Dropping 100mV exactly

[elektrinis]

So I have this project, where I need to drop 100mV (±20mV): from 4.3V to 4.2V.
Current around 50mA, only one-way.
Max Iq= 50μA.
The drop should stay within ±20mV in wide current range, from 1μA to 50mA.
One extra problem: free space is very limited. I have space for maybe single SOT23 package...

Was thinking maybe find an opamp with BJT output with correct rail-to-rail drop, but could not find a suitable one.
Was also thinking about series voltage references and LDOs, but no luck as well.

What are my options??

[JPortici]

Check out TS922

[elektrinis]

Check out TS922

Iq = 3mA ...

[JPortici]

uh! i forgot about the Iq

[RES]

A difference amplifier (high current >=100mA opamp) plus a multi turn pot or adjustable voltage reference?

AD8603

[magic]

I don't think PNP will get you that close to the rail. PMOS is the way to go, driven by an opamp, or just a CMOS opamp with rail to rail output if you can find a suitable one at such low Iq.

To derive the Vin-100mV reference:
For 20% tolerance you could probably just use a 43:1 resistive divider to ground if you trust your input voltage to be right.
Otherwise, LM385/LM285 are 10µA shunt references - hang one from Vin to ground with a suitable load resistor in series, divide to get 100mV.

[hli]

Whats the important specification here? Dropping exactly 100mV, regardless of the input voltage (e.g. deliver also 3.2V from 3.3V)? Or deliver 4.2V with an input as low as 4.3V? For the latter, look for an adjustable LDO with really low dropout (e.g. a LTC1844 - SOT23, needs two resistors and maybe two caps, depending on what you already have in your circuit).

[elektrinis]

Good question. Yes, I need to limit the voltage to 4.2V regardless of input, so was looking in to LDOs. Would love to avoid passives, but might be able to fit a couple of 0402s...

[SiliconWizard]

Is it part of a LiPo battery charger or something?

[elektrinis]

Yes it is. There is a product designed to use 3.7/4.3V battery, but those are not available in consumer quantities and that size. Modification of that product is another route, but it needs firmware hacking and I'm looking in to it separately.
I have already found an ideal diode in sot23 package for discharging mode, and now need something for charging.
Will look deeper in to LDOs.

[SiliconWizard]

That looked like it.

Finding an appropriate LDO is tough. I've found a few that meet some of your requirements (fixed 4.2V output, capless, low iq...) but never quite all. Some have higher iq than 50µA, some have higher max./ drop-out voltage than 100mV, and most of all - most won't fit your +/-20mV requirement... but you may eventually find something.
(Edit: example: LDCL015 from ST. Higher iq and voltage accuracy that's probably not good enough.)

One point that comes to mind here is the behavior of the charging circuit. I suppose you know how LiPo chargers work: constant current phase/constant voltage phase. Adding an LDO between the charger's output and the battery may confuse it in a couple of ways. That may be workable, but certainly something to consider IMO.

[hli]

If you actually need a LiPo charger for a lower voltage, why not looking at LiPo charger circuits? Since you are worrying about voltage drop, maybe a LTC4071 can be an option? Its a shunt-based LiPo charger.

[elektrinis]

The product already includes a fancy firmware-controlled charger, so adding one more on top would result in some unpredictable behavior. I have found an LDO with Iq below 1μA, will try that for now.

[exe]

I have already found an ideal diode in sot23 package

Interesting, what was the part number?

[elektrinis]

Ideal diode: MAX40200

[SiliconWizard]

I may have missed something, but the MAX40200 is not an LDO, it's an ideal diode. Besides, its quiescent current is given at 7µA typ. (not under 1µA). And how are you going to drop exactly 100mV with it?