Modulation using an MTS110-a3-vis AOM.

[-bart-]

I recently obtained a MTS 110 a3 vis AOM from gogu. Satisfied as I was, I ordered a second one and decided to write a review on its modulation characteristics. This AOM is suitable to modulate a CW laser that has no or poor modulation means by itself. I successfully modulated a 632 HeNe, a 543 HeNe, a 488 single line argon and a 532 dpss. It is driven by a matching AA driver.

An AOM in general is not a magic device that will cure all your modulation issues in an instant.
There are a few issues you will have to deal with:
-This AOM has a theoretical static extinction rate of 2000, so for display purposes you will need an additional shutter.
-The linearity has a slight S-curve.
-The positional and angular stability are critical. When a dichro gets de-adjusted by 1mrad, you'll get a slight error in the color representation in the far field. 1mrad de-adjustment of an AOM can cost you an efficiency drop of 25% or worse.
-The AOM is polarisation sensitive, resulting in bad modulation efficiency when used with random polarised sources.

The optical window is 3mmx3mm, allowing to modulate pretty fat beams. Inevitably, when using beams this big, you can't try and find the sweet spot on the crystal.

The deflection angle is 90mRad at 532nm, so a minimum distance of 80 mm has to be maintained between the AOM and the galvos. More distance is needed if you plan to terminate the 0th order beam into a light trap of some sort.

To implement a AOM successfully you will need at least:
A braggmount to meticulously adjust its position.
A optical power meter to optimise the efficiency of the AOM.
Preferably a quick photocell to optimise the RF-driver settings for a proper Uin to Pout curve.

The driver consumes 300mA at 24VDC. So a simple 10VA SMPS is sufficient to power this AOM.
The input is not balanced and its ground is connected to chassis, so you'll have to make yourself a differential receiver to connect it to ILDA pin 5 and pin 18. (Alternatively you could use pin 5 and pin 25)
I got best results by setting the driver at maximum power, then attenuating the input level to obtain the power level needed for optimal modulation.


Modulation results:
1 division equals 5mW

square wave 10Hz
square wave 100Hz
square wave 1kHz
square wave 10kHz
square wave 50kHz
square wave 100kHz

square wave burst 8x 500µS 10Hz repetition rate
square wave burst 8x 500µS 100Hz repetition rate
square wave burst 8x 50µS 100Hz repetition rate
square wave burst 8x 5µS 100Hz repetition rate

Vertical axis contains calculated efficiency
Analog input linearity response at 1kHz typical ramp sample
Analog input linearity response at 1 kHz average of 128 ramps

Analog input sine modulation 1kHz

Equipment used: Scope Tektronix TDS 210, Function generator HP 8116A-001, Photocell BPW 34, Powermeter Ophir AN2, 20 mW 632nm HeNe

[tocket]

Neat! I actually have one of these on the way from gogu as well. I decided to get this one over the isomet because of the larger separation angle.

What does this static extinction number mean? Is it that 1/2000th (or less) of the input leaks into the first order even with no RF applied?

Do you have any tips on where to get good (and cheap) bragg mounts?

I don't think the non-linearity of the curve is that big a problem, especially not for me as I only need to get ~150 mW out of a 400 mW laser for color balance. By not going to full power you get better linearity. According to the datasheet you get 80% of the radiant power at 50% RF power and a linearity that's not much worse than that of direct injection lasers.

[FourDee]

Nice
Modulation, as it should be ...

[mixedgas]

Nice
Modulation, as it should be ...

Stay tuned,I'm about to release a modulation linearizer board for AOMs, allowing a TTL on/off AOM to be modulated for intensity control.

Steve

[-bart-]

Is it that 1/2000th (or less) of the input leaks into the first order even with no RF applied?

I think so. To measure this figure properly I would need a stable >1W laser and a lowpower pyro head.

There is a trade-off between low-light leakage and a good modulation curve. The trade-off-pot is called "biais" on the isomet driver

[sugeek]

Thank you for the review, If you can go over your adjustments of the driver I would appreciate it. In Steve's review he said not to touch the power setting, but to adjust the gain, so I was curious about that.

-Adam

[mixedgas]

lets rephrase that shall we? These drivers are NOT unconditionally stable.
Dont touch the bias and gain unless you have a RF power meter, you can end up hammering the crystal.

I'll make up a dummy load schematic with a calibrated diode for you in a day or two. reasoning is thus: The RF amp used in most drivers is a class AB, and you can blow the crap out of the input stage if you overdrive it. If you drive it too hard, you can whack the bond wires right off the crystal. A key phrase with the RF going into the XTAL is clean sine wave. Overdrive results in squaring of the waveform and "bouncing" of the wave inside the xtal.

Get yourself 4 noninductive 200 ohm 1/4 watt or 1/2watt resistors, carbon or metal film, but NOT wirewound, and a male SMA or BNC. A few 1n914s or 1n4148s and a 30-50V, .1 uf ceramic cap, and I'll show you how to make a dummy load and do measurements off of mine. Some things are best done right or not at all.

A scope probe less then 10:1 (10X low capacitance probe) across the RF output is high VSWR and can whack the driver. I speak from experience. I whacked one when I was a kid, and once had a contract to rebond the crystals on whacked units. The problem is the matching with the scope probe, if the xtal is not a purely 50 ohm load (and it aint) transfer measurements from one user to another are whacked at best.

read: http://www.techlib.com/files/detect.pdf

I will build one, measure the RF power levels and get back to you. Green thing in the sketch is the ceramic cap.

Steve

[DZ]

-bart-, awesome review! Just need to set your lasers for CW and stick an AOM in front of it for the absolute best modulation characteristics. Out of curiosity, what type of losses do you have on that AOM? Judging from the information you provided, it looks like 20mW in and roughly 17mW out? That's really good.

[-bart-]

Judging from the information you provided, it looks like 20mW in and roughly 17mW out? That's really good.

That's about right. Keep in mind that I used the most stable TEM00, polarised HeNe with the tight-ass beam I could find. And after half a day of fiddling I barely touched the claimed >85% mark, the typical 90% remains illusive.

If I didn't sold my coherent compass to Jem, I would have used that one, but I doubt that it would have yield significant better results.

[mixedgas]

That's about right. Keep in mind that I used the most stable TEM00, polarised HeNe with the tight-ass beam I could find. And after half a day of fiddling I barely touched the claimed >85% mark, the typical 90% remains illusive.

The key to the remaining 5% is use a multi axis mount. ie tip-tilt and brag angle.

Steve