Microsceond simultaneous exposures with rolling shutters


Bruce Griffiths
 

If a rolling shutter camera has a cold shoe or equivalent for an external flash, then its possible to use the contact closure (originally an actual mechanical contact closure now typically an open drain switch) to trigger a pulsed laser synchronously with the front shutter curtain. A simple monostable based on a 555 timer can be used to set the laser pulse width. For lasers diodes with no monitor photodiode the laser diode can be driven from the output of the monostable via a resistor. For laser diodes with a monitor photodiode, feedback from the monitor photodiode can be used to ensure satisfactory operation over a wide temperature range. A suitable driver circuit can either use a number of wideband opamps or a few inexpensive bipolar transistors to limit the maximum laser current and achieve sufficiently fast turn on and turn off (switching delay of around 1 microsecond or so). This delay can be reduced to a few tens of nanosconds using Schottky diode camps to keep the bipolar transistors out of deep saturation. For most interferometric applications 4D Technology found that a pulse width of around 100 mircoseconds with a laser power of 10 milliwatts was sufficient for testing uncoated test surfaces. 

I've used LTSpice to simulate the operation of a bipolar transistor based driver circuit for laser diodes with monitor photodiodes.  

With such a pulsed laser even a digital camera with a mechanical focal plane shutter could be used to capture 100 microsecond exposure interferograms.

Bruce


George Roberts (Boston)
 

Maybe post the circuit here?  I might build it some day.


Bruce Griffiths
 

George

Here's the driver circuit for a direct green laser diode with a monitor photodiode that is rated for a 10mW output and a corresponding monitor photodiode photocurrent of 100 microamp for that 10mW laser diode output.
Q1 acts as a controlled current sink drive for the laser diode. Q2 limits the maximum laser diode current. R1 sets the laser diode current limit. Q3 regulates the monitor diode photocurrent. R2 sets the monitor diode photocurrent.
Q4 switches the laser driver on when Q4 is turned on by the input signal.
The small signal Schottky diodes D4, D5 clamp the collectors of Q3, Q4 respectively preventing deep saturation so that transistor turnoff delay is reduced from 1 microscond or more (when these transistors are saturated) to a few tens of nanoseconds. The compensation capacitor C1 minimises overshoot and slows the laser turn on to around 300 nanoseconds or so.

The drive signal for Q4 is a positive pulse the pulsewidth of which controls the laser output pulse duration. A pulse amplitude of 5V rather than 12V shown could be used if the value of R6 were reduced. 

When using with a DSLR or similar with a cold shoe a 555 monostable or equivalent can be driven by the contact closure. It may be best to locate the timer section in an enclosure that mates with the cold shoe (or with a hot shoe just use the cold shoe contact closure and ignore the other contacts/signals). An added complication is the need with a 555 type monostable to differentiate the trigger signal if its longer than the output pulse width. Also, provision to turn the laser on via switch is desirable during interferometer alignment and when adjusting the power output with the help of a laser power sensor.

The driver will maintain a stable laser output power over a wide temperature range unlike simpler drivers used without feedback from a monitor photodiode.

A simple charge pump can be used to produce the negative supply voltage which has a maximum load current of a few hundred microamp.

Bruce

On 22/10/2022 12:46 George Roberts (Boston) <bb@...> wrote:


Maybe post the circuit here?  I might build it some day.


Bruce Griffiths
 

George

Here's a driver optimised for driving a direct green laser diode such as am OSRAM PLT5 522.

The turn on delay is dominated by the base voltage of the npn emitter follower slewing from the off voltage to the on voltage (<1V difference). This is determined by the current flowing in R2 (sets the monitor photodiode operating current and hence the laser output when turned on) and the value of the compensation capacitor (minimises overshoot and ringing when the laser diode turns on or off) C1.

The driver is idle (laser off) current is only a few milliamps.
If an extra Schottky diode is added together with a switch, the switch can be used to turn the laser on irrespective of the state of the input signal (shown as V2 for which +5V corresponds to laser off and 0V corresponds to laser on).
A CMOS inverter or a Schottky clamped transistor could be used to invert the input signal if required. 

All parts, or at least functional equivalents, are readily available.
However, most of these will be surface mount parts.
Apart from the laser diode there are no expensive parts.
The driver circuit and the laser diode should be mounted on the same circuit board.

The turn on delay of this driver as shown is about 200 nanoseconds.
The turn of delay is a few tens of nanoseconds.
Using an opamp based driver will not reduce these switching delays.

Bruce

On 22/10/2022 17:28 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's the driver circuit for a direct green laser diode with a monitor photodiode that is rated for a 10mW output and a corresponding monitor photodiode photocurrent of 100 microamp for that 10mW laser diode output.
Q1 acts as a controlled current sink drive for the laser diode. Q2 limits the maximum laser diode current. R1 sets the laser diode current limit. Q3 regulates the monitor diode photocurrent. R2 sets the monitor diode photocurrent.
Q4 switches the laser driver on when Q4 is turned on by the input signal.
The small signal Schottky diodes D4, D5 clamp the collectors of Q3, Q4 respectively preventing deep saturation so that transistor turnoff delay is reduced from 1 microscond or more (when these transistors are saturated) to a few tens of nanoseconds. The compensation capacitor C1 minimises overshoot and slows the laser turn on to around 300 nanoseconds or so.

The drive signal for Q4 is a positive pulse the pulsewidth of which controls the laser output pulse duration. A pulse amplitude of 5V rather than 12V shown could be used if the value of R6 were reduced. 

When using with a DSLR or similar with a cold shoe a 555 monostable or equivalent can be driven by the contact closure. It may be best to locate the timer section in an enclosure that mates with the cold shoe (or with a hot shoe just use the cold shoe contact closure and ignore the other contacts/signals). An added complication is the need with a 555 type monostable to differentiate the trigger signal if its longer than the output pulse width. Also, provision to turn the laser on via switch is desirable during interferometer alignment and when adjusting the power output with the help of a laser power sensor.

The driver will maintain a stable laser output power over a wide temperature range unlike simpler drivers used without feedback from a monitor photodiode.

A simple charge pump can be used to produce the negative supply voltage which has a maximum load current of a few hundred microamp.

Bruce
On 22/10/2022 12:46 George Roberts (Boston) <bb@...> wrote:


Maybe post the circuit here?  I might build it some day.


Bruce Griffiths
 

George

Here's a simple circuit for driving a RED laser diode that doesn't have a monitor photodiode.
The pnp transistor switches the laser diode on and off within a few tens of nanosconds (with good PCB layout). The input RC circuit slows the input signal at the base of the pnp transistor to reduce the turn on overshoot of the laser diode.

The 555 monostable/timer isnt a good choice for this application as its not a true edge triggerd monostable and if the input pulse is longer than the design value of the monostable output the output pulse width will be equal to the input pulse width.

A true edge triggered monostable such as a 74LVC123 or equivalent is probably a better choice.   

Bruce

On 24/10/2022 14:45 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's a driver optimised for driving a direct green laser diode such as am OSRAM PLT5 522.

The turn on delay is dominated by the base voltage of the npn emitter follower slewing from the off voltage to the on voltage (<1V difference). This is determined by the current flowing in R2 (sets the monitor photodiode operating current and hence the laser output when turned on) and the value of the compensation capacitor (minimises overshoot and ringing when the laser diode turns on or off) C1.

The driver is idle (laser off) current is only a few milliamps.
If an extra Schottky diode is added together with a switch, the switch can be used to turn the laser on irrespective of the state of the input signal (shown as V2 for which +5V corresponds to laser off and 0V corresponds to laser on).
A CMOS inverter or a Schottky clamped transistor could be used to invert the input signal if required. 

All parts, or at least functional equivalents, are readily available.
However, most of these will be surface mount parts.
Apart from the laser diode there are no expensive parts.
The driver circuit and the laser diode should be mounted on the same circuit board.

The turn on delay of this driver as shown is about 200 nanoseconds.
The turn of delay is a few tens of nanoseconds.
Using an opamp based driver will not reduce these switching delays.

Bruce
On 22/10/2022 17:28 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's the driver circuit for a direct green laser diode with a monitor photodiode that is rated for a 10mW output and a corresponding monitor photodiode photocurrent of 100 microamp for that 10mW laser diode output.
Q1 acts as a controlled current sink drive for the laser diode. Q2 limits the maximum laser diode current. R1 sets the laser diode current limit. Q3 regulates the monitor diode photocurrent. R2 sets the monitor diode photocurrent.
Q4 switches the laser driver on when Q4 is turned on by the input signal.
The small signal Schottky diodes D4, D5 clamp the collectors of Q3, Q4 respectively preventing deep saturation so that transistor turnoff delay is reduced from 1 microscond or more (when these transistors are saturated) to a few tens of nanoseconds. The compensation capacitor C1 minimises overshoot and slows the laser turn on to around 300 nanoseconds or so.

The drive signal for Q4 is a positive pulse the pulsewidth of which controls the laser output pulse duration. A pulse amplitude of 5V rather than 12V shown could be used if the value of R6 were reduced. 

When using with a DSLR or similar with a cold shoe a 555 monostable or equivalent can be driven by the contact closure. It may be best to locate the timer section in an enclosure that mates with the cold shoe (or with a hot shoe just use the cold shoe contact closure and ignore the other contacts/signals). An added complication is the need with a 555 type monostable to differentiate the trigger signal if its longer than the output pulse width. Also, provision to turn the laser on via switch is desirable during interferometer alignment and when adjusting the power output with the help of a laser power sensor.

The driver will maintain a stable laser output power over a wide temperature range unlike simpler drivers used without feedback from a monitor photodiode.

A simple charge pump can be used to produce the negative supply voltage which has a maximum load current of a few hundred microamp.

Bruce
On 22/10/2022 12:46 George Roberts (Boston) <bb@...> wrote:


Maybe post the circuit here?  I might build it some day.


Bruce Griffiths
 

George

Attached circuit illustrates how the contact closure (S1) can be used to trigger a monostable (74XX123) to generate a 100 microsecond wide pulse to drive the laser driver input.
The pulse width is set by R1C1. Choose either the positive or the negative pulse output as required by the laser driver pulse input.
A few refinements such as a somewhat longer power on reset time constant (R3C2) are required but the circuit illustrates the essentials.

Bruce

On 25/10/2022 17:33 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's a simple circuit for driving a RED laser diode that doesn't have a monitor photodiode.
The pnp transistor switches the laser diode on and off within a few tens of nanosconds (with good PCB layout). The input RC circuit slows the input signal at the base of the pnp transistor to reduce the turn on overshoot of the laser diode.

The 555 monostable/timer isnt a good choice for this application as its not a true edge triggerd monostable and if the input pulse is longer than the design value of the monostable output the output pulse width will be equal to the input pulse width.

A true edge triggered monostable such as a 74LVC123 or equivalent is probably a better choice.   

Bruce
On 24/10/2022 14:45 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's a driver optimised for driving a direct green laser diode such as am OSRAM PLT5 522.

The turn on delay is dominated by the base voltage of the npn emitter follower slewing from the off voltage to the on voltage (<1V difference). This is determined by the current flowing in R2 (sets the monitor photodiode operating current and hence the laser output when turned on) and the value of the compensation capacitor (minimises overshoot and ringing when the laser diode turns on or off) C1.

The driver is idle (laser off) current is only a few milliamps.
If an extra Schottky diode is added together with a switch, the switch can be used to turn the laser on irrespective of the state of the input signal (shown as V2 for which +5V corresponds to laser off and 0V corresponds to laser on).
A CMOS inverter or a Schottky clamped transistor could be used to invert the input signal if required. 

All parts, or at least functional equivalents, are readily available.
However, most of these will be surface mount parts.
Apart from the laser diode there are no expensive parts.
The driver circuit and the laser diode should be mounted on the same circuit board.

The turn on delay of this driver as shown is about 200 nanoseconds.
The turn of delay is a few tens of nanoseconds.
Using an opamp based driver will not reduce these switching delays.

Bruce
On 22/10/2022 17:28 Bruce Griffiths <bruce.griffiths@...> wrote:


George

Here's the driver circuit for a direct green laser diode with a monitor photodiode that is rated for a 10mW output and a corresponding monitor photodiode photocurrent of 100 microamp for that 10mW laser diode output.
Q1 acts as a controlled current sink drive for the laser diode. Q2 limits the maximum laser diode current. R1 sets the laser diode current limit. Q3 regulates the monitor diode photocurrent. R2 sets the monitor diode photocurrent.
Q4 switches the laser driver on when Q4 is turned on by the input signal.
The small signal Schottky diodes D4, D5 clamp the collectors of Q3, Q4 respectively preventing deep saturation so that transistor turnoff delay is reduced from 1 microscond or more (when these transistors are saturated) to a few tens of nanoseconds. The compensation capacitor C1 minimises overshoot and slows the laser turn on to around 300 nanoseconds or so.

The drive signal for Q4 is a positive pulse the pulsewidth of which controls the laser output pulse duration. A pulse amplitude of 5V rather than 12V shown could be used if the value of R6 were reduced. 

When using with a DSLR or similar with a cold shoe a 555 monostable or equivalent can be driven by the contact closure. It may be best to locate the timer section in an enclosure that mates with the cold shoe (or with a hot shoe just use the cold shoe contact closure and ignore the other contacts/signals). An added complication is the need with a 555 type monostable to differentiate the trigger signal if its longer than the output pulse width. Also, provision to turn the laser on via switch is desirable during interferometer alignment and when adjusting the power output with the help of a laser power sensor.

The driver will maintain a stable laser output power over a wide temperature range unlike simpler drivers used without feedback from a monitor photodiode.

A simple charge pump can be used to produce the negative supply voltage which has a maximum load current of a few hundred microamp.

Bruce
On 22/10/2022 12:46 George Roberts (Boston) <bb@...> wrote:


Maybe post the circuit here?  I might build it some day.


Franz
 

Bruce,
Is that the circuit you wanted to post?


Bruce Griffiths
 

Franz

No, this one.

For most modern consumer digital cameras with a cold shoe the contact closure is an open drain switch to ground with an open circuit voltage rating of 5V or so.
This is simulated by S1 which triggers the monostable.
The output pulse width is independent of the switch closure deration.
For a mechanical switch closure a more complex circuit is required to holdoff retriggering whilst the switch contact bounces. An additional monostable with a relatively long output pulse width can be used to do this.

Bruce

On 25/10/2022 22:54 Franz <franz.hagemann@...> wrote:


Bruce,
Is that the circuit you wanted to post?


Sorin
 

Bruce,

what are the numbers  (0 5 10u 100n 100n 10m) and (-1 1 500u 1n 1n 1u)?



On Tuesday, October 25, 2022 at 01:51:32 PM GMT+3, Bruce Griffiths <bruce.griffiths@...> wrote:


Franz

No, this one.

For most modern consumer digital cameras with a cold shoe the contact closure is an open drain switch to ground with an open circuit voltage rating of 5V or so.
This is simulated by S1 which triggers the monostable.
The output pulse width is independent of the switch closure deration.
For a mechanical switch closure a more complex circuit is required to holdoff retriggering whilst the switch contact bounces. An additional monostable with a relatively long output pulse width can be used to do this.

Bruce

On 25/10/2022 22:54 Franz <franz.hagemann@...> wrote:


Bruce,
Is that the circuit you wanted to post?


Bruce Griffiths
 

Sorin

Its a screenshot from an LTspice simulation.
These are the parameters of a pulsed voltage source.
The first one
0 = off voltage
5= on voltage
100u = turn on delay of 100 microseconds
100n = 100 nanosecond risetime
100n = 100 nanosecond fall time
10m = 10 milliseconds on time
This is the power supply for the monostable.
rise and fall times are somewhat unrealistic but can be modified to more accurately reflect a real supply.
The second pulsed voltage source is just a convenient way of turning the switch on and off. It could be modified to simulate switch bounce if required. The parameters are
-1 = off voltage
1 = on voltage
500u = 500 microseconds turn on delay
1n = 1 nanoscond risetime
1n = 1 nanosecond fall time
1u = 1 microsecoond on time

For the closed loop and open loop laser drivers various circuits were simulated using shunt mosfets, opamps etc and the simplest that had fast switching times and low laser diode turnon and turn off currents were chosen.

For typical direct green laser diodes the laser anode is typically connected to the case. A driver with the laser anode connected to ground simplifies use of copper heatsinks required for continuous operation when adjusting the laser power output.
The Schottky diodes shown are rated at 5V reverse voltage and the model schottky diode undergoes zener breakdown at about 6V reverse voltage which introduced some issues with a couple of driver circuits. In practice a higher breakdown schottky signal diode like a 1N5711 or equivalent would be used but I didn't have a spice model for a 1N5711 on this PC.

Mounting the monostable powered by a lithium coin cell in a small enclosure that mates with the camera shoe is one option with a piece of thin coax (RG174 or similar) connecting it to the actual laser and its driver.
For some drivers disconnecting the monostable pulser from the driver circuit turns the laser on which may be useful in setting up.
 
Bruce

On 26/10/2022 00:53 Sorin via groups.io <sorin_sfartz@...> wrote:



Bruce,

what are the numbers  (0 5 10u 100n 100n 10m) and (-1 1 500u 1n 1n 1u)?



On Tuesday, October 25, 2022 at 01:51:32 PM GMT+3, Bruce Griffiths <bruce.griffiths@...> wrote:


Franz

No, this one.

For most modern consumer digital cameras with a cold shoe the contact closure is an open drain switch to ground with an open circuit voltage rating of 5V or so.
This is simulated by S1 which triggers the monostable.
The output pulse width is independent of the switch closure deration.
For a mechanical switch closure a more complex circuit is required to holdoff retriggering whilst the switch contact bounces. An additional monostable with a relatively long output pulse width can be used to do this.

Bruce
On 25/10/2022 22:54 Franz <franz.hagemann@...> wrote:


Bruce,
Is that the circuit you wanted to post?


Bruce Griffiths
 

On Fri, Oct 21, 2022 at 08:41 PM, Bruce Griffiths wrote:
applications 4D Technology found that a pulse width of around 100 mircoseconds with a laser power of 10 milliwatts was sufficient for testing uncoated test surfaces. 

I've used LTSpice to simulate the operation of a bipolar transistor based driver circuit for laser diodes with monitor photodiodes.  

With such a pulsed laser even a digital camera with a mechanical focal plane shutter could be used to capture 100 microsecond exposure interferograms.

Bruce

 

An ISO518 hot shoe (not a cold shoe as previously stated) or equivalent is actually required.
The centre and outer contacts are connected by an electronic switch to trigger the flash in modern cameras,
Whilst ISO10330 allows trigger voltages up to 24 volts, some cameras limit the open circuit contact voltage to 5V.
For those cameras with proprietary shoes an adapter may be required. For older cameras with mechanical flash trigger contacts the open circuit voltage rating can be much higher.

Bruce

Bruce