# Ultra Low-Power H-Bridge (micro Amp)

I'm working on a project to drive a pair of electrodes, one at 12 V and one at 0 V, with a very low-frequency polarity change (every 10 minutes). The current draw across the electrodes peaks at 200 uA, and is typically closer to 40 uA. I've been trying to fins a way to do this with an H-bridge IC, but since they are typically used to operate much higher current draws the leakage current is unacceptable (the most I can afford is around 25 uA). Any suggestions for ICs or transistor arrangements that would help accomplish this would be much appreciated!

• Does the leakage current really matter? If on one side the low-side switch is turned on and the high-side is turned off, but the high-side switch then leaks a little, it will only drain straight through the low-side switch to ground, it won't affect the load. Or is it straight up power consumption that is the issue? Commented Feb 1, 2022 at 14:26
• By asking for an H-bridge, I assume you wanted to REVERSE the polarity of a voltage every few minutes. Another interpretation of your question could be that you want to apply and remove a voltage difference every few minutes. Please clarify which option you are looking for in your project. Commented Feb 1, 2022 at 15:46
• Correct, I want to reverse the polarity of the voltage every few minutes, and in this case it's the the power consumption that's limiting. Commented Feb 1, 2022 at 15:55

You can realise the H-Bridge with regular small-signal MOSFET such as 2N7002 (NMOS) or BSS84 (PMOS). Their leakage is way below uA... even below nA typically.

Using PMOSFET for the top MOSFETs has the advantage that you can easily control them from a controller (use a common-source NMOS or NPN to turn them on/off). Bottom NMOSFETs can be controlled directly.

One leg would look like this:

simulate this circuit – Schematic created using CircuitLab

• Thanks!! Just to confirm, you could tie M3 and M1 to the same signal, correct? And then is there a reason R1 in the inverter needs to be 100 kOhm? Since that would lead to 120 uA of current drain through the resistor when M3 is ON. Or could that be a higher value to limit power consumption? Commented Feb 1, 2022 at 15:53
• @OctaviusPublius 100k Ohm can be "anything".. true. But too low values will draw a lot of current when M3 is on, so I kept the value rather high because I guess you are looking at a low power application. Tying the two inputs together is dangerous: if that signal goes high, M1 will flick on basically immediately, but M2 will only slowly switch off (due to R1). So there is a period of shootthrough. Either control them separately with dead time from the controller, or provide additional dead-time generation circuitry. Commented Feb 1, 2022 at 16:09
• I appreciate it, that makes a lot of sense. With both sides set up, could I omit the inverter and tie together the high signal from one side with that of the other? (HIGH to HIGH and LOW to LOW in this diagram here: imgur.com/a/mXdhF8d Commented Feb 1, 2022 at 17:06
• @OctaviusPublius if you omit the inverter M3 in one leg you will not be able to properly control the PMOS (unless your controller has open-drain type outputs available, then it could work). But what you can do, is to add one more inverter to the second leg, one each for its Bottom and Top signal. Then you could tie the two Top signals and the two Bottom signals of the two legs (one of which will have inverted logic) together. Commented Feb 1, 2022 at 17:55

With very low frequency and low current and leakage issues, you may want to use a DPDT relay. Switch the relay's coil with a simple transistor as switch. The following is with a 1Hz input to the transistor base. You'll need to adjust base resistor for your signal source.

Positive power...

Reverse power...

• Make sure to use a Reed relay or other hermetic relay or solid-state relay. The unsealed generic relays can have unrealiable contact at µAs stuff. Also this solution doesn't allow to disconnect the supply from the load, which may be necessary. But the implicit shoot-through impossibility is good :) (At least you can achieve it, by putting the load on the common terminals at 12/0V at the switches terminals) Commented Feb 1, 2022 at 14:43