Speeding optoisolators

[John Larkin]

[snip]

View in a fixed-width font such as Courier.

.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.

At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.

...Jim Thompson

I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

True. But you do get a nice r-r logic swing, pretty fast, at very low
power consumption, from a simple circuit. For speed, use a pin diode
and a TIA.

OK, why not use the push-pull phototransistors into a tia? Somewhere,
somehow, there may be a use for that. Or even p-p pin diodes?


How about this?

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +
| OPTO |
.--. .---|---------/--.
| | | --- |/ |
--' '-- | \ / ---> | |
4v p-p | -+- |>. |
| | | |
'---|----------|-'
- >-------------' |
| comparator
+------------- +
| out------->
| thr------ -
R
|
|
L
|
|
gnd

Which is sort of like something I may do one day soon, in real life.

John

 

[John Larkin]

Jim Thompson wrote:
[snip]

View in a fixed-width font such as Courier.

.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.

At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.

...Jim Thompson

I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

Fred's idea of reducing LED current is a good one; I'd originally set
the LED current a bit high for CTR and aging allowance, but I was too
conservative. Lower LED current is better.

Here's a faster two-transistor version that minimizes the voltage
swing across the opto's output transistor, and thus manages to pass

1MHz according to the simulator:

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R45
4v p-p | -+- |>. | / 4.7k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R42 | | tf=90nS, tr=125nS
4.7k |<' Q41 +------------------->
+5>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q42
R43 \ +------| 2n2369
2.2k / | |>.
\ \ |
| R44 / |
| 560 \ |
=== | |
GND | |
=== ===
GND GND

It uses more parts, so we're straying from the initial charter of
simple, low-power, and medium-fast.

Common-base transistor Q41 is pretty close to a TIA, so we're
pushing the limits of easy improvement for this opto. In the interest
of speed I've left out any anti-smoke resistors. Besides the speed
advantage, this version uses half the LED current and the opto sees
less voltage, so it's even safer than its predecessors.

Best to all,
James Arthur

Wonder what might be done with those devices where the opto-transistor
base is brought out?

...Jim Thompson

A b-e resistor, ballpark 47k to 1M, really speeds them up, at the cost
of ctr. Given reasonable led drives, in the 5 mA ballpark, it's worth
it.

How about some sort of feedback loop, driving the base such as to keep
the collector current constant? Light increases c-b leakage (I think
that's what happens), the loop backs off external base current, and
Vbe hardly changes at all. That might be fast.

John

 

[Jim Thompson]

On Sun, 07 Oct 2007 08:27:30 -0700, Jim Thompson
[snip]

Wonder what might be done with those devices where the opto-transistor
base is brought out?

...Jim Thompson

A b-e resistor, ballpark 47k to 1M, really speeds them up, at the cost
of ctr. Given reasonable led drives, in the 5 mA ballpark, it's worth
it.

How about some sort of feedback loop, driving the base such as to keep
the collector current constant? Light increases c-b leakage (I think
that's what happens), the loop backs off external base current, and
Vbe hardly changes at all. That might be fast.

John

I'll give that a whirl. Don't know if my Spice models are any good
though.

...Jim Thompson

 

[James Arthur]

Jim Thompson wrote:
[snip]
View in a fixed-width font such as Courier.
.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.
At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.
...Jim Thompson
I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

Fred's idea of reducing LED current is a good one; I'd originally set
the LED current a bit high for CTR and aging allowance, but I was too
conservative. Lower LED current is better.

Here's a faster two-transistor version that minimizes the voltage
swing across the opto's output transistor, and thus manages to pass

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R45
4v p-p | -+- |>. | / 4.7k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R42 | | tf=90nS, tr=125nS
4.7k |<' Q41 +------------------->
+5>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q42
R43 \ +------| 2n2369
2.2k / | |>.
\ \ |
| R44 / |
| 560 \ |
=== | |
GND | |
=== ===
GND GND

It uses more parts, so we're straying from the initial charter of
simple, low-power, and medium-fast.

Common-base transistor Q41 is pretty close to a TIA, so we're
pushing the limits of easy improvement for this opto. In the interest
of speed I've left out any anti-smoke resistors. Besides the speed
advantage, this version uses half the LED current and the opto sees
less voltage, so it's even safer than its predecessors.

Best to all,
James Arthur

Wonder what might be done with those devices where the opto-transistor
base is brought out?

...Jim Thompson

Could be quite a bit faster. As John pointed out already, photodiode
mode into a TIA is about as good as it gets.

My common-base circuit still allows about 35mV of swing across the
phototransistor, so even here a true TIA would help some.

Cheers,
James Arthur

 

[James Arthur]

How would these circuits handle 3.0V power rails ??

What changes would need to be made ??

Thanks

don

Like many "simple" circuits, it's not entirely simple
in that several parameters (drive, CTR, storage time,
transimpedance gain) interact.

Here's the design procedure:
1) Set R41 for desired LED current.
2) Set R42-R43 divider for ~1v.
3) Choose R44 for reliable operation of Q42 at the lowest
anticipated photocurrent. Allow for LED aging and
CTR variation.

The speed--over which you have no more control--is then
determined by how quickly R44 can pull the opto's
phototransistor out of saturation.

Increasing photocurrent increases the phototransistor's
storage time, slowing the circuit, but considerable overdrive
is necessary to cover CTR variations and LED aging.

Here's my go at a lower-voltage version that solidly passes
a 650kHz squarewave:

FIGURE 5.
========== +2.7v
-+-
R51 |
3.6k 1/2 |
+ >----/\/\/----. MOCD217 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R55
4v p-p | -+- |>. | / 2.2k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R52 | | tf=150nS, tr=175nS
2.7k |<' Q51 +------------------->
+2.7v>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q52
R53 \ +------| 2n2369
1.2k / | |>.
\ \ |
| R54 / |
| 1k \ |
=== | |
GND | |
=== ===
GND GND

Please note that the circuit relies on photocurrent to bias
the phototransistor linear. If the LED is normally "off"
it'll take a few uS of light to bias up that phototransistor.
(You can't just whack it with a short "ON" pulse out of
the blue and expect full speed. For that you'd have to
have the LED normally "ON" and pulse it "OFF.")

So, continuous waveforms at full speed are fine, as well
as inverted (low-going) pulses.

All these constraints are much relaxed at lower speeds.

HTH,
James Arthur

 

[Donald]

Like many "simple" circuits, it's not entirely simple
in that several parameters (drive, CTR, storage time,
transimpedance gain) interact.

Here's the design procedure:
1) Set R41 for desired LED current.
2) Set R42-R43 divider for ~1v.
3) Choose R44 for reliable operation of Q42 at the lowest
anticipated photocurrent. Allow for LED aging and
CTR variation.

The speed--over which you have no more control--is then
determined by how quickly R44 can pull the opto's
phototransistor out of saturation.

Increasing photocurrent increases the phototransistor's
storage time, slowing the circuit, but considerable overdrive
is necessary to cover CTR variations and LED aging.

Here's my go at a lower-voltage version that solidly passes
a 650kHz squarewave:

FIGURE 5.
========== +2.7v
-+-
R51 |
3.6k 1/2 |
+ >----/\/\/----. MOCD217 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R55
4v p-p | -+- |>. | / 2.2k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R52 | | tf=150nS, tr=175nS
2.7k |<' Q51 +------------------->
+2.7v>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q52
R53 \ +------| 2n2369
1.2k / | |>.
\ \ |
| R54 / |
| 1k \ |
=== | |
GND | |
=== ===
GND GND

Please note that the circuit relies on photocurrent to bias
the phototransistor linear. If the LED is normally "off"
it'll take a few uS of light to bias up that phototransistor.
(You can't just whack it with a short "ON" pulse out of
the blue and expect full speed. For that you'd have to
have the LED normally "ON" and pulse it "OFF.")

So, continuous waveforms at full speed are fine, as well
as inverted (low-going) pulses.

All these constraints are much relaxed at lower speeds.

HTH,
James Arthur

Thanks

don

 

[John Larkin]

Like many "simple" circuits, it's not entirely simple
in that several parameters (drive, CTR, storage time,
transimpedance gain) interact.

Here's the design procedure:
1) Set R41 for desired LED current.
2) Set R42-R43 divider for ~1v.
3) Choose R44 for reliable operation of Q42 at the lowest
anticipated photocurrent. Allow for LED aging and
CTR variation.

The speed--over which you have no more control--is then
determined by how quickly R44 can pull the opto's
phototransistor out of saturation.

Increasing photocurrent increases the phototransistor's
storage time, slowing the circuit, but considerable overdrive
is necessary to cover CTR variations and LED aging.

Here's my go at a lower-voltage version that solidly passes
a 650kHz squarewave:

FIGURE 5.
========== +2.7v
-+-
R51 |
3.6k 1/2 |
+ >----/\/\/----. MOCD217 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R55
4v p-p | -+- |>. | / 2.2k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R52 | | tf=150nS, tr=175nS
2.7k |<' Q51 +------------------->
+2.7v>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q52
R53 \ +------| 2n2369
1.2k / | |>.
\ \ |
| R54 / |
| 1k \ |
=== | |
GND | |
=== ===
GND GND

Please note that the circuit relies on photocurrent to bias
the phototransistor linear. If the LED is normally "off"
it'll take a few uS of light to bias up that phototransistor.
(You can't just whack it with a short "ON" pulse out of
the blue and expect full speed. For that you'd have to
have the LED normally "ON" and pulse it "OFF.")

Well, run a little quiescent current through the led, just enough to
bias up the base some.

John

 

[Fred Bloggs]

[snip]

View in a fixed-width font such as Courier.

.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.

At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.

...Jim Thompson

I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

Fred's idea of reducing LED current is a good one; I'd originally set
the LED current a bit high for CTR and aging allowance, but I was too
conservative. Lower LED current is better.

Here's a faster two-transistor version that minimizes the voltage
swing across the opto's output transistor, and thus manages to pass

1MHz according to the simulator:

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R45
4v p-p | -+- |>. | / 4.7k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R42 | | tf=90nS, tr=125nS
4.7k |<' Q41 +------------------->
+5>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q42
R43 \ +------| 2n2369
2.2k / | |>.
\ \ |
| R44 / |
| 560 \ |
=== | |
GND | |
=== ===
GND GND

It uses more parts, so we're straying from the initial charter of
simple, low-power, and medium-fast.

Common-base transistor Q41 is pretty close to a TIA, so we're
pushing the limits of easy improvement for this opto. In the interest
of speed I've left out any anti-smoke resistors. Besides the speed
advantage, this version uses half the LED current and the opto sees
less voltage, so it's even safer than its predecessors.

Best to all,
James Arthur

View in a fixed-width font such as Courier.

..
..
..
..
.. FIGURE 5. R22
.. ========== 1k
.. ---/\/\---
.. | |
.. | |
.. | |\ |
.. 1/2 +-----|-\ | |\
.. + >-------------. MOCD217 | | >-+---|+\ .--.
.. | OPTO | 3V--|+/ | >-> | |
.. .--. .---|---------/-- |/ 3.5V-|-/ --' '--
.. | | | --- |/ | |/
.. --' '-- | \ / ---> | |
.. 1.5mA | -+- |>. | dual comp
.. | | | |
.. '---|----------|-'
.. - >-------------' |
.. |
.. ===
.. GND
..
..
..
..
..
..
..

 

[Fred Bloggs]

Jim Thompson wrote:

[snip]

View in a fixed-width font such as Courier.

.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.

At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.

...Jim Thompson

I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

Fred's idea of reducing LED current is a good one; I'd originally set
the LED current a bit high for CTR and aging allowance, but I was too
conservative. Lower LED current is better.

Here's a faster two-transistor version that minimizes the voltage
swing across the opto's output transistor, and thus manages to pass

1MHz according to the simulator:

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R45
4v p-p | -+- |>. | / 4.7k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R42 | | tf=90nS, tr=125nS
4.7k |<' Q41 +------------------->
+5>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q42
R43 \ +------| 2n2369
2.2k / | |>.
\ \ |
| R44 / |
| 560 \ |
=== | |
GND | |
=== ===
GND GND

It uses more parts, so we're straying from the initial charter of
simple, low-power, and medium-fast.

Common-base transistor Q41 is pretty close to a TIA, so we're
pushing the limits of easy improvement for this opto. In the interest
of speed I've left out any anti-smoke resistors. Besides the speed
advantage, this version uses half the LED current and the opto sees
less voltage, so it's even safer than its predecessors.

Best to all,
James Arthur

Wonder what might be done with those devices where the opto-transistor
base is brought out?

...Jim Thompson

That allows you to manipulate the base drive through a reasonably low
impedance and use positive feedback. The old NEC Optocoupler Handbook
contained at least one Schmitt trigger circuit using a single external
transistor to do this; it was as fast as you get using these opto's.

 

[Vladimir Vassilevsky]

That allows you to manipulate the base drive through a reasonably low
impedance and use positive feedback. The old NEC Optocoupler Handbook
contained at least one Schmitt trigger circuit using a single external
transistor to do this; it was as fast as you get using these opto's.

Here is another idea: negative feedback through the other optocoupler.
As soon as the first optocoupler starts to conduct, it cuts of the LED
current via the other optocoupler. So both optocouplers are always in
the barely conducting condition; there are only the slight variations of
the currents and voltages.

Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

 

[Fred Bloggs]

Here is another idea: negative feedback through the other optocoupler.
As soon as the first optocoupler starts to conduct, it cuts of the LED
current via the other optocoupler. So both optocouplers are always in
the barely conducting condition; there are only the slight variations of
the currents and voltages.

Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

That's a good idea, and he has a dual anyway. It may tricky with the
time delay ( feedback transport delay) inherent to oc's though, but it
can be worked.

 

[John Larkin]

Jim Thompson wrote:

[snip]

View in a fixed-width font such as Courier.

.
.
.
. +5V +5v
. FIGURE 3. -+- -+-
. ========== | |
. \ |
. R22 / |
. 10k \ .'
. R21 | |< Q21
. 1.5k 1/2 +--------| 2n3906
. + >----/\/\/----. MOCD217 | |\
. | OPTO | \
. .--. .---|---------/--. +------>
. | | | --- |/ | |
. --' '-- | \ / ---> | | /
. 4v p-p | -+- |>. | \ R23
. | | | | / 3,9k
. '---|----------|-' |
. - >-------------' | |
. '-----------+
. |
. /
. \ R33
. / 1k0
. |
. ===
. GND
.
.

At first glance I thought that might work, but I think the negative
feedback will add Miller thru the opto NPN and keep it slow.

...Jim Thompson

I don't have the MOCD217, but SPICE simulation on the MOC223 shows tr/tf
as about equal at 550ns. It easily handles 100KHz and he can back that
input drive off to 0.5mA or so, 2mA is way too much, and there is very
little danger of exceeding 100% CTR at that drive level. Yep, so NFB
does make the transitions slower.

Inspired by Fred's use of simulation, I plugged an MOC207 into LTspice
and found it surprisingly faithful in modelling my original Fig. 2
circuit.

So inspired, I explored some more.

Attractive as John's totem-pole idea was, I couldn't adjust things to
jerk the opto output transistors 'off.' Maybe I missed something, but
once they're saturated, they're stubborn.

Fred's idea of reducing LED current is a good one; I'd originally set
the LED current a bit high for CTR and aging allowance, but I was too
conservative. Lower LED current is better.

Here's a faster two-transistor version that minimizes the voltage
swing across the opto's output transistor, and thus manages to pass

1MHz according to the simulator:

FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +---------.
| OPTO | |
.--. .---|---------/--. |
| | | --- |/ | |
--' '-- | \ / ---> | | \ R45
4v p-p | -+- |>. | / 4.7k
| | | | \
'---|----------|-' | --. .--
- >-------------' | | | |
| | '--'
R42 | | tf=90nS, tr=125nS
4.7k |<' Q41 +------------------->
+5>--/\/\--+------| 2n3906 |
| |\ /
| | |/ Q42
R43 \ +------| 2n2369
2.2k / | |>.
\ \ |
| R44 / |
| 560 \ |
=== | |
GND | |
=== ===
GND GND

It uses more parts, so we're straying from the initial charter of
simple, low-power, and medium-fast.

Common-base transistor Q41 is pretty close to a TIA, so we're
pushing the limits of easy improvement for this opto. In the interest
of speed I've left out any anti-smoke resistors. Besides the speed
advantage, this version uses half the LED current and the opto sees
less voltage, so it's even safer than its predecessors.

Best to all,
James Arthur

View in a fixed-width font such as Courier.

.
.
.
.
. FIGURE 5. R22
. ========== 1k
. ---/\/\---
. | |
. | |
. | |\ |
. 1/2 +-----|-\ | |\
. + >-------------. MOCD217 | | >-+---|+\ .--.
. | OPTO | 3V--|+/ | >-> | |
. .--. .---|---------/-- |/ 3.5V-|-/ --' '--
. | | | --- |/ | |/
. --' '-- | \ / ---> | |
. 1.5mA | -+- |>. | dual comp
. | | | |
. '---|----------|-'
. - >-------------' |
. |
. ===
. GND
.
.
.
.
.
.
.

Dual comp? The first stage should be an opamp.

But do that, and add a little LED pre-bias current, and it looks
pretty good.

John

 

[James Arthur]

True. But you do get a nice r-r logic swing, pretty fast, at very low
power consumption, from a simple circuit. For speed, use a pin diode
and a TIA.

OK, why not use the push-pull phototransistors into a tia? Somewhere,
somehow, there may be a use for that. Or even p-p pin diodes?

How about this?





FIGURE 4.
========== +5v
-+-
R41 |
2.7k |
+ >----/\/\/----. MOC207 +
| OPTO |
.--. .---|---------/--.
| | | --- |/ |
4v p-p | -+- |>. |
| | | |
'---|----------|-'
- >-------------' |
| comparator
+------------- +
| out------->
| thr------ -
R
|
|
L
|
|
gnd

Which is sort of like something I may do one day soon, in real life.

John

The current waveform from the opto is a ramp, so you could get a
pretty nice flat-topped impulse output this way. The tradeoff is
you'd have to keep the amplitude pretty low to avoid Miller slowing
the opto's transistor. If you've got a fast comparator handy it's not
bad though.

James

 

[James Arthur]

On Sun, 07 Oct 2007 13:32:59 -0700, James Arthur wrote:








Well, run a little quiescent current through the led, just enough to
bias up the base some.

John

That does help if your LED is normally off and you really need that
first pulse response. There is still some charge-up time though, as
the LED "bias current" isn't the full amount, so the photobase still
has to charge up a little to pass full current when the signal pulse
arrives.

Using an LT1818 I've got a TIA simulted similar to the one Fred's
posted. With the LED pre-biasing, the TIA output is nearly up to full
swing by the 2nd pulse @ 2MHz, versus 8 pulses required when run
without the improvement.

At 1MHz, the 3rd pulse is usable even without pre-biasing the LED.

At this level of speed and complexity though, the simplest solution
would be to invert the driving signal, thus ensuring the
phototransistor's biased and ever-ready for quick action.

James
~~~~~~~~~~~~~~~~~~~

Version 4
SHEET 1 1184 724
WIRE 16 80 -80 80
WIRE 288 80 96 80
WIRE -240 208 -288 208
WIRE -144 208 -160 208
WIRE -80 208 -80 80
WIRE -64 208 -80 208
WIRE 224 208 128 208
WIRE 288 208 288 80
WIRE 400 208 288 208
WIRE 1104 208 400 208
WIRE -288 224 -288 208
WIRE 688 224 560 224
WIRE 400 240 400 208
WIRE 688 240 688 224
WIRE 1104 240 1104 208
WIRE 176 272 128 272
WIRE -64 304 -96 304
WIRE 176 304 176 272
WIRE 560 336 560 224
WIRE 688 336 688 320
WIRE -288 352 -288 304
WIRE -96 352 -96 304
WIRE 224 352 224 208
WIRE 224 352 176 352
WIRE 528 352 224 352
WIRE 992 368 592 368
WIRE 1104 368 1104 320
WIRE 400 384 400 320
WIRE 528 384 400 384
WIRE 784 400 688 400
WIRE 400 416 400 384
WIRE 688 416 688 400
WIRE 784 432 784 400
WIRE 400 528 400 496
WIRE 560 528 560 400
WIRE 688 528 688 496
WIRE 688 528 560 528
WIRE 480 576 432 576
WIRE 576 576 544 576
WIRE 176 656 176 352
WIRE 432 656 432 576
WIRE 432 656 176 656
WIRE 464 656 432 656
WIRE 576 656 576 576
WIRE 576 656 544 656
WIRE 992 656 992 368
WIRE 992 656 576 656
FLAG -288 352 0
FLAG -96 352 0
FLAG 1104 368 0
FLAG 784 432 0
FLAG 688 336 0
FLAG 176 304 0
FLAG 400 528 0
SYMBOL Optos\\MOC207 32 272 R0
SYMATTR InstName U1
SYMBOL voltage -288 208 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 3 -414 101 Left 0
SYMATTR Value PULSE(0 4.6 10uS 50nS 50nS .2uS 500nS)
SYMATTR InstName V1
SYMBOL res -144 192 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R3
SYMATTR Value 4.7k
SYMBOL voltage 1104 224 R0
SYMATTR InstName V2
SYMATTR Value 3.3v
SYMBOL diode -144 224 R270
WINDOW 0 32 32 VTop 0
WINDOW 3 0 32 VBottom 0
SYMATTR InstName D1
SYMATTR Value 1N914
SYMBOL res 112 64 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R6
SYMATTR Value 10k
SYMBOL res 560 640 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R4
SYMATTR Value 1k
SYMBOL cap 544 560 R90
WINDOW 0 0 32 VBottom 0
WINDOW 3 32 32 VTop 0
SYMATTR InstName C1
SYMATTR Value 10pF
SYMBOL res 384 224 R0
SYMATTR InstName R1
SYMATTR Value 10k
SYMBOL voltage 688 224 R0
SYMATTR InstName V3
SYMATTR Value 5v
SYMBOL voltage 688 400 R0
SYMATTR InstName V4
SYMATTR Value 5v
SYMBOL Opamps\\LT1818 560 368 R0
SYMATTR InstName U2
SYMBOL res 384 400 R0
SYMATTR InstName R2
SYMATTR Value 10k
TEXT -144 520 Left 0 !.tran 0 20uS 0 100nS
TEXT -720 104 Left 0 ;This version pre-biases the LED and the summing
node,\nfor faster initial response.