Modulation efficiency

[ElektroFreak]

445 is only an option if you don't mind sloppy beams. If you want good beam quality with minimal divergence, a TEM00 beam is necessary. To me, buying a PCAOM and running your DPSS lasers CW is the best option. The losses might not be much more than if you limited the DPSS output to a lower power and used optical feedback.

Honestly, beam quality is not overly important with laser shows, so if you're not concerned with it you could go with diodes for both blue and red, but you'll still have green DPSS to contend with, and they can be just as noisy as 473nm DPSS under modulation in many cases.

The good news is that there are a bunch of new direct injection diode designs that give a flat-top profile. It's not TEM00, but it is fairly even power distribution across a circular profile. TEM00 is more desirable to people with a desire for high-quality beams, but a flat top profile is MUCH cleaner than the rectangular multimode diode output many of us are used to.

TTL modulation can and in my experience does cause more power fluctuations than analog. Allowing the current to ramp up rather than suddenly pop on seems to smooth the fluctuations. They don't disappear by any means, but they do seem less noticeable. Galvo blanking or an AOM does do TTL nicely and gets rid of these issues, but in my case I just use direct modulation since my TTL systems are only one color.

[DZ]

So, if this is essentially a thermal issue, which part of the laser is having the problem? The pump? Could the pump be changing wavelength while it's heating up? If that's the case, then I wonder what would happen if you added a heater around the laser head, forcing it to remain "hot" while not modulating.

[ElektroFreak]

What we're discussing is caused by thermal lensing in the crystals. When the extremely powerful 808nm pump beam is focused into the crystals, it causes localized heating, which causes expansion, which in turn causes power fluctuations since the alignment of the cavity has changed minutely. When you modulate, this happens over and over with each pulse. When the laser is cold, the amount of thermal lensing that occurs with each pulse is large, but as it warms up it gradually reaches sort of a thermal equilibrium (and is kept there by a TEC) and the amount of lensing per pulse is reduced, but never eliminated. This is why each successive pulse in the scope shots above become larger, and the fluctuations remain no matter what the temperature. If there is a pause in the modulation, even for a second, it allows the crystals to cool slightly which in turn resets the whole process slightly. I hope that makes sense.. In essence, the crystal (and consequently the entire cavity) are oscillating physically as the laser is modulated. If the very shape of a laser cavity is physically vibrating, even at nano scale, expect fluctuations.

Your idea of using a heater is already employed in many high-end frequency-doubled lasers. Doubling crystals (such as KTP in the case of green DPSS lasers) prefer to be kept at a precise temperature for optimum stability. In the case of KTP, it likes to be warm, so a KTP oven is used to heat the crystal to improve stability.

[DZ]

Your idea of using a heater is already employed in many high-end frequency-doubled lasers. Doubling crystals (such as KTP in the case of green DPSS lasers) prefer to be kept at a precise temperature for optimum stability. In the case of KTP, it likes to be warm, so a KTP oven is used to heat the crystal to improve stability.

Ah, that's right, forgot about that. And in that case the heat is applied very specifically whereas the idea I came up with would be for the entire unit, in which case it's only effect might have been to reduce the life of the laser.

[danielbriggs]

Precise control of a TEC or oven can stabilize CW output no problem, but it cannot react anywhere nearly fast enough to cope with these sub-millisecond fluctuations. I don't know of a single option currently available that can.

*Thinking out loud*

Am I correct in assuming the instability is because the TEC doesn't know what will happen next as to regards the power output of the diode? Well what if the TEC(s) knew ahead of time, before the diode was modulated to that state?

For a pre-programmed fixed show, what if it were possible to somehow feed the same colour modulation signals to the TEC's at a time constant ahead of modulation of the diode?
No good for say use with LivePro or on the go stuff, but surely the TEC's would be able to "predict" what's about to hit it with more or less absolute precision? It may be possible to even apply this to a "live" output if the time offset required was small, by buffering the colour modulation for a millisecond or so?
Kind of analogous to a prefetcher if you will...

However I appreciate this would be hard to implement and take a lot of time and effort to research and perfect in practice, as well as some hefty hardware modifications of DAC and laser driver...

What do you reckon?

Again, just a thought experiment

[ElektroFreak]

@DZ: Depending on how hot you're talking, you would be reducing the life of the laser, yes.

Also, I really should add that you're partially correct about the diode influencing this process as well. As I said in a previous post, multimode diodes do not respond linearly to a change in current, but they are much better than DPSS as a whole. Since pump diodes are almost always multimode, their own nonlinearity is transferred to the crystals which adds to the overall potential for instability. As you can see, DPSS technology is quite hairy due to the complexity of the physics involved. That in a nutshell is why the really good cavities cost $$$$$, and it's also why I personally believe it is impossible to directly modulate a DPSS laser with 100% stability. Some good ingenuity will lead to improvements over time I have no doubt, but 100% stability under direct modulation is out of the question.


Regarding TECs, they respond very slowly to a change in current, so they could never keep up with the fluctuations. The only good method that exists presently is optical feedback, but I think it needs to be implemented such that it reduces the pump diode output when the green output spikes, rather than increase the pump output when the green output sags. Increasing the current rapidly is what is responsible for the instability in the first place, and decreasing the current is less stressful to the diode. Even then, however, there will still be fluctuations that cannot be controlled.

[DZ]

Just for clarification, earlier it was mentioned that the viasho had a photodiode integrated, that actually isn't true. Ben added it to the post as a way to improve the product.

[ElektroFreak]

Ahh.. that's a big clarification! Thanks for that..

[danielbriggs]

Regarding TECs, they respond very slowly to a change in current, so they could never keep up with the fluctuations.

Again, I'm just thinking of alternatives for the sake of it...
What if you were not to use TEC's then to heat the crystal. What if you were to use another IR diode (say pointing at the crystal 90 deg to surface) (and say of a different wavelength that does not interfere with the 532 lasing) then pump the crystal with the difference of the modulation input. i.e. if the 808nm diode is instantaneously modulated to 40%, the other IR diode would be modulated to 60%, i.e. the sum of the two diodes power would always equal effectively keeping the crystal at ideal temperature.
Thus the KTP always receives a constant "100%" pumping power, just the ratio of lasing and non-lasing wavelengths differ as per the modulation input?
On paper this looks like it would solve the modulation problem, while keeping the cavity at effective 100% CW operation?

[ElektroFreak]

I can't say I've ever heard that one before. Interesting idea.

My thoughts are that you'll probably run into a host of new problems implementing such a system. A big one would be that doubling crystals double more than one wavelength, so you'd need to find a wavelength that isn't doubled at all by the crystal, otherwise it would throw off your green output. 1064nm into KTP gives us 532nm. 1065 would give us 532.5nm, 1100 would give us 550nm and so on. KTP is only efficient above ~1000nm pump wavelength, and I'm not sure off the top of my head how high it will go, but you can see why you couldn't use just any diode. It would need to be carefully chosen depending on the type of doubling crystal.

And that's just the beginning..

Now I'm not saying it wouldn't work, but it would be a real bitch to figure out.