Multi-chemistry support

[skobkin]

What?

I saw some requests for LiFePo4 batteries.

But I think it would be great to integrate something like CN3795 which support wider options of battery types.

Why?

Because there are regions where winter is very cold (-30-40C) and charging usual LiPo batteries is simply unsafe.

Here come LTO batteries which are not only safer, but also more reliable over time because the can do much more charge cycles.

Of course LTO isn't a universal solution for everything and that's why using multi-chemistry IC is a good choice, so people who live in other places could just use LiPo or LiFePO4 batteries instead.

[K-Kloppstock]

Currenlty working on this, i take the CN3082 in account and make it availible for LI-Ion LiFePo4 LTO and Na-Ion,

sadly it needs an additonal TPS631000 to convert single cell LTO of Na-Ion Voltage into 3.3 Volts, i expect the design to be finished in around 2 weeks and will publish it here

[skobkin]

@K-Kloppstock Such a wonderful news!

Good luck!

[K-Kloppstock]

Update :

I spent some time on how to set up the whole charger and now came to the fitting parts

Charger will be a CN3082 : quite cheap affordable , easy to set up , no external mosfet needed CN3795 sound nice but needs alot more space and parts. and if you need to charger your solar node with 4A something is supposed to be wrong

contrary to my other charger already posted i will not use FS8205a or any other mosfet to disconnect the battery, after a long time of search it came to my mind to use a TPS37000 to monitor the voltage and disable the the buck-boost converter supposed to be TP63031 wich is also easy to implement and offers ~ 92 to 95 % eff. for our field of use.

this design has a slight disadvantage it will consume ~ 12 instead of ~ 6 µA per hour if there is a low voltage lockout, means another 1 mAh will go away each 3 days to keep the charger in sleep, i think this will be okay for solar nodes

another disadvantage : since this will be multi chemistry there will be no dedicated IC for over current / over voltage protection
regarding over voltage : this is not supposed to happen if charger is set correct

overcurrent : maybe i will add a PTC fuse to keep the current below ~ 200 mA, but those NRF52 should not drain that much.

Next Step : Build up a Demo in TINA and do PCB design.

BR

[K-Kloppstock]

So Simulation went well

Specs for Power of / Power Reset , Charge Limit will be

LTO
Off ~ 2,1 V
Reset ~2,13 V
Charge Max ~ 2,69 V

Na+
Off ~ 2,2 V
Reset ~2,25 V
Charge Max ~ 3,97 V

LiFePo4
Off ~ 2,9 V
Reset ~2,95 V
Charge Max ~ 3,39 V

NMC / Li-Ion
Off ~ 3,38 V
Reset ~ 3,43 V
Charge Max 4,12 V

Battery Specs taken from : https://www.batterydesign.net/electrical/open-circuit-voltage/

I have LiFePo4 & NMC @ home, so i will run a full cycle with a voltage logger ( as long mine still works) with them after i finished the PCB
On 18650 NA+ or LTO Cells : These are like ~ 10€ each on Aliexpress, i do not plan to buy these just for testing yet. So someone else need to try it out. Battery select will work with a 4 Pos sliding switch, also a Jumper will be present to switch between 5V & 3.3V Output, just another resistor to add.

I am positive to keep it in 20x40 to 20x55 board dimension with ~ 1mm space on each side for slot in mounting. Also will stick do Standard SOT etc. components so it will be somewhat solderable at home, for TI-TINA testing i choose components with same specs but availible for testing.

Gonna call this charger later : NaToLiFe , i ll keep you in loop for next updates

[oldmanav]

Just thinking, if you wanted to reduce BOM cost, the battery selection could be multiple solder jumpers.

I've seen SMD DIP switches used on other designs for battery selection.

Andy

[K-Kloppstock]

Just thinking, if you wanted to reduce BOM cost, the battery selection could be multiple solder jumpers.

I've seen SMD DIP switches used on other designs for battery selection.

Andy

I Thought about solder jumpers at first too, but i have one voltage divider for charge-stop and one for undervoltage
both needs to be set, i personally think it will offer more access to people without soldering skill if i add a slide switch, still the first Proto will be in THT since its cheaper to order it at ALI and add id myself instead of paying the JLCPCB extra fee for non standard parts.

Meanwhile is started to do the PCB works as Easyeda.

To Community : is there anyone with a LTO or NA+ Cell at hand willing to test it ?

BR

[skobkin]

@K-Kloppstock I'll order LTO from Aliexpress gladly to test, but it'd be problematic for me to order the board production since I'm in Russia. There are local manufacturers here, but they're more expensive than JLPCB. And I'm not qualified enough to do it myself from just a PCB, unfortunately.
But I'll definitely invest some time when later versions would be ready since this is what our local community needs for outdoor deployment in winter.

[K-Kloppstock]

so i finished on PCB Design for now this will do for testing.

Big issue is : CN3082 is not availible at JLCPCB therefore i need to add it myself later. As of now i am not sure if that EasyEDA footprint is correct, therefore i will order the CN3082 first, wait for it to arrive then modify the PCB if needed and order it.

Also for testing i will use as little as possible of extended parts therefore everything i can replace with THT for home assembly will be done.

Will update if I have news

[K-Kloppstock]

Update : Parts arrived, modified PCB accordingly , now ordering PCB

[Wulffi-Germany]

@K-Kloppstock
I can provide some Na Ion cells (18650 type).

May I ask which country you live in?

[K-Kloppstock]

Also Germany,

Meanwhile 4pcs of LTO and Na Ion are on the way Via AliExpress. Nevertheless thanks for your offer. I also got my hand on a USB DAQ therefore I can measure charge and discharge curve as soon as the PCB arrives and I have it assembled

[SimonRZz]

any news on this?

[K-Kloppstock]

Yes :
-> PCB arrived, need to add CN3082 by hand on weekend.
-> DAQ : https://de.aliexpress.com/item/1005008925558504.html?gatewayAdapt=glo2deu
got his card, wrote a app in Lazarus to get a measurement every second into a csv file.
App is now doing a 4h testrun to see if it works fine

Will solder CN3082 on weekend and will do testrun with first chemistry

BR

[K-Kloppstock]

Update : i royally fucked up my circuit . my first design included a LM393 for voltage level checking then i moved to TPS3700DDCR , but still kept the connections for LM393 therefore : i can test if charging works the way i want, but not the 5v or 3,3 v output.

Edit 1 : i think i can manage to fix it with wire bride, currently testing lifepo4 chargeging

Edit 2 : LiFePo4 Charg works just fine

For Anyone interested in that software : https://github.com/K-Kloppstock/ADS1256-STM32-readout/