cause the laser to output at potentially two different wavelengths, leading to super interesting uses like Raman spectroscopy
Do you know what kind of precision can you measure with the range finder? I'm assuming the scope bandwidth will limit it, but assuming that's not the limiting factor, can you measure distance down to micrometers?
Measuring small distances woudl become technically challenging due to pulse width, risetime and bandwidth constraints. Centimetre precision will be good I imagine, but for microns a different approach is needed. You could try interferometry, or, if you want a super cheap, and very cool solution check out this video by Applied Science Laser diode self-mixing: Range-finding and sub-micron vibration measurement: th-cam.com/video/MUdro-6u2Zg/w-d-xo.html
This is a phenomenal result. Do you have any idea if the same technique can be used for LEDs? I feel like LEDs are getting to the power levels of being able to pump lasers like flashlamps are able to.
Hi. I've built the same schematic as you did and I encountered two issues: everything is working fine when using a common Led diode (red or white) till I reach over 180V. If I try to go above, the resistor R13, 1.2kohm and the transistor (I used 2N5192G) is getting hot for whatever reason. seccon issue is that I tested the board with a laser diode (blue and red) and even at 200V it couldn't reach the treshold. I didn't damage the diodes though, they still work well in CW mode. Do you have any suggestions regarding my situation? Thank you!
That's odd. The transistor shouldn't even get warm, the duty cycle with nanosecond pulses approaches zero! Sounds like the transistor in yours is always on, effectively sinking the HV supply. How are you driving its base? Does the drive signal go from 0 to 5v? (Some arbs have an offset output, i.e zero is not really zero)
$2 bucks for 5 boards is great. I think they panelize at the factory to reduces costs. I hand populated these though because of the experimental nature of it.
@@LesLaboratory those boards you made where very nice, ive started using jl for my little laser pointer boards. have you seen the nichia diode arrays, ive got one here rated at 128 watts, its scary, ive not experimented with it yet due to having damaged laser eyewear, ime not risking it untill my new set arrive from thorlabs
Did you notice any appreciable differences between the ZTX415 and the 2N5192G? You made drivers with both, but left us hanging on what, if any, differences there are, and if one is better than the other in any way!
Yes, I probably should have mentioned it. The 2N5192 results in a higher threshold and is obsolete, though still available. The ZTX415 has a lower threshold and is currently available, but expensive at 10 bucks each. Both work just fine, and I am sure others will too.
Maybe try a GaN FET? The EPC boards you showed has full design files (I've made some using JLC, which was the EPC9126, and others), and they are very serious about minimizing inductance, you could probably get nicer edges and more peak power!
The output is some two orders of magnitude lower that the SC setup. I have tried it and seen no broadening. With that all said, it might be possible to get it to do something interesting, perhaps if I use kilometers of fiber, and it is a work in progress
Excellent vid; thanks for sharing. I didn’t know diodes could be driven quite that hard without COD etc. The dye pumping sounds very exciting; I’m currently (slowly) building a very high power rig myself. Can you share what the specific dye is that’s suitable for the pulsed ~450nm pumping? Another thought: you varied the pulse repetition rate, but did you get any effects / curves for a fixed rep rate but varying the length of the pulses? Keep up the good work!
Back in 2007 I wrote about COD on Sam's Laser faq, and I was fairly convinced it was unavoidable for Laser diodes. There must be thermal effects involved in this process as well as sheer number of photons. The Dye in the papers was DCM (4-(Dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)- 4H-pyran) and Lases in the red. It's about 160 bucks a gram where I am! The length of the pulses is dependent on the PFN and is not variable. It is possible to lengthen the pulse by adding coax to the capacitor bank, but I do not have any text data for this yet.
@@LesLaboratory Thanks for the reply. Oh DCM(!!!), I have some of that in the garage - let me know if you'd like a small amount of it for your experiments. :) (I'm UK based).
@@danielbriggs2 Thanks! I will more then likely take you up on that when I get round to a proper build. I have a very short path Cuvette, I just need to get round to making a mount and a resonator for it.
I find this all interesting, even if i don't understand everything, however i am left with one question .... At both +- 12:51 and +- 17:59 we can see the pulsed blue laser.... it looks like the light starts as kind of a vertical line and then evolves into a more circular shape.... What's happening here ? Is this maybe an effect by the camera ? Or is this something else that i don't understand ?
This is because of the awful beam shape. Laser diodes have different rates of divergence, that is not easily corrected with a single lens. They have a "slow axis" where divergence is lowest, and a "fast axis" where divergence is higher. When collimating for a far field spot, you end up with one axis being collimated, and the other being focused, hence the change in beam shape along its path. This can be corrected with an external cylinder lens pair, as seen in the rangefinder experiment later in the video.
Very interesting video, thanks for sharing. I have two questions: 1. What was the energy per pulse that you got? The paper showed pulse energies in the mJ range...did you get anywhere near to that? 2. Do you think the pulsed laser diodes (905nm) can be overdriven in a similar fashion to get higher pulse energies? The datasheets all say up to 100W peak with 0.1% duty cycle and 100ns pulse widths, but i haven't seen anything about overdriving them. Pls share if you know about this.
Thanks! Pulse energy is about 468nJ (calculated) so almost half a mJ, not too bad. The IR diodes, (I assume) are already at their max output in pulsed mode. I suspect with cooling, you might get a couple of watts CW. It might be interesting to try though with very short pulses!
At 12:42 to 12:56, you are definitely not imaging the 5 kHz rep rate train of nanosecond pulses since the length in air of a nanosecond light pulse is about 30 cm (1 foot), and the separation distance between pulses at 5 kHz rep rate is about 60 km. I don't know what is producing the light dots you are imaging, but it cannot be the nanosecond long pulses at 5 kHz rep rate. Below some conjectures as to what may be causing the light dots you are imaging, but these are only guesses. It is difficult to tell exactly from the images, but the separation between the light dots looks like a few centimeters or so. Using 3 cm for the separation means the light dots are separated by 0.1 ns of light travel time over that separation distance and the corresponding repetition frequency is 10 GHz. If the separation is 6 cm, then the time between light dots is 0.2 ns and the corresponding repetition frequency is 5 GHz. Thus, you may be imaging some high frequency (multi-GHz) intensity modulation of the luminescence, intensity modulation of the amplified spontaneous emission (ASE) or intra-pulse intensity modulation of the laser pulses emitted from the laser diode. This modulation may be due to some RFI or EMI picked up in the circuit, perhaps from a 5 GHz wifi signal, mid-band 5G cell phone signal, or some other multi-GHz electronic device you may have. Another possibility is that back reflections of the emitted laser light from something may be causing external optical feedback into the diode laser cavity which may cause high frequency intensity modulation of the output. It is well known that a major problem with semiconductor lasers, both FP and DFB types, is that these semiconductor lasers are highly sensitive to the laser light which re-enters the laser cavity after being reflected by an external reflector. External optical feedback of the laser light usually causes instability of operation of a laser diode and increased laser intensity fluctuations. A variety of optical elements, including lenses, fiber endfaces, etc. can be the source of unwanted optical feedback. Rayleigh backscatter from fiber can be another source of optical feedback. Packaged laser diodes may also receive optical feedback because of the external cavity formed between the laser diode chip and a transparent window of the package.
I was also confused about this photography. I tried to drive with 5KHz rep rate and found these pulses when I move the laser viewing card very fast in the path of the beam. But I am unable to capture using camera, I don't know The ISO, shutter speed required to capture the effect. Although a 5,6 ns pulse will travel around 5-6 feet, so it is some illusion I guess?
@@kumar-qb2pe As you move the laser card across the beam, if you move it fast enough, each laser pulse will hit on a different spot on the card separated from the spot that the previous pulse hit, so you'll see a series of spots across the laser card. The fluorescence from the laser card induced by the incident few nanosecond laser pulse will last on the order of microseconds to tens of milliseconds depending on the particular fluorescent material the laser card uses. The typical human eye integration time is on the order of 100 milliseconds, so your eye is integrating the fluorescence from all the spots over a time much longer than the fluorescence lasts. To have a camera see the same thing as your eye when recording the laser card pulled through the laser beam, set the shutter speed to 1/10 th second which is 100 ms. On some cameras the closest setting may be 1/8 th second which is 125 ms, which is close enough. The ISO required at that shutter speed will depend on the aperture size you have. In general you want to have an ISO as low as possible to minimize noise. You'll have to experiment with ISO and aperture. You'll want as large an aperture as possible, but the larger the aperture, the lower the depth of field. I suggest starting with the largest available aperture that has just enough depth of field for the image you want to capture. Then start with the lowest available ISO and crank it up until you see the spots in the image. Good Luck!
How well does this work with LEDs? I want to make a high power flash for night time aerial photography with UAVs. I calculated that a 400J flashlamp flash is enough for my purposes but the cap bank for it would be heavy, making the use of a larger multi rotor necessary. So how much can white LEDs be overdriven in pulses? Edit: NVM A xenon flash lamp is still far better and cheaper than equivalent overdriven LED flash. A 400J cap bank discharging in 1ms means a power of about 400kW, if 20% of that is converted to light that's still 80kW of light vs a couple of kW for a very expensive and heavy array of over driven LEDs.
I have tried with LED's in place of the laser during testing, and it works ok, but I have not measured the brightness. I don't think you would anywhere near the light output you are needing with any amount of LED's. Xenon for those applications is exactly the right choice as you say. Somebody somewhere must be producing compact capacitors for flash applications though.
Hi! Very interesting video and I learned a lot. I might have seen every single one of your uploads by now. I'm working on a project to generate two picosecond pulse-width beams from diodes whose peaks are apart on the order of picoseconds. Any suggestion on how to go about this would be extremely appreciated!!!
Honestly i'm not sure at this point. ~7 nanoseconds is where I am at just now. I might be possible to get into the picosecond range by re-spinning the board with lower inductance and capacitance. The ZTX415 is capable of picosecond rise times, but I am sure there are better alternatives. What is your application?
You could try driving the laser driver with an impulse generator based on diodes/TL. Check herotek and Macom. But,they generator pulse >30ps. if you come across any technique to generate a few ps.Please let me know
@@LesLaboratory, I really appreciate your feedback. For ultra-fast molecular spectroscopy, you need two ultra-fast pump and probe pulses. I'm trying to do national lab stuff on a budget, and this video is the best starting point I've seen.
I wanted to mess around with the values of those to see what effect it had on pulse width. Larger values, a few Ohms in total, lengthens the pulse, however the waveform looked particularly ugly, so I reverted back to 0 Ohm.
The signal looks horrible. I was hoping it would lengthen the pulse a little. To lengthen the pulse and lowed the repak power, a short length of coax soldered on to the capacitor bank (As In AN47) will do it much more sensiblly. The resistors have been removed in the latest design, and I will be testing again soon.
Could this be used to make a supercontinuum laser, replacing the complicated nitrogen+dye pump? If I remember correctly, your dye lased around 426nm in the supercontinuum video. Sharp makes a 425nm diode that'll do 2W continuous.
It might, and that is one of the motivations of this project. For SC generation (like with N2 and Dye Lasers), you need a fairly wide bandwidth of Laser light to begin with, and the bandwidth from diodes is probably too narrow. The power levels are very low as well, about two orders of magnitude lower than the nitrogen Laser pumped dye laser. With all that said, it might, just might, be possible to use very narrow core fiber, and multiple pump diodes to get something to happen. So far I have not observed any broadening in fiber, but I have only tried a couple of Laser diodes.
So Les, I was just thinking how similar the circuit is that you are using is to the famous spark gap capacitor discharge of the DIY Nitrogen laser. And I wondered could you put a laser diode across the plates of a nitrogen laser? Does this make sense?
It's very similar. I think a diode placed across the rails of a Nitrogen Laser would be instantly destroyed though. There is a huge difference between ~200v and 20kV, before the diode even emitted light, I suspect several kV woudl appear across it. With all that said, I have been toying with the idea of building a Nitrogen Laser (for the most part) on a PCB. It would be so cool if they could be manufactured really small and inexpensively.
Hello Les! :) Great job with your channel - many of the topics align with my research! I decided to comment on this video because a large part of my PhD thesis (defended about a year ago) focuses on advanced electronic drivers for generating nano- and subnanosecond laser pulses from semiconductor lasers. Unfortunately, it's written in Polish, so the text may be difficult for you to understand. :( Anyway, if you ever need assistance on the topic, or if you'd like to share ideas or brainstorm, feel free to reach out! :)
@@LesLaboratory Would also be cool to have it blink 1 billion digits of Pi within a second (using its binary expansion, or variable length pulses for each digit) :)
Hello!! This is a most educational video to improve solid state laser diode designs. I need help designing a depth measurement, hand held, range finder. I have been experimenting with 405 nm lasers to measure from the surface to the bottom of shallow ocean environments (under 100 meters). Could your systems be used for time-of-flight measurements? Your set up looks so good for improving my designs! Any suggestions?
@@Forensic1Man is could be adapted for time of flight as described in the video. 405nm has a shallow penetration depth in seawater though, you might be better off with a wavelength of around 480-490nm for underwater ranging.
According to academic papers, yes, but it seems to require a rather precise setup. Pumping a Dye laser with one of these is one of the motivations for this project.
I just wondering, why use a powerful diode laser for making pulses. Is it useful for data transmitting or it can be useful for engraving some objects??
Can you please explain how the photo was made. This isnt obviosly how you photographed distinct and uniformly spaced spots. It isnt obvious how the paper sees discrete spots also. isnt the light pulse itself is a foot wide ? All I can surmise is that the spacings must be a frame rate strobe of the camera off the card but the ending photo without a card is a mystery. What reflected the pulses at those unique spaced posistions??????
A nanosecond is slightly under 30cm. at lightspeed. I suspect your thumbnail is off the mark. 😏 Computer pioneer Grace Hopper used to hand out nanosecond lengths of wire during her computing lectures. They were just under 30cm. long.
This is what is ask GPT about "Can FWHM solve traditional TOF Lidar interference issue?" ChatGPT said： ChatGPT FWHM (full width at half maximum) can help address interference issues in traditional time-of-flight (TOF) Lidar systems by providing a measure of the laser pulse width. A narrower FWHM can improve range resolution and reduce overlapping signals, which minimizes interference from multiple returns or background noise. However, it’s not a standalone solution; optimizing the system design, including detector sensitivity and signal processing algorithms, is also crucial for effectively mitigating interference
I am into free energy. I already got the a small working prototype. To scale up and apply to actual products. Your expertise is the final missing piece. If you are interest in changing the world, please let me know.
Great show. I think we are halfway through the process. Could you finish this process and make a video about FEEDBACK AND AMPLIFICATION of the laser beam? I would like links to buy things like lenses, mirrors, SESAM etc. Please provide such a version of the video and description because I have little knowledge on this subject. I found such a video but there is no explanation th-cam.com/video/aOTLhk6GuB4/w-d-xo.html If it would not be a problem for you. Best regards
So…I developed a nanosecond pulse driver for LED (the 3W Nichia “star” variety that were brand new at the time) back in 2006-7 to observe ink droplet formation in high speed printing systems. LED because it was safe for operators and didn’t require a special curtain.
I built the driver with a totem output. The top side connected to 60V and the bottom to -5V on the theory that removing any charge on the LED die was essential to eliminating tail current and speeding up the output.
This arrangement was successful. My employer applied for the patent but another inventor had just filed before me. Such is life!
@@envisionelectronics sweet, that sounds really cool! Shame about the patent fiasco though!
Fantastic, this is really cool to see. I work for Lightware Lidar, and our bread and butter is pulsing laser diodes really hard for miniscule periods of time (for the purposes of range finding), so this hits really close to home for me. I also love that long exposure of the beam at different distances, you can really make out its profile.
That sounds awesome! I'm curious what kind of pulse widths and peak powers do they get out of those for Lidar?
@@LesLaboratory We use 905 nm diodes with pulses of ~20 ns @ 20 kHz. I think our peak power is ~10 W, with the average being 2.5 mW.
Something like that was part of my masters work 40 years ago. I used coax cable as load lines. This gives an almost perfect rectangular waveform of the diode current. The length of the load lines defines the pulse width.
Awesome! There is a screenshot of a paper towards the end of the video, where a coax is used to customise the pulse width.
Ah the famous coaxial line pulse forming network. Try making it out of foil sheet and roll it up like a jelly roll capacitor. Add some extra turns of foil in the center and it transforms the current wave to a very high voltage quickly. It doubles the input voltage with every quarter turn. ❤
My pulsed range finder did run at 10kHz and its range was 6km. Used in the SAAB Training System weapons simulator BT 41, 53, 56.
Sweet! I have an ex MOD rangefinder under the bench. One day I will dust it off an do something with it!
I believe range-finding technology was used for military technology and weaponry before it was ever applied to the automotive industry.
For instance: Collision avoidance technology.
You've done a very nice design, assembly and experimental testing project with overdriving visible laser diodes.
This video brings back memories of when the Laser Sensors Team that I worked on at the U.S. Army Night Vision and Electronic Sensors Directorate (NVESD) did some testing of overdriving inexpensive commercially available continuous wave (cw) near infrared (NIR) diode lasers to produce nanosecond scale pulses for rangefinding, lidar, and free space optical communications applications back in the late 1980s.
It was fun in the initial experiments cranking up the peak current until we blew up each diode laser to determine the overdriving limits. We would always say when we reached the limit for each diode "It blowed up real good!" That was a phrase that two farmer characters, played by John Candy and Joe Flaherty on SCTV in the '70s, would say in the "Farm Film Report," in which the two farmers discuss the latest movies and film theory - a routine parodying Siskal and Ebert. Supposedly the best part of a film is when things get "blowed up real good." They even had some celebrity guests (other cast members dressed up as Dustin Hoffman, Bernadette Peters, etc) who were "blowed up real good" during special "Celebrity Blow Up" segments.
Hi, I'm a test engineer and I find your experimental setup and your entire experiment incredibly useful! I have been trying to puzzle out for months how to measure pulsed laser diodes and what equipment I can use there (I haven't been able to find very short rise time transimpedance amplifiers), so the average power method and back-calculation of peak power seems very insightful, thank you!
However, for the love of God (and I know I'm just a random person on the internet but I'm also an electronics engineer with a passion for signal integrity and proper layout techniques), do NOT split the ground planes and place just a single pour ground plane over your entire board. It will improve your noise a lot. I don't know who in the literature suggested to put these R7-R11 resistors there and for what reason, but the split plane underneath the resistors is gonna create a very strong EMI that will radiate everywhere. Moreover it will add inductance to your dI/dT loop which in your case is from point C to H1.2 to the laser diode, then back through H1.1 all the way to the GND of Q1. This loop should ideally have a *uniform* ground plane underneath and be kept as short as possible to minimise EMI (both radiation and pickup).
Excellent, you are welcome!
I am not a test engineer, so good advice like this is always welcome! The bottom plane under the PFN likely contributes significant capacitance, slowing things down as well I imagine, diode leads could be shorter, and the FMMT415 could be used in place of the ZTX415.
@@LesLaboratory the bottom plane doesn't add as much capacitance as you'd think because you have a two layer board so the ground plane is really far away from the signal layer (since this is a two layer board and the gnd is in the bottom layer which is usually 1.2-1.6mm thick) and because the ground plane split that you have is very minimal. That said, from an electrical engineering perspective the most important thing for a pulse forming network is not necessarily low impedance or inductance, but impedance matching of the transmission line and the source impedance (and ideally the load impedance as well). So my point is, you can put a solid ground plane underneath and then calculate via your EDA software the trace width that you would need to achieve a specific impedance (say 50Ω or sth).
Having a signal pass over a split plane is a big no no in PCB design because that spot there is acting as an antenna and it both receives and emits radiation (which is harmful both to other devices nearby and to your device if it picks up stray signals). Moreover the extra zero ohm resisors are adding inductance to the loop and that both increases the rise/fall time and it can create a gnd bounce (where your "ground" is not really at 0V during the transient but at a slightly higher voltage than the actual gnd plane exactly due to the zero ohm resistors which have some small resistance and additional inductance). Also, important thing, I wouldn't use zero ohm resistors in the PFN even if it's just for stitching (which imho is a bad idea to begin with as I said), but I would rather use some specific resistance value like 0.5Ohm or whatever because then you have a controlled value, zero ohms are just some random low value with no effort in controlling the value.
In any case, I would suggest at a future iteration to just ditch all of these parallel low side resistors at the very least. Second step would be to try to calculate a target impedance to hit for the trace that delivers the pulse.
@@InTimeTraveller Agree. Also the width of the traces determines the absolute high current value delivearble to the load. Same with cables.
Yes, you can, but keep it cool. Every 10 C above the rated temperature (up to a limit, of course) about half of the life is reduced.
Thanks! I was hoping someone who has played with these in CW mode would comment on lifetimes :-)
I guess heating is less an issue here, rather bonding wire ripping themselves apart from magnetic forces.
@@rolmie I'm sure there are a bunch of other failure modes as well, such as dendrite growth, and photon damage to the facets, but so far so good! I should just let it run at high current for a few days or weeks and see what happens! The results in the literature seem good though in terms of lifetime.
@@LesLaboratory @Breaking Taps might have the equipment to answer this :)
@@rolmie @BreakingTaps might even have a use for it. With such short duration pulses, thermal effects will be absent, but the number of photons from each pulse, might, possibly, be able to micro machine on the order of atoms or molecules at a time!
Awesome ! this just keeps getting better ! The shot with the pulse train output is mindblowing, cheers for your work !
Thanks! Yeah that picture is pretty neat. I built a motorized rig out of old printer parts, just so I could get that shot!
@@LesLaboratory :)
Jim Williams lives on in many ways ,good to see another. Great Video!
Yep, I'm just standing on the shoulders of giants. Thanks!
@@LesLaboratory ...another fan of Jim Williams here...actually spoke to him briefly about "flyback switching regulators"...must have been back in the 80's...brilliant guy!!
Essentially the design of the pulse circuit is to keep the length of the current path as short as possible, I made a pulse laser with bonded chips that gave the same result at 70V as with discrete components and PCB at 350V . I changed from avalanche Transistor to FET. It is still produces today with no improvement.
Make a video about it.
I've binged a bunch of your videos now and you should have at least 100k subscribers. Great content!
@@inshortspaceisbig3349 Thanks! Maybe some day!
As a former PCB layout professional, I can say that your HV traces need much more clearance away from ground. Even a small amount of moisture in the air or using it at a bit higher altitude, (I live at 6,000 feet,) could cause arc-over of the HV. A spacing between the conductors and ground of nearly .15 inch would be the minimum spacing for generic PCB material (without any sort of conformal coating) that I would use under any circumstances with just over 230 V applied, while .2 inch would be much better since your PCB layout clearly has the room for it. With the possibility of an unknown higher voltage being applied for experiments I would design the bd with a .25 inch gap between the HV and Ground wherever I possibly could, (including a wider footprint for the transistor and the LASER connector,) with a painted-on HV dielectric where the traces cannot be given that clearance in order to ensure reliability of operation. Any sort of contamination between the traces, from moisture to dust reduces the dielectric insulation between the traces, increasing to probability of arc-over. Circuits used near the saltiness of the oceans also require extra clearance. Even altitude has a pronounced effect, so with all of those unknowns it's better to just provide as much clearance as you can afford to in the creation of a HV bd design made available to your users.
0.2 inches, is 5mm.
Don't think this is of concern here with avalanche breakthrough from the transistor of a little more than 100V. Inductance on the other hand plays a major role at these frequencies...
For sure, the design is not optimal, nor hardened against the elements, it's just experimental. IPC-2221 suggests 1.25mm minimum clearance up to 300v between conductors on an un-coated board (0-10007ft), but it is generic and probably superseded by more rigorous standards.
The pin pitch on the ZTX415 is really quite small, and so creepage could occur on the package itself as well. Far better to use the FMMT415 and ensure sufficient clearance on the rest of the board.
For this I just wanted proof of concept. For another project I have been working on I have taken much more care, since I have up to 1kV on the PCB!
I think creepage distance is much less of an issue than you imagine.
First creepage distance is much less of a risk on the parts of the circuit with nanosecond pulses since there will not be enough time to get a breakdown propagating unless the field strength is very high.
On the DC part it can be an issue but not as big as if it was 230 mains voltage. When designing for 230 V mains the nominal peak voltage is 325 V and then you need to be able to handle voltage spikes of a few kilovolts lasting for tens of microseconds from switching transients and distant lightning. In a small DC-supply like this there will be much less transients.
More clearance is better but not if it causes more inductance in the pulls path. The loop area needs to be as small as possible. To spread out the legs of e.g. the laser diode will reduce the peak power since the inductance increases. It is better to improve the insulation if needed by coating, poting or oil immersion.
Another thing to consider is the threshold current of the diodes. They typically won't lase until they are at a couple of hundred ma of current. You might want to think about being able to apply a dc ofset to get to threshold, then modulate on top of that. Of course now you will have to improve the heat sinking of the diode. I would also be curious what the diode is doing thermally.
I have seen biasing like before, and this was even done in earlier efforts, but didn't seem to be required at these extreme currents. It would still be an interesting thing to try out, now that all the other issues are ironed out. Someone suggested there is a paper where they were getting picosecond pulses out of an LD that was biased and fed with RF pulses.
There is way more work to be done here!
Id love to see more quantum entanglement experiments and wish i could test a theory with using a GAGG crystal instead of dye
It's something I aim to get around to, probably with BBO (if I can afford one) but after that, I will be looing at less expensive materials to try.
This is _very_ cool! At that pulse duration and peak power it might be starting into the range of micromachining (when coupled to a microscope objective for a small spot). Especially in something like silicon, the absorption depth will be very small, around 1um iirc? Really love that long-exposure shot too, neat to see the individual pulses!
Thanks! Somewhere in that region, it is very shallow on silicon. This should work with 405nm diodes as well. There are even 375nm diodes in production now but they are $$$!
I'm glad you got jlc as a sponsor! I'm a long time customer, they always deliver very nice PCBs! Really changes electronics prototyping for me!
Thank they are superb. Process tracking, fast delivery, and I always seem to rack up coupons for SMD assembly as well, which is nice for when I need it.
When I worked at a National Lab, I designed and built a laser system that would gain-switch a seed laser diode using rf, resulting in a 50ps FWHM pulse at various rep rates. This would be sent to a 5W fiber laser amplifier. The average power out of this system would remain 5 watts, but the peak power was dependent on the rf duty cycle. I reached peak powers of 1.6kW in this system.
@@JohnHansknecht That sounds awesome! I have read about these, very very nice! I'm going to have to aim for picoseconds with these diodes, things could get interesting.
@@LesLaboratory You won't get power out of seed lasers that are gained switched. Just about any laser diode can be gain-switched. No special diode needed, but you need to bias it near threshold through an impedance block (resistor and sometimes inductor) and capacitively couple an rf energy of about a watt. A typical 10mW seed diode will put out barely a milliwatt when gain-switched, but if it is fed into an amplifier, this is where the results can get interesting. Takes a very fast (many Ghz) scope with photodiode input to setup the parameters for proper switching. This is usually out of reach of most hobbyists.
I did wonder if some of the microwave PTFE laminates and GaN microwave devices might give faster rise times, but how about a machined coaxial cavity? That might give a really serious and near-constant discharge current when compared with coaxial cables. If the RF impedance is a good match, then it shouldn't have any overshoot or oscillation. You'd still need a hyperfast pulse generator to drive the microwave device of course. Once you get past 30 GHz, you are really into chip-bonding territory, but there you are well into the picosecond domain. Not much point as the laser packages are hugely suboptimal and Stone Age - unless you can hack them up and get within a millimetre or two of the diode chip. Very cool work, fascinating as ever.
You are actually correct about the cavity method. This was actually used already for high-Q filters, and for Piston-tuners. Use of any kind of high-Q cavity structure, with lowest losses, creates the most compact bandspread (aka, sharpest waveshape) of the rebound signal. Now actually being used for quantum physics, as "vacancy devices" to create optical manipulations between goezinta and goezouta photonic driving impulses. Works great, and is highly efficient as opposed to other methods.
Interesting, this is similar with a few other comments here. Microwave engineering is not my forte, but I will have a read around and see if there is anything suitable. Picoseconds is a really tempting domain.
I bought some cheap laser glasses from amazon and I wanted to make sure they actually worked. To test them I dawned an oculus quest headset. This essentially air gapped my eyes from the laser so that there was 100% no way I could get the laser light back into my eyes. Then I projected the laser onto a target where I could see it, and then held the glasses I had purchased in front of the target.
Would be interesting to try with a 405nm diode too.
Awesome work. Love ya man
Thanks! Oh totally. 445nm is cheap 405 not so much, but now its been well tested, its time to try I think. There are UV laser diodes now as well, so things are getting real interesting!
I second the idea to use a GaN MOSFET take a look at the LMG1020 and his datasheet.
Yep! There is some interesting tech out there now for this particular field. I'm also interested in playing with SiC stuff for high voltage work as well.
Awesome job, Les. This is likely the best video of yours thus far. I am impressed.
Thanks! Hopefully people will be able to make use of it!
I've been waiting for someone to do this,thankyou for this beautiful madness
@@Greg-l3j You are welcome! I have just ordered a couple of high power 405nm and 520nm diodes to try as well!
Cool that you used 2 diodes from Roithern Laser, its an Austrian company but they seem to have the widest range of diodes - worldwide! We are also using them ...
I don't quite understand the "long exposure" photo at 12:55. What is the beam reflecting off in mid-air?
The earlier shot with the flapping paper is essentially using this laser as a glorified strobe lamp. The size and spacing of the spots would be the paper's speed times the pulse width and pulse repetition period.
The pulses themselves would be spatially a couple of feet long (using the rough 1ns/foot speed of light). You could observe this using smoke and a picosecond-ish camera shutter.
I was also wondering how a long exposure shot resulted in discrete laser dots like that - I don't understand... I guess there has to be a "moving piece of paper" or something in that photo too, but you just can't see it in the exposure, and you only see the dots
This is probably explained aliasing of the camera internal sampling rate vs the function generator frequency used to trigger the laser driver. This could explain the spacing between the spots and the length of those spots. Normally, those spots should be 4 or 5 feet long because the optical power metered was about 6ns.
Ah! I thought about videoing how this shot was done, but didn't think anyone would find it very interesting! Maybe I should have!
I tried doing a long exposure, whilst quickly wafting a piece of paper along the beam length, and it works, as on camera, but not particularly well.
Right behind the Laser you can see a pair of rails. This was a Linear motorised rail from a printer. I mounted an arm to it with a white target on the end, and that sits in the path of the Laser beam. The camera shutter is opened and the target moved from one end to the other, quite quickly. This was driven using an Arduino.
To answer your initial question, I would believe you. Some IR laser diodes worked like this. It was in an oldish book, "more electronic projects for the evil genius" chapter 4 (I had to look it up) where I saw this. Uses a thyristor to dump a capacitor through an IR laser diode. It is on the Internet Archive now, but I still have the physical book somewhere.
I had a thought of a cool technique for lidar, but I couldn't wrap my head around how to do it with a pulse:
Use Radar ideas:
Analoguely:
modulate the beam's intensity with a frequency ramp, say 100-1KHz and back down.
Mix the PD return with the outgoing beam (some lasers used to have PDs in them), and get a beat frequency out of it; the frequency of the beat will tell you the range to the object without needing to solve ns timing for TOF sensing. If you want to get even more accurate, you can start figuring out the phase difference (some guys did that with a Red Pitaya).
I'm sure you could figure out how do that with say a 100-1KHz pulse interval and figure that out, but the math is making my head hurt... and "phase" is...?
Interesting idea but I don't see how the beat frequency will remain constant for a non- linear frequency sweep; nonlinear because the sweep goes up and comes down. Due to this, the difference between the outgoing and incoming frequencies will not remain constant for a given distance.
@@akarshchaturvedi2803 Good point. I'd probably make it be a constant speed sweep from 1KHz to 20KHz and then reset back to 1KHz. You should have a nice constant split except at the end and that would be an out-of-bounds step versus a ramp to zero and a ramp back up to steady.
Spectacular! Thanks Les!❤
Thanks!
This vid came just in the right time! I've been working on a small project to make dToF LiDAR. So far the digital part (tapped delay line on FPGA) is straight forward, but I'm looking for a fast LD driver that capable of 10ns pulse width.
Great! You can extend the pulse width by attaching a length of un-terminated coax in parallel with the capacitor bank (or even in place of it!) to widen the pulse width (see application note 47). There is a nice video on this by Dave Jones @EEVblog here: th-cam.com/video/uBYMePUFinQ/w-d-xo.html
Another fantastic video Les. Thanks so much for putting in all that work and sharing it with us all.
Thanks! You are welcome!
les is more
@@gsestream 😄
🤣
Don't forget the old Tunnel Diodes. We built our first sub-nanosecond pulser, using these, along with some IBM digital transistors, capable of 1GHz bandwidths. The problems were horrible. Temperature unstable, moving or squeezing the coax changed the speeds and shapes, device failure due to avalanche breakdowns, absolutely NOT-repeatable. Other than that, it worked great.
So it's a maybe :-D
Oh man do I have to comment on this. You mentioned pumping dye lasers with diodes, but in my opinion there would be an even more prominent example that might be of interest, Ti:Sapphire. With the power of blue/green LDs rising and getting ever closer to the absorption peak at 488 I think that could make for a wicked laser, especially which the prospect of modelocking in mind.
Another way of generating short pulses from a diode that came to mind was having them DC biased and then modulated with RF. I would have to look that paper up again to a| how short those pulses got. Great video as always!
@@jobidi99 oh for sure, Ti:Sapphire is the answer. I'm hoping some company in China starts producing the optics at a low enough price for everyone to have fun!
Interesting, if you find the link please share it. One notable observation though, we might be at the limit for pulse width out of these things, certainly at this power level. The input pulse is only 2 nanoseconds, but the output is closer to 7 nanoseconds. I'm not entirely sure why. Perhaps enough energy is dumped in the die that it just continues to glow and Lase, or perhaps the capacitance of the die is storing energy.
@@LesLaboratory Sharing links on yt is always a bit difficult, but the paper i was thinking of is "Nonresonant self-injection seeding of a gain-switched diode laser" and available on reserchgate. This apparently manages to reach the picosend regime. Another paper I can't find anymore this technique was used to seed a more powerful solid state laser
@@jobidi99 got it thanks! I will have a read and see if it is something that can be applied! Cheers!
This is really spectacular work. Love it.
Thanks! :-)
I love your videos, and this was another hit!
@@tolkienfan1972 Thanks! It has done better than I thought, perhaps because there is no unobtanoum for the build.
@@LesLaboratory more likely google "tweaking" their algorithm arbitrarily
What I wonder immediately: Would a SiC-based semiconductor as switching device as well as an extremely low-rise-time driver bevore the transistor significantly reduce the pulse-width and thus the "overclocking" capability? A suitable inductor between R13 and the collector could also help with that.
Once again a awesome Video 🤩
Great job! ❤
Thanks!
Wow. Outstanding, mate. Subscription well earned!
Spectacular, thanks!
Great video. Thanks!
nanosecond gain switching of LDs is *all* about the driver. and yes... $2500 seems to be about the bottom of the price point. the upside, as always, is control and "turnkeyability".... these drivers have buttons, displays, standard connectors, etc.
When you took the average power measurement, is the power meter you're using a photodiode or a thermal sensor?
If these are high peak power pulses, the photodiode will saturate and result in a false average power reading.... something to consider. Or consider using the pyroelectric sensor you made in a previous video since it more equipped for peak power
The power meter is a silicon sensor, but, the integration time is about 2 seconds, so the pulses are ignored and we get a fairly accurate average power.
Unfortunately because the output is in the nano joule range, my Pyroelectric sensors wont even detect it, which is a shame. Like you I would rather see a direct measurement than a derived one.
I'm considering studying optics & lasers next year (MSEE) and videos like this make me very very excited. Amazing content as always Les!
I was wondering if you could cover pico and femto as well in the future?
Best, TT
Thanks! I would love to get down to pico and pemto. It has already been remarked that the Raman Laser build probably resulted in fs pulses, but at the moment I have not way to detect them...yet.
As this channel grows, I'm sure I will find a way to do it inexpensively :-)
@@LesLaboratoryAutocorrelator is the technique used to measure pulse width in fs-ps. It requires an interferometer setup with very good delay line
@@michaeljordan494 yea but thats easier said than done. Check out the *MPQ 268 report* from Max Planck Institute of Quantum Optics
I was watching a TH-cam channel FUEL TECH .Luis was demonstrating his company's ignition systems (internal combustion engines) . Spark plugs haven't changed in100years. Then I saw you light a match in a laser beam,and I got the idea of, laser beam spark plugs you should call Luis seldom at Fuel Tech. You guys could make millions.
Very cool!
Thanks!
I'm not sure how likely it is for the laser diode to fail, but it may be a good idea to mount a screw terminal where the laser diode normally gets soldered, that way it's easier to swap the diode. It looks like the 2.54mm terminals would fit as-is on this PCB.
You can get those dyes from Highlighter pens.
GREAT VIDEO...again!
The image starting at about timestamp 12;52 absolutely stunning!! If I understand things correctly (that the pulses re 7ns wide at 5KHz bursts, then each 'segment' of those flashes should contain about 30,000 bursts of light?
As for the retroreflector, I have built some from three plane mirrors and found that the beam returns "PARALLEL" , but not directly to the source if struck off-center...is this true in your manufactured reflector?
THANKS MUCH!!!!!!
ps:
Are you still planning to do a video on an improved series-injection pulse transformer, as in the 1,000,000 Watt laser video??
I buit a similar curcuit with the first "high power" laser diode i played with. It was a 30W 915nm pulse laser diode used for a military rangefinder. Despite the high rating it only made pulses of 4 to 8nS at the 30W and the beam was truly invisible. One other use of avalache transistor arrays is for driving pockels cells used for laser beam q switching.❤
Nice! In a future video, I will be doing an avalanche Pockels cell driver. I have already designed and tested a PCB, I just need a suitable Pockels cell to test it. Well, I mean I have a commercial one, but I want to be able to build one!
@LesLaboratory you can get the laser and all out of a surplus Tornado aircraft laser rangefinder and target seeker.They swapped them out for newer diode and erbium based "eye safe" versions in 2018 Lots of other fun bits to play with in those too. ❤️
@@christopherleubner6633 I already have one 😉
What exactly did the long exposure photo at 12:53 capture? Were you moving an object at constant speed to intersect the beam and capture each dot?
Yes. Just behind it a metal rail is visible. This is the linear carriage from an old printer, driven by a stepper motor connected to an Arduino. The carriage had a white plastic target mounted to it.
Opened the shutter on the camera, and turned on the actuator. A real neat effect.
Have you looked into using GaN FETs? They are capable of extremely short pulses, and the ones I've looked at from EPC have packages that are designed for absolute minimum inductance.
Hmm, I will look into that. I was thinking for the next board using an FMMT415, I only used through hold because it is easier to re-work the prototype. Not thought about GaN FET's so thanks or the tip, I will go look those up!
Can you better explain the long exposure image that has the pulse train of laser light? What exactly am I looking at?
*Nanosecond Pulsed Visible Laser Diode Driver: Pushing the Limits*
* *0:24** Powerful Visible Laser Diodes:* The video focuses on the NUBM44 and NUBM47 blue laser diodes, the most powerful visible single emitter diodes available, capable of 6 watts (and potentially up to 7-8 watts with reduced lifespan) of optical output power.
* *0:54** Pulsed vs. Continuous Wave:* While high power infrared pulsed laser diodes are common in LiDAR applications, visible laser diodes are traditionally operated in continuous wave mode with current limiting to prevent damage.
* *1:40** Scientific Literature:* Research indicates that pulsed operation of high power visible laser diodes can achieve peak powers of several watts, with pulse durations under 10 ns and currents of 50-60 amps. [From bussi7859's comment] Optimal performance requires minimizing the length of the current path.
* *4:33** Achieving High Peak Power:* The video highlights experiments demonstrating peak output powers approaching 30 watts from commercially available visible laser diodes using a simple avalanche pulsed driver circuit. The 445nm laser diodes performed exceptionally well, reaching over 200 nanojoules and 25 watts peak power.
* *6:12** Avalanche Driver Circuit:* The video explores using a Jim Williams avalanche driver, known for its nanosecond rise times, as a simpler and potentially more cost-effective alternative to commercial MOSFET-based pulsed laser drivers. [From laserdan's comment] GaN FETs might be an interesting alternative due to their short pulse capabilities and low inductance packaging.
* *6:50**-**7:00** Driver Design and PCB Layout:* The video showcases a custom-designed PCB for the avalanche driver, emphasizing a large ground plane and thick traces to minimize inductance, drawing inspiration from commercial designs.
* *11:28** High Voltage Power Supply:* The driver requires a high voltage power supply (around 188 volts in the demonstrated setup).
* *12:35** Nanosecond Pulses:* The driver successfully generates nanosecond pulses, visible as individual flashes when a piece of paper is moved rapidly through the beam path.
* *13:13** Power Measurement:* Measurements reveal a pulse width of 6.97 ns at a repetition rate of 5 kHz, with an average output power of 2.34 mW.
* *14:47** Peak Power Calculation:* Based on the measurements, the calculated peak power is an impressive 67.1 watts at a drive current exceeding 80 amps.
* *15:18** Laser Rangefinder Application:* The pulsed laser is used to demonstrate a simple rangefinder using a beam splitter, a corner cube retroreflector, and a fast photodiode. The time delay between the transmitted and received pulses accurately measures the distance to a target. [From bussi7859's comment] Rangefinders operating at 10kHz have demonstrated ranges up to 6km.
* *17:00** Fluorescent Dyes and Future Applications:* The pulsed laser's potential for fluorescence analysis and use in confocal microscopy setups are discussed. Future work includes experimenting with different avalanche transistors, laser diodes, and pulse forming network parameters to further optimize performance. [From MathewMoss-fp9ju's comment] There's interest in exploring quantum entanglement experiments with the pulsed laser and crystals like GAGG. [From arktik75's comment] There's also interest in exploring the possibility of pumping dye lasers with this pulsed laser.
I used gemini-1.5-pro-exp-0801 on rocketrecap dot com to summarize the transcript.
Cost (if I didn't use the free tier): $0.08
Input tokens: 21048
Output tokens: 756
Cool! Never though of using tools like this.
Thank you for the video! I assume you made sure the oscilloscope actually has 50 ohms input impedance for your measurements? As you may be aware, most of them have a high input impedance by default. Some can be switched into 50 ohms input mode.
Thanks! Yes, its a TDS784 and has a switchable impedance of 50ohms. For fast measurements like this, it's the only way. On my little 100MHz scope I use a 50 Ohm terminator right before the input.
Do you think it's possible to create a pulsing circuit that alternates between 2 voltages? This way you can cause the laser to output at potentially two different wavelengths, leading to super interesting uses like Raman spectroscopy
Do you know what kind of precision can you measure with the range finder?
I'm assuming the scope bandwidth will limit it, but assuming that's not the limiting factor, can you measure distance down to micrometers?
Measuring small distances woudl become technically challenging due to pulse width, risetime and bandwidth constraints. Centimetre precision will be good I imagine, but for microns a different approach is needed. You could try interferometry, or, if you want a super cheap, and very cool solution check out this video by Applied Science Laser diode self-mixing: Range-finding and sub-micron vibration measurement: th-cam.com/video/MUdro-6u2Zg/w-d-xo.html
Hahahaha I've been planning to do this for a long time now!! Glad you tried it!!
Your wiring gives about 1nH/mm of length, keep the lengths SHORT
For sure, it could do with improvement. Lead lengths on the Laser diode are too long as well. For what it is 2nS isn't too bad.
Could you elaborate what this means? (New to laser stuff)
Hi Less. What kind of trigger pulse you feed the board with? Sin, triangular or square?
Thank you.
This is a phenomenal result. Do you have any idea if the same technique can be used for LEDs? I feel like LEDs are getting to the power levels of being able to pump lasers like flashlamps are able to.
Ah, I just saw your other comments about driving LEDs. Still curious about any applications you can think of here!
Hi. I've built the same schematic as you did and I encountered two issues:
everything is working fine when using a common Led diode (red or white) till I reach over 180V. If I try to go above, the resistor R13, 1.2kohm and the transistor (I used 2N5192G) is getting hot for whatever reason.
seccon issue is that I tested the board with a laser diode (blue and red) and even at 200V it couldn't reach the treshold. I didn't damage the diodes though, they still work well in CW mode.
Do you have any suggestions regarding my situation? Thank you!
That's odd. The transistor shouldn't even get warm, the duty cycle with nanosecond pulses approaches zero! Sounds like the transistor in yours is always on, effectively sinking the HV supply. How are you driving its base? Does the drive signal go from 0 to 5v? (Some arbs have an offset output, i.e zero is not really zero)
Would this make a good laser cutter?
For the function generator 1us Pulse how much voltage is required, can we use an Arduino to generate 1us pulses
wow thats an insane price for such a small circuit, how do they justify that ? Great experiment, over 60 watts is really quite something.
$2 bucks for 5 boards is great. I think they panelize at the factory to reduces costs. I hand populated these though because of the experimental nature of it.
@@LesLaboratory those boards you made where very nice, ive started using jl for my little laser pointer boards. have you seen the nichia diode arrays, ive got one here rated at 128 watts, its scary, ive not experimented with it yet due to having damaged laser eyewear, ime not risking it untill my new set arrive from thorlabs
@@LesLaboratoryI think he was talking about the £340 boards.
Very nice!
Thanks!
Did you notice any appreciable differences between the ZTX415 and the 2N5192G? You made drivers with both, but left us hanging on what, if any, differences there are, and if one is better than the other in any way!
Yes, I probably should have mentioned it. The 2N5192 results in a higher threshold and is obsolete, though still available. The ZTX415 has a lower threshold and is currently available, but expensive at 10 bucks each. Both work just fine, and I am sure others will too.
Maybe try a GaN FET? The EPC boards you showed has full design files (I've made some using JLC, which was the EPC9126, and others), and they are very serious about minimizing inductance, you could probably get nicer edges and more peak power!
Yeah a couple of people have mentioned GaN FET's its definitely worth looking into.
I am wondering if this could be used to pump the fiber optic super continuum laser thus avoiding the more complicated N2 + dye laser set up?
The output is some two orders of magnitude lower that the SC setup. I have tried it and seen no broadening. With that all said, it might be possible to get it to do something interesting, perhaps if I use kilometers of fiber, and it is a work in progress
Excellent vid; thanks for sharing.
I didn’t know diodes could be driven quite that hard without COD etc.
The dye pumping sounds very exciting; I’m currently (slowly) building a very high power rig myself. Can you share what the specific dye is that’s suitable for the pulsed ~450nm pumping?
Another thought: you varied the pulse repetition rate, but did you get any effects / curves for a fixed rep rate but varying the length of the pulses?
Keep up the good work!
Back in 2007 I wrote about COD on Sam's Laser faq, and I was fairly convinced it was unavoidable for Laser diodes. There must be thermal effects involved in this process as well as sheer number of photons.
The Dye in the papers was DCM (4-(Dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)- 4H-pyran) and Lases in the red. It's about 160 bucks a gram where I am!
The length of the pulses is dependent on the PFN and is not variable. It is possible to lengthen the pulse by adding coax to the capacitor bank, but I do not have any text data for this yet.
@@LesLaboratory Thanks for the reply. Oh DCM(!!!), I have some of that in the garage - let me know if you'd like a small amount of it for your experiments. :)
(I'm UK based).
@@danielbriggs2 Thanks! I will more then likely take you up on that when I get round to a proper build. I have a very short path Cuvette, I just need to get round to making a mount and a resonator for it.
I find this all interesting, even if i don't understand everything, however i am left with one question .... At both +- 12:51 and +- 17:59 we can see the pulsed blue laser.... it looks like the light starts as kind of a vertical line and then evolves into a more circular shape.... What's happening here ? Is this maybe an effect by the camera ? Or is this something else that i don't understand ?
This is because of the awful beam shape. Laser diodes have different rates of divergence, that is not easily corrected with a single lens. They have a "slow axis" where divergence is lowest, and a "fast axis" where divergence is higher. When collimating for a far field spot, you end up with one axis being collimated, and the other being focused, hence the change in beam shape along its path.
This can be corrected with an external cylinder lens pair, as seen in the rangefinder experiment later in the video.
Very interesting video, thanks for sharing. I have two questions:
1. What was the energy per pulse that you got? The paper showed pulse energies in the mJ range...did you get anywhere near to that?
2. Do you think the pulsed laser diodes (905nm) can be overdriven in a similar fashion to get higher pulse energies? The datasheets all say up to 100W peak with 0.1% duty cycle and 100ns pulse widths, but i haven't seen anything about overdriving them. Pls share if you know about this.
Thanks!
Pulse energy is about 468nJ (calculated) so almost half a mJ, not too bad.
The IR diodes, (I assume) are already at their max output in pulsed mode. I suspect with cooling, you might get a couple of watts CW. It might be interesting to try though with very short pulses!
I was hoping you might offer some of those PCB's for sale!
@@laserhobbyist9751 interesting. Populated or unpopulated, and with or without laser diode?
At 12:42 to 12:56, you are definitely not imaging the 5 kHz rep rate train of nanosecond pulses since the length in air of a nanosecond light pulse is about 30 cm (1 foot), and the separation distance between pulses at 5 kHz rep rate is about 60 km.
I don't know what is producing the light dots you are imaging, but it cannot be the nanosecond long pulses at 5 kHz rep rate.
Below some conjectures as to what may be causing the light dots you are imaging, but these are only guesses.
It is difficult to tell exactly from the images, but the separation between the light dots looks like a few centimeters or so. Using 3 cm for the separation means the light dots are separated by 0.1 ns of light travel time over that separation distance and the corresponding repetition frequency is 10 GHz. If the separation is 6 cm, then the time between light dots is 0.2 ns and the corresponding repetition frequency is 5 GHz.
Thus, you may be imaging some high frequency (multi-GHz) intensity modulation of the luminescence, intensity modulation of the amplified spontaneous emission (ASE) or intra-pulse intensity modulation of the laser pulses emitted from the laser diode. This modulation may be due to some RFI or EMI picked up in the circuit, perhaps from a 5 GHz wifi signal, mid-band 5G cell phone signal, or some other multi-GHz electronic device you may have.
Another possibility is that back reflections of the emitted laser light from something may be causing external optical feedback into the diode laser cavity which may cause high frequency intensity modulation of the output.
It is well known that a major problem with semiconductor lasers, both FP and DFB types, is that these semiconductor lasers are highly sensitive to the laser light which re-enters the laser cavity after being reflected by an external reflector. External optical feedback of the laser light usually causes instability of operation of a laser diode and increased laser intensity fluctuations. A variety of optical elements, including lenses, fiber endfaces, etc. can be the source of unwanted optical feedback. Rayleigh backscatter from fiber can be another source of optical feedback. Packaged laser diodes may also receive optical feedback because of the external cavity formed between the laser diode chip and a transparent window of the package.
I was also confused about this photography. I tried to drive with 5KHz rep rate and found these pulses when I move the laser viewing card very fast in the path of the beam.
But I am unable to capture using camera, I don't know The ISO, shutter speed required to capture the effect.
Although a 5,6 ns pulse will travel around 5-6 feet, so it is some illusion I guess?
@@kumar-qb2pe As you move the laser card across the beam, if you move it fast enough, each laser pulse will hit on a different spot on the card separated from the spot that the previous pulse hit, so you'll see a series of spots across the laser card. The fluorescence from the laser card induced by the incident few nanosecond laser pulse will last on the order of microseconds to tens of milliseconds depending on the particular fluorescent material the laser card uses. The typical human eye integration time is on the order of 100 milliseconds, so your eye is integrating the fluorescence from all the spots over a time much longer than the fluorescence lasts.
To have a camera see the same thing as your eye when recording the laser card pulled through the laser beam, set the shutter speed to 1/10 th second which is 100 ms. On some cameras the closest setting may be 1/8 th second which is 125 ms, which is close enough.
The ISO required at that shutter speed will depend on the aperture size you have. In general you want to have an ISO as low as possible to minimize noise. You'll have to experiment with ISO and aperture. You'll want as large an aperture as possible, but the larger the aperture, the lower the depth of field.
I suggest starting with the largest available aperture that has just enough depth of field for the image you want to capture. Then start with the lowest available ISO and crank it up until you see the spots in the image.
Good Luck!
tech people rediscover strobe effect
How well does this work with LEDs? I want to make a high power flash for night time aerial photography with UAVs. I calculated that a 400J flashlamp flash is enough for my purposes but the cap bank for it would be heavy, making the use of a larger multi rotor necessary. So how much can white LEDs be overdriven in pulses? Edit: NVM A xenon flash lamp is still far better and cheaper than equivalent overdriven LED flash. A 400J cap bank discharging in 1ms means a power of about 400kW, if 20% of that is converted to light that's still 80kW of light vs a couple of kW for a very expensive and heavy array of over driven LEDs.
I have tried with LED's in place of the laser during testing, and it works ok, but I have not measured the brightness.
I don't think you would anywhere near the light output you are needing with any amount of LED's.
Xenon for those applications is exactly the right choice as you say. Somebody somewhere must be producing compact capacitors for flash applications though.
I'm looking for a cheap but good visible tem00 >1w laser for holography on photopolymer film, any ideas? Thx
Hi! Very interesting video and I learned a lot. I might have seen every single one of your uploads by now. I'm working on a project to generate two picosecond pulse-width beams from diodes whose peaks are apart on the order of picoseconds. Any suggestion on how to go about this would be extremely appreciated!!!
Honestly i'm not sure at this point. ~7 nanoseconds is where I am at just now. I might be possible to get into the picosecond range by re-spinning the board with lower inductance and capacitance. The ZTX415 is capable of picosecond rise times, but I am sure there are better alternatives. What is your application?
You could try driving the laser driver with an impulse generator based on diodes/TL. Check herotek and Macom.
But,they generator pulse >30ps. if you come across any technique to generate a few ps.Please let me know
@@LesLaboratory, I really appreciate your feedback. For ultra-fast molecular spectroscopy, you need two ultra-fast pump and probe pulses. I'm trying to do national lab stuff on a budget, and this video is the best starting point I've seen.
@@michaeljordan494 Thank you. I will definitly update you if I have any success.
Hi Les. What is the use of R1 to R6 resistor if they are 0 ohm? Why not connect the LD directly to the capacitors bank? Thanks.
I wanted to mess around with the values of those to see what effect it had on pulse width. Larger values, a few Ohms in total, lengthens the pulse, however the waveform looked particularly ugly, so I reverted back to 0 Ohm.
Many thanks Les!Then I can simplify the board by eliminating the resistors and also the resistors for oscilloscope since I don't have one anymore😊.
Is it possible to make Diy Powerful fiber laser using some Diode lasers and fiber?
Such Lasers already exist, however I have not attempted it from a home-brew standpoint.
@@LesLaboratory will you try?
what happens when you increase the current limiting resistor so that pulse duration is longer and peak current is lower?
The signal looks horrible. I was hoping it would lengthen the pulse a little. To lengthen the pulse and lowed the repak power, a short length of coax soldered on to the capacitor bank (As In AN47) will do it much more sensiblly. The resistors have been removed in the latest design, and I will be testing again soon.
Could this be used to make a supercontinuum laser, replacing the complicated nitrogen+dye pump? If I remember correctly, your dye lased around 426nm in the supercontinuum video. Sharp makes a 425nm diode that'll do 2W continuous.
It might, and that is one of the motivations of this project. For SC generation (like with N2 and Dye Lasers), you need a fairly wide bandwidth of Laser light to begin with, and the bandwidth from diodes is probably too narrow. The power levels are very low as well, about two orders of magnitude lower than the nitrogen Laser pumped dye laser. With all that said, it might, just might, be possible to use very narrow core fiber, and multiple pump diodes to get something to happen. So far I have not observed any broadening in fiber, but I have only tried a couple of Laser diodes.
So Les, I was just thinking how similar the circuit is that you are using is to the famous spark gap capacitor discharge of the DIY Nitrogen laser. And I wondered could you put a laser diode across the plates of a nitrogen laser? Does this make sense?
It's very similar. I think a diode placed across the rails of a Nitrogen Laser would be instantly destroyed though. There is a huge difference between ~200v and 20kV, before the diode even emitted light, I suspect several kV woudl appear across it. With all that said, I have been toying with the idea of building a Nitrogen Laser (for the most part) on a PCB. It would be so cool if they could be manufactured really small and inexpensively.
@@LesLaboratory what about depositing silver chemically on two sides of a piece of mica? That would be very tidy!
@@LesLaboratory could one close the spark gap so it triggered at 200v instead?
@@LesLaboratory I would love to see this!
Hello Les! :) Great job with your channel - many of the topics align with my research! I decided to comment on this video because a large part of my PhD thesis (defended about a year ago) focuses on advanced electronic drivers for generating nano- and subnanosecond laser pulses from semiconductor lasers. Unfortunately, it's written in Polish, so the text may be difficult for you to understand. :( Anyway, if you ever need assistance on the topic, or if you'd like to share ideas or brainstorm, feel free to reach out! :)
wow!
Can the function generator generate arbitrary predefined nanosecond-length signals?
@@codacoder no it's just an inexpensive one. It would be nice to have variable pulse width for this.
@@LesLaboratory Would also be cool to have it blink 1 billion digits of Pi within a second (using its binary expansion, or variable length pulses for each digit) :)
Hello!! This is a most educational video to improve solid state laser diode designs. I need help designing a depth measurement, hand held, range finder. I have been experimenting with 405 nm lasers to measure from the surface to the bottom of shallow ocean environments (under 100 meters). Could your systems be used for time-of-flight measurements? Your set up looks so good for improving my designs! Any suggestions?
@@Forensic1Man is could be adapted for time of flight as described in the video. 405nm has a shallow penetration depth in seawater though, you might be better off with a wavelength of around 480-490nm for underwater ranging.
@@LesLaboratory You are correct concerning 490 nm transmission. I miss-typed, being focused on your 405 nm work.
Beautiffull, Les....can a dye laser be pumped with this laser? Thanks.
According to academic papers, yes, but it seems to require a rather precise setup. Pumping a Dye laser with one of these is one of the motivations for this project.
I thought your aiming for that. I'l try to buid the same setup. I have the diode already. The PCB is two layers or more? How much will cost to order?
@@arktik75 two layers and $2 for five boards.
how can I order the boards Les?
I just wondering, why use a powerful diode laser for making pulses. Is it useful for data transmitting or it can be useful for engraving some objects??
Can you please explain how the photo was made. This isnt obviosly how you photographed distinct and uniformly spaced spots. It isnt obvious how the paper sees discrete spots also. isnt the light pulse itself is a foot wide ? All I can surmise is that the spacings must be a frame rate strobe of the camera off the card but the ending photo without a card is a mystery. What reflected the pulses at those unique spaced posistions??????
he said it's a long exposure shot
i think it's just a strobe effect on a suspiciously consistently-moved piece of paper
i read that title wrong at first
What's wrong with comically-sized beam splitters?
@@graealex for demos like this, nothing at all, but compared to the tiny laser diode, it's kinda fun 😀
@@LesLaboratory Love the videos and love the list of supporters grow!
@@graealex Thanks! Yes, it's really great, and I have the encouragement from supporters like you to thank for it!
A nanosecond is slightly under 30cm. at lightspeed. I suspect your thumbnail is off the mark. 😏 Computer pioneer Grace Hopper used to hand out nanosecond lengths of wire during her computing lectures. They were just under 30cm. long.
Ha, more than likely 😀 Just read up on that, this was followed by handing out grains of ground pepper for picoseconds!
@@LesLaboratory we both learned something! lol.
watched a youtuber make a powerful lazer gun aimed at fuel pile Boom
lite the neiybors BBQ 1 block away @O
This is what is ask GPT about "Can FWHM solve traditional TOF Lidar interference issue?"
ChatGPT said：
ChatGPT
FWHM (full width at half maximum) can help address interference issues in traditional time-of-flight (TOF) Lidar systems by providing a measure of the laser pulse width. A narrower FWHM can improve range resolution and reduce overlapping signals, which minimizes interference from multiple returns or background noise. However, it’s not a standalone solution; optimizing the system design, including detector sensitivity and signal processing algorithms, is also crucial for effectively mitigating interference
Oh but why not KiCAD :'(
@@cwbh10 I'm always busy and moving about for work. I can access easyeda from any computer and any browser, so it's a convenience thing.
What is a daod?
I'm guessing this is the auto closed caption attempt at diode!
I am into free energy. I already got the a small working prototype. To scale up and apply to actual products. Your expertise is the final missing piece. If you are interest in changing the world, please let me know.
999th like and this is the 111th comment too 😅👌
Sweet!
Great show. I think we are halfway through the process. Could you finish this process and make a video about FEEDBACK AND AMPLIFICATION of the laser beam? I would like links to buy things like lenses, mirrors, SESAM etc. Please provide such a version of the video and description because I have little knowledge on this subject. I found such a video but there is no explanation th-cam.com/video/aOTLhk6GuB4/w-d-xo.html
If it would not be a problem for you. Best regards