HBT- wave noise

[George Herold]

So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise
I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

When I dial this all up in the lab I see about 10 nA
of photocurrent. (1%) The number one problem with this
measurement is knowing when the LD starts to lase.
(I'm thinking some real-time optical spectrum measurement
would help...)
I also dream of scraping the back reflective bit off the laser diode
There's a gold wire bond there and I need some good optics!

So I'm trying to think of some 'proof of concept' type experiment.
(I was first tempted to give up... measuring a ~1% increase in the noise
is not easy. )

But here's my 'crazy' idea. I measure the noise as a function of L
keeping the DC photo-current constant. (OK there are still lotsa issues**)
but keeping the current constant cancels a lot of uncertainties.
(Noise in current source driving led.
Electrical BW of measurement!)

George H.

*where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..
(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.

** Mechanical vibrations!! Can I keep all those at low frequency?
I'm not sure what (noise) layer lies below mechanical vibrations...
But that is a number one issue...
Oh, I've also seen 1/f noise from LED's (worse when driven hard.)
(Again a reason to filter out LF noise.)

Oh and the other huge flaw with this is that as I crank up the
current my light source turns into a laser.. with lots more noise
during the transition. So the first thing to do is measure noise
as function of current from the LD/ diode light source.
I bet that will suck... (putting on my eeyore hat.)

Noise everywhere, and I want to find and measure
just this little bit.

GH.














* where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..
(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.

 

[Phil Hobbs]

On 2020-02-14 14:45, George Herold wrote:

So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.

Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.

When I dial this all up in the lab I see about 10 nA
of photocurrent. (1%) The number one problem with this
measurement is knowing when the LD starts to lase.

Stimulated emission effects start showing up well below threshold,
unfortunately.

(I'm thinking some real-time optical spectrum measurement
would help...)
I also dream of scraping the back reflective bit off the laser diode
There's a gold wire bond there and I need some good optics!

You can make superluminescent diodes (SLDs) in the lab by gouging the
back facet of a laser diode with a scriber. Of course this destroys any
passivation back there, which doesn't help reliability. Also there are
still important feedback effects even from a damaged back facet.

So I'm trying to think of some 'proof of concept' type experiment.
(I was first tempted to give up... measuring a ~1% increase in the noise
is not easy. )

But here's my 'crazy' idea. I measure the noise as a function of L
keeping the DC photo-current constant. (OK there are still lotsa issues**)
but keeping the current constant cancels a lot of uncertainties.
(Noise in current source driving led.
Electrical BW of measurement!)

George H.

*where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..
(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.

** Mechanical vibrations!! Can I keep all those at low frequency?
I'm not sure what (noise) layer lies below mechanical vibrations...
But that is a number one issue...
Oh, I've also seen 1/f noise from LED's (worse when driven hard.)
(Again a reason to filter out LF noise.)

Oh and the other huge flaw with this is that as I crank up the
current my light source turns into a laser.. with lots more noise
during the transition. So the first thing to do is measure noise
as function of current from the LD/ diode light source.
I bet that will suck... (putting on my eeyore hat.)

Noise everywhere, and I want to find and measure
just this little bit.

GH.


* where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..

Not unreasonable for a LED.

(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.

My main issue with this scheme is that it assumes that a laser diode not
all that far below threshold is a thermal source to good accuracy, which
is IMO fairly far from obvious.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[George Herold]

On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:

On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.

Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.

Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when entendue is mentioned. )

When I dial this all up in the lab I see about 10 nA
of photocurrent. (1%) The number one problem with this
measurement is knowing when the LD starts to lase.

Stimulated emission effects start showing up well below threshold,
unfortunately.

Hmm.. Yeah that will be a problem. The wave noise will be a function
of the det. source geometry.. so maybe can be disentangled.

(I'm thinking some real-time optical spectrum measurement
would help...)
I also dream of scraping the back reflective bit off the laser diode
There's a gold wire bond there and I need some good optics!

You can make superluminescent diodes (SLDs) in the lab by gouging the
back facet of a laser diode with a scriber. Of course this destroys any
passivation back there, which doesn't help reliability. Also there are
still important feedback effects even from a damaged back facet.

OK this is what I was thinking of. These two wavelength laser diodes have
their active element exposed.
(I can break the gold wire bonds with my finger.. or finger nail.)
Any idea How much scratching I need? Will some hunk of metal work?
Or by scribber do you mean a diamond tipped thingie?

So I'm trying to think of some 'proof of concept' type experiment.
(I was first tempted to give up... measuring a ~1% increase in the noise
is not easy. )

But here's my 'crazy' idea. I measure the noise as a function of L
keeping the DC photo-current constant. (OK there are still lotsa issues**)
but keeping the current constant cancels a lot of uncertainties.
(Noise in current source driving led.
Electrical BW of measurement!)

George H.

*where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..
(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.

** Mechanical vibrations!! Can I keep all those at low frequency?
I'm not sure what (noise) layer lies below mechanical vibrations...
But that is a number one issue...
Oh, I've also seen 1/f noise from LED's (worse when driven hard.)
(Again a reason to filter out LF noise.)

Oh and the other huge flaw with this is that as I crank up the
current my light source turns into a laser.. with lots more noise
during the transition. So the first thing to do is measure noise
as function of current from the LD/ diode light source.
I bet that will suck... (putting on my eeyore hat.)

Noise everywhere, and I want to find and measure
just this little bit.

GH.


* where opt_BW (in Hz) is the optical bandwidth
of the source. (I'm not sure what the BW is. 10 nm is a guess..

Not unreasonable for a LED.

(lamda is 780 nm ~ 1um, freq=c/lamda ~3x10^14, so 1% of that is
3E12 Hz.. let's say 5E12 'cause of my fudges times the charge of the
electron is (1.6X10^-19) 8 x 10^-7 amps. Which I've been rounding up
to 1 uA.


My main issue with this scheme is that it assumes that a laser diode not
all that far below threshold is a thermal source to good accuracy, which
is IMO fairly far from obvious.

Right. The next thing to do is to measure the laser noise as a function of
current. .. I guess that gets kinda tricky if the 'stimulated emmision
noise' at 1% (shot noise) level is bad.... OK at some point measuring
this type of photon correlation (stimulated emission) would also be good.

As always thanks for the thoughts/ ideas.

George H.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 2020-02-14 20:59, George Herold wrote:

On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.

When I dial this all up in the lab I see about 10 nA
of photocurrent. (1%) The number one problem with this
measurement is knowing when the LD starts to lase.

Stimulated emission effects start showing up well below threshold,
unfortunately.
Hmm.. Yeah that will be a problem. The wave noise will be a function
of the det. source geometry.. so maybe can be disentangled.

(I'm thinking some real-time optical spectrum measurement
would help...)
I also dream of scraping the back reflective bit off the laser diode
There's a gold wire bond there and I need some good optics!

You can make superluminescent diodes (SLDs) in the lab by gouging the
back facet of a laser diode with a scriber. Of course this destroys any
passivation back there, which doesn't help reliability. Also there are
still important feedback effects even from a damaged back facet.
OK this is what I was thinking of. These two wavelength laser diodes have
their active element exposed.
(I can break the gold wire bonds with my finger.. or finger nail.)
Any idea How much scratching I need? Will some hunk of metal work?
Or by scribber do you mean a diamond tipped thingie?

I'd certainly use a diamond or carbide scriber to reduce the amount of
force required.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[George Herold]

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:

On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.

Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)

When I dial this all up in the lab I see about 10 nA
of photocurrent. (1%) The number one problem with this
measurement is knowing when the LD starts to lase.

Stimulated emission effects start showing up well below threshold,
unfortunately.
Hmm.. Yeah that will be a problem. The wave noise will be a function
of the det. source geometry.. so maybe can be disentangled.

(I'm thinking some real-time optical spectrum measurement
would help...)
I also dream of scraping the back reflective bit off the laser diode
There's a gold wire bond there and I need some good optics!

You can make superluminescent diodes (SLDs) in the lab by gouging the
back facet of a laser diode with a scriber. Of course this destroys any
passivation back there, which doesn't help reliability. Also there are
still important feedback effects even from a damaged back facet.
OK this is what I was thinking of. These two wavelength laser diodes have
their active element exposed.
(I can break the gold wire bonds with my finger.. or finger nail.)
Any idea How much scratching I need? Will some hunk of metal work?
Or by scribber do you mean a diamond tipped thingie?

I'd certainly use a diamond or carbide scriber to reduce the amount of
force required.

Hmm my diamond scribe is the size of a bic pen.
Sand paper on the end of some tweezers/ toothpick?

George H.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[George Herold]

On Sunday, February 16, 2020 at 7:52:10 PM UTC-5, jla...@highlandsniptechnology.com wrote:

On Sun, 16 Feb 2020 16:36:00 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.
Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)


ROSAs, fiber-coupled detectors with fast TIAs, are dirt cheap. They
usually have AGC, but at low light levels they'll crank themselves to
max gain.

What's a ROSA? As usual I'm most likely searching for the wrong thing.

Any 'fiber coupled photodiodes' lead to spendy things ~$100
The TT OPF470 is also sold in what looks like a bnc
connector in front for ~$20 each from mouser.
(OPF472)?
https://www.mouser.com/datasheet/2/414/OPF422-1530602.pdf

--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

 

On Sun, 16 Feb 2020 16:36:00 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.
Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)

ROSAs, fiber-coupled detectors with fast TIAs, are dirt cheap. They
usually have AGC, but at low light levels they'll crank themselves to
max gain.



--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

 

[Michael Terrell]

On Sunday, February 16, 2020 at 7:36:05 PM UTC-5, George Herold

Hmm my diamond scribe is the size of a Bic pen.
Sand paper on the end of some tweezers/ toothpick?

A grain of sand, glued to the end of a small dowel rod? It's left to the student to align it, point down. ;-)

(Hint: Make a depression in the center with a small knife point, or cuticle scissors. then put a drop of glue in the hole. drop the grain of sand into it, and use a thin needle or an Exacto knife point to turn it the right way.)

 

On Sun, 16 Feb 2020 17:34:41 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

On Sunday, February 16, 2020 at 7:52:10 PM UTC-5, jla...@highlandsniptechnology.com wrote:
On Sun, 16 Feb 2020 16:36:00 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.
Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)


ROSAs, fiber-coupled detectors with fast TIAs, are dirt cheap. They
usually have AGC, but at low light levels they'll crank themselves to
max gain.

What's a ROSA? As usual I'm most likely searching for the wrong thing.

Any 'fiber coupled photodiodes' lead to spendy things ~$100
The TT OPF470 is also sold in what looks like a bnc
connector in front for ~$20 each from mouser.
(OPF472)?
https://www.mouser.com/datasheet/2/414/OPF422-1530602.pdf

A 10 gbps SFP module typically has a ROSA, a TOSA, and an ASIC inside,
for around $20 or so.

Google rosa tosa





--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

 

[Phil Hobbs]

On 2020-02-16 21:57, jlarkin@highlandsniptechnology.com wrote:

On Sun, 16 Feb 2020 17:34:41 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Sunday, February 16, 2020 at 7:52:10 PM UTC-5, jla...@highlandsniptechnology.com wrote:
On Sun, 16 Feb 2020 16:36:00 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.
Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)


ROSAs, fiber-coupled detectors with fast TIAs, are dirt cheap. They
usually have AGC, but at low light levels they'll crank themselves to
max gain.

What's a ROSA? As usual I'm most likely searching for the wrong thing.

Any 'fiber coupled photodiodes' lead to spendy things ~$100
The TT OPF470 is also sold in what looks like a bnc
connector in front for ~$20 each from mouser.
(OPF472)?
https://www.mouser.com/datasheet/2/414/OPF422-1530602.pdf

A 10 gbps SFP module typically has a ROSA, a TOSA, and an ASIC inside,
for around $20 or so.

Google rosa tosa

And you can get the 1.25 Gb/s ones for a buck or so on eBay.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[John Larkin]

On Mon, 17 Feb 2020 15:24:50 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-16 21:57, jlarkin@highlandsniptechnology.com wrote:
On Sun, 16 Feb 2020 17:34:41 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Sunday, February 16, 2020 at 7:52:10 PM UTC-5, jla...@highlandsniptechnology.com wrote:
On Sun, 16 Feb 2020 16:36:00 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

On Saturday, February 15, 2020 at 11:21:22 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:59, George Herold wrote:
On Friday, February 14, 2020 at 3:23:23 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 14:45, George Herold wrote:
So to see this correlation/ overlap between photons, I need to meet
two criteria.
1.) photon flux, to have intensity (wave) noise equal to shot noise

You're shooting for a fairly dramatic effect there. HB's Narrabri
observatory worked down at the 1% level IIRC. Of course you want
students to actually see it, but getting an effect that large is asking
a lot. Shot noise is nearly always the ultimate limit in optical
measurements.
Oh the equivalent to shot noise is just a measure. I'm getting
a numbers of a few percent... which looks hard.

I need I_dc = e * opt_BW,* which I guesstimate to be about 1 uA.
2.) spacial 'uncertainty' I need the quantity,

D*d/(lamda*L) <=~1 (is less than or about equal to one.)

Source needs to be unresolved at the detector, check.


Where D is detector diameter, d is source diameter,
L is distance from source to detector and lamda is the
wavelength of the light.

So I don't know the optical BW nor the source size.
I can measure the Laser diode.. ~0.015" ~350 um
(the source size might be a lot less than this?)

A single transverse mode has an etendue of lambda**2/2 in each
polarization. With lenses you can have any tradeoff of solid angle vs
area you like.
Yeah, but if I stick a lens in front of my detector I'm changing the
detector area.. so I don't really gain anything.. (maybe that is what you
are saying. I get confused when etendue is mentioned. )

Etendue is the product of area and projected solid angle. It's related
to the thermodynamic idea of phase space volume.
Cue a fav. xkcd
https://what-if.xkcd.com/145/

It's not that I don't understand the concept...

Anyway for the HBT thing seems the smaller the PD the better.

There's a littel one from vishay VEMD1060
https://datasheet.octopart.com/VEMD1060X01-Vishay-datasheet-63587238.pdf

And this one for fibers looks interesting. OPF470
https://www.mouser.com/datasheet/2/414/OPF470-1530569.pdf
(more expensive.. but cheaper that other fiber coupled
detectors!)


ROSAs, fiber-coupled detectors with fast TIAs, are dirt cheap. They
usually have AGC, but at low light levels they'll crank themselves to
max gain.

What's a ROSA? As usual I'm most likely searching for the wrong thing.

Any 'fiber coupled photodiodes' lead to spendy things ~$100
The TT OPF470 is also sold in what looks like a bnc
connector in front for ~$20 each from mouser.
(OPF472)?
https://www.mouser.com/datasheet/2/414/OPF422-1530602.pdf

A 10 gbps SFP module typically has a ROSA, a TOSA, and an ASIC inside,
for around $20 or so.

Google rosa tosa

And you can get the 1.25 Gb/s ones for a buck or so on eBay.

Cheers

Phil Hobbs

There must be some neat physics experiment one could do with a couple
SFP modules. Or even one.



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[David Nadlinger]

On 17.02.20 10:39 pm, John Larkin wrote:

There must be some neat physics experiment one could do with a couple
SFP modules. Or even one.

SFPs can make pretty decent detectors for locking two lasers together at
some GHz offset. Of course, that's mostly a prerequisite for other
experiments rather than one by itself.

— David

 

On Tue, 18 Feb 2020 01:42:45 +0000, David Nadlinger
<david@klickverbot.at> wrote:

On 17.02.20 10:39 pm, John Larkin wrote:
There must be some neat physics experiment one could do with a couple
SFP modules. Or even one.

SFPs can make pretty decent detectors for locking two lasers together at
some GHz offset. Of course, that's mostly a prerequisite for other
experiments rather than one by itself.

— David

One could resolve the prop delay of a long piece of fiber to ps or fs,
but I don't know how useful that could be. Maybe that experiment where
acceleration drags light. Or stress. Or something.



--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

 

[George Herold]

On Monday, February 17, 2020 at 8:42:49 PM UTC-5, David Nadlinger wrote:

On 17.02.20 10:39 pm, John Larkin wrote:
There must be some neat physics experiment one could do with a couple
SFP modules. Or even one.

SFPs can make pretty decent detectors for locking two lasers together at
some GHz offset. Of course, that's mostly a prerequisite for other
experiments rather than one by itself.

— David

Yeah one problem with 'neat physics/ engineering thing'
Is that I always want some good experiment to hang on the end
of it. I did this little demo of locking our diode laser
to a spectral line.. but really only useful if you have
something 'good' to do with it once it's locked. So it was
just. 'gee whiz' it's locked.

George H.

 

[George Herold]

On Tuesday, February 18, 2020 at 9:33:15 AM UTC-5, George Herold wrote:

On Monday, February 17, 2020 at 8:42:49 PM UTC-5, David Nadlinger wrote:
On 17.02.20 10:39 pm, John Larkin wrote:
There must be some neat physics experiment one could do with a couple
SFP modules. Or even one.

SFPs can make pretty decent detectors for locking two lasers together at
some GHz offset. Of course, that's mostly a prerequisite for other
experiments rather than one by itself.

— David

Yeah one problem with 'neat physics/ engineering thing'
Is that I always want some good experiment to hang on the end
of it. I did this little demo of locking our diode laser
to a spectral line.. but really only useful if you have
something 'good' to do with it once it's locked. So it was
just. 'gee whiz' it's locked.

George H.

"The other way to HBT noise"
So I've also built Rb lamps*. Which you can filter
down to one of the ~10 GHz wide 'D' lines at 780 or 795 nm.
Lamda ~0.8 um (I just use 1^-3 mm and remember to make L a bit longer)

Having measured ~100 with a bulb/source diameter of ~1/4"
(6mm") and the same detector area filtered to one line.
(D = d = 6 mm)
With a 6" (L = 150 mm)
A typical photocurrent was 2 uA, which (assuming I did the math
right.) gave me ~30pA of photo current at the needed area/ distance.
(d*D)/(L*lamda) ~1 (36 mm^2/0.15 mm^2 +~ 240) (2uA/ 240^2 = 34.7pA)
Oh and e*10 GHz is 1.6 nA... so maybe a several percent signal.

That is out of the range of ordinary photodiodes IMO.
And into PMT's or avalanche PD's.

George H.
*one problem with lamps is they can be noisy..
a lot can depend on how much Rb is put in 'em.
(they also drift..)