Noise reduction by use of a reference channel in pulsed lasers

[nedy...@googlemail.com]

Hi all,

I've recently come across a couple of Phil Hobb's papers on reaching the shot noise limit by using a reference channel and some clever electronics (http://www.opticsinfobase.org/opn/abstract.cfm?uri=OPN-2-4-17) which is a remarkable achievement and I'm amazed it's not more widely used. I guess a lot of physicists are scared of circuits (I certainly am)!

I'd like to apply this to reduce the effect of pulse-to-pulse variation of a QCL that I'm using (rep. rate around 1MHz, duty cycle 5% ish).

I have 2 specific questions:

1. My understanding is that the reference channel in this setup needs to be fed a "significant" portion of the beam. Is this 10%, 50% or 90% of the incoming beam (or does it depend?)

2. Would this setup even work with a pulse train? Would it be able to deal with the 0/0 calculations would pop up when the laser isn't on? If this setup isn't suitable, can you recommend one that would be?

I've looked at a simple voltage division circuit using something like the Analog Devices AD539, but I'm running into issues when the laser is off (0 divided by 0 again).

Thanks in advance,

Ned

[Phil Hobbs]

On 05/14/2013 12:56 PM, nedy...@googlemail.com wrote:
> Hi all,
>
> I've recently come across a couple of Phil Hobb's papers on reaching
> the shot noise limit by using a reference channel and some clever
> electronics
> (http://www.opticsinfobase.org/opn/abstract.cfm?uri=OPN-2-4-17) which
> is a remarkable achievement and I'm amazed it's not more widely used.
> I guess a lot of physicists are scared of circuits (I certainly am)!
>
> I'd like to apply this to reduce the effect of pulse-to-pulse
> variation of a QCL that I'm using (rep. rate around 1MHz, duty cycle
> 5% ish).
>
> I have 2 specific questions:
>
> 1. My understanding is that the reference channel in this setup needs
> to be fed a "significant" portion of the beam. Is this 10%, 50% or
> 90% of the incoming beam (or does it depend?)

The comparison beam has to be bigger than the signal beam, because the
circuit works by throwing away some of the comparison photocurrent until
the two are exactly equal.

>
> 2. Would this setup even work with a pulse train? Would it be able to
> deal with the 0/0 calculations would pop up when the laser isn't on?
> If this setup isn't suitable, can you recommend one that would be?

Noise cancellers work reasonably well with pulsed lasers because the
feedback loop bandwidth drops to zero between pulses, so the balance
tends to be preserved. With a 5% duty cycle, you're pretty safe from
the main worry, namely nonlinearity.

>
> I've looked at a simple voltage division circuit using something like
> the Analog Devices AD539, but I'm running into issues when the laser
> is off (0 divided by 0 again).

Analog dividers are noisy, and either slow or inaccurate or both,
depending on the model. Also the noise suppression depends on the loop
gain, so if your divider has a bandwidth of 1 MHz, at 10 kHz you only
get 40 dB noise suppression (that's electrical dB, i.e. 100 times in
voltage terms).

I do a fair number of custom noise canceller designs for folks, up to
100 MHz so far, so give me a shout if you need something faster than the
few-MHz version in the "without tears" paper,
http://electrooptical.net/www/canceller/WithoutTears.pdf .

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

[George Herold]

On May 14, 4:08 pm, Phil Hobbs
<pcdhSpamMeSensel...@electrooptical.net> wrote:

Great I'll read that again. I was at a party the other day talking
with my post-doc advisor. His students are looking for little signal
changes with a noisy laser and I suggested just a simple subtraction
circuit.

George H.

>
> Cheers
>
> Phil Hobbs
>
> --
> Dr Philip C D Hobbs
> Principal Consultant
> ElectroOptical Innovations LLC
> Optics, Electro-optics, Photonics, Analog Electronics
>
> 160 North State Road #203
> Briarcliff Manor NY 10510
>

> hobbs at electrooptical dot nethttp://electrooptical.net- Hide quoted text -
>
> - Show quoted text -