I have only 20mA in my circuit and after trying to charge a battery with TP4056 it doesn't work. TP4056 is a linear charger so I need to try with LTC4002 or even better to improve efficiency. Anyway, my first attempt was to implement a charge pump like this:

My goal is to obtain 100mA or more (8v) pulses from 20mA @ 8V input. To do this, I used a supercapacitor connected to a hysteresis comparator so when Vcap is >5V output is enabled, so pulse can be done and feed the li-ion charger, and when Vcap is bellow 1V this output is disabled so supercapacitor can charge again.

This is the schematic: schematic (Some resistor values was changed over the tests...)

and this is the expected output: expected output

I tested it with a load and circuit seems to be successfully enabled, but if load is present seems like is not able to disable comparator output and some weird waveforms happens.

Maybe this schematic is not the best approach to solve the problem: any suggestions?

------- EDIT -------------------------------------------------------------------

This power supply is for a GSM device with peaks of 2A but average of 60mA. The device will sleep some days and some hours a day, so can be "recharged" in this moments. Supercapacitor can do the job of charge/discharge, but since consumption is higher that the one who a supercapatitor can give, I considered a battery.

Solar chargers have similar pulsed charge currents, but higher than mine.

  • \$\begingroup\$ If your source can only supply 20 mA, you can't get an average charging current of more than 20 mA, and you'll probably waste a significant amount of power in your pulse charge circuit, leaving even less current for charging the battery.. \$\endgroup\$ Nov 8, 2019 at 17:05
  • \$\begingroup\$ I dont want to get more than 20mA average, but peaks of 100mA. Maybe pulse charging circuit waste a significant amount of power.. any suggestions? \$\endgroup\$ Nov 8, 2019 at 17:09
  • \$\begingroup\$ Please don’t! Why not boosting the voltage and limit the charge current, and probably use some other charge IC? Charge termination will be a challenge. \$\endgroup\$
    – winny
    Nov 8, 2019 at 17:26
  • \$\begingroup\$ boosting the voltage gives me less current... ¿? LTC4002 is a switch mode charger instead linear, so it can be better, but I don't know if support low current charging. Charge termination needs constant current, right? \$\endgroup\$ Nov 8, 2019 at 17:45
  • \$\begingroup\$ Sorry, buck. I mean buck. Sort of. Read the datasheet for the battery. It’s usually detecting when the current falls below 100 mA in CV mode. \$\endgroup\$
    – winny
    Nov 8, 2019 at 18:07

1 Answer 1


I think pulsing charge to a Li-ion battery isn't a good idea, since they have a limited number of charge cycles. I also doubt that pulsing a high current gives faster charging than constantly charging at a current your power supply can provide.

The datasheet for TP4056 says "absolute maximum rating -0.3 - 8 V", so I wouldn't recommend to power it with a nominal 8 V that might fluctuate higher, if that's your case. The datasheet also specifies a programmable current between 130-1000 mA. Since you only have 20 mA @ 8V, I'd look at some other charger that can provide lower charge current, for example the common MCP73831.

I would also look at some regulator to get 5V from your 8V. The choice depends on the requirements of your project. Some reasonable choices are a linear regulator, a ready-made switching module (such as from Pololu) or a switching regulator IC with your own supporting components. A switching regulator would be most efficient in your case.

This suggestion to regulate the voltage plus programming a charger to stay within your power budget will give a simpler and more reliable solution than your original idea.

Alternatively, LTC4002 seems to be more efficient, as you say. At a first reading, I don't see what charge current limit that can be programmed. EDIT: The current is limited by Rsense, with the formula I = 100mV/Rsense. I don't see what limits there are and I'm not clear what will happen if LTC4002 tries draw more current than available.

  • \$\begingroup\$ MCP73831 is a linear mode charger. I choose LTC4002 because it is switch mode charger and will be better than TP4056 linear mode charger. It this concept works i will try to change (again) to BQ25713 because it is buck-boost (highest Vin range) and includes protection and gas gauge into the same chip. But I dont know if this is the best way to solve my problem because infinite low current charge or pulsed normal current charge. \$\endgroup\$ Nov 8, 2019 at 17:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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