10 pin Li-Po battery connector experiment

A place to talk about all the Operating Systems, Software & Hardware
User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Mon Jan 07, 2019 10:58 pm

I am testing the 10-pin connector as documented recently.
First, here is wiring (see pic)
pin 1, 2, 3 goes to Red wire as positive (I used 20 AWG silicone wire).
I use 60W soldering iron to do the wiring (PCB's big copper backplane spread the heat too fast if using smaller soldering iron.)
and direct-soldering on the socket side to avoid tiny components.

pin 8, 9, 10 go to black wire as Ground.

I then soldered a 10 K ohm resistor on the 5th test pad (which is NTC) and Gnd (black wire) as the NTC sit-in.

The unused two pins (SCL and SCK) are for battery's fuel gauge IC. LP's doc does not use it thus I don't use them yet.
Attachments
20190107_141.jpg
20190107_141.jpg (380.7 KiB) Viewed 2851 times

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Mon Jan 07, 2019 11:09 pm

The Red-Black pair goes to my lab power supply so I can start the experiment.

I found a slightly different behavior than LP Alpha's USB type-C power input or DC (4-pin) power input.
- if supply voltage is kind of low (say, less than 10V), no LED on LPA will lit;
However, if I briefly push the power button, the Red LED will blink (approx 3 times a second) for about 6 seconds then shut itself off.
If I press the Power button continuously instead of briefly, I will see the stable Red LED light and Blue LED light coming up, which means LPA is booting up.

How low such power supply can be?
I found at 5.2V input, I can reliably trigger the booting up sequence (goes to UEFI prompt, I don't want to play live OS at such marginal operation condition.) Note that this is not really good since at such low voltage, the current draw will be very high. I.e., not really for stable, reliable operations.

I had tested 7.4V, 9V, and 12V as the power feeding into the LPA. They all work fine. (Sorry, did not test the 15V but according to LP's documentation, it should be fine.)

The above should help folks doing the Li-Po (with its own BMS) design.

Please note as input can go as low as 5.2V (actually after boot, I had tried bringing it down to 4.6V and it still works. I wouldn't want to say "reliably".)
This means if using your own Li-Po battery pack, itself much have a under-voltage shutoff function.

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Mon Jan 07, 2019 11:14 pm

One last test, can such connector be used for charging the Li-Po battery pack?

I am using the original LPA's USB type-C PD power supply to supply the power to LPA board.
Then I tested the Red-Black pair of wires. It shows stable 8.4V.
Bingo, it is the unloaded 2s Li-Po charging voltage (when 2s batteries are full.)

I did not investigate further to see if there is any charge control circuit in LPA board (such as minimum the constant current.)
I am not sure LPA has a full 2s LiPo charging controller (in fact I doubt it has, or the CC value is set fairly high.)
I would assume the charge controller probably should be located on the battery pack side (as part of the BMS/charge control combo.)

The concludes my findings.

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Tue Jan 08, 2019 12:42 am

I'd like to emphasis again:

This 10-pin Li-Po in connector is meant for
well-protected (BMS + balancer at minimum) battery pack, not for raw Li-Po/Li-Ion/LiFePO4 cells.
(e.g., latter setup will drain below the minimum Li-Po cell voltage which will significantly/irreversibly degrade the life expectancy of cell.)

Also charging without proper CC, constant current, (then later CV) charging circuit is outright dangerous.
< edited to clarify >
CC's value, if set very high, will not be suitable. How to set (by a resistor or by a register? I don't know.)
This is set by laptop factory, since it has the full control of the laptop motherboard and the laptop Li-Po pack.
For maker board like LP Alpha, it's a variable.

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Tue Jan 08, 2019 2:46 am

Adding some partial info and waiting for people who have time to drill down and thoroughly exercise the battery options.

1. If I put a 2s Li-Po in that battery connector, indeed it can be charged, it seems to be no CC or CC value set very high
circuit on LPA board itself (need to study more.)

2. Try to check the dual/concurrent power option and tried to remove one of the power sources to see it causes PC to hang. It seems to be okay for low and (up to) medium current drawn applications. Under heavy load, don't know.
It's sort of like the laptop (with built-in battery pack) behavior.

3. Note that (I am running Win10 Pro) Windows detected the battery power in place and seems to
trying to calibrate battery capacity and energy flow-in/flow-out.
Without battery pack providing fuel gauge info thru I2C (that SCL and SCK lines unused in LPA doc), the estimation may be pretty crude.

4. Power source from Li-Po 10-pin, appears to be backfeeding the DC-in 4-pin connector (with 0.6V voltage drop.)

This entire area should be a factory design topic, I am simple an observer lightly scratching the surface.

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Tue Jan 08, 2019 4:36 am

There indeed no CC circuit (or CC value set too high for my test condition) for Li-Po charging so it is battery pack's own responsibility (or someone needs to investigate how to set CC per your own battery spec.)
Actually, I found the power drain (if battery is being charged* while LP Alpha is operating) if
using DC power-IN (or even 15V 3A type-C PD) can be too much (this is wiring dependent) to affect the system stability.

* in one case, I've seen 12W or more being drawn and just used to charge the battery itself. This is pretty high.

Anyway, careful planning and understand what you are doing is ultimately important.

User avatar
Lutz
Posts: 211
Joined: Mon Apr 18, 2016 4:30 am
Location: Germany, NRW

Tue Jan 08, 2019 9:56 am

ccs_hello wrote:
Tue Jan 08, 2019 2:46 am
(...)
3. Notice (I am running Win10 Pro) Windows detected the power in existence and seems to
trying to calibrate battery capacity and energy flow-in/flow-out.
Without battery pack providing fuel gauge info thru I2C (that SCL and SCK lines unused in LPA doc), the estimation may be pretty crude.
(...)
That would suggest that this circuit
https://www.lattepanda.com/topic-p26075.html#p26075
is supported by "Windows10".
Why isn't the battery already available? Or at least enough documentation ?
Would be a purchase argument for me if I didn't have to test or even develop all this myself.

User avatar
Boring
Posts: 88
Joined: Sat Apr 16, 2016 7:14 pm

Tue Jan 08, 2019 11:23 am

Cool, so it is possible to hook up a battery and use it as an UPS?

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Tue Jan 08, 2019 1:41 pm

Boring wrote:
Tue Jan 08, 2019 11:23 am
Cool, so it is possible to hook up a battery and use it as an UPS?
Poorman's version. Similar to a laptop (with built-in battery pack) behavior.

User avatar
ccs_hello
Posts: 410
Joined: Sat Oct 13, 2018 9:16 pm

Tue Jan 08, 2019 1:53 pm

@ Lutz
re: that specific laptop battery pack's circuit
https://www.lattepanda.com/topic-p26075.html#p26075

A laptop factory can tune the entire system
- set and read the battery fuel gauge IC
- preset the battery charger's CC value (if charger is located on laptop's motherboard)
- integrate these and be supported with Windows (or Linux) < -- may be there is a low level driver design/configuration guide for that specific OS

LP Alpha is a maker board and the user can source his/her own battery pack (with various spec. from multiple models.)
This is why I mentioned end user will have to know and design his own.
The cited post only suggests a battery pack "hardware wise, from wiring point of view" (sort of*) compatible with LP Alpha's 10 pin connector.

* The battery pack has I2C interface for fuel gauge IC but LP Alpha's doc does not show it's being taken advantages of.
(The LP Alpha PCB seems to have circuit paths and two protection (or pull-up) resistors for that so it's still a secret, work in progress, or wait to be reverse-engineered area.)

Post Reply