Ah, the smell of solder,

G

George Herold

Guest
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\>
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.
 
On Fri, 14 Feb 2020 10:23:37 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)
Click to expand...

R1 does add some phase shift, which is bad for loop stability.

Put a cap from the opamp output to the inverting input. 50 nF maybe.

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.
Click to expand...

You're probably heating the LED with too much current. Hit it with a
zot of freeze spray. That can be dramatic.
--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
 
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.
Click to expand...

You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.

You can make way sub-Poissonian current sources that way.

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
 
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Click to expand...
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.
You can make way sub-Poissonian current sources that way.
Click to expand...
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

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
Click to expand...
 
On 15/02/2020 01:25, George Herold wrote:
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

George H.
Click to expand...

Collector hi-z vs emitter lo-z ?

piglet
 
On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.
Click to expand...

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

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
 
On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?
Click to expand...

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

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
 
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs
Click to expand...

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?




--

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 2020-02-15 13:47, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 12:16:08 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs

You could return the last lowpass cap of the Vbe multiplier to the top
of the LED. Use one of Win's super low noise NPNs.
Click to expand...

The impedance of the LED is way too low for that to work equivalently,
though. It's way easier to do the simulated inductor thing.

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
 
On Sat, 15 Feb 2020 12:16:08 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs
Click to expand...

You could return the last lowpass cap of the Vbe multiplier to the top
of the LED. Use one of Win's super low noise NPNs.

Pity that the photons are still Poissonian. Somebody should fix that.







--

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 Saturday, February 15, 2020 at 2:00:38 AM UTC-5, piglet wrote:
On 15/02/2020 01:25, George Herold wrote:
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

George H.

Collector hi-z vs emitter lo-z ?

piglet
Click to expand...

Oh like just moving the diode load to the collector side of the
circuit I drew above. I guess I'll have to try it and see.

thx
George H.
 
On Saturday, February 15, 2020 at 7:22:14 AM UTC-5, Phil Hobbs wrote:
On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.
Click to expand...
OK, I like big R's in current supplies 'cause of lower
Johnson current noise.
5V and 0.1A = a hefty R... a volt should be enough for any man. :^)

George H.
BJT beta is the really low noise thing in analogue.

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
Click to expand...
 
On Saturday, February 15, 2020 at 12:03:37 PM UTC-5, jla...@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?
Click to expand...
I did that last time. (~2nV/ rtHz at ~10V/100mA)* So something
a bit different. (a cap mult somewhere in the output might be good.)

George H.
*could have been better... too much noise from upstream voltage regulator.
opamps and transistors are better (even a big tip31)

--

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"
Click to expand...
 
On Saturday, February 15, 2020 at 12:16:14 PM UTC-5, Phil Hobbs wrote:
On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs
Click to expand...
OK this:
"A stiff sub-Poissonian current source"

Can I use that phrase in advertising copy? :^)

GH.

--
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
Click to expand...
 
On Sat, 15 Feb 2020 12:16:08 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs
Click to expand...

OK, help me think this out.

If I put 1 volt across a 1K metal film resistor, I get 1 mA, or 6e15
electrons per second. If the electrons arrive randomly at the end of
the resistor, we'd see 18 pA RMS shot noise per root Hz bandwidth.

But we don't see that noise, because the electrons arrive more
regularly than randomly.

But if I split the resistor in half, two separate 2K resistors in
parallel, the electrons from each resistor are uncorrelated, so I
should get a bunch of current noise. But I don't. I think.

If electrons line up through some proximity effect, sort of like
Cooper pairs, using two resistors should increase noise. Or scribe a
cut down the middle. Or just use a wider resistor. The resistors that
I buy are already pretty wide compared to an electron.

Diode current has shot noise because the conductive element is wide
and short. So at some point distance prevents electrons from seeing
one another.

The electron correlation can't be a bulk effect of noise on the end
terminals, because the current noise is there if the circuit is low
impedance.




--

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 Saturday, February 15, 2020 at 1:11:43 PM UTC-8, jla...@highlandsniptechnology.com wrote:

If I put 1 volt across a 1K metal film resistor, I get 1 mA, or 6e15
electrons per second. If the electrons arrive randomly at the end of
the resistor, we'd see 18 pA RMS shot noise per root Hz bandwidth.

But we don't see that noise, because the electrons arrive more
regularly than randomly.
Click to expand...

Metals keep an electron in an undisclosable location; cathode
rays have shot noise, but metal conductors ought not (nor
MOS, under most circumstances).
 
George Herold wrote...
"A stiff sub-Poissonian current source"
Can I use that phrase in advertising copy? :^)
Click to expand...

The word stiff may have other connotations.


--
Thanks,
- Win
 
On 2020-02-15 16:11, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 12:16:08 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs

OK, help me think this out.

If I put 1 volt across a 1K metal film resistor, I get 1 mA, or 6e15
electrons per second. If the electrons arrive randomly at the end of
the resistor, we'd see 18 pA RMS shot noise per root Hz bandwidth.

But we don't see that noise, because the electrons arrive more
regularly than randomly.

But if I split the resistor in half, two separate 2K resistors in
parallel, the electrons from each resistor are uncorrelated, so I
should get a bunch of current noise. But I don't. I think.
Click to expand...

The correlations are enforced by electron scattering--the shot noise
suppression is a factor of the mean free path divided by the length of
the resistor. That works the same in both resistors.

If electrons line up through some proximity effect, sort of like
Cooper pairs, using two resistors should increase noise. Or scribe a
cut down the middle. Or just use a wider resistor. The resistors that
I buy are already pretty wide compared to an electron.

Diode current has shot noise because the conductive element is wide
and short. So at some point distance prevents electrons from seeing
one another.
Click to expand...

It's more the very low carrier density in the depletion zone, I think.

The electron correlation can't be a bulk effect of noise on the end
terminals, because the current noise is there if the circuit is low
impedance.
Click to expand...

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
 
On Sunday, February 16, 2020 at 12:26:59 PM UTC-5, Phil Hobbs wrote:
On 2020-02-15 16:11, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 12:16:08 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-15 12:03, jlarkin@highlandsniptechnology.com wrote:
On Sat, 15 Feb 2020 07:22:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-02-14 20:25, George Herold wrote:
On Friday, February 14, 2020 at 3:28:25 PM UTC-5, Phil Hobbs wrote:
On 2020-02-14 13:23, George Herold wrote:
.... and burnt plastic in the morning.

In which George makes his first home circuit,
has fun, but is ultimately disappointed in the numbers.
(with hat tip to A.A. Milne :^)


I was going to post of pic of circuit, home lab...
(but drop box keeps losing it.)
it's amazing what a mess of things I can make in a day.

I built a little current source to drive diodes.
___Vs +9-12V
|\ /
-Vin-|+\ |/
| >--R1(1k)--| (2n4401)
+-|-/ |\
| |/ |
| Diode load
| |
+------R2 10k------+
|
R3 10 ohms
|
GND
Vin 0-1V
worked fine with the diode, but I had a switch around the diode.
and with no (diode) load it oscillated. Till I shorted R1.
Is R1 a bad idea?

Anyway I kept R1 but put a bypass cap across it. (0.1uf)

Oh! and the poor 2n4401 got kinda hot at 50-100 mA.
I put in a helper (2nd 2n4401) and added 1 ohm in each
emitter arm to help balance current (not shown in circuit.)

I put a green led in first. And here was a weird thing.
I was monitoring the photocurrent* at the same time and the
light went up... (5 uA to 7 uA) at ~80 mA.. overdriving this
T- 3/4(?) plastic lens through hole. photo current decreased with more
drive current after that point.

OK and then the main attraction is the photo-current from
this panasonic laser diode. (Below threshold)
Well first having a laser diode with both 650 and 780 nm
co-inear is way cool... made things easy to line up.

But the numbers are not at all good. I measure 10-20 nA of photocurrent
(or less, alot depends on hwo close I get to the laser threshold...
more later) Which I guesstimate is only 1-2% of the photons will overlap
and 'correlate'. I'm going to start a new thread about this.

George H.


You'd be much better off driving the diode from the collector than the
emitter. If you change your circuit to use a PNP and put the sense
resistor in series with its emitter, life gets better. You'll probably
want to bypass the base to the positive supply, and add a capacitor from
the op amp's output to its inverting input.
Hmm OK. High side sensing? I've done that.. keeps the diode grounded,
which is nice. I haven't thought about a low noise drive yet.

You can make way sub-Poissonian current sources that way.
OK. Is there some reason one topology is quieter than the other?
Seems both would be about the same.

A resistor in the emitter reduces the shot noise current by a factor of
IR/26 mV due to negative feedback (emitter degeneration), until base
current shot noise becomes important. Five volts gets you 46 dB in a
high-beta transistor.

BJT beta is the really low noise thing in analogue.

Cheers

Phil Hobbs

For low noise LED drive, why not a low noise voltage source (aka Vbe
multiplier) and a resistor?

You can do that too, but it takes more headroom and relies much more on
the diode characteristics. A stiff sub-Poissonian current source can
force the diode current to be whatever it likes.

Cheers

Phil Hobbs

OK, help me think this out.

If I put 1 volt across a 1K metal film resistor, I get 1 mA, or 6e15
electrons per second. If the electrons arrive randomly at the end of
the resistor, we'd see 18 pA RMS shot noise per root Hz bandwidth.

But we don't see that noise, because the electrons arrive more
regularly than randomly.

But if I split the resistor in half, two separate 2K resistors in
parallel, the electrons from each resistor are uncorrelated, so I
should get a bunch of current noise. But I don't. I think.

The correlations are enforced by electron scattering--the shot noise
suppression is a factor of the mean free path divided by the length of
the resistor. That works the same in both resistors.


If electrons line up through some proximity effect, sort of like
Cooper pairs, using two resistors should increase noise. Or scribe a
cut down the middle. Or just use a wider resistor. The resistors that
I buy are already pretty wide compared to an electron.

Diode current has shot noise because the conductive element is wide
and short. So at some point distance prevents electrons from seeing
one another.

It's more the very low carrier density in the depletion zone, I think.
Click to expand...
I attribute shot noise in diodes to be because the electrons get excited
(thermally) one at a time over the barrier. (And the charge goes from one
side to the other. I like this image of a resistor as a body with
conducting plates on both ends. As a single electron scatters through the
resistor some fraction of an electron charge is 'moving' from one end plate
to the other.)_

George H.


The electron correlation can't be a bulk effect of noise on the end
terminals, because the current noise is there if the circuit is low
impedance.

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
Click to expand...
 
Winfield Hill <winfieldhill@yahoo.com> wrote in news:r2bt8h0b39
@drn.newsguy.com:

George Herold wrote...

"A stiff sub-Poissonian current source"
Can I use that phrase in advertising copy? :^)

The word stiff may have other connotations.
Click to expand...

So too could 'current source'.
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top