Wireless switch for power

Okay, I will be more verbose, maybe I can snow us both under, it certainly is blanco outside here in the Chi!

An opto-isolator is a device that can be described in many ways, I don't want my POV to seem exclusive.  It isn't even with me.

What one gets with an opto-isolator is the ability to separate power supplies.  Your processor/controller is a slab of semiconductor, unless you are poor, Soviet, and in the disco era.  Then it is spread about on four slabs.  Nowadays its the other way, sometimes, multiple processors on a slab.  Your power transistor, nearly invariably, nowadays, an N-channel MOSFET.  This is if you are trying to replicate an SPST switch.  Is also made of thousands of transistors.  'cept these are all alike and in parallel and series with their gates (control terminals) bussed.  Still, its an LSI.  The currents and potentials and loads that modern power transistors are rather amazing.  Since our Q is in the high-current regime, we are sure to have a positive temperature coefficient in our main conduction channel, the clappers on the switch.  This means if we want even more current, we can parallel our power transistors.  Now the only limits are the wife's pocketbook.  Even a fart or hiccup from a power system can entropize sissy-pants CPUs in this situation.  Let's look at the situation where we have a simple stepper motor driver driving a stepping motor with a mechanically reactive load.  Let's say our load is asymmetrical and out of the gravitationally normal plane.  Now, our motor must move and sometimes restrain the load.  We energize and refrain from energizing our coils in our stepper motor.  Now, normally, if one was using a solenoid-actuated-valve to irrigate the rosebushes when one was on vacation, one wouldn't care much about spike dynamics.  When the plunger is de-energized, it snaps back violently, changing (water flow) state given a first-order valve design.  We don't really care about this as a mechanical proposition 'til we are dealing with valves profoundly out of proportion with rosebushes.  This implies that our anti-spike could be a first-order constructed of a diode.  But with the stepper motor in this particular case, I am motivated to let the spike-back current out of the circuit more slowly.  There are  circuits made with Rs. Ds, Zds, Cs and even Ls and Ts, if one is a conservation-freak, that can do this.  This conflicts with my desire to avoid soft- and hard- faults in my controller, however.  I want my controller to be a confident little dictator, in an electrical lair.  The first way to do this is to make sure you have a fast uF-ish scaled decoupling capacitor across your device.  If you put any bulky-bulk on your board, consider a backwards shottky diode across it, for the power-down circumstance.  Really bulk users, like large audio PAs have to use relays and resistors and thermistors during power up/down to minimize current and dissipate stored energy, I digress.

When I impose an opto-isolator in my circuit, I get to use another power-supply.  Generally an isolator would go to a different form of energy than electricity as an intermediate:  A transformer uses the magnetic (thank Thoradson) field.  The energy inserted and extracted from this device is, ideally, in phase.  Transductor is the same thing,'cept the extraction is 180 degrees, ideally with the imposition.  Capacitive transformers use acoustic energy, but seem to have fallen out of fashion.  An acoustic pair can use ultrasound and be bidirectional.  RF could work although conceptually, it is just using a weekly-coupled transformer.  Light is an obvious medium.  So an opto is something like a photo-diode aimed at a photo-transistor in an encapsulant that is opaque.  Susceptible to radio-active decay, but opaque.  The LED is a two-terminal device, the photo-transistor is presented as a three-terminal device.  The original version wasn't DS fast.  You could make it fast.  You can current feed the devices, you can keep them out of saturation, and you can use positive feedback.  If the winter keeps up, I might be doing that for  grins.  Market progress has integrated the follow-on discretes:  Now our device, in this case is saturating (read digital -- or Schmidt) and features an ideal topology for an output, push-pull.  If you miss TTL, you can always put a diode and resistor!  The output stage can handle a fair amount of current, and it can source or sink current.  This is ideal for jocking the control terminal of an SPST device, a push-pull pair or half of an H-bridge.  Because, the control terminal of a big MOSFET, or three, is essentially capacitive if we are doing things rightly.  So, we could construct, say, our own solid-state relay out of just the opto, the powerQ and a few miscellaneous passives and connectors.

=============Don't read this if you already know how to ballast an led or you will get a headache, and you already know how==========

Okay, I will be more verbose, maybe I can snow us both under, it certainly is blanco outside here in the Chi!

An opto-isolator is a device that can be described in many ways, I don't want my POV to seem exclusive.  It isn't even with me.

What one gets with an opto-isolator is the ability to separate power supplies.  Your processor/controller is a slab of semiconductor, unless you are poor, Soviet, and in the disco era.  Then it is spread about on four slabs.  Nowadays its the other way, sometimes, multiple processors on a slab.  Your power transistor, nearly invariably, nowadays, an N-channel MOSFET.  This is if you are trying to replicate an SPST switch.  Is also made of thousands of transistors.  'cept these are all alike and in parallel and series with their gates (control terminals) bussed.  Still, its an LSI.  The currents and potentials and loads that modern power transistors are rather amazing.  Since our Q is in the high-current regime, we are sure to have a positive temperature coefficient in our main conduction channel, the clappers on the switch.  This means if we want even more current, we can parallel our power transistors.  Now the only limits are the wife's pocketbook.  Even a fart or hiccup from a power system can entropize sissy-pants CPUs in this situation.  Let's look at the situation where we have a simple stepper motor driver driving a stepping motor with a mechanically reactive load.  Let's say our load is asymmetrical and out of the gravitationally normal plane.  Now, our motor must move and sometimes restrain the load.  We energize and refrain from energizing our coils in our stepper motor.  Now, normally, if one was using a solenoid-actuated-valve to irrigate the rosebushes when one was on vacation, one wouldn't care much about spike dynamics.  When the plunger is de-energized, it snaps back violently, changing (water flow) state given a first-order valve design.  We don't really care about this as a mechanical proposition 'til we are dealing with valves profoundly out of proportion with rosebushes.  This implies that our anti-spike could be a first-order constructed of a diode.  But with the stepper motor in this particular case, I am motivated to let the spike-back current out of the circuit more slowly.  There are  circuits made with Rs. Ds, Zds, Cs and even Ls and Ts, if one is a conservation-freak, that can do this.  This conflicts with my desire to avoid soft- and hard- faults in my controller, however.  I want my controller to be a confident little dictator, in an electrical lair.  The first way to do this is to make sure you have a fast uF-ish scaled decoupling capacitor across your device.  If you put any bulky-bulk on your board, consider a backwards shottky diode across it, for the power-down circumstance.  Really bulk users, like large audio PAs have to use relays and resistors and thermistors during power up/down to minimize current and dissipate stored energy, I digress.

When I impose an opto-isolator in my circuit, I get to use another power-supply.  Generally an isolator would go to a different form of energy than electricity as an intermediate:  A transformer uses the magnetic (thank Thoradson) field.  The energy inserted and extracted from this device is, ideally, in phase.  Transductor is the same thing,'cept the extraction is 180 degrees, ideally with the imposition.  Capacitive transformers use acoustic energy, but seem to have fallen out of fashion.  An acoustic pair can use ultrasound and be bidirectional.  RF could work although conceptually, it is just using a weekly-coupled transformer.  Light is an obvious medium.  So an opto is something like a photo-diode aimed at a photo-transistor in an encapsulant that is opaque.  Susceptible to radio-active decay, but opaque.  The LED is a two-terminal device, the photo-transistor is presented as a three-terminal device.  The original version wasn't DS fast.  You could make it fast.  You can current feed the devices, you can keep them out of saturation, and you can use positive feedback.  If the winter keeps up, I might be doing that for  grins.  Market progress has integrated the follow-on discretes:  Now our device, in this case is saturating (read digital -- or Schmidt) and features an ideal topology for an output, push-pull.  If you miss TTL, you can always put a diode and resistor!  The output stage can handle a fair amount of current, and it can source or sink current.  This is ideal for jocking the control terminal of an SPST device, a push-pull pair or half of an H-bridge.  Because, the control terminal of a big MOSFET, or three, is essentially capacitive if we are doing things rightly.  So, we could construct, say, our own solid-state relay out of just the opto, the powerQ and a few miscellaneous passives and connectors.

=============Don't read this if you already know how to ballast an led or you will get a headache, and you already know how==========