Author Topic: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy  (Read 1946 times)

0 Members and 1 Guest are viewing this topic.

Offline mccTopic starter

  • Contributor
  • Posts: 10
  • Country: de
Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« on: October 21, 2018, 09:54:43 am »
Hi,

I bought one of these:
https://www.aliexpress.com/item/1PCS-Raspberry-Pi-Zero-UPS-power-board-integrated-serial-port-power-detection/32893660585.html?spm=a2g0s.9042311.0.0.323e4c4ds8xUsP
for my Raspberry Pi Zero W.
It is partly documented here:
https://github.com/linshuqin329/UPS-Lite
I translated the chinese spoken manual via google and some guesswork to english but didn't find a answer to my question.
The circuit consists of a MAX17040 fuel gauge, a SIL CP2104 serial-to-usb interface chip, a Fitipower FP6717 boots converter chip (with
supposedly about 90% efficiency...) and a mysterious little 6pin SOT23-6.
The according datasheets of the known chips doesn't tell anything of a charging functionality.
I did a measureing of the current and voltage which flows while charging the LiION cell (LiION not LiPoly according to the writing of the cell
itsself. I did the current measureing with a USB "charging doctor" (nothing fancy) and voltage measureing with a BRYMEN BM867 at the cell
itsself.
The voltage rises to 4.23 V and stays there and the starting current is about 500mA which slowly falls down to 0.5mA
in the second half of the circle. Looks quite normal. 4.23V looks a little high for my beginners eyes, though.

My question is: What is this SOT23-6 6-pin something on the PCB. A transistor? A charging IC (one pin goes directly
to the (+) of the LiION cell).
Or what does charging?

Another question:
One can read via I2C the SOC and the voltage as measured by the MAX17040.
As fas a I know, it is recommended to charge up to 75% and discharge no further than 25% charge of the cell.
The voltage measured by the MAX17040 and what I measure with my DMM match.
And the amount of mAh I have to recharge to reach 0.5mA cutoff current and the reported SOC match.
But SOC and voltage reported by the MAX17040 compared to a Voltage vs. SOC table I found on the
internet does really not match.

Is this Raspi power supply any good?
Or am I badly wrong at some point with my assumptions?
Is there any reasonable SOC vs. Voltage chart for LiION-cells?

I have attached the fotos I made from the PCB.

The datasheets are here:
https://datasheets.maximintegrated.com/en/ds/MAX17040-MAX17041.pdf
https://www.silabs.com/documents/public/data-sheets/cp2104.pdf
https://pdf-datasheet-datasheet.netdna-ssl.com/pdf-down/F/P/6/FP6717-Fiti.pdf

Thank you very much in advance for any help in exploring and evaluating this circuit!
Cheers!
Meino

 

Offline kjr18

  • Regular Contributor
  • *
  • Posts: 197
  • Country: pl
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #1 on: October 22, 2018, 10:06:12 am »
Indeed this sot23-6 looks like TP4057 or some other clone, set to 400mA charge current. These chips are very simple but quite good li-ion cc\cv charge ic's. You can try to compare your circuit to what shows up in datasheet.
 
The following users thanked this post: mcc

Offline mccTopic starter

  • Contributor
  • Posts: 10
  • Country: de
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #2 on: October 22, 2018, 07:17:18 pm »
Hi kjr18,

WHOW! The first "I understand this" on my side here...! Great !
I used the continuity tester of my DMM to verify the schematic in this datasheet:
https://datasheet.lcsc.com/szlcsc/Nanjing-Extension-Microelectronics-TP4057_C12044.pdf
The only difference seems to be, that the LEDs on my board get an R extra each to
limit the current...which seems reasonable for a LiPo-powered board.

The next, what I need to understand is the MAX-chip.
Currently I am in doubt, whether the value read via I2C is the percentage of charge
remaining in the LiPo (this is - I think - what the datasheet says) or the amount
already dicharged from the LiPo.

I definetly need some valid chart, which shows a voltage <=> SOC dependency.
As I understand LiPo (and that does not mean much) the battery voltage
measured to estimate its SOC is the voltage without load....
Correct?

Does anyone has experiences with this MAX17040?
Is he worth it or ... ?

Thanks a lot again for any help in advance!
Cheers!
Meino

 

Offline stmdude

  • Frequent Contributor
  • **
  • Posts: 479
  • Country: se
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #3 on: October 23, 2018, 05:06:50 am »
I shipped a few million devices with the MAX17040 in it.

If you're talking about the "SoC" registers in it, it's the percentage of charge _remaining_ in the battery.

Please note though. The MAX17040 works in mysterious ways, and is powered by bullshit and unicorn farts. It's not a columb counting device, hence it's really just guessing. It's pretty good at guessing, but it's still guessing.  (Hint: It's just reading the voltage, and what happens to a battery when you apply a load?)

Also, the MAX needs a few complete cycles to self-calibrate itself before the SoC readings are somewhat reliable.

Lastly, if you're going for maximal accuracy, the MAX needs to be calibrated (not self-calibration, manual calibration), which if I recall correctly is done by Maxim. (Its been a few years)
 
The following users thanked this post: mcc

Offline mccTopic starter

  • Contributor
  • Posts: 10
  • Country: de
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #4 on: October 23, 2018, 05:28:43 pm »
Hi stmdude,

Thank you very much for your help! :)

Currently this happens after three charge/discharge cylces:
When I start the Raspi and let it run idle ("running idle"-what
did I wrote there......? :)
I mean: Only the kernel is running and a little script doing
some I2C to get the voltage and the percentage off the chip
from time to time) this happens:
The voltage level falls and I can verify the value with my DMM.
So no problem here.
But the percentage rises first and falls then but towards the
point I would shutdown/switch off, the percentage is way to high.
I use this table (posting #6):
https://www.rcgroups.com/forums/showthread.php?2366548-LiPo-80-rule-and-my-shallow-charge-rates
According to that chart, 3.69V is about 10% left charge.
But my board reports:
++++++++++++++++++++
Voltage: 3.65V
Battery:   30%
++++++++++++++++++++

You wrote, calibration is done by...MAXIM? So...do I need to send
them my board? Is a joke...or? OR? ;)

What voltages would you suggest for a LiPo cell, which is currently
be charged AND slightly under load (95mAh - the linux is running reading the
voltage and SoC), has 1000mAh and is supposedly a Li*Po* cell despite
the writing on it saying "LiIon", as upper and lower limits to not to
do harmful things to the lifespan of the LiPo?
How many cycles -- roughly guessed -- I need before the Max17040 from
"wildly guessed mode" to an elaborated guess of the SoC?
And why LiPos are such delicate beasts?...;)

Thanks a lot in advance for any help!
Really appreciated! :)
Cheers!
Meino





 

Offline stmdude

  • Frequent Contributor
  • **
  • Posts: 479
  • Country: se
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #5 on: October 23, 2018, 06:11:09 pm »
Ok, lots of questions in there.. :)

So, the rule of thumb is: Never go below 3.0V unloaded on a LiPo cell.

Since the unloaded voltage will be higher than the loaded voltage, you can run your battery all the way down to 3.0V, and when you stop drawing power, the voltage will rise above 3.0V.

A medium quality LiPo has about 10mOhm internal resistance. Meaning, if you pull 95mA from it, you will get 0.010*0.095 volts of drop. (<1mV)

As for the 17040, three cycles should get it somewhat accurate. The calibration from Maxim is not required unless you need higher accuracy, and I don't think they offer it unless you ship a lot of devices.

As for 30% SOC @ 3.65V, it sounds somewhat reasonable actually.  I'd run it down to 0% and keep an eye on a voltmeter to make sure the battery doesn't drop below 3.0V.

As for the RCGroups sheets/charts, since they're not specifying the internal resistance or the load they're putting on the cells, they're kind of useless to look at.

As a bonus: Keep in mind that LiPos self-discharge. So if you run a cell down to 3.0V or close to it, you should put some charge into it fairly soon.
 
The following users thanked this post: mcc

Offline mccTopic starter

  • Contributor
  • Posts: 10
  • Country: de
Re: Exploring a LiPo-Charger/Booster/Fuel-Gauge-thingy
« Reply #6 on: October 23, 2018, 06:46:10 pm »
Hi stmdude,

thanks a lot for the info material and the bonus track...I like it! ;)

...that's why I posted it under "beginners"...if my question would contain words like "bosons",  "quarks", "myon" and "LHC" I more
likely had posted it under "quantum physics"...but doubt to get any answer there... ;)
Back to the topic...
I read some articles about "extending the lifespan of your LiPos" and such and always misses any information, whether the given
values for different voltage levels are measured under load or without it. Additionally often there is no time stamp on such documents.
LiPos did evolve over time I think...
Those documents say: Discharge down to 3.5 volt (25%) and charge up to 75% (no voltage found here, I think because it is in the CV area
of the charge process, hehehehe).

Will the difference that great: Under load 3.0 V and without load 3.5V

Cheers!
Meino

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf