# lowest loss for switch

Are there any other options for an extremely low loss switching circuit with extremely low switching frequency (about 4 times per day) and voltages from 0-3V with currents <2mA.

Current ideas include:

• latching relays (<100mOhm, but use ~50mW for 5ms to switch) => would only consume power during switching, which is infrequent
• analog signal switches (have about 100mOhm resistance and leakage/quiescent current in the tens of nA range) (eg. TI ts3a24159) => continuous quiescent power consumption

Does anyone have other solutions?

FYI: the switches are used in a "rectifier" that only changes the polarity of its input voltage about 4 times per day. It will be part of a TEG (thermal energy generator) that scavenges energy from a temperature difference and stores it in a battery. Since the temperature difference can be of both polarities the TEG will also output two different voltage polarities. So basically I want to minimize the total energy loss over a long period of time.

• What exactly do you care about? Conductance losses when the switch is closed, or power used by the switch? If the latter, why? – marcelm Apr 8 '16 at 15:15
• Iq for the TI switch is 15na typically at room temperature. Is that really significant? If so, why? A battery probably has >100x that much self-discharge current. – Spehro Pefhany Apr 8 '16 at 15:28
• @marcelm I care about the total losses in the system. Conductance losses are the main concern here, but if the power used by the switch itself is too large (for example with a regular relay) it too factors in heavily. It will be part of a TEG (thermal energy generator) that scavenges energy from a temperature difference and stores it in a battery. Since the temperature difference can be of both polarities the TEG will also output two different voltage polarities. So basically I want to minimize the total energy loss over a long period of time. – studentforlife Apr 10 '16 at 8:24

MOSFETs are indeed an option, and if you look around you can get P channel devices (easiest to switch because you need no boost circuit) with $R_ds(on)$ as low as 1.5 m$\Omega$ with 1.8V Vgs which is, quite honestly, astounding for a P channel device.
There does appear to be a mismatch between the page and datasheet, but I can certainly get P channel devices with about 3 m$\Omega$ from Vishay.