Low Power High Voltage Flip Flop

You should be able to build a J/K flip-flop with discrete transistors - there are N-channel and P-channel MOSFETs that can take that kind of voltage.

The gate-oxide isn't that robust, so you might need level shifters to make the logic work. It won't have to be that fast, so the level shifters won't need a lot of current.

Classic CMOS is good for logic rails up to about 18V, but if there a 25V version I haven't heard of it.

You may need to build a debounce circuit for the switch, else the response may be erratic, as the F/F will count the parity of the contact bounces.

There are a number of simple circuits using a pair of transistors, which will do what you need. They take advantage of charge being stored on a capacitor to change the state of the circuit with the same push button. A Google search turned up a number of possibilities. The FET based designs can be very low power when off, however, you will need to find FETs with gate to source voltage ratings to match your power source. I would look for at least 30V.

Here's a page with FET devices. The link is long, so you may need to patch it up from this post.

Figure 3 has the basic circuit. Other variations add different features. Figure 9 works with higher power source voltages. If you don't have a low voltage regulator, you can use resistors to step down the output voltage.

Is a latching relay too simple a solution:

You may not want to use this

but it's a fun circuit.

A version with two fets would stay on/off forever.

Yeah, that is my notion. I need some details.

I guess I failed to mention it is not a simple switch. It is a rotary encoder with a momentary push button. If there is a rotary encoder of roughly the same size (~ 25mm with a ~6mm shaft) with an on/off switch rather than a momentary switch, I would love to use it. The design uses an EC11 rotary encoder.

What is your notion? 4000 series CMOS won't handle the battery voltages you are talking about. You would need to add two transistors and a voltage regulator to make a 4000 series part compatible with high side switching your 25V power rail. You can do the same job with just the two transistors, no regulator and no 4000 series CMOS. I provided links to some sample circuits.

Maybe you could provide a bit more feedback of what you see here that you like and don't like, and why? Feedback would help a lot.

I don't know if there is a better solution than the opto-coupler, but it is definitely a quick and easy solution to interface with the Arduino.

Won't the phototransistor get unhappy with reverse bias? A better optocoupler would use a symmetric photofet. Something like the H11F1 would work, but there are others.

John

If isolation is not required, and a common ground is used, there are analog switches which should handle 25V. But I would not say an opto is overkill. At least a transistor short won't fry the Arduino.

I'm not seeing any way C1 can pull down the phototransistor emitter; the DC bias doesn't work with the connections as shown.

I don't follow. The only thing the phototransistor can do is pull down the gate of X2 when activated. That will let the Arduino turn off the power source. Because the phototransistor is bipolar, it won't be able to pull up on the X2 gate to turn on the power source. I'm not sure how C1 limits anything. It provides the current to drive the X2 gate (over driving R3 for a moment) until the circuit can switch. It also debounces the push button.

If the C1 capacitor starts at 25V bias, when does it ever drop to lower voltage? How much lower? Do you expect to pull down that emitter on the optoisolator through the load, with reverse-breakdown of the base-emitter?

When the power controller is off, the cap is charged to the incoming voltage. When the controller is on, the gate of X1 is pulled low by X2 and the cap is discharged to ground through X2.

This is essentially the same circuit as one made with a pair of inverters in a loop, forming a bistable device. One inverter charges a cap, and the switch loops that back to that inverter's input, forcing the inverter to change state. Then the circuit is stable in the other state. No matter which state the circuit is in, the switch forces it to the opposite state, because it "transports" the state from the output of one inverter to the input of the same inverter.

I think you are looking at R1 as if it controls the voltage on the battery. X2 controls that voltage. Keep in mind that X1, X2 and R3 form a bistable element.

Those polarities look wrong?

The opto coupler transistor will reverse zener so circuit will always switch on when supply is above circa 8-9V. One solution might be to add a diode in series with either emitter or collector. If the intention is for arduino to only switch off and only manual switch on allowed then you could return the opto coupler emitter to ground instead. If so and the arduino 0V is same as switch ground then you could even substitute the opto with an NPN or Nch mosfet.

piglet

AHA, I see it now. I was thinking this was a flip-flop of the usual type, two PMOS or two NMOS, rather than one of each. I didn't see all the arrows.