So A forum member "soubitos" was kind enough to give me one of his products to review.
It started when he identified his product solved my issue within this thread.
https://www.eevblog.com/forum/beginners/battery-operated-device-power-supply/msg1427165I had a device, a binary clock which I hoped would run mostly on batteries. Powering it from an 18650 alone was okay, but I risked running the 18650 too low, I also had to disconnect everything to charge the 18650.
So I originally purchased a TP4056 board and a 5V boost converter board. Alas, the boost converter efficiency was terrible although the TP4056 did fairly well.
So the TP4056FlexAdv board arrived in the post from soubitos, all the way from Greece.
It's 5cm x 5cm, quite compact and full of interesting features. It, as the name suggests contains a TP4056 for charging the 18650, but also a cut off protection chip for the 18650. On top of that it has a boost converter with a variable output.
Power input comes from a multitude of connectors, a set of screw terminals, a DC jack and micro USB.
For power outputs the choice is either direct battery connection or the variable boost output.
Charge current is set via a jumper from 60mA through to 1Amp with 8 different selections.
The boost output voltage is set via a 10 (20?) turn potentiometer.
Although I did not test it, you can also enable a remote signal line to control the boost voltage from a 2 pin JST connector. To enable this requires a jumper selection.
So does it work?
Yes. I put it straight into service running my binary clock. The binary clock pulls between 50 and 100mA depending on how many of it's LEDs are lit. With an 18650 (2600mAH generic Chinese laptop cell) and TP4056 alone I was getting just under 2 days, call it 40 hours before the oLED browned out crashing the ATmega328p which was followed a few hours later by the TP4056 pulling the 18650 into protection.
With the TP4056FlexADV and it's boost converter in the mix obviously the screen and LEDs were brighter, but run time fell to around 26 hours. It is hard to judge the outright efficiency from this due to the binary clock drawing different amount of current at the higher voltage, but it's still a lot better than the 5V boost module solution I had previously, which died in 12 hours.
As to it's function with battery management, I powered the setup from the USB port on the back of my monitor which is only on while the monitor is functioning. When it goes to standby or gets switched off when I leave the house or go to bed the 18650 takes over. In this configuration the TP4056FlexADV didn't faulter. When the monitor powered up it began charging the battery, a few hours later it indicated the 18650 was fully charged, when the monitor went off it went back to running the binary clock on battery again. I ran it like this for several weeks without problem.
Here is the circuit with the temporary addition of my multi-meter showing the load current and the USB monitor showing the power supply current. The 18650 in this case has been charging already for about an hour since I got home after running all day.
The good:
* Does what it's designed to do. Stable 5V (I didn't test it at higher or lower voltages)
* Charges the 18650 to around 4.18V on the occasion I tested.
* Cuts out shortly after the oLED display crashes the ATmega when it browns out, though this could be a i2c comms issue. So probably a little below 2.9V.
* Doesn't need any help, support or poking or prodding at to work continuously in service. So ideal to put into a product and leave it to do it's job.
The bad - to be fair there are a few rough edges:
* There is noise on the output. From the FlexADV itself switching noise on the boost module is around 30mVpp at 1KHz, but it seems to pass through any power supply noise readily. I was measuring the same noise on the USB input as I was seeing on the output. Powering it from a noisy switch mode mains adapter showed much higher ripply on the output, but I think my little min scope was getting freaked out by the mains leakage as 4Vpp 50Hz on a 5V supply doesn't sound right, I also measured the same on the power supply itself.
* I might have got an engineering prototype, but the board still had a bit of flex here and there which took the look off it a little.
* The output voltage adjust pot suffers, as any such adjust pot does from not being very linear, it does very little for several turns and then start to ramp up quickly and was quite difficult to get it to settled on 5.00V, I settled for 4.95V.
* The instructions could probably use a "Quick start guide" with bullet points on getting it powered up first time.
* I'm not fond of the bi-color LED, a red and a green separately would look better, but I'm just being picky now.
Anyway, I'm glad soubitos sent me this and I'm happy for it to be the power supply for my binary clock, or possibly other little semi-battery powered projects.
I gather for example it would run something like a ESP8266 happily on a few 18650s for a week. I have not tested it with any ultra low power devices.
Anyway, I believe this is one of soubitos's ebay listings for the module. Which is very cheap for what it is! Definitely worth under a fiver!
https://www.ebay.co.uk/itm/Lithium-Battery-Charger-DC-Step-Up-Converter-Booster-Solar-Li-ion-Li-Po/152960893936?hash=item239d2e03f0:g:6FgAAOSweRhauhJsAny questions, feel free and I'll try and get measurements if you need them.