Precision current from variable voltage -> ?

BADHABIT

I would like to regulate current to a device(.09mA)·regardless of the voltage (from 3.2 to 2.6v).

Me thinks it is possible, that I have somewhere read how. There is a battery operated 3v system with devices that will operate down to 2.6v,·with some·devices·requiring precise·current, that needs to operate for extended periods. Having the ability to provide proper current at dwindling voltages would provide the system with the ability to stay alive longer and make me a happy camper.

··············································· ············· ·········

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Phil Pilgrim (PhiPi)

Take a look at the LM317L. The app notes illustrate its use as a current regulator. You will lose 1.2V of your input voltage in the process, however. The required resistor will be about 13.3K for the 90 microamps you cited. (That is what you meant, right?)

-Phil

BADHABIT

It is.

I'm not sure the 1.2v loss could be tolerated. The system is run from a Li·coin cell and I'm not sure that a charge pump/regulator to get back to voltage·would be fuel efficient. Unless there is a dif't way to re-establish 3.2-2.6v.

If I switched to a dif't battery voltage, then it would work. But if I did that I could set the regulator @ the lowest useable voltage and negate the issue altogether. The coin cell offers a package that has a good size as well as being affordable and available.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Phil Pilgrim (PhiPi)

If you can tolerate a 0.6V loss, a circuit like this might work:



The "zener" is actually a micropower 1.2V precision shunt voltage reference, such as the Analog Devices ADR280. It only requires 10µA to operate, so its bias resistor can be fairly large (e.g. 100K or more). You may see some temperature dependencies in the transistor, but you didn't say how accurate the source had to be. Anyway, it might be worth trying.

-Phil

Tracy Allen

Does the device that requires the 90 microAmps of current have to be grounded? If not, it can be wrapped in the feedback circuit of a rail to rail op-amp, shown as the "load" in the diagram.

If the reference is 0.4 volts, and R is a 5 kOhm pot, then when you adjust the pot to 4444 Ohms, the current will be regulated at 90 µA. If the op-amp can swing to within a few mV of its positive rail, then the voltage overhead of the circuit will be 0.4 volts, dropped across the sense resistor, R.

-- The LT6650 is a combination of a 0.4 volt reference and rail to rail op-amp in one SOT23-5 package, and operates on power supplies of 1.4 to 18 volts at Iq of 6 µA

-- If 0.4 volt overhead is too much, it is possible to make the same circuit with a chopper op-amp like the ISL28133 or LT2054, and use an external low voltage reference, say 20 mV instead of 400 mV.
-- If the device needs to be ground referenced, a transistor can be included in the feedback loop, but that adds ~0.5 volt of overhead as seen in the circuit Phil just posted.
-- And if the device to be both grounded and near zero overhead, that can be done with a couple of op-amps, a voltage references and a few resistors.
So, where in the circuit, and how accurate does it need to be?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

BADHABIT

Thanks for the replies guys. I really appreciate them. Sorry it took so long to get back.

For pure simplicity the zener wins. The op-amp looks like a very solid, fairly easy to implement s/u also. I esp. like the sot-23 pkg - can't stand all that big DIP $h!t.

I think I'm gong to use a buck/boost - charge pump and LDO from TI - p/n:TPS61121. From what I can tell I will be able to eek the most amount of juice from the cell while still powering the thing. The Iq is 4uA and it handles 1.8-5.5v

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH