This circuit is a latching relay that is activated by passing current through the coil to charge and discharge a capacitor. Coil resistance controls maximum current and capacitor charge/discharge rate. There is never a sudden power collapse across the coil and the coil in this instance does not require a protection diode.
Last edited:
You are switching an inductor - you will get inductive kick back when you switch off the current, doesn't matter what's connected on the other side. Its the high dI/dt that generates the high voltage that cooks your driver chip. I suspect the reason you've been lucky is usually you switch off after the capacitor's charged and the current is already very small. That's a precarious situation that can be resolved with two schottky diodes.
OK, one last reply. It appears that you do not understand microprocessors and logic-controlled circuits. The only way that an inductor can cause a voltage spike is if current flow is suddenly interrupted with an open-circuit and the current has nowhere to go. That results in theoretically infinite voltage until a path for the current is found. The inductor in my circuit never sees an open circuit. The source processor pin is held either high (+5V) or low (power ground), it never floats. Current flows only long enough for the capacitor to charge or discharge, far less than one second for all practical purposes and maximum current is always less than 20mA maximum. Switching time between states is approximately 5 micro seconds -- 5 ten thousands of a second). This is done under software control. The processor pin is held at the state when power was removed until there is no power remaining in the circuit. Now, please stop harassing me over my design.
Here is a proposed modification. Board size stays the same at 2.5" x 1.4" but two trim pots are added.
One trimmer sets the turn-on delay in a range from 1 second to 60 seconds in two ranges. The first range -- half the pot rotation -- provides granular resolution from 1 second to 10 seconds, and the second half of rotation provides less granular resolution from 10 seconds to 60 seconds.
The second trimmer sets the shutdown point as a percentage of full power. Currently, the board is hard-coded to shut down when the supply voltage drops to 75% of normal, but I would like to make it variable. I'm thinking that probably 85~90% of normal would be a better value but prefer to leave that to the user. A tightly regulated supply would probably permit a value as high as 95%, but that's just an opinion.
A second modification adds a pad at the control output. The status of the relay can be sensed to perhaps drive a more robust relay to switch power amp output, for example. The load must be kept at a minimum (voltage sense only, no current draw) to avoid affecting the relay operation.
One trimmer sets the turn-on delay in a range from 1 second to 60 seconds in two ranges. The first range -- half the pot rotation -- provides granular resolution from 1 second to 10 seconds, and the second half of rotation provides less granular resolution from 10 seconds to 60 seconds.
The second trimmer sets the shutdown point as a percentage of full power. Currently, the board is hard-coded to shut down when the supply voltage drops to 75% of normal, but I would like to make it variable. I'm thinking that probably 85~90% of normal would be a better value but prefer to leave that to the user. A tightly regulated supply would probably permit a value as high as 95%, but that's just an opinion.
A second modification adds a pad at the control output. The status of the relay can be sensed to perhaps drive a more robust relay to switch power amp output, for example. The load must be kept at a minimum (voltage sense only, no current draw) to avoid affecting the relay operation.
Attachments
So if there's a brown-out, then when the microcontroller enters reset and sets all its pins high-Z - what happens then? Or the power drops instantly (say by a short across the supply). Then all the CMOS FETs turn off.... dead chip.
Or when programming the microcontroller what happens to the pins then? Have you checked?
Never use an inductive load without kick-back protection, its basic stuff and will save your bacon one day...
In case anyone is interested, I've enhanced this design to incorporate user-selectable delay (0-60 seconds) and user-selectable percent-of-PS-nominal for cutoff when power supply voltage drops. Also added an input voltage range of 10~30VDC with automatic calibration so that the preset delay and shutdown settings do not change at any voltage in the range. The new thread is here: user-programmable delay.