LiFePO4 cell level monitoring devices

[alctel]

The Orion JR is quite a bit cheaper and does more than enough, I have one. I have it driving a relay for general disconnect (the oh **** one), one for HVC, one for LVC and one that disconnects the power wire for the alt relay a bit before the HVC

[mbartosch]

Originally Posted by mikedefieslife

@mbartosch did you ever find anything? Low voltage is easy there are tons of alarms around and in anycase you could always disconnect the pack at 12.6/12.8 and be safe.

Following up on some ideas and recommendations of this thread I did some market research and decided on a modular approach to the whole maintenance and balancing issue.
My EE nerd and battery maintenance toolkit now consists of the following components:

- Cell monitor: ISDT BG-8s
- Balancing charger: ISDT T8
- Switching Power Supply: 600 W 230 V AC to 0-26 V DC
- 400 W 12V DC to 24 V DC step-up converter
- a DIY 4 * 50 W resistor array for manual balancing the pack, should that be necessary
- A DPS 5020 USB 50 V 20 A DC CV CC power supply for manual/EE purposes
- precision digital multimeter

The ISDT BG-8S was recommended by John and is actually pretty neat. I bought two units and connect them to the main bank or our two portable banks when needed. Which is basically when performing the yearly load test or rebalancing.

The ISDT T8 charges charges with 30 A and completely automates cell balancing (2.2 A max) very neatly. In the past I performed balancing manually with the 4 * 50 W resistor box which was totally annoying and time consuming.

The switching power supply is used to power the charger when shore power is available.

The 12 to 24 V DC-DC converter can be used to step up a 12 V battery pack voltage and balance-charge another one when not connected to shore power. I can use this to balance charge our portable dinghy packs if this is necessary.

The DPS 5020 lab power supply is a nerdy addition to the mix, basically gives me the possibility to quickly dial any voltage or CC I need. Only downside is that it does not isolate ground.

For high voltage disconnect there is nothing that ive seen short of buying a full BMS. I would assume that with an ardunio/ESP32 it might be possible to make something that will look at voltages and send a signal once any one of the cells reaches 3.65 or whatever cut off you want. You can then take that signal and sound an alarm/open a relay or anything else.

On Entropy we use a BMS which monitors cell and pack voltage for LVC/HVC but does not balance (I removed the balancing shunts). That's all I need, balancing and capacity test is done once a year manually.

And yes, of course it's possible to DIY a cell monitor with an embedded system and a I2C precision quad differential voltage sensor, but I'd ask myself why I don't just simply buy one...

[john61ct]

Hundreds of cheap chinese dual-port BMSs can be configured to drive whatever relays + contactors you like.

In going that route, you don't have to pay for the (claimed) superior control FETs since current is not passing through the BMS.

Of course you need external OCP, but everyone's fusing the power circuits anyway right?

Info displays or BT comms with your smartphone, user adjustability included.

I'm not advocating this approach, but I have to believe an amateur trying to duplicate all that functionality by kludging around themselves programming MCUs, is not going to get a more reliable result.

[john61ct]

Originally Posted by mbartosch

The DPS 5020 lab power supply is a nerdy addition to the mix, basically gives me the possibility to quickly dial any voltage or CC I need. Only downside is that it does not isolate ground.

For charging serial-connected packs at mid-point taps, I know to look for isolation. . .

Can you ELI5 why that would be important otherwise?

[john61ct]

Originally Posted by mbartosch

we use a BMS which monitors cell and pack voltage for LVC/HVC but does not balance

Link?

reliable so far?

do you use a failsafe at bank level voltages?

[CatNewBee]

https://youtu.be/R6QQkNXN-9k
... If you stil think a BMS is not required...

[john61ct]

Are you saying he did not have one?

Often relying on "a BMS"

as in a single product you just buy and hook up

is what causes fires, probably as much as them preventing them.

It is your whole system - design, component selection, implementation, ongoing care and maintenance - that ensures safety, reliability and longevity, not reliance on some magic bullet device.

And finally, I wish idiots would stop acting as if the label "lithium" batteries designated any particular chemistry out of the many hundreds falling under that general umbrella term.

This was not an LFP pack, but scrapped EV LI cells from a Leaf.

Which is why it was in the backyard shed, I hope just a science experiment, would be completely idiotic to consider for use on a boat.

Which is the actual takeaway from the video.

[mbartosch]

Originally Posted by john61ct

For charging serial-connected packs at mid-point taps, I know to look for isolation. . .

Can you ELI5 why that would be important otherwise?

As mentioned I have a resistor box for manually shunting current when balancing, allowing to reduce SoC of a cell in the pack.

If the DPS5020 DC lab supply had an isolated ground one could attach DC of the whole pack to DC in of the DC Lab supply and - now it gets evil - output to the one cell of the same pack that needs higher SoC with respect to the other cells.
Note: Do NOT do that unless you absolutely know what you are doing! (In particular with non-isolated PS such as the mentioned DPS5020)

[mbartosch]

Originally Posted by john61ct

Link?

reliable so far?

do you use a failsafe at bank level voltages?

I documented the entire design on our blog, see https://www.entropypool.de/engineering/

It‘s a House Power BMS which is no longer sold. It does work fine and absolutely does what it is supposed to do.
The BMS monitors cell level voltage and pack voltage, it triggers HVC, LVC and a separate complete shutdown via the main contactor.

I added pack level warning thresholds using a Victron BMV 700, programmed to generate alarms at high and low pack level voltage as well as a programmed alarm at 20 % SoC (calculated via Coulomb counting)

[mikedefieslife]

I have the ISDT IG8 already for checking voltages. I actually find that it's not particularly accurate, but it is at least constant. The BMS I run is basically a BMS of last resort, all current runs through it. It will do HVC/LVC itself, but won't control an external relay.

I would like to keep it that i.e as a last resort, and have another device trigger relays at whatever settings I prescribe. Contrary to John's post I've not noticed any on the usual Chinese shopping websites.

The Orion Jr looks like an expensive way of doing that. I'm not impressed with what I've read considering practically everything is an additional add-on for it.