uCurrent v2.0

[lai3IeSh]

Depends on the Rdson you get for 2.5 to 3V Vgs.

[Psi]

What about LIR2032 batteries. 3.7v nominal lithium ion coin cell batteries. I've already built a lithium charger using a microchip MCP73831 powered by a mini-USB socket.

They have about 35mAh of capacity, so around 50hrs of runtime. Charging at ~17mA (0.5C rate) would take 2 hours. Minimal extra cost for the charging IC and the usb socket.

Yeah, that would be functionally better.

I just get a kick out of using green technology and ultra capacitors due to their lifespan compared to batteries. 

[lai3IeSh]

What about LIR2032 batteries. 3.7v nominal lithium ion coin cell batteries. I've already built a lithium charger using a microchip MCP73831 powered by a mini-USB socket.

They have about 35mAh of capacity, so around 50hrs of runtime. Charging at ~17mA (0.5C rate) would take 2 hours. Minimal extra cost for the charging IC and the usb socket.

Charging via USB is certainly an option, but I think it is abuse of USB
You could use standard AA NiMHs or LiPos and charge them in your standard battery charger - no extra cost except for the battery socket. And you probably would charge them seldom, because they have a high capacity. And they are cheap.

[lai3IeSh]

I just get a kick out of using ultra capacitors due to their lifespan compared to batteries. 

I like that. Thinking green
You could also use a royer converter to transfer energy wireless, so you would have some kind of "charging cradle".
I wanted to do this with my phone (I have a big Sony Ericcson, so lots of space) for a while now, but I had no spare time.

[Psi]

What about LIR2032 batteries. 3.7v nominal lithium ion coin cell batteries. I've already built a lithium charger using a microchip MCP73831 powered by a mini-USB socket.

They have about 35mAh of capacity, so around 50hrs of runtime. Charging at ~17mA (0.5C rate) would take 2 hours. Minimal extra cost for the charging IC and the usb socket.

Charging via USB is certainly an option, but I think it is abuse of USB
You could use standard AA NiMHs or LiPos and charge them in your standard battery charger - no extra cost except for the battery socket. And you probably would charge them seldom, because they have a high capacity. And they are cheap.

You could keep the supercapacitor, they charge up insanely quick (~instant) if you have enough input current.
No need to wait for ya batteries to charge before you can use your ucurrent.

[lai3IeSh]

They have about 35mAh of capacity, so around 50hrs of runtime. Charging at ~17mA (0.5C rate) would take 2 hours. Minimal extra cost for the charging IC and the usb socket.

Charging via USB is certainly an option, but I think it is abuse of USB

You could keep the supercapacitor, they charge up insanely quick (~instant) if you have enough input current.
No need to wait for ya batteries to charge before using your ucurrent.

But watch out! USB allows you to take only 100mA, so you have to regulate your inrush current.

[lai3IeSh]

A Bat54 has a forward voltage drop of less than 250mV @50uA. That is 12.5uW. I don't know if you can beat that.

I just found the BAT754, which has a Vf of maybe 180mV @100uA. The reverse leakage current is <150nA @3V Vr

[Psi]

You could keep the supercapacitor, they charge up insanely quick (~instant) if you have enough input current.
No need to wait for ya batteries to charge before using your ucurrent.

But watch out! USB allows you to take only 100mA, so you have to regulate your inrush current.

hehe yeah, going to need much more current than what USB is able to do if we want to charge it instantly.
Probably need like a laptop powerbrick

[lai3IeSh]

Well, here is what I could come up with.
The LTC3535 is a step-up with Vin > Vout operation, so you can plug anything between 0.8 and 5V in.
If you know your minimum battery voltage is above 3.6V or so, it would obviously better to use step-down converters. However, I started out with a single 1.2V NiMh or LiPo cell. The LTC3535 has two outputs with independent shutdown pins.
1.8V for the micro and 3.6 with an LDO for the analog circuitry. Maybe I'll use a single 3.6V step-up for digital and analog (analog still with 3.3V LDO).
The micro can shutdown the power supply for the analog part and itself. Using a tact-switch, /SHDN is pulled above 0.8V and the step-up for 1.8V micro is active. The micro then pulls /SHDN1 high and also /SHDN2, activating the analog power supply.
There is no off-switch as of yet, the micro auto-shuts device off.

The analog circuitry is a straight forward transimpedance amp, with selectable feedback resistor Rf (5k, 50k, 1M).
All currents are converted to a 2mV to 200mV range as noted on the circuit. I didn't want to make the output of the transimpedance amp too small, so the offset voltage of the gain amplifier wouldn't matter much.

I'm not happy with that many switches, maybe it is possible to reduce the number by smart routing of the signals. I think I'll use analog muxes/switches, with less than 5 Ohm on-resistance, so it is negligible. I know, I'm a hypocrite, because the analog switches/muxes cost mor than the mechanical ones.

The switches are controlled by the micro, which also has 3 LEDs indicating the current range, so you can interpret your DMM output.

I don't know what to do with the 2mA-range. Using a 1Ohm feedback resistor forbids an analog switch. Using 500 or 50 Ohm Rf requires another voltage divider.

Also, I guess in the uA/pA range guarding is required, and I didn't look at the leakage of the protection diodes either, nor did I calculate the required power.

I'd like some thoughts/suggestions on the feedback resistor values and the corresponding gains.

Edit: I just realised, the voltage divider has to buffered, or it will form a voltage divider with the feedback resistor. I think it is possible to move the divider behind the gain amplifier.