Why dont they make a "Light Emitting TRIODE"

The LED ("Light Emitting DIODE") is common place now. In fact it seems
like they went from being simple "power on" indicator lamps to the
lighting of the future for darn near everything.

But what I dont understand is why they have never made a "Light Emitting
TRIODE"? (LET). It seems there would be many advantages!

 

[Tom Biasi]

On 1/24/2015 4:13 PM, electron206@online.com wrote:

The LED ("Light Emitting DIODE") is common place now. In fact it seems
like they went from being simple "power on" indicator lamps to the
lighting of the future for darn near everything.

But what I dont understand is why they have never made a "Light Emitting
TRIODE"? (LET). It seems there would be many advantages!

Can you list some of the advantages they might have?

 

[Jim Thompson]

On Sat, 24 Jan 2015 16:23:06 -0500, Tom Biasi <tombiasi@optonline.net>
wrote:

On 1/24/2015 4:13 PM, electron206@online.com wrote:
The LED ("Light Emitting DIODE") is common place now. In fact it seems
like they went from being simple "power on" indicator lamps to the
lighting of the future for darn near everything.

But what I dont understand is why they have never made a "Light Emitting
TRIODE"? (LET). It seems there would be many advantages!

Can you list some of the advantages they might have?

You could sell them to the crowd that believes in global warming ;-)

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[whit3rd]

On Saturday, January 24, 2015 at 1:23:08 PM UTC-8, Tom Biasi wrote:

On 1/24/2015 4:13 PM, electron206@online.com wrote:
The LED ("Light Emitting DIODE") is common...
But what I dont understand is why they have never made a "Light Emitting
TRIODE"? (LET). It seems there would be many advantages!

Can you list some of the advantages they might have?

One immediately comes to mind: power, ground, and modulation input
are the three wires required for a laser transmitter. I generally see
these in SFP form, with receiver wires as well... the seventeen other
wires kinda obscure the triode nature of the beast.

 

[Phil Allison]

Michael Black wrote:

It's a different case with an bipolar transistor that is light activated,
I can't remember the name suddenly.

** Phototransistor.

Often photo Darlingtons are used in opto-couplers.



.... Phil

 

[Michael Black]

On Sat, 24 Jan 2015, electron206@online.com wrote:

The LED ("Light Emitting DIODE") is common place now. In fact it seems
like they went from being simple "power on" indicator lamps to the
lighting of the future for darn near everything.

But what I dont understand is why they have never made a "Light Emitting
TRIODE"? (LET). It seems there would be many advantages!


Because what's the use?

If nothing else, it wouldn't be a triode in this day and age, while a
"diode" did morph from describing a two element tube to a two element
semiconductor device, other terms have been used to describe the terminal
solid state devices.

A diode is the way the device generates light. A "triode" implies some
level of control, and that can be had with a separate device. It's not
like in the old days where a magic eye tube was put in a glass envelope
along with a triode, back then if you needed a triode to control the magic
eye, leaving it separate would mean another tube, which were expensive and
bulky, and would need another tube socket.

There may be LEDs with seemingly built in controlling devices, but they
would then be on the level of an integrated circuit, a transistor feeding
an LED, not some device that combines a bipolar transistor and LED.

Some optoisolators give the same function, I imagine. If the put a
transistor in to drive the LED, then that's the same effect, you never
notice if it's one device or not.

It's a different case with an bipolar transistor that is light activated,
I can't remember the name suddenly. There, the base is light sensitive,
and since they figured out some functions for that (they needed to detect
light, and needed some amplification) then they exist. Indeed, they are
the receivers in those optoisolators.Sometimes the base is even brought
out, so you can adjust the bias on the base, along with it being light
controlled.

Michael

 

>"Can you list some of the advantages they might have? "

That's what I was thinking. The only thing I can think of is to lower the load on whatever drives thee LED. They already pull little enough that microprocessors can usually drive the anyway but there is always the wattage race. You could have power and ground busses where the LEDS are mounted and the processor coud drive the base which would presumably need less current. The base resistors could be built in like a digital transistor and it also xould be confihued as a current regualtor eliminating the limiting resistor and making the device tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of them tried the idea but nobody bought them. I can only think of that one advantage to it, and that would not be worth the extra cost IMO. Also, if the transistor is used as a current regulator then you have dissipation. Unless you get wild and put a switcher in there. Then you are talking a coil csuae you can't pulse the LED with overcurrent and expect it to last. The cost might end up being a dollar insrtead of the few cents a regular LED and resistor cost.

 

[Phil Hobbs]

On 1/26/2015 10:41 AM, jurb6006@gmail.com wrote:

"Can you list some of the advantages they might have?"

That's what I was thinking. The only thing I can think of is to lower
the load on whatever drives thee LED. They already pull little enough
that microprocessors can usually drive the anyway but there is always
the wattage race. You could have power and ground busses where the
LEDS are mounted and the processor coud drive the base which would
presumably need less current. The base resistors could be built in
like a digital transistor and it also xould be confihued as a current
regualtor eliminating the limiting resistor and making the device
tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of
them tried the idea but nobody bought them. I can only think of that
one advantage to it, and that would not be worth the extra cost IMO.
Also, if the transistor is used as a current regulator then you have
dissipation. Unless you get wild and put a switcher in there. Then
you are talking a coil csuae you can't pulse the LED with overcurrent
and expect it to last. The cost might end up being a dollar insrtead
of the few cents a regular LED and resistor cost.

Making bipolar transistors out of III-V semiconductors (e.g. GaAs, InAs,
AlGaAs, InGaAs, InP, GaN, etc) is difficult, and they don't work very
well. Plus the extra constraints would make the LET much harder to build.

The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently. If you just use a regular GaAs
PN diode, all the light gets absorbed by the top semiconductor layer.

Then there's the 95% or so that you lose to total internal reflection
(taking into account the Fresnel reflection of the light that does make
it out), and the 50% you lose due to half the light starting out going
the wrong way....

Modern LEDs are heterojunctions, with transparent semiconductor layers
top and bottom to reduce absorption, and controlled roughening of the
surfaces to let the light rattle around till it escapes.

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

 

[Jim Thompson]

On Mon, 26 Jan 2015 15:14:06 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 1/26/2015 10:41 AM, jurb6006@gmail.com wrote:
"Can you list some of the advantages they might have?"

That's what I was thinking. The only thing I can think of is to lower
the load on whatever drives thee LED. They already pull little enough
that microprocessors can usually drive the anyway but there is always
the wattage race. You could have power and ground busses where the
LEDS are mounted and the processor coud drive the base which would
presumably need less current. The base resistors could be built in
like a digital transistor and it also xould be confihued as a current
regualtor eliminating the limiting resistor and making the device
tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of
them tried the idea but nobody bought them. I can only think of that
one advantage to it, and that would not be worth the extra cost IMO.
Also, if the transistor is used as a current regulator then you have
dissipation. Unless you get wild and put a switcher in there. Then
you are talking a coil csuae you can't pulse the LED with overcurrent
and expect it to last. The cost might end up being a dollar insrtead
of the few cents a regular LED and resistor cost.


Making bipolar transistors out of III-V semiconductors (e.g. GaAs, InAs,
AlGaAs, InGaAs, InP, GaN, etc) is difficult, and they don't work very
well. Plus the extra constraints would make the LET much harder to build.

The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently. If you just use a regular GaAs
PN diode, all the light gets absorbed by the top semiconductor layer.

Then there's the 95% or so that you lose to total internal reflection
(taking into account the Fresnel reflection of the light that does make
it out), and the 50% you lose due to half the light starting out going
the wrong way....

Modern LEDs are heterojunctions, with transparent semiconductor layers
top and bottom to reduce absorption, and controlled roughening of the
surfaces to let the light rattle around till it escapes.

Cheers

Phil Hobbs

SiGe transistors are super-duper ;-)

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Phil Hobbs]

On 1/26/2015 3:40 PM, Jim Thompson wrote:

On Mon, 26 Jan 2015 15:14:06 -0500, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 1/26/2015 10:41 AM, jurb6006@gmail.com wrote:
"Can you list some of the advantages they might have?"

That's what I was thinking. The only thing I can think of is to lower
the load on whatever drives thee LED. They already pull little enough
that microprocessors can usually drive the anyway but there is always
the wattage race. You could have power and ground busses where the
LEDS are mounted and the processor coud drive the base which would
presumably need less current. The base resistors could be built in
like a digital transistor and it also xould be confihued as a current
regualtor eliminating the limiting resistor and making the device
tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of
them tried the idea but nobody bought them. I can only think of that
one advantage to it, and that would not be worth the extra cost IMO.
Also, if the transistor is used as a current regulator then you have
dissipation. Unless you get wild and put a switcher in there. Then
you are talking a coil csuae you can't pulse the LED with overcurrent
and expect it to last. The cost might end up being a dollar insrtead
of the few cents a regular LED and resistor cost.


Making bipolar transistors out of III-V semiconductors (e.g. GaAs, InAs,
AlGaAs, InGaAs, InP, GaN, etc) is difficult, and they don't work very
well. Plus the extra constraints would make the LET much harder to build.

The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently. If you just use a regular GaAs
PN diode, all the light gets absorbed by the top semiconductor layer.

Then there's the 95% or so that you lose to total internal reflection
(taking into account the Fresnel reflection of the light that does make
it out), and the 50% you lose due to half the light starting out going
the wrong way....

Modern LEDs are heterojunctions, with transparent semiconductor layers
top and bottom to reduce absorption, and controlled roughening of the
surfaces to let the light rattle around till it escapes.

Cheers

Phil Hobbs

SiGe transistors are super-duper ;-)

...Jim Thompson

Yup, the BFP640 is a fave. But the hole mobility of III-Vs is so
horrible, and the minority carrier lifetime so short, that you can't get
any beta no matter what you do. Those pesky direct bandgaps again.

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

 

[Maynard A. Philbrook Jr.]

In article <54C6A00E.6010708@electrooptical.net>,
pcdhSpamMeSenseless@electrooptical.net says...

On 1/26/2015 10:41 AM, jurb6006@gmail.com wrote:
"Can you list some of the advantages they might have?"

That's what I was thinking. The only thing I can think of is to lower
the load on whatever drives thee LED. They already pull little enough
that microprocessors can usually drive the anyway but there is always
the wattage race. You could have power and ground busses where the
LEDS are mounted and the processor coud drive the base which would
presumably need less current. The base resistors could be built in
like a digital transistor and it also xould be confihued as a current
regualtor eliminating the limiting resistor and making the device
tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of
them tried the idea but nobody bought them. I can only think of that
one advantage to it, and that would not be worth the extra cost IMO.
Also, if the transistor is used as a current regulator then you have
dissipation. Unless you get wild and put a switcher in there. Then
you are talking a coil csuae you can't pulse the LED with overcurrent
and expect it to last. The cost might end up being a dollar insrtead
of the few cents a regular LED and resistor cost.


Making bipolar transistors out of III-V semiconductors (e.g. GaAs, InAs,
AlGaAs, InGaAs, InP, GaN, etc) is difficult, and they don't work very
well. Plus the extra constraints would make the LET much harder to build.

The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently. If you just use a regular GaAs
PN diode, all the light gets absorbed by the top semiconductor layer.

Then there's the 95% or so that you lose to total internal reflection
(taking into account the Fresnel reflection of the light that does make
it out), and the 50% you lose due to half the light starting out going
the wrong way....

Modern LEDs are heterojunctions, with transparent semiconductor layers
top and bottom to reduce absorption, and controlled roughening of the
surfaces to let the light rattle around till it escapes.

Cheers

Phil Hobbs

Rattle around, that's so sophisticated of you, Phil!

Jamie

 

"The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently."

IIRC that was the holdup with the blue LEDs in the beginning. the rest were two cents and the blue ones were a buck.

One thing I found out is that the white ones mostly have a phosphor. That is probably what made them suitable for use in LCD TVs. If you had red, green and blue LEDs the output was too spikey for good colorimetry.

That was probably a bad move for TV engineers because dropping the TV would bereak the CCFLs in there sometimes. I guess they'll just have to make the PC boards thinner and brittle.

 

[Phil Hobbs]

On 1/26/2015 9:53 PM, Maynard A. Philbrook Jr. wrote:

In article <54C6A00E.6010708@electrooptical.net>,
pcdhSpamMeSenseless@electrooptical.net says...

On 1/26/2015 10:41 AM, jurb6006@gmail.com wrote:
"Can you list some of the advantages they might have?"

That's what I was thinking. The only thing I can think of is to lower
the load on whatever drives thee LED. They already pull little enough
that microprocessors can usually drive the anyway but there is always
the wattage race. You could have power and ground busses where the
LEDS are mounted and the processor coud drive the base which would
presumably need less current. The base resistors could be built in
like a digital transistor and it also xould be confihued as a current
regualtor eliminating the limiting resistor and making the device
tolerant of more of a range of voltage.

I wouldn't be surprised if of all the companies in the world one of
them tried the idea but nobody bought them. I can only think of that
one advantage to it, and that would not be worth the extra cost IMO.
Also, if the transistor is used as a current regulator then you have
dissipation. Unless you get wild and put a switcher in there. Then
you are talking a coil csuae you can't pulse the LED with overcurrent
and expect it to last. The cost might end up being a dollar insrtead
of the few cents a regular LED and resistor cost.


Making bipolar transistors out of III-V semiconductors (e.g. GaAs, InAs,
AlGaAs, InGaAs, InP, GaN, etc) is difficult, and they don't work very
well. Plus the extra constraints would make the LET much harder to build.

The main issue in LED design is that the primary photogeneration is very
efficient, but the photons are generated deep inside the die, and it's
really hard to get them out efficiently. If you just use a regular GaAs
PN diode, all the light gets absorbed by the top semiconductor layer.

Then there's the 95% or so that you lose to total internal reflection
(taking into account the Fresnel reflection of the light that does make
it out), and the 50% you lose due to half the light starting out going
the wrong way....

Modern LEDs are heterojunctions, with transparent semiconductor layers
top and bottom to reduce absorption, and controlled roughening of the
surfaces to let the light rattle around till it escapes.

Cheers

Phil Hobbs

Rattle around, that's so sophisticated of you, Phil!

Jamie

Well, that's pretty much what happens. Listen carefully and you'll hear
it.

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

 

[whit3rd]

On Wednesday, January 28, 2015 at 5:32:50 AM UTC-8, elect...@online.com wrote:

On Tue, 27 Jan 2015 18:49:52 -0800 (PST), jurb6006@gmail.com wrote:

... the holdup with the blue LEDs in the beginning. the rest were
two cents and the blue ones were a buck.

Why were the blue ones so costly?

They have a large bandgap (an energy range where the quantum mechanics
of the solid material does not allow any solution to the electron wave equation).
Only a single-event hop, across a large bandgap, by an electron, can create
light emission of the required type.

Large bandgap => high frequency light emission (blue end of spectrum)
Small bandgap => low frequency light emission (infrared)

That means most materials are unsuitable, right off the bat.

Then, it implies that the possible impurities that would create solutions in
the bandgap range are many (and the semiconductor junction won't emit
light if those are allowed, or even if microscopic strains and voids are allowed).
The earliest usable semiconductor devices were Ge based, because the low
bandgap made impurity sensitivity less troublesome than with Si. It was only after
decades of Si device production that the first 'coppermine' chips used copper
for wiring, because copper is an important impurity and would ruin silicon
devices if it weren't carefully contained/controlled.

Another problem, that's hard to explain, is the three-dimensional nature of
the 'bandgap'; some materials (silicon being one) have an "indirect bandgap",
meaning that an electron transition from lowest-energy-of-high-band
to highest-energy-of-low-band requires a momentum change. There's
no (not much) momentum in the photon, so another particle must be
involved: the complication here, is that the 'single-event hop of an electron'
is a fast event, while a dance with other particles involved is slow (and
for practical purposes, that means it's disallowed).

Those guys with the Nobel prize for getting it all right, explored many
paths of the maze before they found a useful blue glow.