DIY portable, useful LiFePO4

[markopolo]

I've had confirmation from the vendor now that using B-,B1,B2,B3 and B+ does the same as using BC0,BC1,BC2,BC3 and BC4 on this particular BMS. I think B-,B1,B2,B3 and B+ is the better way.

------------------
Just some additional notes about the BMS that might turn up in internet searches and help others who want to purchase or already have this and other Smart BMS made by Jiabaida ( DGJBD or JDB ). These BMS use the App with the Baby Elephant icon ( Xiaoxiang app ) and the PC program JBDTools ( JBD Tools ).

The most recent versions of the App with full access and the PC program are available on jiabaida.com - View the site through Google Translate. A view only version of the App is in the Play Store and Apple's equivalent store. You might have to pay for the Apple version but I think it gives full access to settings.

There are many versions of the App and PC program out there going back for years I think. Newer versions of the App give you full access and let you set a password but be warned that setting the password may prevent the PC program from being able to read the EEPROM. It is fixable. You just need a version of the PC program that also lets you set the password there. If the App gives you full access and the PC program won't read the EEPROM then the password is most likely the problem. Many users won't need the PC program now. You did in the past because some settings were only accessed through the PC program. You would need it to load firmware if ever required and you can save the EEPROM programming and restore those settings.

I purchased from XJR Technology - http://www.rjxzstech.com/Products.html . That's RJXZS Tech through their AliExpress RJXZS store - https://www.aliexpress.com/store/1821822

Order processing was fast, packaging was adequate, shipping was fast, communication has been fast and complete. Shawn is multilingual and provides clear instructions when asked. I've asked for help twice and received it promptly. I'd have no hesitation purchasing from RJXZS again.

[israndy]

What BMS are we talking about, I know the manufacturer's link, but which one? And where did you find the software? I know the version numbers, but where was the software actually downloaded from?

Interesting thread, but very difficult to follow w/o specific info.

[markopolo]

The BMS is on the last post of the first page of this topic. It'll be obvious once you've seen it: Products Check them all out though some might be better for your particular needs.

I'm not at the computer that has things bookmarked. Just use Google Translate as I suggested or even just plain old Google for the Apk. Jiabaida Xiaoxiang apk- it's not hard, don't give up. Remember the domain name. Use your antivirus to scan what you download.

No point in posting a direct link as it could change at anytime - might as well learn how to actually find it. You'll no doubt need to know how to find it in the future if you are heading down this path.

You'll need some type of apk explorer to see real version numbers as they're often not in the filename. That's how you'll really know what's old and what's new.

Good luck!

[israndy]

Originally Posted by markopolo

Good luck!

Aaaaaaand... goodbye, geez!

[markopolo]

Good luck to everyone doing a project like this. I mean it. You're going to need it.

Yesterday was a project low point with news that the second order of batteries was refunded after over 60 days wait. The first order was previously refunded. Probably 65 days since initial orders were made and no batteries. It was looking like I really shouldn't have done this.

Later in the day another order did show up. It's only 20% of the total needed but at least I finally have some cells to take actual fitment measurements and do some testing etc. Things are looking up.

Then reality hit again. It's going to be a long process charging these cells and testing them to see if any are bad and having to prove that to try to get a refund etc.

Then I started wondering if the 10 chargers I made will work or go up in smoke or damage the batteries .........

There's always a bright side to things though. The way items trickled in did let me slow down & do more research etc. I'll probably get a bit more excited later today if I get some batteries charged up nicely. Fingers crossed that the capacity of the cells turns out better than expected - that would give me a boost for sure.

[markopolo]

Here are the initial steps I'll do to process these cells:

1. Video opening the packages just in case the wrong batteries are received.
2. Number them to keep track of them.
3. Apply insulation / isolation gaskets on the positive end of the cells. That's where a short is most likely to happen I think.
4. Charge them using the automatic charge controllers I assembled.

All of the batteries received came at a low state of charge (3.225V). That's to be expected and it's a good thing - it means the seller likely following IATA transport rules.

All lithium ion cells and batteries (UN 3480 only) must be shipped at a state of charge (SoC) not exceeding 30% of their rated capacity.

The ten tiny charge controllers are working as they should! 6 are CN3058E chip controlled and 4 are TP5000 chip controlled. I'm using the 5V rail from a PC PSU for power.

Charging is automatic so I don't have to sit and watch them. A Green LED light will indicate a charged cell. The CN3058E boards had LEDs on the board. I had to solder the LEDs onto the TP5000 boards Charge termination and resumption is automatic.

I expect the cells will take 4 or 5 hours to fully charge. Current is almost at an amp for each charge controller. Termination voltage will be between 3.6V & 3.65V. I'm hoping for the upper range but expect that there will be some variance between modules.
gasket.JPG

CN3058E.JPG

CN3058E_.JPG

TP5000 4.JPG

TP5000.JPG

[ikanode]

Did you build or buy the tiny charge controllers?
If built, would you consider making the circuit board files available? (Although, I'm not sure I can solder a board that small.)

Thanks,
Irv

[markopolo]

Bought on AliExpress using a few different vendors. I think they're on Amazon also.

The TP5000 controllers include the LED in the package but you have to solder it on. The CN3058E controllers come fully assembled.

I've charged 12 cells now. The TP5000 boards seem to fully charge the cells. Those cells were at 3.4x volts after an overnight rest. The CN3058E boards get the cells close to 100%, maybe 99% or 99.5% (my guess). Those cells were 3.38 or 3.37 volts after an overnight rest. The difference probably doesn't matter though.

My meter seems to be accurate (enough). It's close to what the EDB-M05 tester shows on my PC. EDB-M05 3.318V, my multi-meter 3.314V when I compared the two today.

For this project the TP5000 boards were the slightly better choice IMO. If I was caring for a rechargeable flashlight for example then the CN3058E boards would probably result in longer battery life over time (not per use).

The CN3058E boards came with pins attached so I used some old PC speaker connectors I had. I do have to revisit/redo those connectors though as one or two were loose and caused intermittent connectivity. I'll try bending the pins a little bit. The CN3058E based boards were much easier to get ready for use.

[markopolo]

Cell #1 is on the tester now for a full capacity test and it's probably/hopefully a 24hr+ test. I numbered the cells as I unpacked them. The test is 0.25A load with a 2.75V cut-off.

40 cells each contributing 0.25A would be a 10A pack load. I think the math is as easy as that.

I won't test every cell like that but will run the same test on Cell #11 tomorrow.

As mentioned in the previous post, 12 cells have been charged. 1 to 10 were charged yesterday and 11 & 12 were charged this morning. I wanted to see how full they all ended up so did some very quick tests charging to 3.65V with current dropping to 0.02C.

These were very quick tests and you can see which cells came off the TP5000 controllers and which came off the CN3058E controllers in the following image. I was observing and controlling the tests via my phone and this is a portion of that screen capture:
Cell 11.png

The amps in rapidly tapers off forming something like a reverse J trailing tail. Cell 11 didn't accept/need much current to reach the voltage target and the current in looked more like an L rather than the reverse trailing J. It was like it was blocked & made me think of the memory effect discussions. The bumps you see were caused by me adjusting voltage up or down. Bumping the voltage up to 3.7V seemed to help get things going again.

Cell 12 tested like the others so I put Cell 11 on again and it looked like the others with that distinctive reverse J trailing tail.

I think it's worth doing a full capacity test on Cell #11 tomorrow after Cell #1's test completes. It'll be useful to compare the two tests.

[markopolo]

To convert the information I'm posting to apply it to a 12V nominal system multiply the volts by 4 and multiply the amps by the number of parallel batteries.

My 4S10P Example:
3.1v * 4 = 12.4v
0.25A * 10 paralleled cells = 2.5A

For capacity in Ah also multiply the Ah delivered by the number of parallel batteries.

Example: 6.382Ah * 10 = 63.82 Amp Hours

The first cell tested supplied 6.382Ah. That's close to what I hoped/wished for and a bit better than I expected. The cell had come off charge approx 45 minutes before the test started. It would not be unusual for an RVer to arrive at a campsite with a fully charged battery fresh off charging.

2020-2-3-8-34-9-EBD-M05.jpg

The plateaus or steps or support levels look like Battle Borns low rate discharge chart ( https://battlebornbatteries.com/comp...teries-series/ )


BB10012 curve steps.png

I added the arrows. Perhaps it is typical of low current discharges of cylindrical cells.

98% of the energy was delivered at or above 3.0V so only 2% supplied after that. I chose a 2.75V cutoff. After maybe 45 minutes, voltage recovered to only 2.8V.

If you have a large enough battery bank or only need to supply light to moderate loads there would seem to be little point to using a 2.5V cutoff. This all appears to make the 2.3V under-voltage protection at 2.30V listed in the Renogy LFP manual look very low (to me).

[Boxster1971]

Some what related. Noticed that LG Chem has started a Lithium-ion battery safety campaign. Full page ad in local paper. Here is link to their web site associated with campaign.

lgchem

Mostly targeting use of bare cells in vaping devices and flashlights.

[markopolo]

Just wanted to add that after 20 hours at rest the average voltage of these cells is 3.443V. 3.443V*4 = 13.772V.


I think 13.6V resting voltage for nominal 12V LiFePO4 system would be close to being full and a reasonable target if you want to get the battery near fully charged. There probably is not much energy stored between actually fully charged and 13.6V.

[markopolo]

Originally Posted by Boxster1971

Some what related. Noticed that LG Chem has started a Lithium-ion battery safety campaign. Full page ad in local paper. Here is link to their web site associated with campaign.

lgchem

Mostly targeting use of bare cells in vaping devices and flashlights.

It's a good reminder. People have been killed by cells in vaping devices exploding.

I have a good idea of the tremendous energy stored in batteries like this. It does apply to LiFePO4 and even the larger prismatic cells often seen in DIY RV systems. Short them by dropping a tool or some similar event and the result would likely be spectacular if not deadly.

[markopolo]

Boxster's post reminded me that I wanted to link to a video. These are the cells that I have - that's how/why I came across it. I don't know exactly what going on in the video but basically it gives an idea of what happens with a shorted cell and why you don't want to go there. The tabs might be OK for 5A/10A or so, no where near enough to start a truck or jump start a truck.

Skip ahead to around 13 minute point here: https://youtu.be/QtSyTboXG5o?t=807

https://www.youtube.com/watch?v=QtSy...youtu.be&t=806

[booster]

Originally Posted by markopolo

Just wanted to add that after 20 hours at rest the average voltage of these cells is 3.443V. 3.443V*4 = 13.772V.


I think 13.6V resting voltage for nominal 12V LiFePO4 system would be close to being full and a reasonable target if you want to get the battery near fully charged. There probably is not much energy stored between actually fully charged and 13.6V.

Very interesting and falls in line to a large degree with what we have been seeing in the industry recommendations lately concerning charge voltage and leaving headroom in the SOC, based on the knee of the current curve in charging. 13.8v as a cutoff has been seen repeatedly IIRC, although I also think we are seeing a bit notes stating that lower rate charging is best, and I think that would affect the cutoff point voltage. Perhaps stopping the voltage climb at 13.8v and having a bit a hold there would be best, or maybe even a bit lower voltage.


I think a very interesting thing to watch is what the drop in battery manufacturers do about charge voltage as things mature. AFAIK, they are all still holding on to the 14.6v point for charging cutoff or even holding there.

[markopolo]

My post re: 13.6V was probably not clear. I did charge them to the 14.6V equivalent. I should have typed:


I think 13.6V resting voltage for nominal 12V LiFePO4 system would be close to being full and a reasonable target (to see after charging and a rest period) if you want to get the battery near fully charged.

Booster makes an interesting point. When I next have an opportunity to charge a cell I'll try 13.8V/4 and up to see the acceptance rate.

[markopolo]

Some thoughts about LiFePO4 charging voltage.

I've owned LiFePO4 cells for like 3 or 4 days now so, by today's standards, that likely makes me an expert! (that's meant as a joke in case nobody gets it)

After replying to Booster's post I remembered that I had briefly tried lower voltage charging. It's instant feedback when you just have to move a dial on your phone to see how voltage affects current. There's an article on lower voltage charging on Powerstreams (sp) website that sums it up better than I ever could. It makes sense to me but, at this point, I like 14.6V to low current acceptance. I'd even go to 14.8V if needed to force something to happen. 14.8V is the upper limit for me. It's slightly into overcharge territory from the current research papers I've read.

Many on the forum will disagree with higher voltage charging and I'm certainly not suggesting that anyone do it. We have to make our own decisions.

There have been some stellar results reported by some who do lower voltage charging. I do wonder if it causes the memory effect we've heard of.

There is a distinction to be made between opportunity charging / everyday type charging and specifically near fully charging the battery periodically as part of preventative maintenance plan for care of the battery.

I also think some of the lower voltage charging stuff originates as an overreaction to very poor results the early LiFePO4 adopters had following the recommendations provided by manufacturers of the day. I've seen several posts about the early days of LiFePO4 charging recommendations being over 4VPC / 16V.

Some links and quotes about the actual upper voltage limit of LiFePO4:

https://www.redarc.com.au/news/colly...hium-batteries

LiFePO4 charging totally ceases at about 16.8 volts but most charger makers’ cut off at 14.65-14.7 volts.

https://www.solacity.com/how-to-keep...tteries-happy/

For LiFePO4 that window is about 8.0V (2.0V per cell) to 16.8 Volt (4.2V per cell). The built-in BMS should take care to keep the battery well within those limits.

Here's a link to a 2007 instruction manual from Thunder Sky - https://www.thunderstruck-ev.com/Man...t%20Manual.pdf

maximum charging voltage is 4.30V

I think LiFePO4 was referred to as LiFFePO4 back then:

https://enacademic.com/dic.nsf/enwiki/10954863

LFP typically refers to their Lithium Iron Phosphate (LiFFePO4) batteries.

https://www.powerstream.com/LLLF.htm

A LiFePO4 battery has a much wider overcharge tolerance of about 0.7V from its charging voltage plateau of 3.5V per cell.

A LiFePO4 battery can be safely overcharged to 4.2 volts per cell, but higher voltages will start to break down the organic electrolytes.

[booster]

Interesting stuff, thanks Marko.


We have seen some stuff on the memory effect on the batteries, and most said you needed a "periodic" high voltage charge/discharge is needed to prevent memory loss (in the batteries) from becoming permanent loss, but I don't recall ever seeing and actual time or frequency for it. Others have claimed that operating in the midrange will make the batteries last forever so who knows what is real.


Wouldn't it be interesting to find out that charging lithium was similar to lead cells with charge at voltage, hold, and then add a higher voltage quick cycle periodically

[markopolo]

It's a new day and a new definition of 100% charged LiFePO4 for me. It is 3.6V OCV for a new cell (BOL - Beginning of life) from what I've read. Nice to know, might be important later. It probably applies to top grade cells.

And, my opinion about various voltages has changed. 3.7V might be useful during the pack assembly stage if used on a single cell but caution and maybe luck would be needed if applying 14.8V to an assembled pack. This battery chemistry will absorb current to zero amps. If you charge at a voltage higher than 3.6VPC (14.4V) you run the risk of ending up with a OCV over 3.6VPC and apparently you don't want to do that.

Here are some links and quotes re: 3.6V OCV and 100% SOC: (link followed by quote from source)

https://arxiv.org/pdf/1810.09014.pdf

The accumulative charge was run continuously from 100% SoC at 3.6V to empty at 2.0V over several cycles in a discharge process

https://formula-hybrid.org/wp-conten...ical-Cells.pdf

and: http://www.drypower.com.au/files/spl...M12-90E3_2.pdf

apply a constant voltage hold at this voltage until the current decays to near-zero. This process charges the cells to 100% state of charge (SOC).

https://www.solarpaneltalk.com/forum...-battery-setup

The OCV should rest after 12 hours between 3.4 and 3.6v. Below 3.4v, and your cells are getting tired, or were not allowed to absorb enough.

https://electronics.stackexchange.co...epo4-batteries

A fully charged LiFePo4 battery will stay at 3.6v (or close to it) even after resting for 24hrs. Cells that have gone through many cycles may fall to 3.33-3.4v due to cell degradation.

More links:

https://www.evworks.com.au/page/tech...-lithium-batt/

In the case of LiFePO4 chemistry, the absolute maximum is 4.2V per cell, though it is recommended that you charge to 3.5-3.6V per cell, there is less than 1% extra capacity between 3.5V and 4.2V.

We're told that you get less cycles if you discharge 100% than if you discharge 80% for example.

Does it matter where the 80% comes from? Is it only from 100% and stopping at 20% that provides for longer battery life? Does a draw down from 90% to 10% provide the same additional battery life? How about 80% down 0%?

You need to know 100% in order to know 95% or 80% or 20% (for example).

I don't know how different A123's LiFePO4 Nanophosphate is compared to the plain old LiFePO4 that most of us probably have. Their instructions to charge to 3.6VPC (14.4V) with current decaying to near zero and their 3.5VPC float (14.0V) sure is interesting.

As of today & for the pack I'm assembling, I now like charging to 3.6VPC (14.4V) and very low current for a "periodic" full charge. Fingers crossed. That will keep the cells below 3.6VPC OCV.

My opinion might change again.

[markopolo]

I targeted 3.5VPC OCV yesterday and ended up with 3.51 VPC average after 9 hours minimum at rest. I've logged it all in a spreadsheet and will check them again in 24 hrs.

I got there by holding 3.6V until almost zero current (0.00x amps). It wasn't a long absorption - I did it on a per cell basis (24 cells now). Once these cells get to 3.6V then current in falls fairly quickly.

3.5VPC * 4 = 14V

Looks to me like that's the maximum voltage I should do balancing at.

I happen to have a Atwood onboard battery maintainer that I picked up for $5 or so from a pile of clearance stuff at Walmart a few years ago. It was cheap enough that I took a chance on it. It is a 1.5A charger with a bit of an odd profile IMO that charges to 14.2 and then floats at 14V so I have never had a use for it. It might be perfect to bring all of the cells in the assembled pack to exactly 3.5VPC in conjunction with the balancing function of the BMS.

This is not advice, It's just my experiment.

I do think that you need to know what 100% is to know what 80% is though.