0
\$\begingroup\$

I am looking for a linear regulator that would follow these rules:

  • input voltage from 3.3 - 4.2. (18650 battery or x3 AAA batteries)
  • output voltage 3.3V @ 200 mA max. (for ESP8266).
  • small footprint - ideally would not require extra components except capacitors

Main concern is that in most cases Vout = Vin + 1V which would not work at all. (What is the name of this characteristic, what should I be looking for ?)

I'm looking for a linear regulator that would allow to get as much juice from battery, idealy Vout=~Vin or some small value that is would not increase price.

P.S. I'm aware of switching / DC-DC / buck regulators, but schematic would contain as few components as possible and would not require all the benefits from them.

P.P.S regarding load and discharge rate. According to picture below, discharge rate would be something like 0.15C, so 80% of battery could be used before dropping voltage to 3.6 Volts.

enter image description here

\$\endgroup\$
8
  • 2
    \$\begingroup\$ The name of the characteristic is dropout voltage. There are many linear regulators that can support a very low dropout voltage. It is not possible for a linear regulator (of standard design) to support a true Vout = Vin as the pass element will always have some voltage across it. \$\endgroup\$ May 12 '20 at 11:30
  • 2
    \$\begingroup\$ If you require 3.3 volts out with only 3 volts in then it's not a linear regulator. Be clear about what you need please. Output current is required and not mAh - mAh is related to battery capacity and is irrelevant for the spec of a regulator. \$\endgroup\$
    – Andy aka
    May 12 '20 at 11:33
  • 2
    \$\begingroup\$ If your output needs to vary from less than your input to more than your input you need a switch-mode regulator. Take a look at 'buck-boost' devices. Specific shopping recommendations are off topic here but there are several available. The circuit will consist of an inductor and switching device several off the shelf modules exist. \$\endgroup\$ May 12 '20 at 11:51
  • 3
    \$\begingroup\$ The 3.3V input for the ESP8266 is in the range of 1.7V to 3.6V according to some online documentation; why not use a linear regulator with the output set at 2.7V which would require a dropout voltage of <= 0.3V @ 200mA. Search manufacturers parametric tables and I am sure there will be plenty of them to choose from. \$\endgroup\$ May 12 '20 at 11:58
  • 2
    \$\begingroup\$ When you say 200 mAh max, do you really mean 200mA max? \$\endgroup\$ May 12 '20 at 11:59
1
\$\begingroup\$

1) What voltage does your chip require?

Looking at the datasheet page 4, the analog parts of ESP8266 require 2.5-3.6V and the digital/IO parts require 1.8-3.6V. Thus it will work with a minimum voltage of 2.5V or 3V.

enter image description here

So you can use a 3.0V LDO or even a 3.3V LDO which will run in dropout mode (without regulation) when battery voltage drops below what is required to achieve 3.3V output. It is best to select a PMOS based LDO instead of a PNP one as the latter will have increased ground current in dropout when the pass transistor saturates. A PMOS based LDO in dropout mode will behave like a resistor, the RdsON of the internal FET.

Another feature you may find useful is a Power Good output to turn off ESP8266 when battery drops below 3V to protect it from overdischarge, but you may also do this in software if you use the ADC.

2) Power dissipation

With 4.2V in, 3.3V out, and 200mA current it will dissipate 180mW so a package capable of high dissipation is not necessary. A SOT-23 or SOT-89 package will be fine, small, and easy to solder.

3) Dropout voltage and regulation

Since ESP8266 will still work at 2.5V, for a 3.0V LDO you should aim for 2-300mV or lower dropout voltage at 200mA, leaving some margin for transient output voltage drop when the chip draws pulsed current. A fast transient response LDO will allow a smaller output cap.

Other criteria: fixed regulator since you say you want minimum parts count, low dropout, low idle current for battery life, stable with ceramic cap for low parts count and small size.

I've entered all this stuff in the mouser search and we have some candidates. Feel free to look at the list.

For example TLV70030 fits most criteria and has a fast transient response, you can check on the graph that the output voltage stays in the allowed range when output current is pulsed.

enter image description here

Some ultra low quiescent current LDOs look like they'd work but if you check the transient response, you will notice they are too slow so output voltage will drop below 2.5V unless you use a huge output cap. For example this is MCP1700, with very low 1.6µA quiescent current:

enter image description here

Notice the output voltage is displayed at 500mV/div, so it drops by 500mV on a 100mA output current pulse. Also the scope is set to 200µs/div. Compare these values to the TLV70030 above, for this one the scope is set to only 10µs/div and 50mv/div. So the curves look somewhat identical, but the result is completely different: one LDO is at least 10x faster than the other (at the cost of higher quiescent current), which means the output voltage stays in range with a tiny output cap. When the load draws current pulses, always check transient response with a scope. If supply voltage gets out of allowed range, your microcontroller will misbehave, crash or reboot.

A LDO with input voltage vs quiescent current curve like this should be avoided as it will draw extra current when the battery is discharged:

enter image description here

NCV8114 doesn't have this problem, it should be a good choice too.

enter image description here

So as you see it isn't all about dropout, it is about keeping the output voltage in the allowed range considering input voltage and output current variations. Please double check the rest of the parameters before choosing a chip, of course.

\$\endgroup\$
2
  • \$\begingroup\$ Can you please check table below if I missed some crucial parameters in selected components that would prevent them to be used in my solution ? Thank you. \$\endgroup\$
    – saikek
    May 12 '20 at 21:05
  • 1
    \$\begingroup\$ They should all work fine. Personally I'd go with TLV70030 because they specified transient response properly, ie step from 0mA to 100mA at 10µs/div whereas the others step from 10mA to 100mA at 500µs/div which means there is nothing interesting to see on the graph... If you want to save 7c you can use the cheaper ones too. \$\endgroup\$
    – bobflux
    May 13 '20 at 10:12
1
\$\begingroup\$

Unfortunately, there is no Vout = Vin voltage regulator. Because regulator is an active component and it also requires an energy to operate. Therefore, you may use DC-DC convertor before a low dropout voltage regulator.

\$\endgroup\$
0
\$\begingroup\$

Thank you all for comments - I've ended up with this table of regulators from LCSC inventory, that might actually do a job (in case that it would be needed for anyone):

Also, I'm already using LiPo protection provided by XB3303A (reverse connection, overcurrent, etc), so it would take care of battery.

I would probably be using TI solution, recommended by @peufeu.

enter image description here

\$\endgroup\$
0
\$\begingroup\$

Instead of an LDO you could look at a step-down converter.

some of these, such as the TPS62840 have a feature called 100% mode, where

The duty-cycle increases as the input voltage comes closer to the output voltage. Once the input voltage decreases to near 100% duty cycle, the output voltage set point is increased by +30 mV. As the input voltage decreases further, the device enters 100% duty-cycle mode and keeps the high-side MOSFET on continuously.

Which means the only drop you will see is the drop over the internal resistance of the converter.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.